diff options
31 files changed, 3191 insertions, 12754 deletions
diff --git a/doc/doc-txt/ChangeLog b/doc/doc-txt/ChangeLog index ddc559685..198e991a2 100644 --- a/doc/doc-txt/ChangeLog +++ b/doc/doc-txt/ChangeLog @@ -1,10 +1,11 @@ -$Cambridge: exim/doc/doc-txt/ChangeLog,v 1.428 2006/11/07 15:56:17 ph10 Exp $ +$Cambridge: exim/doc/doc-txt/ChangeLog,v 1.429 2006/11/07 16:50:36 ph10 Exp $ Change log file for Exim from version 4.21 ------------------------------------------- Exim version 4.64 ----------------- + TK/01 Bugzilla #401. Fix DK spooling code so that it can overwrite a leftover -K file (the existence of which was triggered by #402). While we were at it, introduced process PID as part of the -K @@ -259,6 +260,8 @@ PH/39 If -R or -S was given with -q<time>, the effect of -R or -S was ignored, and queue runs started by the daemon processed all messages. This has been fixed so that -R and -S can now usefully be given with -q<time>. +PH/40 Import PCRE release 6.7 (fixes some bugs). + Exim version 4.63 ----------------- diff --git a/doc/doc-txt/pcrepattern.txt b/doc/doc-txt/pcrepattern.txt index b638a00db..9712c86b4 100644 --- a/doc/doc-txt/pcrepattern.txt +++ b/doc/doc-txt/pcrepattern.txt @@ -1,5 +1,5 @@ This file contains the PCRE man page that describes the regular expressions -supported by PCRE version 6.2. Note that not all of the features are relevant +supported by PCRE version 6.7. Note that not all of the features are relevant in the context of Exim. In particular, the version of PCRE that is compiled with Exim does not include UTF-8 support, there is no mechanism for changing the options with which the PCRE functions are called, and features such as @@ -110,9 +110,9 @@ BACKSLASH If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the pattern (other than in a character class) and characters between a - # outside a character class and the next newline character are ignored. - An escaping backslash can be used to include a whitespace or # charac- - ter as part of the pattern. + # outside a character class and the next newline are ignored. An escap- + ing backslash can be used to include a whitespace or # character as + part of the pattern. If you want to remove the special meaning from a sequence of charac- ters, you can do so by putting them between \Q and \E. This is differ- @@ -148,7 +148,7 @@ BACKSLASH \t tab (hex 09) \ddd character with octal code ddd, or backreference \xhh character with hex code hh - \x{hhh..} character with hex code hhh... (UTF-8 mode only) + \x{hhh..} character with hex code hhh.. The precise effect of \cx is as follows: if x is a lower case letter, it is converted to upper case. Then bit 6 of the character (hex 40) is @@ -156,26 +156,24 @@ BACKSLASH becomes hex 7B. After \x, from zero to two hexadecimal digits are read (letters can be - in upper or lower case). In UTF-8 mode, any number of hexadecimal dig- - its may appear between \x{ and }, but the value of the character code - must be less than 2**31 (that is, the maximum hexadecimal value is - 7FFFFFFF). If characters other than hexadecimal digits appear between - \x{ and }, or if there is no terminating }, this form of escape is not - recognized. Instead, the initial \x will be interpreted as a basic - hexadecimal escape, with no following digits, giving a character whose - value is zero. + in upper or lower case). Any number of hexadecimal digits may appear + between \x{ and }, but the value of the character code must be less + than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8 mode (that is, + the maximum hexadecimal value is 7FFFFFFF). If characters other than + hexadecimal digits appear between \x{ and }, or if there is no termi- + nating }, this form of escape is not recognized. Instead, the initial + \x will be interpreted as a basic hexadecimal escape, with no following + digits, giving a character whose value is zero. Characters whose value is less than 256 can be defined by either of the - two syntaxes for \x when PCRE is in UTF-8 mode. There is no difference - in the way they are handled. For example, \xdc is exactly the same as - \x{dc}. + two syntaxes for \x. There is no difference in the way they are han- + dled. For example, \xdc is exactly the same as \x{dc}. - After \0 up to two further octal digits are read. In both cases, if - there are fewer than two digits, just those that are present are used. - Thus the sequence \0\x\07 specifies two binary zeros followed by a BEL - character (code value 7). Make sure you supply two digits after the - initial zero if the pattern character that follows is itself an octal - digit. + After \0 up to two further octal digits are read. If there are fewer + than two digits, just those that are present are used. Thus the + sequence \0\x\07 specifies two binary zeros followed by a BEL character + (code value 7). Make sure you supply two digits after the initial zero + if the pattern character that follows is itself an octal digit. The handling of a backslash followed by a digit other than 0 is compli- cated. Outside a character class, PCRE reads it and any following dig- @@ -187,9 +185,11 @@ BACKSLASH Inside a character class, or if the decimal number is greater than 9 and there have not been that many capturing subpatterns, PCRE re-reads - up to three octal digits following the backslash, and generates a sin- - gle byte from the least significant 8 bits of the value. Any subsequent - digits stand for themselves. For example: + up to three octal digits following the backslash, ane uses them to gen- + erate a data character. Any subsequent digits stand for themselves. In + non-UTF-8 mode, the value of a character specified in octal must be + less than \400. In UTF-8 mode, values up to \777 are permitted. For + example: \040 is another way of writing a space \40 is the same, provided there are fewer than 40 @@ -209,16 +209,15 @@ BACKSLASH Note that octal values of 100 or greater must not be introduced by a leading zero, because no more than three octal digits are ever read. - All the sequences that define a single byte value or a single UTF-8 - character (in UTF-8 mode) can be used both inside and outside character - classes. In addition, inside a character class, the sequence \b is - interpreted as the backspace character (hex 08), and the sequence \X is - interpreted as the character "X". Outside a character class, these - sequences have different meanings (see below). + All the sequences that define a single character value can be used both + inside and outside character classes. In addition, inside a character + class, the sequence \b is interpreted as the backspace character (hex + 08), and the sequence \X is interpreted as the character "X". Outside a + character class, these sequences have different meanings (see below). Generic character types - The third use of backslash is for specifying generic character types. + The third use of backslash is for specifying generic character types. The following are always recognized: \d any decimal digit @@ -229,56 +228,83 @@ BACKSLASH \W any "non-word" character Each pair of escape sequences partitions the complete set of characters - into two disjoint sets. Any given character matches one, and only one, + into two disjoint sets. Any given character matches one, and only one, of each pair. These character type sequences can appear both inside and outside char- - acter classes. They each match one character of the appropriate type. - If the current matching point is at the end of the subject string, all + acter classes. They each match one character of the appropriate type. + If the current matching point is at the end of the subject string, all of them fail, since there is no character to match. - For compatibility with Perl, \s does not match the VT character (code - 11). This makes it different from the the POSIX "space" class. The \s - characters are HT (9), LF (10), FF (12), CR (13), and space (32). + For compatibility with Perl, \s does not match the VT character (code + 11). This makes it different from the the POSIX "space" class. The \s + characters are HT (9), LF (10), FF (12), CR (13), and space (32). (If + "use locale;" is included in a Perl script, \s may match the VT charac- + ter. In PCRE, it never does.) A "word" character is an underscore or any character less than 256 that - is a letter or digit. The definition of letters and digits is con- - trolled by PCRE's low-valued character tables, and may vary if locale- - specific matching is taking place (see "Locale support" in the pcreapi - page). For example, in the "fr_FR" (French) locale, some character - codes greater than 128 are used for accented letters, and these are + is a letter or digit. The definition of letters and digits is con- + trolled by PCRE's low-valued character tables, and may vary if locale- + specific matching is taking place (see "Locale support" in the pcreapi + page). For example, in the "fr_FR" (French) locale, some character + codes greater than 128 are used for accented letters, and these are matched by \w. - In UTF-8 mode, characters with values greater than 128 never match \d, + In UTF-8 mode, characters with values greater than 128 never match \d, \s, or \w, and always match \D, \S, and \W. This is true even when Uni- - code character property support is available. + code character property support is available. The use of locales with + Unicode is discouraged. Unicode character properties When PCRE is built with Unicode character property support, three addi- - tional escape sequences to match generic character types are available + tional escape sequences to match character properties are available when UTF-8 mode is selected. They are: - \p{xx} a character with the xx property - \P{xx} a character without the xx property - \X an extended Unicode sequence + \p{xx} a character with the xx property + \P{xx} a character without the xx property + \X an extended Unicode sequence The property names represented by xx above are limited to the Unicode - general category properties. Each character has exactly one such prop- - erty, specified by a two-letter abbreviation. For compatibility with - Perl, negation can be specified by including a circumflex between the - opening brace and the property name. For example, \p{^Lu} is the same - as \P{Lu}. - - If only one letter is specified with \p or \P, it includes all the - properties that start with that letter. In this case, in the absence of - negation, the curly brackets in the escape sequence are optional; these - two examples have the same effect: + script names, the general category properties, and "Any", which matches + any character (including newline). Other properties such as "InMusical- + Symbols" are not currently supported by PCRE. Note that \P{Any} does + not match any characters, so always causes a match failure. + + Sets of Unicode characters are defined as belonging to certain scripts. + A character from one of these sets can be matched using a script name. + For example: + + \p{Greek} + \P{Han} + + Those that are not part of an identified script are lumped together as + "Common". The current list of scripts is: + + Arabic, Armenian, Bengali, Bopomofo, Braille, Buginese, Buhid, Cana- + dian_Aboriginal, Cherokee, Common, Coptic, Cypriot, Cyrillic, Deseret, + Devanagari, Ethiopic, Georgian, Glagolitic, Gothic, Greek, Gujarati, + Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Inherited, Kannada, + Katakana, Kharoshthi, Khmer, Lao, Latin, Limbu, Linear_B, Malayalam, + Mongolian, Myanmar, New_Tai_Lue, Ogham, Old_Italic, Old_Persian, Oriya, + Osmanya, Runic, Shavian, Sinhala, Syloti_Nagri, Syriac, Tagalog, Tag- + banwa, Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, + Ugaritic, Yi. + + Each character has exactly one general category property, specified by + a two-letter abbreviation. For compatibility with Perl, negation can be + specified by including a circumflex between the opening brace and the + property name. For example, \p{^Lu} is the same as \P{Lu}. + + If only one letter is specified with \p or \P, it includes all the gen- + eral category properties that start with that letter. In this case, in + the absence of negation, the curly brackets in the escape sequence are + optional; these two examples have the same effect: \p{L} \pL - The following property codes are supported: + The following general category property codes are supported: C Other Cc Control @@ -324,8 +350,17 @@ BACKSLASH Zp Paragraph separator Zs Space separator - Extended properties such as "Greek" or "InMusicalSymbols" are not sup- - ported by PCRE. + The special property L& is also supported: it matches a character that + has the Lu, Ll, or Lt property, in other words, a letter that is not + classified as a modifier or "other". + + The long synonyms for these properties that Perl supports (such as + \p{Letter}) are not supported by PCRE, nor is it permitted to prefix + any of these properties with "Is". + + No character that is in the Unicode table has the Cn (unassigned) prop- + erty. Instead, this property is assumed for any code point that is not + in the Unicode table. Specifying caseless matching does not affect these escape sequences. For example, \p{Lu} always matches only upper case letters. @@ -378,24 +413,23 @@ BACKSLASH However, if the startoffset argument of pcre_exec() is non-zero, indi- cating that matching is to start at a point other than the beginning of the subject, \A can never match. The difference between \Z and \z is - that \Z matches before a newline that is the last character of the - string as well as at the end of the string, whereas \z matches only at - the end. - - The \G assertion is true only when the current matching position is at - the start point of the match, as specified by the startoffset argument - of pcre_exec(). It differs from \A when the value of startoffset is - non-zero. By calling pcre_exec() multiple times with appropriate argu- + that \Z matches before a newline at the end of the string as well as at + the very end, whereas \z matches only at the end. + + The \G assertion is true only when the current matching position is at + the start point of the match, as specified by the startoffset argument + of pcre_exec(). It differs from \A when the value of startoffset is + non-zero. By calling pcre_exec() multiple times with appropriate argu- ments, you can mimic Perl's /g option, and it is in this kind of imple- mentation where \G can be useful. - Note, however, that PCRE's interpretation of \G, as the start of the + Note, however, that PCRE's interpretation of \G, as the start of the current match, is subtly different from Perl's, which defines it as the - end of the previous match. In Perl, these can be different when the - previously matched string was empty. Because PCRE does just one match + end of the previous match. In Perl, these can be different when the + previously matched string was empty. Because PCRE does just one match at a time, it cannot reproduce this behaviour. - If all the alternatives of a pattern begin with \G, the expression is + If all the alternatives of a pattern begin with \G, the expression is anchored to the starting match position, and the "anchored" flag is set in the compiled regular expression. @@ -403,68 +437,81 @@ BACKSLASH CIRCUMFLEX AND DOLLAR Outside a character class, in the default matching mode, the circumflex - character is an assertion that is true only if the current matching - point is at the start of the subject string. If the startoffset argu- - ment of pcre_exec() is non-zero, circumflex can never match if the - PCRE_MULTILINE option is unset. Inside a character class, circumflex + character is an assertion that is true only if the current matching + point is at the start of the subject string. If the startoffset argu- + ment of pcre_exec() is non-zero, circumflex can never match if the + PCRE_MULTILINE option is unset. Inside a character class, circumflex has an entirely different meaning (see below). - Circumflex need not be the first character of the pattern if a number - of alternatives are involved, but it should be the first thing in each - alternative in which it appears if the pattern is ever to match that - branch. If all possible alternatives start with a circumflex, that is, - if the pattern is constrained to match only at the start of the sub- - ject, it is said to be an "anchored" pattern. (There are also other + Circumflex need not be the first character of the pattern if a number + of alternatives are involved, but it should be the first thing in each + alternative in which it appears if the pattern is ever to match that + branch. If all possible alternatives start with a circumflex, that is, + if the pattern is constrained to match only at the start of the sub- + ject, it is said to be an "anchored" pattern. (There are also other constructs that can cause a pattern to be anchored.) - A dollar character is an assertion that is true only if the current - matching point is at the end of the subject string, or immediately - before a newline character that is the last character in the string (by - default). Dollar need not be the last character of the pattern if a - number of alternatives are involved, but it should be the last item in - any branch in which it appears. Dollar has no special meaning in a - character class. + A dollar character is an assertion that is true only if the current + matching point is at the end of the subject string, or immediately + before a newline at the end of the string (by default). Dollar need not + be the last character of the pattern if a number of alternatives are + involved, but it should be the last item in any branch in which it + appears. Dollar has no special meaning in a character class. The meaning of dollar can be changed so that it matches only at the very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This does not affect the \Z assertion. The meanings of the circumflex and dollar characters are changed if the - PCRE_MULTILINE option is set. When this is the case, they match immedi- - ately after and immediately before an internal newline character, - respectively, in addition to matching at the start and end of the sub- - ject string. For example, the pattern /^abc$/ matches the subject - string "def\nabc" (where \n represents a newline character) in multi- - line mode, but not otherwise. Consequently, patterns that are anchored - in single line mode because all branches start with ^ are not anchored - in multiline mode, and a match for circumflex is possible when the - startoffset argument of pcre_exec() is non-zero. The PCRE_DOL- - LAR_ENDONLY option is ignored if PCRE_MULTILINE is set. - - Note that the sequences \A, \Z, and \z can be used to match the start - and end of the subject in both modes, and if all branches of a pattern - start with \A it is always anchored, whether PCRE_MULTILINE is set or - not. + PCRE_MULTILINE option is set. When this is the case, a circumflex + matches immediately after internal newlines as well as at the start of + the subject string. It does not match after a newline that ends the + string. A dollar matches before any newlines in the string, as well as + at the very end, when PCRE_MULTILINE is set. When newline is specified + as the two-character sequence CRLF, isolated CR and LF characters do + not indicate newlines. + + For example, the pattern /^abc$/ matches the subject string "def\nabc" + (where \n represents a newline) in multiline mode, but not otherwise. + Consequently, patterns that are anchored in single line mode because + all branches start with ^ are not anchored in multiline mode, and a + match for circumflex is possible when the startoffset argument of + pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if + PCRE_MULTILINE is set. + + Note that the sequences \A, \Z, and \z can be used to match the start + and end of the subject in both modes, and if all branches of a pattern + start with \A it is always anchored, whether or not PCRE_MULTILINE is + set. FULL STOP (PERIOD, DOT) Outside a character class, a dot in the pattern matches any one charac- - ter in the subject, including a non-printing character, but not (by - default) newline. In UTF-8 mode, a dot matches any UTF-8 character, - which might be more than one byte long, except (by default) newline. If - the PCRE_DOTALL option is set, dots match newlines as well. The han- - dling of dot is entirely independent of the handling of circumflex and - dollar, the only relationship being that they both involve newline - characters. Dot has no special meaning in a character class. + ter in the subject string except (by default) a character that signi- + fies the end of a line. In UTF-8 mode, the matched character may be + more than one byte long. When a line ending is defined as a single + character (CR or LF), dot never matches that character; when the two- + character sequence CRLF is used, dot does not match CR if it is immedi- + ately followed by LF, but otherwise it matches all characters (includ- + ing isolated CRs and LFs). + + The behaviour of dot with regard to newlines can be changed. If the + PCRE_DOTALL option is set, a dot matches any one character, without + exception. If newline is defined as the two-character sequence CRLF, it + takes two dots to match it. + + The handling of dot is entirely independent of the handling of circum- + flex and dollar, the only relationship being that they both involve + newlines. Dot has no special meaning in a character class. MATCHING A SINGLE BYTE Outside a character class, the escape sequence \C matches any one byte, - both in and out of UTF-8 mode. Unlike a dot, it can match a newline. - The feature is provided in Perl in order to match individual bytes in - UTF-8 mode. Because it breaks up UTF-8 characters into individual + both in and out of UTF-8 mode. Unlike a dot, it always matches CR and + LF. The feature is provided in Perl in order to match individual bytes + in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes, what remains in the string may be a malformed UTF-8 string. For this reason, the \C escape sequence is best avoided. @@ -513,9 +560,11 @@ SQUARE BRACKETS AND CHARACTER CLASSES PCRE is compiled with Unicode property support as well as with UTF-8 support. - The newline character is never treated in any special way in character - classes, whatever the setting of the PCRE_DOTALL or PCRE_MULTILINE - options is. A class such as [^a] will always match a newline. + Characters that might indicate line breaks (CR and LF) are never + treated in any special way when matching character classes, whatever + line-ending sequence is in use, and whatever setting of the PCRE_DOTALL + and PCRE_MULTILINE options is used. A class such as [^a] always matches + one of these characters. The minus (hyphen) character can be used to specify a range of charac- ters in a character class. For example, [d-m] matches any letter @@ -616,11 +665,10 @@ VERTICAL BAR matches either "gilbert" or "sullivan". Any number of alternatives may appear, and an empty alternative is permitted (matching the empty - string). The matching process tries each alternative in turn, from - left to right, and the first one that succeeds is used. If the alterna- - tives are within a subpattern (defined below), "succeeds" means match- - ing the rest of the main pattern as well as the alternative in the sub- - pattern. + string). The matching process tries each alternative in turn, from left + to right, and the first one that succeeds is used. If the alternatives + are within a subpattern (defined below), "succeeds" means matching the + rest of the main pattern as well as the alternative in the subpattern. INTERNAL OPTION SETTING @@ -666,12 +714,9 @@ INTERNAL OPTION SETTING the effects of option settings happen at compile time. There would be some very weird behaviour otherwise. - The PCRE-specific options PCRE_UNGREEDY and PCRE_EXTRA can be changed - in the same way as the Perl-compatible options by using the characters - U and X respectively. The (?X) flag setting is special in that it must - always occur earlier in the pattern than any of the additional features - it turns on, even when it is at top level. It is best to put it at the - start. + The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA + can be changed in the same way as the Perl-compatible options by using + the characters J, U and X respectively. SUBPATTERNS @@ -683,18 +728,18 @@ SUBPATTERNS cat(aract|erpillar|) - matches one of the words "cat", "cataract", or "caterpillar". Without - the parentheses, it would match "cataract", "erpillar" or the empty + matches one of the words "cat", "cataract", or "caterpillar". Without + the parentheses, it would match "cataract", "erpillar" or the empty string. - 2. It sets up the subpattern as a capturing subpattern. This means - that, when the whole pattern matches, that portion of the subject + 2. It sets up the subpattern as a capturing subpattern. This means + that, when the whole pattern matches, that portion of the subject string that matched the subpattern is passed back to the caller via the - ovector argument of pcre_exec(). Opening parentheses are counted from - left to right (starting from 1) to obtain numbers for the capturing + ovector argument of pcre_exec(). Opening parentheses are counted from + left to right (starting from 1) to obtain numbers for the capturing subpatterns. - For example, if the string "the red king" is matched against the pat- + For example, if the string "the red king" is matched against the pat- tern the ((red|white) (king|queen)) @@ -702,50 +747,75 @@ SUBPATTERNS the captured substrings are "red king", "red", and "king", and are num- bered 1, 2, and 3, respectively. - The fact that plain parentheses fulfil two functions is not always - helpful. There are often times when a grouping subpattern is required - without a capturing requirement. If an opening parenthesis is followed - by a question mark and a colon, the subpattern does not do any captur- - ing, and is not counted when computing the number of any subsequent - capturing subpatterns. For example, if the string "the white queen" is + The fact that plain parentheses fulfil two functions is not always + helpful. There are often times when a grouping subpattern is required + without a capturing requirement. If an opening parenthesis is followed + by a question mark and a colon, the subpattern does not do any captur- + ing, and is not counted when computing the number of any subsequent + capturing subpatterns. For example, if the string "the white queen" is matched against the pattern the ((?:red|white) (king|queen)) the captured substrings are "white queen" and "queen", and are numbered - 1 and 2. The maximum number of capturing subpatterns is 65535, and the - maximum depth of nesting of all subpatterns, both capturing and non- + 1 and 2. The maximum number of capturing subpatterns is 65535, and the + maximum depth of nesting of all subpatterns, both capturing and non- capturing, is 200. - As a convenient shorthand, if any option settings are required at the - start of a non-capturing subpattern, the option letters may appear + As a convenient shorthand, if any option settings are required at the + start of a non-capturing subpattern, the option letters may appear between the "?" and the ":". Thus the two patterns (?i:saturday|sunday) (?:(?i)saturday|sunday) match exactly the same set of strings. Because alternative branches are - tried from left to right, and options are not reset until the end of - the subpattern is reached, an option setting in one branch does affect - subsequent branches, so the above patterns match "SUNDAY" as well as + tried from left to right, and options are not reset until the end of + the subpattern is reached, an option setting in one branch does affect + subsequent branches, so the above patterns match "SUNDAY" as well as "Saturday". NAMED SUBPATTERNS - Identifying capturing parentheses by number is simple, but it can be - very hard to keep track of the numbers in complicated regular expres- - sions. Furthermore, if an expression is modified, the numbers may - change. To help with this difficulty, PCRE supports the naming of sub- - patterns, something that Perl does not provide. The Python syntax - (?P<name>...) is used. Names consist of alphanumeric characters and - underscores, and must be unique within a pattern. - - Named capturing parentheses are still allocated numbers as well as + Identifying capturing parentheses by number is simple, but it can be + very hard to keep track of the numbers in complicated regular expres- + sions. Furthermore, if an expression is modified, the numbers may + change. To help with this difficulty, PCRE supports the naming of sub- + patterns, something that Perl does not provide. The Python syntax + (?P<name>...) is used. References to capturing parentheses from other + parts of the pattern, such as backreferences, recursion, and condi- + tions, can be made by name as well as by number. + + Names consist of up to 32 alphanumeric characters and underscores. + Named capturing parentheses are still allocated numbers as well as names. The PCRE API provides function calls for extracting the name-to- - number translation table from a compiled pattern. There is also a con- - venience function for extracting a captured substring by name. For fur- - ther details see the pcreapi documentation. + number translation table from a compiled pattern. There is also a con- + venience function for extracting a captured substring by name. + + By default, a name must be unique within a pattern, but it is possible + to relax this constraint by setting the PCRE_DUPNAMES option at compile + time. This can be useful for patterns where only one instance of the + named parentheses can match. Suppose you want to match the name of a + weekday, either as a 3-letter abbreviation or as the full name, and in + both cases you want to extract the abbreviation. This pattern (ignoring + the line breaks) does the job: + + (?P<DN>Mon|Fri|Sun)(?:day)?| + (?P<DN>Tue)(?:sday)?| + (?P<DN>Wed)(?:nesday)?| + (?P<DN>Thu)(?:rsday)?| + (?P<DN>Sat)(?:urday)? + + There are five capturing substrings, but only one is ever set after a + match. The convenience function for extracting the data by name + returns the substring for the first, and in this example, the only, + subpattern of that name that matched. This saves searching to find + which numbered subpattern it was. If you make a reference to a non- + unique named subpattern from elsewhere in the pattern, the one that + corresponds to the lowest number is used. For further details of the + interfaces for handling named subpatterns, see the pcreapi documenta- + tion. REPETITION @@ -954,8 +1024,10 @@ ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS meaning or processing of a possessive quantifier and the equivalent atomic group. - The possessive quantifier syntax is an extension to the Perl syntax. It - originates in Sun's Java package. + The possessive quantifier syntax is an extension to the Perl syntax. + Jeffrey Friedl originated the idea (and the name) in the first edition + of his book. Mike McCloskey liked it, so implemented it when he built + Sun's Java package, and PCRE copied it from there. When a pattern contains an unlimited repeat inside a subpattern that can itself be repeated an unlimited number of times, the use of an @@ -996,31 +1068,41 @@ BACK REFERENCES it is always taken as a back reference, and causes an error only if there are not that many capturing left parentheses in the entire pat- tern. In other words, the parentheses that are referenced need not be - to the left of the reference for numbers less than 10. See the subsec- - tion entitled "Non-printing characters" above for further details of - the handling of digits following a backslash. + to the left of the reference for numbers less than 10. A "forward back + reference" of this type can make sense when a repetition is involved + and the subpattern to the right has participated in an earlier itera- + tion. - A back reference matches whatever actually matched the capturing sub- - pattern in the current subject string, rather than anything matching + It is not possible to have a numerical "forward back reference" to sub- + pattern whose number is 10 or more. However, a back reference to any + subpattern is possible using named parentheses (see below). See also + the subsection entitled "Non-printing characters" above for further + details of the handling of digits following a backslash. + + A back reference matches whatever actually matched the capturing sub- + pattern in the current subject string, rather than anything matching the subpattern itself (see "Subpatterns as subroutines" below for a way of doing that). So the pattern (sens|respons)e and \1ibility - matches "sense and sensibility" and "response and responsibility", but - not "sense and responsibility". If caseful matching is in force at the - time of the back reference, the case of letters is relevant. For exam- + matches "sense and sensibility" and "response and responsibility", but + not "sense and responsibility". If caseful matching is in force at the + time of the back reference, the case of letters is relevant. For exam- ple, ((?i)rah)\s+\1 - matches "rah rah" and "RAH RAH", but not "RAH rah", even though the + matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original capturing subpattern is matched caselessly. - Back references to named subpatterns use the Python syntax (?P=name). + Back references to named subpatterns use the Python syntax (?P=name). We could rewrite the above example as follows: - (?<p1>(?i)rah)\s+(?P=p1) + (?P<p1>(?i)rah)\s+(?P=p1) + + A subpattern that is referenced by name may appear in the pattern + before or after the reference. There may be more than one back reference to the same subpattern. If a subpattern has not actually been used in a particular match, any back @@ -1109,8 +1191,8 @@ ASSERTIONS does find an occurrence of "bar" that is not preceded by "foo". The contents of a lookbehind assertion are restricted such that all the strings it matches must have a fixed length. However, if there are sev- - eral alternatives, they do not all have to have the same fixed length. - Thus + eral top-level alternatives, they do not all have to have the same + fixed length. Thus (?<=bullock|donkey) @@ -1223,12 +1305,18 @@ CONDITIONAL SUBPATTERNS tives in the subpattern, a compile-time error occurs. There are three kinds of condition. If the text between the parentheses - consists of a sequence of digits, the condition is satisfied if the - capturing subpattern of that number has previously matched. The number - must be greater than zero. Consider the following pattern, which con- - tains non-significant white space to make it more readable (assume the - PCRE_EXTENDED option) and to divide it into three parts for ease of - discussion: + consists of a sequence of digits, or a sequence of alphanumeric charac- + ters and underscores, the condition is satisfied if the capturing sub- + pattern of that number or name has previously matched. There is a pos- + sible ambiguity here, because subpattern names may consist entirely of + digits. PCRE looks first for a named subpattern; if it cannot find one + and the text consists entirely of digits, it looks for a subpattern of + that number, which must be greater than zero. Using subpattern names + that consist entirely of digits is not recommended. + + Consider the following pattern, which contains non-significant white + space to make it more readable (assume the PCRE_EXTENDED option) and to + divide it into three parts for ease of discussion: ( \( )? [^()]+ (?(1) \) ) @@ -1241,12 +1329,16 @@ CONDITIONAL SUBPATTERNS tern is executed and a closing parenthesis is required. Otherwise, since no-pattern is not present, the subpattern matches nothing. In other words, this pattern matches a sequence of non-parentheses, - optionally enclosed in parentheses. + optionally enclosed in parentheses. Rewriting it to use a named subpat- + tern gives this: + + (?P<OPEN> \( )? [^()]+ (?(OPEN) \) ) - If the condition is the string (R), it is satisfied if a recursive call - to the pattern or subpattern has been made. At "top level", the condi- - tion is false. This is a PCRE extension. Recursive patterns are - described in the next section. + If the condition is the string (R), and there is no subpattern with the + name R, the condition is satisfied if a recursive call to the pattern + or subpattern has been made. At "top level", the condition is false. + This is a PCRE extension. Recursive patterns are described in the next + section. If the condition is not a sequence of digits or (R), it must be an assertion. This may be a positive or negative lookahead or lookbehind @@ -1273,8 +1365,8 @@ COMMENTS at all. If the PCRE_EXTENDED option is set, an unescaped # character outside a - character class introduces a comment that continues up to the next new- - line character in the pattern. + character class introduces a comment that continues to immediately + after the next newline in the pattern. RECURSIVE PATTERNS @@ -1304,15 +1396,19 @@ RECURSIVE PATTERNS tion.) The special item (?R) is a recursive call of the entire regular expression. - For example, this PCRE pattern solves the nested parentheses problem - (assume the PCRE_EXTENDED option is set so that white space is - ignored): + A recursive subpattern call is always treated as an atomic group. That + is, once it has matched some of the subject string, it is never re- + entered, even if it contains untried alternatives and there is a subse- + quent matching failure. + + This PCRE pattern solves the nested parentheses problem (assume the + PCRE_EXTENDED option is set so that white space is ignored): \( ( (?>[^()]+) | (?R) )* \) First it matches an opening parenthesis. Then it matches any number of substrings which can either be a sequence of non-parentheses, or a - recursive match of the pattern itself (that is a correctly parenthe- + recursive match of the pattern itself (that is, a correctly parenthe- sized substring). Finally there is a closing parenthesis. If this were part of a larger pattern, you would not want to recurse @@ -1393,8 +1489,14 @@ SUBPATTERNS AS SUBROUTINES (sens|respons)e and (?1)ibility is used, it does match "sense and responsibility" as well as the other - two strings. Such references must, however, follow the subpattern to - which they refer. + two strings. Such references, if given numerically, must follow the + subpattern to which they refer. However, named references can refer to + later subpatterns. + + Like recursive subpatterns, a "subroutine" call is always treated as an + atomic group. That is, once it has matched some of the subject string, + it is never re-entered, even if it contains untried alternatives and + there is a subsequent matching failure. CALLOUTS @@ -1431,5 +1533,5 @@ CALLOUTS gether. A complete description of the interface to the callout function is given in the pcrecallout documentation. -Last updated: 28 February 2005 -Copyright (c) 1997-2005 University of Cambridge. +Last updated: 06 June 2006 +Copyright (c) 1997-2006 University of Cambridge. diff --git a/doc/doc-txt/pcretest.txt b/doc/doc-txt/pcretest.txt index 695741dae..dfa03b80b 100644 --- a/doc/doc-txt/pcretest.txt +++ b/doc/doc-txt/pcretest.txt @@ -12,8 +12,7 @@ NAME SYNOPSIS - pcretest [-C] [-d] [-dfa] [-i] [-m] [-o osize] [-p] [-t] [source] - [destination] + pcretest [options] [source] [destination] pcretest was written as a test program for the PCRE regular expression library itself, but it can also be used for experimenting with regular @@ -55,6 +54,12 @@ OPTIONS per API is used to call PCRE. None of the other options has any effect when -p is set. + -q Do not output the version number of pcretest at the start of + execution. + + -S size On Unix-like systems, set the size of the runtime stack to + size megabytes. + -t Run each compile, study, and match many times with a timer, and output resulting time per compile or match (in millisec- onds). Do not set -m with -t, because you will then get the @@ -76,53 +81,54 @@ DESCRIPTION ber of data lines to be matched against the pattern. Each data line is matched separately and independently. If you want to - do multiple-line matches, you have to use the \n escape sequence in a - single line of input to encode the newline characters. The maximum - length of data line is 30,000 characters. + do multi-line matches, you have to use the \n escape sequence (or \r or + \r\n, depending on the newline setting) in a single line of input to + encode the newline characters. There is no limit on the length of data + lines; the input buffer is automatically extended if it is too small. - An empty line signals the end of the data lines, at which point a new - regular expression is read. The regular expressions are given enclosed - in any non-alphanumeric delimiters other than backslash, for example + An empty line signals the end of the data lines, at which point a new + regular expression is read. The regular expressions are given enclosed + in any non-alphanumeric delimiters other than backslash, for example: /(a|bc)x+yz/ - White space before the initial delimiter is ignored. A regular expres- - sion may be continued over several input lines, in which case the new- - line characters are included within it. It is possible to include the + White space before the initial delimiter is ignored. A regular expres- + sion may be continued over several input lines, in which case the new- + line characters are included within it. It is possible to include the delimiter within the pattern by escaping it, for example /abc\/def/ - If you do so, the escape and the delimiter form part of the pattern, - but since delimiters are always non-alphanumeric, this does not affect - its interpretation. If the terminating delimiter is immediately fol- + If you do so, the escape and the delimiter form part of the pattern, + but since delimiters are always non-alphanumeric, this does not affect + its interpretation. If the terminating delimiter is immediately fol- lowed by a backslash, for example, /abc/\ - then a backslash is added to the end of the pattern. This is done to - provide a way of testing the error condition that arises if a pattern + then a backslash is added to the end of the pattern. This is done to + provide a way of testing the error condition that arises if a pattern finishes with a backslash, because /abc\/ - is interpreted as the first line of a pattern that starts with "abc/", + is interpreted as the first line of a pattern that starts with "abc/", causing pcretest to read the next line as a continuation of the regular expression. PATTERN MODIFIERS - A pattern may be followed by any number of modifiers, which are mostly - single characters. Following Perl usage, these are referred to below - as, for example, "the /i modifier", even though the delimiter of the - pattern need not always be a slash, and no slash is used when writing - modifiers. Whitespace may appear between the final pattern delimiter + A pattern may be followed by any number of modifiers, which are mostly + single characters. Following Perl usage, these are referred to below + as, for example, "the /i modifier", even though the delimiter of the + pattern need not always be a slash, and no slash is used when writing + modifiers. Whitespace may appear between the final pattern delimiter and the first modifier, and between the modifiers themselves. The /i, /m, /s, and /x modifiers set the PCRE_CASELESS, PCRE_MULTILINE, - PCRE_DOTALL, or PCRE_EXTENDED options, respectively, when pcre_com- - pile() is called. These four modifier letters have the same effect as + PCRE_DOTALL, or PCRE_EXTENDED options, respectively, when pcre_com- + pile() is called. These four modifier letters have the same effect as they do in Perl. For example: /caseless/i @@ -130,99 +136,111 @@ PATTERN MODIFIERS The following table shows additional modifiers for setting PCRE options that do not correspond to anything in Perl: - /A PCRE_ANCHORED - /C PCRE_AUTO_CALLOUT - /E PCRE_DOLLAR_ENDONLY - /f PCRE_FIRSTLINE - /N PCRE_NO_AUTO_CAPTURE - /U PCRE_UNGREEDY - /X PCRE_EXTRA - - Searching for all possible matches within each subject string can be - requested by the /g or /G modifier. After finding a match, PCRE is + /A PCRE_ANCHORED + /C PCRE_AUTO_CALLOUT + /E PCRE_DOLLAR_ENDONLY + /f PCRE_FIRSTLINE + /J PCRE_DUPNAMES + /N PCRE_NO_AUTO_CAPTURE + /U PCRE_UNGREEDY + /X PCRE_EXTRA + /<cr> PCRE_NEWLINE_CR + /<lf> PCRE_NEWLINE_LF + /<crlf> PCRE_NEWLINE_CRLF + + Those specifying line endings are literal strings as shown. Details of + the meanings of these PCRE options are given in the pcreapi documenta- + tion. + + Finding all matches in a string + + Searching for all possible matches within each subject string can be + requested by the /g or /G modifier. After finding a match, PCRE is called again to search the remainder of the subject string. The differ- ence between /g and /G is that the former uses the startoffset argument - to pcre_exec() to start searching at a new point within the entire - string (which is in effect what Perl does), whereas the latter passes - over a shortened substring. This makes a difference to the matching + to pcre_exec() to start searching at a new point within the entire + string (which is in effect what Perl does), whereas the latter passes + over a shortened substring. This makes a difference to the matching process if the pattern begins with a lookbehind assertion (including \b or \B). - If any call to pcre_exec() in a /g or /G sequence matches an empty - string, the next call is done with the PCRE_NOTEMPTY and PCRE_ANCHORED - flags set in order to search for another, non-empty, match at the same - point. If this second match fails, the start offset is advanced by - one, and the normal match is retried. This imitates the way Perl han- + If any call to pcre_exec() in a /g or /G sequence matches an empty + string, the next call is done with the PCRE_NOTEMPTY and PCRE_ANCHORED + flags set in order to search for another, non-empty, match at the same + point. If this second match fails, the start offset is advanced by + one, and the normal match is retried. This imitates the way Perl han- dles such cases when using the /g modifier or the split() function. + Other modifiers + There are yet more modifiers for controlling the way pcretest operates. - The /+ modifier requests that as well as outputting the substring that - matched the entire pattern, pcretest should in addition output the - remainder of the subject string. This is useful for tests where the + The /+ modifier requests that as well as outputting the substring that + matched the entire pattern, pcretest should in addition output the + remainder of the subject string. This is useful for tests where the subject contains multiple copies of the same substring. - The /L modifier must be followed directly by the name of a locale, for + The /L modifier must be followed directly by the name of a locale, for example, /pattern/Lfr_FR For this reason, it must be the last modifier. The given locale is set, - pcre_maketables() is called to build a set of character tables for the - locale, and this is then passed to pcre_compile() when compiling the - regular expression. Without an /L modifier, NULL is passed as the - tables pointer; that is, /L applies only to the expression on which it + pcre_maketables() is called to build a set of character tables for the + locale, and this is then passed to pcre_compile() when compiling the + regular expression. Without an /L modifier, NULL is passed as the + tables pointer; that is, /L applies only to the expression on which it appears. - The /I modifier requests that pcretest output information about the - compiled pattern (whether it is anchored, has a fixed first character, - and so on). It does this by calling pcre_fullinfo() after compiling a - pattern. If the pattern is studied, the results of that are also out- + The /I modifier requests that pcretest output information about the + compiled pattern (whether it is anchored, has a fixed first character, + and so on). It does this by calling pcre_fullinfo() after compiling a + pattern. If the pattern is studied, the results of that are also out- put. The /D modifier is a PCRE debugging feature, which also assumes /I. It - causes the internal form of compiled regular expressions to be output + causes the internal form of compiled regular expressions to be output after compilation. If the pattern was studied, the information returned is also output. The /F modifier causes pcretest to flip the byte order of the fields in - the compiled pattern that contain 2-byte and 4-byte numbers. This - facility is for testing the feature in PCRE that allows it to execute + the compiled pattern that contain 2-byte and 4-byte numbers. This + facility is for testing the feature in PCRE that allows it to execute patterns that were compiled on a host with a different endianness. This - feature is not available when the POSIX interface to PCRE is being - used, that is, when the /P pattern modifier is specified. See also the + feature is not available when the POSIX interface to PCRE is being + used, that is, when the /P pattern modifier is specified. See also the section about saving and reloading compiled patterns below. - The /S modifier causes pcre_study() to be called after the expression + The /S modifier causes pcre_study() to be called after the expression has been compiled, and the results used when the expression is matched. - The /M modifier causes the size of memory block used to hold the com- + The /M modifier causes the size of memory block used to hold the com- piled pattern to be output. - The /P modifier causes pcretest to call PCRE via the POSIX wrapper API - rather than its native API. When this is done, all other modifiers - except /i, /m, and /+ are ignored. REG_ICASE is set if /i is present, - and REG_NEWLINE is set if /m is present. The wrapper functions force - PCRE_DOLLAR_ENDONLY always, and PCRE_DOTALL unless REG_NEWLINE is set. + The /P modifier causes pcretest to call PCRE via the POSIX wrapper API + rather than its native API. When this is done, all other modifiers + except /i, /m, and /+ are ignored. REG_ICASE is set if /i is present, + and REG_NEWLINE is set if /m is present. The wrapper functions force + PCRE_DOLLAR_ENDONLY always, and PCRE_DOTALL unless REG_NEWLINE is set. - The /8 modifier causes pcretest to call PCRE with the PCRE_UTF8 option - set. This turns on support for UTF-8 character handling in PCRE, pro- - vided that it was compiled with this support enabled. This modifier + The /8 modifier causes pcretest to call PCRE with the PCRE_UTF8 option + set. This turns on support for UTF-8 character handling in PCRE, pro- + vided that it was compiled with this support enabled. This modifier also causes any non-printing characters in output strings to be printed using the \x{hh...} notation if they are valid UTF-8 sequences. - If the /? modifier is used with /8, it causes pcretest to call - pcre_compile() with the PCRE_NO_UTF8_CHECK option, to suppress the + If the /? modifier is used with /8, it causes pcretest to call + pcre_compile() with the PCRE_NO_UTF8_CHECK option, to suppress the checking of the string for UTF-8 validity. DATA LINES - Before each data line is passed to pcre_exec(), leading and trailing - whitespace is removed, and it is then scanned for \ escapes. Some of - these are pretty esoteric features, intended for checking out some of - the more complicated features of PCRE. If you are just testing "ordi- - nary" regular expressions, you probably don't need any of these. The + Before each data line is passed to pcre_exec(), leading and trailing + whitespace is removed, and it is then scanned for \ escapes. Some of + these are pretty esoteric features, intended for checking out some of + the more complicated features of PCRE. If you are just testing "ordi- + nary" regular expressions, you probably don't need any of these. The following escapes are recognized: \a alarm (= BEL) @@ -230,6 +248,8 @@ DATA LINES \e escape \f formfeed \n newline + \qdd set the PCRE_MATCH_LIMIT limit to dd + (any number of digits) \r carriage return \t tab \v vertical tab @@ -238,7 +258,9 @@ DATA LINES \x{hh...} hexadecimal character, any number of digits in UTF-8 mode \A pass the PCRE_ANCHORED option to pcre_exec() + or pcre_dfa_exec() \B pass the PCRE_NOTBOL option to pcre_exec() + or pcre_dfa_exec() \Cdd call pcre_copy_substring() for substring dd after a successful match (number less than 32) \Cname call pcre_copy_named_substring() for substring @@ -262,41 +284,58 @@ DATA LINES ated by next non-alphanumeric character) \L call pcre_get_substringlist() after a successful match - \M discover the minimum MATCH_LIMIT setting + \M discover the minimum MATCH_LIMIT and + MATCH_LIMIT_RECURSION settings \N pass the PCRE_NOTEMPTY option to pcre_exec() + or pcre_dfa_exec() \Odd set the size of the output vector passed to pcre_exec() to dd (any number of digits) \P pass the PCRE_PARTIAL option to pcre_exec() or pcre_dfa_exec() + \Qdd set the PCRE_MATCH_LIMIT_RECURSION limit to dd + (any number of digits) \R pass the PCRE_DFA_RESTART option to pcre_dfa_exec() \S output details of memory get/free calls during matching \Z pass the PCRE_NOTEOL option to pcre_exec() + or pcre_dfa_exec() \? pass the PCRE_NO_UTF8_CHECK option to - pcre_exec() + pcre_exec() or pcre_dfa_exec() \>dd start the match at offset dd (any number of digits); this sets the startoffset argument for pcre_exec() + or pcre_dfa_exec() + \<cr> pass the PCRE_NEWLINE_CR option to pcre_exec() + or pcre_dfa_exec() + \<lf> pass the PCRE_NEWLINE_LF option to pcre_exec() + or pcre_dfa_exec() + \<crlf> pass the PCRE_NEWLINE_CRLF option to pcre_exec() + or pcre_dfa_exec() - A backslash followed by anything else just escapes the anything else. - If the very last character is a backslash, it is ignored. This gives a - way of passing an empty line as data, since a real empty line termi- - nates the data input. + The escapes that specify line endings are literal strings, exactly as + shown. A backslash followed by anything else just escapes the anything + else. If the very last character is a backslash, it is ignored. This + gives a way of passing an empty line as data, since a real empty line + terminates the data input. If \M is present, pcretest calls pcre_exec() several times, with dif- - ferent values in the match_limit field of the pcre_extra data struc- - ture, until it finds the minimum number that is needed for pcre_exec() - to complete. This number is a measure of the amount of recursion and - backtracking that takes place, and checking it out can be instructive. - For most simple matches, the number is quite small, but for patterns - with very large numbers of matching possibilities, it can become large - very quickly with increasing length of subject string. - - When \O is used, the value specified may be higher or lower than the + ferent values in the match_limit and match_limit_recursion fields of + the pcre_extra data structure, until it finds the minimum numbers for + each parameter that allow pcre_exec() to complete. The match_limit num- + ber is a measure of the amount of backtracking that takes place, and + checking it out can be instructive. For most simple matches, the number + is quite small, but for patterns with very large numbers of matching + possibilities, it can become large very quickly with increasing length + of subject string. The match_limit_recursion number is a measure of how + much stack (or, if PCRE is compiled with NO_RECURSE, how much heap) + memory is needed to complete the match attempt. + + When \O is used, the value specified may be higher or lower than the size set by the -O command line option (or defaulted to 45); \O applies only to the call of pcre_exec() for the line in which it appears. - If the /P modifier was present on the pattern, causing the POSIX wrap- - per API to be used, only \B and \Z have any effect, causing REG_NOTBOL - and REG_NOTEOL to be passed to regexec() respectively. + If the /P modifier was present on the pattern, causing the POSIX wrap- + per API to be used, the only option-setting sequences that have any + effect are \B and \Z, causing REG_NOTBOL and REG_NOTEOL, respectively, + to be passed to regexec(). The use of \x{hh...} to represent UTF-8 characters is not dependent on the use of the /8 modifier on the pattern. It is recognized always. @@ -375,14 +414,15 @@ DEFAULT OUTPUT FROM PCRETEST Note that while patterns can be continued over several lines (a plain ">" prompt is used for continuations), data lines may not. However new- - lines can be included in data by means of the \n escape. + lines can be included in data by means of the \n escape (or \r or \r\n + for those newline settings). OUTPUT FROM THE ALTERNATIVE MATCHING FUNCTION - When the alternative matching function, pcre_dfa_exec(), is used (by - means of the \D escape sequence or the -dfa command line option), the - output consists of a list of all the matches that start at the first + When the alternative matching function, pcre_dfa_exec(), is used (by + means of the \D escape sequence or the -dfa command line option), the + output consists of a list of all the matches that start at the first point in the subject where there is at least one match. For example: re> /(tang|tangerine|tan)/ @@ -391,10 +431,10 @@ OUTPUT FROM THE ALTERNATIVE MATCHING FUNCTION 1: tang 2: tan - (Using the normal matching function on this data finds only "tang".) - The longest matching string is always given first (and numbered zero). + (Using the normal matching function on this data finds only "tang".) + The longest matching string is always given first (and numbered zero). - If /gP is present on the pattern, the search for further matches + If /gP is present on the pattern, the search for further matches resumes at the end of the longest match. For example: re> /(tang|tangerine|tan)/g @@ -406,16 +446,16 @@ OUTPUT FROM THE ALTERNATIVE MATCHING FUNCTION 1: tan 0: tan - Since the matching function does not support substring capture, the - escape sequences that are concerned with captured substrings are not + Since the matching function does not support substring capture, the + escape sequences that are concerned with captured substrings are not relevant. RESTARTING AFTER A PARTIAL MATCH When the alternative matching function has given the PCRE_ERROR_PARTIAL - return, indicating that the subject partially matched the pattern, you - can restart the match with additional subject data by means of the \R + return, indicating that the subject partially matched the pattern, you + can restart the match with additional subject data by means of the \R escape sequence. For example: re> /^?(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)$/ @@ -424,30 +464,30 @@ RESTARTING AFTER A PARTIAL MATCH data> n05\R\D 0: n05 - For further information about partial matching, see the pcrepartial + For further information about partial matching, see the pcrepartial documentation. CALLOUTS - If the pattern contains any callout requests, pcretest's callout func- - tion is called during matching. This works with both matching func- + If the pattern contains any callout requests, pcretest's callout func- + tion is called during matching. This works with both matching func- tions. By default, the called function displays the callout number, the - start and current positions in the text at the callout time, and the + start and current positions in the text at the callout time, and the next pattern item to be tested. For example, the output --->pqrabcdef 0 ^ ^ \d - indicates that callout number 0 occurred for a match attempt starting - at the fourth character of the subject string, when the pointer was at - the seventh character of the data, and when the next pattern item was - \d. Just one circumflex is output if the start and current positions + indicates that callout number 0 occurred for a match attempt starting + at the fourth character of the subject string, when the pointer was at + the seventh character of the data, and when the next pattern item was + \d. Just one circumflex is output if the start and current positions are the same. Callouts numbered 255 are assumed to be automatic callouts, inserted as - a result of the /C pattern modifier. In this case, instead of showing - the callout number, the offset in the pattern, preceded by a plus, is + a result of the /C pattern modifier. In this case, instead of showing + the callout number, the offset in the pattern, preceded by a plus, is output. For example: re> /\d?[A-E]\*/C @@ -459,68 +499,68 @@ CALLOUTS +10 ^ ^ 0: E* - The callout function in pcretest returns zero (carry on matching) by - default, but you can use a \C item in a data line (as described above) + The callout function in pcretest returns zero (carry on matching) by + default, but you can use a \C item in a data line (as described above) to change this. - Inserting callouts can be helpful when using pcretest to check compli- - cated regular expressions. For further information about callouts, see + Inserting callouts can be helpful when using pcretest to check compli- + cated regular expressions. For further information about callouts, see the pcrecallout documentation. SAVING AND RELOADING COMPILED PATTERNS - The facilities described in this section are not available when the + The facilities described in this section are not available when the POSIX inteface to PCRE is being used, that is, when the /P pattern mod- ifier is specified. When the POSIX interface is not in use, you can cause pcretest to write - a compiled pattern to a file, by following the modifiers with > and a + a compiled pattern to a file, by following the modifiers with > and a file name. For example: /pattern/im >/some/file - See the pcreprecompile documentation for a discussion about saving and + See the pcreprecompile documentation for a discussion about saving and re-using compiled patterns. - The data that is written is binary. The first eight bytes are the - length of the compiled pattern data followed by the length of the - optional study data, each written as four bytes in big-endian order - (most significant byte first). If there is no study data (either the + The data that is written is binary. The first eight bytes are the + length of the compiled pattern data followed by the length of the + optional study data, each written as four bytes in big-endian order + (most significant byte first). If there is no study data (either the pattern was not studied, or studying did not return any data), the sec- - ond length is zero. The lengths are followed by an exact copy of the + ond length is zero. The lengths are followed by an exact copy of the compiled pattern. If there is additional study data, this follows imme- - diately after the compiled pattern. After writing the file, pcretest + diately after the compiled pattern. After writing the file, pcretest expects to read a new pattern. A saved pattern can be reloaded into pcretest by specifing < and a file - name instead of a pattern. The name of the file must not contain a < - character, as otherwise pcretest will interpret the line as a pattern + name instead of a pattern. The name of the file must not contain a < + character, as otherwise pcretest will interpret the line as a pattern delimited by < characters. For example: re> </some/file Compiled regex loaded from /some/file No study data - When the pattern has been loaded, pcretest proceeds to read data lines + When the pattern has been loaded, pcretest proceeds to read data lines in the usual way. - You can copy a file written by pcretest to a different host and reload - it there, even if the new host has opposite endianness to the one on - which the pattern was compiled. For example, you can compile on an i86 + You can copy a file written by pcretest to a different host and reload + it there, even if the new host has opposite endianness to the one on + which the pattern was compiled. For example, you can compile on an i86 machine and run on a SPARC machine. - File names for saving and reloading can be absolute or relative, but - note that the shell facility of expanding a file name that starts with + File names for saving and reloading can be absolute or relative, but + note that the shell facility of expanding a file name that starts with a tilde (~) is not available. - The ability to save and reload files in pcretest is intended for test- - ing and experimentation. It is not intended for production use because - only a single pattern can be written to a file. Furthermore, there is - no facility for supplying custom character tables for use with a - reloaded pattern. If the original pattern was compiled with custom - tables, an attempt to match a subject string using a reloaded pattern - is likely to cause pcretest to crash. Finally, if you attempt to load + The ability to save and reload files in pcretest is intended for test- + ing and experimentation. It is not intended for production use because + only a single pattern can be written to a file. Furthermore, there is + no facility for supplying custom character tables for use with a + reloaded pattern. If the original pattern was compiled with custom + tables, an attempt to match a subject string using a reloaded pattern + is likely to cause pcretest to crash. Finally, if you attempt to load a file that is not in the correct format, the result is undefined. @@ -530,5 +570,5 @@ AUTHOR University Computing Service, Cambridge CB2 3QG, England. -Last updated: 28 February 2005 -Copyright (c) 1997-2005 University of Cambridge. +Last updated: 29 June 2006 +Copyright (c) 1997-2006 University of Cambridge. diff --git a/src/scripts/MakeLinks b/src/scripts/MakeLinks index b0b513c8b..a720a160a 100755 --- a/src/scripts/MakeLinks +++ b/src/scripts/MakeLinks @@ -1,5 +1,5 @@ #!/bin/sh -# $Cambridge: exim/src/scripts/MakeLinks,v 1.8 2006/10/16 15:44:36 ph10 Exp $ +# $Cambridge: exim/src/scripts/MakeLinks,v 1.9 2006/11/07 16:50:36 ph10 Exp $ # Script to build links for all the exim source files from the system- # specific build directory. It should be run from within that directory. @@ -43,7 +43,7 @@ ln -s ../../src/pcre/pcre_compile.c pcre_compile.c ln -s ../../src/pcre/pcre_exec.c pcre_exec.c ln -s ../../src/pcre/pcre_fullinfo.c pcre_fullinfo.c ln -s ../../src/pcre/pcretest.c pcretest.c -ln -s ../../src/pcre/pcre_printint.c pcre_printint.c +ln -s ../../src/pcre/pcre_printint.src pcre_printint.src ln -s ../../src/pcre/pcre_study.c pcre_study.c ln -s ../../src/pcre/pcre_tables.c pcre_tables.c ln -s ../../src/pcre/pcre_try_flipped.c pcre_try_flipped.c diff --git a/src/src/pcre/ChangeLog b/src/src/pcre/ChangeLog index b9d74d393..2a1413484 100644 --- a/src/src/pcre/ChangeLog +++ b/src/src/pcre/ChangeLog @@ -1,6 +1,459 @@ ChangeLog for PCRE ------------------ +Version 6.7 04-Jul-06 +--------------------- + + 1. In order to handle tests when input lines are enormously long, pcretest has + been re-factored so that it automatically extends its buffers when + necessary. The code is crude, but this _is_ just a test program. The + default size has been increased from 32K to 50K. + + 2. The code in pcre_study() was using the value of the re argument before + testing it for NULL. (Of course, in any sensible call of the function, it + won't be NULL.) + + 3. The memmove() emulation function in pcre_internal.h, which is used on + systems that lack both memmove() and bcopy() - that is, hardly ever - + was missing a "static" storage class specifier. + + 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns + containing an extended class (one that cannot be represented by a bitmap + because it contains high-valued characters or Unicode property items, e.g. + [\pZ]). Almost always one would set UTF-8 mode when processing such a + pattern, but PCRE should not loop if you do not (it no longer does). + [Detail: two cases were found: (a) a repeated subpattern containing an + extended class; (b) a recursive reference to a subpattern that followed a + previous extended class. It wasn't skipping over the extended class + correctly when UTF-8 mode was not set.] + + 5. A negated single-character class was not being recognized as fixed-length + in lookbehind assertions such as (?<=[^f]), leading to an incorrect + compile error "lookbehind assertion is not fixed length". + + 6. The RunPerlTest auxiliary script was showing an unexpected difference + between PCRE and Perl for UTF-8 tests. It turns out that it is hard to + write a Perl script that can interpret lines of an input file either as + byte characters or as UTF-8, which is what "perltest" was being required to + do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you + can't do is switch easily at run time between having the "use utf8;" pragma + or not. In the end, I fudged it by using the RunPerlTest script to insert + "use utf8;" explicitly for the UTF-8 tests. + + 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at + the end of the subject string, contrary to the documentation and to what + Perl does. This was true of both matching functions. Now it matches only at + the start of the subject and immediately after *internal* newlines. + + 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing + a pointer to an int instead of a pointer to an unsigned long int. This + caused problems on 64-bit systems. + + 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another + instance of the 'standard' template library not being so standard". + +10. There was no check on the number of named subpatterns nor the maximum + length of a subpattern name. The product of these values is used to compute + the size of the memory block for a compiled pattern. By supplying a very + long subpattern name and a large number of named subpatterns, the size + computation could be caused to overflow. This is now prevented by limiting + the length of names to 32 characters, and the number of named subpatterns + to 10,000. + +11. Subpatterns that are repeated with specific counts have to be replicated in + the compiled pattern. The size of memory for this was computed from the + length of the subpattern and the repeat count. The latter is limited to + 65535, but there was no limit on the former, meaning that integer overflow + could in principle occur. The compiled length of a repeated subpattern is + now limited to 30,000 bytes in order to prevent this. + +12. Added the optional facility to have named substrings with the same name. + +13. Added the ability to use a named substring as a condition, using the + Python syntax: (?(name)yes|no). This overloads (?(R)... and names that + are numbers (not recommended). Forward references are permitted. + +14. Added forward references in named backreferences (if you see what I mean). + +15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the + pattern could run off the end of the subject. For example, the pattern + "(?s)(.{1,5})"8 did this with the subject "ab". + +16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if + PCRE_CASELESS was set when matching characters that were quantified with ? + or *. + +17. A character class other than a single negated character that had a minimum + but no maximum quantifier - for example [ab]{6,} - was not handled + correctly by pce_dfa_exec(). It would match only one character. + +18. A valid (though odd) pattern that looked like a POSIX character + class but used an invalid character after [ (for example [[,abc,]]) caused + pcre_compile() to give the error "Failed: internal error: code overflow" or + in some cases to crash with a glibc free() error. This could even happen if + the pattern terminated after [[ but there just happened to be a sequence of + letters, a binary zero, and a closing ] in the memory that followed. + +19. Perl's treatment of octal escapes in the range \400 to \777 has changed + over the years. Originally (before any Unicode support), just the bottom 8 + bits were taken. Thus, for example, \500 really meant \100. Nowadays the + output from "man perlunicode" includes this: + + The regular expression compiler produces polymorphic opcodes. That + is, the pattern adapts to the data and automatically switches to + the Unicode character scheme when presented with Unicode data--or + instead uses a traditional byte scheme when presented with byte + data. + + Sadly, a wide octal escape does not cause a switch, and in a string with + no other multibyte characters, these octal escapes are treated as before. + Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern + /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a + Unicode string. + + I have not perpetrated such confusion in PCRE. Up till now, it took just + the bottom 8 bits, as in old Perl. I have now made octal escapes with + values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they + translate to the appropriate multibyte character. + +29. Applied some refactoring to reduce the number of warnings from Microsoft + and Borland compilers. This has included removing the fudge introduced + seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused + a warning about an unused variable. + +21. PCRE has not included VT (character 0x0b) in the set of whitespace + characters since release 4.0, because Perl (from release 5.004) does not. + [Or at least, is documented not to: some releases seem to be in conflict + with the documentation.] However, when a pattern was studied with + pcre_study() and all its branches started with \s, PCRE still included VT + as a possible starting character. Of course, this did no harm; it just + caused an unnecessary match attempt. + +22. Removed a now-redundant internal flag bit that recorded the fact that case + dependency changed within the pattern. This was once needed for "required + byte" processing, but is no longer used. This recovers a now-scarce options + bit. Also moved the least significant internal flag bit to the most- + significant bit of the word, which was not previously used (hangover from + the days when it was an int rather than a uint) to free up another bit for + the future. + +23. Added support for CRLF line endings as well as CR and LF. As well as the + default being selectable at build time, it can now be changed at runtime + via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to + specify that it is scanning data with non-default line endings. + +24. Changed the definition of CXXLINK to make it agree with the definition of + LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS. + +25. Applied Ian Taylor's patches to avoid using another stack frame for tail + recursions. This makes a big different to stack usage for some patterns. + +26. If a subpattern containing a named recursion or subroutine reference such + as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of + the space required for the compiled pattern went wrong and gave too small a + value. Depending on the environment, this could lead to "Failed: internal + error: code overflow at offset 49" or "glibc detected double free or + corruption" errors. + +27. Applied patches from Google (a) to support the new newline modes and (b) to + advance over multibyte UTF-8 characters in GlobalReplace. + +28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a + difference for some implementation of PCRE in some Windows version. + +29. Added some extra testing facilities to pcretest: + + \q<number> in a data line sets the "match limit" value + \Q<number> in a data line sets the "match recursion limt" value + -S <number> sets the stack size, where <number> is in megabytes + + The -S option isn't available for Windows. + + +Version 6.6 06-Feb-06 +--------------------- + + 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined + in pcreposix.h. I have copied the definition from pcre.h. + + 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree + because pcre.h is no longer a built file. + + 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are + not normally included in the compiled code. + + +Version 6.5 01-Feb-06 +--------------------- + + 1. When using the partial match feature with pcre_dfa_exec(), it was not + anchoring the second and subsequent partial matches at the new starting + point. This could lead to incorrect results. For example, with the pattern + /1234/, partially matching against "123" and then "a4" gave a match. + + 2. Changes to pcregrep: + + (a) All non-match returns from pcre_exec() were being treated as failures + to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an + error message is output. Some extra information is given for the + PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are + probably the only errors that are likely to be caused by users (by + specifying a regex that has nested indefinite repeats, for instance). + If there are more than 20 of these errors, pcregrep is abandoned. + + (b) A binary zero was treated as data while matching, but terminated the + output line if it was written out. This has been fixed: binary zeroes + are now no different to any other data bytes. + + (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is + used to set a locale for matching. The --locale=xxxx long option has + been added (no short equivalent) to specify a locale explicitly on the + pcregrep command, overriding the environment variables. + + (d) When -B was used with -n, some line numbers in the output were one less + than they should have been. + + (e) Added the -o (--only-matching) option. + + (f) If -A or -C was used with -c (count only), some lines of context were + accidentally printed for the final match. + + (g) Added the -H (--with-filename) option. + + (h) The combination of options -rh failed to suppress file names for files + that were found from directory arguments. + + (i) Added the -D (--devices) and -d (--directories) options. + + (j) Added the -F (--fixed-strings) option. + + (k) Allow "-" to be used as a file name for -f as well as for a data file. + + (l) Added the --colo(u)r option. + + (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it + is not present by default. + + 3. A nasty bug was discovered in the handling of recursive patterns, that is, + items such as (?R) or (?1), when the recursion could match a number of + alternatives. If it matched one of the alternatives, but subsequently, + outside the recursion, there was a failure, the code tried to back up into + the recursion. However, because of the way PCRE is implemented, this is not + possible, and the result was an incorrect result from the match. + + In order to prevent this happening, the specification of recursion has + been changed so that all such subpatterns are automatically treated as + atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)). + + 4. I had overlooked the fact that, in some locales, there are characters for + which isalpha() is true but neither isupper() nor islower() are true. In + the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine + and ordfeminine) are like this. This affected the treatment of \w and \W + when they appeared in character classes, but not when they appeared outside + a character class. The bit map for "word" characters is now created + separately from the results of isalnum() instead of just taking it from the + upper, lower, and digit maps. (Plus the underscore character, of course.) + + 5. The above bug also affected the handling of POSIX character classes such as + [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's + permanent tables. Instead, the bit maps for such a class were previously + created as the appropriate unions of the upper, lower, and digit bitmaps. + Now they are created by subtraction from the [[:word:]] class, which has + its own bitmap. + + 6. The [[:blank:]] character class matches horizontal, but not vertical space. + It is created by subtracting the vertical space characters (\x09, \x0a, + \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the + subtraction was done in the overall bitmap for a character class, meaning + that a class such as [\x0c[:blank:]] was incorrect because \x0c would not + be recognized. This bug has been fixed. + + 7. Patches from the folks at Google: + + (a) pcrecpp.cc: "to handle a corner case that may or may not happen in + real life, but is still worth protecting against". + + (b) pcrecpp.cc: "corrects a bug when negative radixes are used with + regular expressions". + + (c) pcre_scanner.cc: avoid use of std::count() because not all systems + have it. + + (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by + "configure" and the latter not, in order to fix a problem somebody had + with compiling the Arg class on HP-UX. + + (e) Improve the error-handling of the C++ wrapper a little bit. + + (f) New tests for checking recursion limiting. + + 8. The pcre_memmove() function, which is used only if the environment does not + have a standard memmove() function (and is therefore rarely compiled), + contained two bugs: (a) use of int instead of size_t, and (b) it was not + returning a result (though PCRE never actually uses the result). + + 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously + large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is + returned instead of calling malloc() with an overflowing number that would + most likely cause subsequent chaos. + +10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag. + +11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled + with this option is matched, the nmatch and pmatch options of regexec() are + ignored. + +12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is + provided in case anyone wants to the the POSIX interface with UTF-8 + strings. + +13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the + C++ linking (needed for some HP-UX environments). + +14. Avoid compiler warnings in get_ucpname() when compiled without UCP support + (unused parameter) and in the pcre_printint() function (omitted "default" + switch label when the default is to do nothing). + +15. Added some code to make it possible, when PCRE is compiled as a C++ + library, to replace subject pointers for pcre_exec() with a smart pointer + class, thus making it possible to process discontinuous strings. + +16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform + much the same function. They were added by different people who were trying + to make PCRE easy to compile on non-Unix systems. It has been suggested + that PCRE_EXPORT be abolished now that there is more automatic apparatus + for compiling on Windows systems. I have therefore replaced it with + PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it + defaults to "extern" for C or "extern C" for C++, which works fine on + Unix-like systems. It is now possible to override the value of PCRE_DATA_ + SCOPE with something explicit in config.h. In addition: + + (a) pcreposix.h still had just "extern" instead of either of these macros; + I have replaced it with PCRE_DATA_SCOPE. + + (b) Functions such as _pcre_xclass(), which are internal to the library, + but external in the C sense, all had PCRE_EXPORT in their definitions. + This is apparently wrong for the Windows case, so I have removed it. + (It makes no difference on Unix-like systems.) + +17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting + of recursive calls to match(). This is different to MATCH_LIMIT because + that limits the total number of calls to match(), not all of which increase + the depth of recursion. Limiting the recursion depth limits the amount of + stack (or heap if NO_RECURSE is set) that is used. The default can be set + when PCRE is compiled, and changed at run time. A patch from Google adds + this functionality to the C++ interface. + +18. Changes to the handling of Unicode character properties: + + (a) Updated the table to Unicode 4.1.0. + + (b) Recognize characters that are not in the table as "Cn" (undefined). + + (c) I revised the way the table is implemented to a much improved format + which includes recognition of ranges. It now supports the ranges that + are defined in UnicodeData.txt, and it also amalgamates other + characters into ranges. This has reduced the number of entries in the + table from around 16,000 to around 3,000, thus reducing its size + considerably. I realized I did not need to use a tree structure after + all - a binary chop search is just as efficient. Having reduced the + number of entries, I extended their size from 6 bytes to 8 bytes to + allow for more data. + + (d) Added support for Unicode script names via properties such as \p{Han}. + +19. In UTF-8 mode, a backslash followed by a non-Ascii character was not + matching that character. + +20. When matching a repeated Unicode property with a minimum greater than zero, + (for example \pL{2,}), PCRE could look past the end of the subject if it + reached it while seeking the minimum number of characters. This could + happen only if some of the characters were more than one byte long, because + there is a check for at least the minimum number of bytes. + +21. Refactored the implementation of \p and \P so as to be more general, to + allow for more different types of property in future. This has changed the + compiled form incompatibly. Anybody with saved compiled patterns that use + \p or \P will have to recompile them. + +22. Added "Any" and "L&" to the supported property types. + +23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode, + but give a compile time error if the value is greater than 0xff. + +24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were + accidentally not being installed or uninstalled. + +25. The pcre.h file was built from pcre.h.in, but the only changes that were + made were to insert the current release number. This seemed silly, because + it made things harder for people building PCRE on systems that don't run + "configure". I have turned pcre.h into a distributed file, no longer built + by "configure", with the version identification directly included. There is + no longer a pcre.h.in file. + + However, this change necessitated a change to the pcre-config script as + well. It is built from pcre-config.in, and one of the substitutions was the + release number. I have updated configure.ac so that ./configure now finds + the release number by grepping pcre.h. + +26. Added the ability to run the tests under valgrind. + + +Version 6.4 05-Sep-05 +--------------------- + + 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines + "--" to be printed when multiple files were scanned, even when none of the + -A, -B, or -C options were used. This is not compatible with Gnu grep, so I + consider it to be a bug, and have restored the previous behaviour. + + 2. A couple of code tidies to get rid of compiler warnings. + + 3. The pcretest program used to cheat by referring to symbols in the library + whose names begin with _pcre_. These are internal symbols that are not + really supposed to be visible externally, and in some environments it is + possible to suppress them. The cheating is now confined to including + certain files from the library's source, which is a bit cleaner. + + 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the + file's purpose clearer. + + 5. Reorganized pcre_ucp_findchar(). + + +Version 6.3 15-Aug-05 +--------------------- + + 1. The file libpcre.pc.in did not have general read permission in the tarball. + + 2. There were some problems when building without C++ support: + + (a) If C++ support was not built, "make install" and "make test" still + tried to test it. + + (b) There were problems when the value of CXX was explicitly set. Some + changes have been made to try to fix these, and ... + + (c) --disable-cpp can now be used to explicitly disable C++ support. + + (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a + backslash in a target when C++ was disabled. This confuses some + versions of "make", apparently. Using an intermediate variable solves + this. (Same for CPP_LOBJ.) + + 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK) + (non-Windows) now includes $(CFLAGS) because these flags are sometimes + necessary on certain architectures. + + 4. Added a setting of -export-symbols-regex to the link command to remove + those symbols that are exported in the C sense, but actually are local + within the library, and not documented. Their names all begin with + "_pcre_". This is not a perfect job, because (a) we have to except some + symbols that pcretest ("illegally") uses, and (b) the facility isn't always + available (and never for static libraries). I have made a note to try to + find a way round (a) in the future. + + Version 6.2 01-Aug-05 --------------------- diff --git a/src/src/pcre/LICENCE b/src/src/pcre/LICENCE index e8eb0d937..daea2e48a 100644 --- a/src/src/pcre/LICENCE +++ b/src/src/pcre/LICENCE @@ -22,7 +22,7 @@ Email domain: cam.ac.uk University of Cambridge Computing Service, Cambridge, England. Phone: +44 1223 334714. -Copyright (c) 1997-2005 University of Cambridge +Copyright (c) 1997-2006 University of Cambridge All rights reserved. @@ -31,7 +31,7 @@ THE C++ WRAPPER FUNCTIONS Contributed by: Google Inc. -Copyright (c) 2005, Google Inc. +Copyright (c) 2006, Google Inc. All rights reserved. diff --git a/src/src/pcre/Makefile b/src/src/pcre/Makefile index 806aa7ca5..067bfcdfb 100644 --- a/src/src/pcre/Makefile +++ b/src/src/pcre/Makefile @@ -1,4 +1,4 @@ -# $Cambridge: exim/src/src/pcre/Makefile,v 1.5 2005/10/03 09:56:42 ph10 Exp $ +# $Cambridge: exim/src/src/pcre/Makefile,v 1.6 2006/11/07 16:50:36 ph10 Exp $ # Makefile for PCRE (Perl-Compatible Regular Expression) library for use by # Exim. This is a tailored Makefile, not the normal one that comes with the @@ -14,7 +14,7 @@ RANLIB = @true ############################################################################## OBJ = pcre_maketables.o chartables.o pcre_fullinfo.o pcre_get.o \ - pcre_globals.o pcre_compile.o pcre_config.o pcre_exec.o pcre_printint.o \ + pcre_globals.o pcre_compile.o pcre_config.o pcre_exec.o \ pcre_study.o pcre_tables.o pcre_try_flipped.o pcre_version.o all: libpcre.a ../pcretest @@ -61,10 +61,6 @@ pcre_globals.o: pcre_globals.c pcre.h config.h pcre_internal.h Makefile @echo "$(CC) pcre_globals.c" $(FE)$(CC) -c $(CFLAGS) pcre_globals.c -pcre_printint.o: pcre_printint.c pcre.h config.h pcre_internal.h Makefile - @echo "$(CC) pcre_printint.c" - $(FE)$(CC) -c $(CFLAGS) pcre_printint.c - pcre_study.o: pcre_study.c pcre.h config.h pcre_internal.h Makefile @echo "$(CC) pcre_study.c" $(FE)$(CC) -c $(CFLAGS) pcre_study.c diff --git a/src/src/pcre/config.h b/src/src/pcre/config.h index 95be42c0e..715471ee1 100644 --- a/src/src/pcre/config.h +++ b/src/src/pcre/config.h @@ -1,11 +1,11 @@ -/* $Cambridge: exim/src/src/pcre/config.h,v 1.1 2004/10/07 13:04:13 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/config.h,v 1.2 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * config.h for PCRE for Exim * *************************************************/ /* The PCRE sources include config.h, which for a free-standing PCRE build gets -set up by autoconf. For the embedded version in Exim, this file, which is +set up by autoconf. For the embedded version in Exim, this file, which is manually maintained, is used. The only configuration thing that matters for the PCRE library itself is @@ -14,22 +14,27 @@ Standard C libraries, but is missing in SunOS4. PCRE expects autoconf to set HAVE_MEMMOVE to 1 in config.h when memmove() is present. If that is not set, it defines memmove() as a macro for bcopy(). -Exim works differently. It handles this case by defining memmove() as a macro -in its os.h-SunOS4 file. We interface this to PCRE by including the os.h file -here, and then defining HAVE_MEMOVE so that PCRE's code in internal.h leaves +Exim works differently. It handles this case by defining memmove() as a macro +in its os.h-SunOS4 file. We interface this to PCRE by including the os.h file +here, and then defining HAVE_MEMOVE so that PCRE's code in internal.h leaves things alone. */ #include "../os.h" #define HAVE_MEMMOVE 1 -/* We also set up directly a number of parameters that, in the freestanding +/* We also set up directly a number of parameters that, in the freestanding PCRE, can be adjusted by "configure". */ #define NEWLINE '\n' #define LINK_SIZE 2 #define MATCH_LIMIT 10000000 +#define MATCH_LIMIT_RECURSION 10000000 #define POSIX_MALLOC_THRESHOLD 10 +#define MAX_NAME_SIZE 32 +#define MAX_NAME_COUNT 10000 +#define MAX_DUPLENGTH 30000 + /* There is some stuff in the PCRE sources for compilation on non-Unix systems and non-ASCII systems. For Exim's purposes, just flatten it all. */ diff --git a/src/src/pcre/dftables.c b/src/src/pcre/dftables.c index 6380d9321..0cb5dc29c 100644 --- a/src/src/pcre/dftables.c +++ b/src/src/pcre/dftables.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/dftables.c,v 1.3 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/dftables.c,v 1.4 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without diff --git a/src/src/pcre/get.c b/src/src/pcre/get.c deleted file mode 100644 index 8edaebe1a..000000000 --- a/src/src/pcre/get.c +++ /dev/null @@ -1,359 +0,0 @@ -/* $Cambridge: exim/src/src/pcre/get.c,v 1.2 2005/06/15 08:57:10 ph10 Exp $ */ - -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - -/* -This is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. See -the file Tech.Notes for some information on the internals. - -Written by: Philip Hazel <ph10@cam.ac.uk> - - Copyright (c) 1997-2003 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - -/* This module contains some convenience functions for extracting substrings -from the subject string after a regex match has succeeded. The original idea -for these functions came from Scott Wimer. */ - - -/* Include the internals header, which itself includes Standard C headers plus -the external pcre header. */ - -#include "internal.h" - - -/************************************************* -* Find number for named string * -*************************************************/ - -/* This function is used by the two extraction functions below, as well -as being generally available. - -Arguments: - code the compiled regex - stringname the name whose number is required - -Returns: the number of the named parentheses, or a negative number - (PCRE_ERROR_NOSUBSTRING) if not found -*/ - -int -pcre_get_stringnumber(const pcre *code, const char *stringname) -{ -int rc; -int entrysize; -int top, bot; -uschar *nametable; - -if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMECOUNT, &top)) != 0) - return rc; -if (top <= 0) return PCRE_ERROR_NOSUBSTRING; - -if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMEENTRYSIZE, &entrysize)) != 0) - return rc; -if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMETABLE, &nametable)) != 0) - return rc; - -bot = 0; -while (top > bot) - { - int mid = (top + bot) / 2; - uschar *entry = nametable + entrysize*mid; - int c = strcmp(stringname, (char *)(entry + 2)); - if (c == 0) return (entry[0] << 8) + entry[1]; - if (c > 0) bot = mid + 1; else top = mid; - } - -return PCRE_ERROR_NOSUBSTRING; -} - - - -/************************************************* -* Copy captured string to given buffer * -*************************************************/ - -/* This function copies a single captured substring into a given buffer. -Note that we use memcpy() rather than strncpy() in case there are binary zeros -in the string. - -Arguments: - subject the subject string that was matched - ovector pointer to the offsets table - stringcount the number of substrings that were captured - (i.e. the yield of the pcre_exec call, unless - that was zero, in which case it should be 1/3 - of the offset table size) - stringnumber the number of the required substring - buffer where to put the substring - size the size of the buffer - -Returns: if successful: - the length of the copied string, not including the zero - that is put on the end; can be zero - if not successful: - PCRE_ERROR_NOMEMORY (-6) buffer too small - PCRE_ERROR_NOSUBSTRING (-7) no such captured substring -*/ - -int -pcre_copy_substring(const char *subject, int *ovector, int stringcount, - int stringnumber, char *buffer, int size) -{ -int yield; -if (stringnumber < 0 || stringnumber >= stringcount) - return PCRE_ERROR_NOSUBSTRING; -stringnumber *= 2; -yield = ovector[stringnumber+1] - ovector[stringnumber]; -if (size < yield + 1) return PCRE_ERROR_NOMEMORY; -memcpy(buffer, subject + ovector[stringnumber], yield); -buffer[yield] = 0; -return yield; -} - - - -/************************************************* -* Copy named captured string to given buffer * -*************************************************/ - -/* This function copies a single captured substring into a given buffer, -identifying it by name. - -Arguments: - code the compiled regex - subject the subject string that was matched - ovector pointer to the offsets table - stringcount the number of substrings that were captured - (i.e. the yield of the pcre_exec call, unless - that was zero, in which case it should be 1/3 - of the offset table size) - stringname the name of the required substring - buffer where to put the substring - size the size of the buffer - -Returns: if successful: - the length of the copied string, not including the zero - that is put on the end; can be zero - if not successful: - PCRE_ERROR_NOMEMORY (-6) buffer too small - PCRE_ERROR_NOSUBSTRING (-7) no such captured substring -*/ - -int -pcre_copy_named_substring(const pcre *code, const char *subject, int *ovector, - int stringcount, const char *stringname, char *buffer, int size) -{ -int n = pcre_get_stringnumber(code, stringname); -if (n <= 0) return n; -return pcre_copy_substring(subject, ovector, stringcount, n, buffer, size); -} - - - -/************************************************* -* Copy all captured strings to new store * -*************************************************/ - -/* This function gets one chunk of store and builds a list of pointers and all -of the captured substrings in it. A NULL pointer is put on the end of the list. - -Arguments: - subject the subject string that was matched - ovector pointer to the offsets table - stringcount the number of substrings that were captured - (i.e. the yield of the pcre_exec call, unless - that was zero, in which case it should be 1/3 - of the offset table size) - listptr set to point to the list of pointers - -Returns: if successful: 0 - if not successful: - PCRE_ERROR_NOMEMORY (-6) failed to get store -*/ - -int -pcre_get_substring_list(const char *subject, int *ovector, int stringcount, - const char ***listptr) -{ -int i; -int size = sizeof(char *); -int double_count = stringcount * 2; -char **stringlist; -char *p; - -for (i = 0; i < double_count; i += 2) - size += sizeof(char *) + ovector[i+1] - ovector[i] + 1; - -stringlist = (char **)(pcre_malloc)(size); -if (stringlist == NULL) return PCRE_ERROR_NOMEMORY; - -*listptr = (const char **)stringlist; -p = (char *)(stringlist + stringcount + 1); - -for (i = 0; i < double_count; i += 2) - { - int len = ovector[i+1] - ovector[i]; - memcpy(p, subject + ovector[i], len); - *stringlist++ = p; - p += len; - *p++ = 0; - } - -*stringlist = NULL; -return 0; -} - - - -/************************************************* -* Free store obtained by get_substring_list * -*************************************************/ - -/* This function exists for the benefit of people calling PCRE from non-C -programs that can call its functions, but not free() or (pcre_free)() directly. - -Argument: the result of a previous pcre_get_substring_list() -Returns: nothing -*/ - -void -pcre_free_substring_list(const char **pointer) -{ -(pcre_free)((void *)pointer); -} - - - -/************************************************* -* Copy captured string to new store * -*************************************************/ - -/* This function copies a single captured substring into a piece of new -store - -Arguments: - subject the subject string that was matched - ovector pointer to the offsets table - stringcount the number of substrings that were captured - (i.e. the yield of the pcre_exec call, unless - that was zero, in which case it should be 1/3 - of the offset table size) - stringnumber the number of the required substring - stringptr where to put a pointer to the substring - -Returns: if successful: - the length of the string, not including the zero that - is put on the end; can be zero - if not successful: - PCRE_ERROR_NOMEMORY (-6) failed to get store - PCRE_ERROR_NOSUBSTRING (-7) substring not present -*/ - -int -pcre_get_substring(const char *subject, int *ovector, int stringcount, - int stringnumber, const char **stringptr) -{ -int yield; -char *substring; -if (stringnumber < 0 || stringnumber >= stringcount) - return PCRE_ERROR_NOSUBSTRING; -stringnumber *= 2; -yield = ovector[stringnumber+1] - ovector[stringnumber]; -substring = (char *)(pcre_malloc)(yield + 1); -if (substring == NULL) return PCRE_ERROR_NOMEMORY; -memcpy(substring, subject + ovector[stringnumber], yield); -substring[yield] = 0; -*stringptr = substring; -return yield; -} - - - -/************************************************* -* Copy named captured string to new store * -*************************************************/ - -/* This function copies a single captured substring, identified by name, into -new store. - -Arguments: - code the compiled regex - subject the subject string that was matched - ovector pointer to the offsets table - stringcount the number of substrings that were captured - (i.e. the yield of the pcre_exec call, unless - that was zero, in which case it should be 1/3 - of the offset table size) - stringname the name of the required substring - stringptr where to put the pointer - -Returns: if successful: - the length of the copied string, not including the zero - that is put on the end; can be zero - if not successful: - PCRE_ERROR_NOMEMORY (-6) couldn't get memory - PCRE_ERROR_NOSUBSTRING (-7) no such captured substring -*/ - -int -pcre_get_named_substring(const pcre *code, const char *subject, int *ovector, - int stringcount, const char *stringname, const char **stringptr) -{ -int n = pcre_get_stringnumber(code, stringname); -if (n <= 0) return n; -return pcre_get_substring(subject, ovector, stringcount, n, stringptr); -} - - - - -/************************************************* -* Free store obtained by get_substring * -*************************************************/ - -/* This function exists for the benefit of people calling PCRE from non-C -programs that can call its functions, but not free() or (pcre_free)() directly. - -Argument: the result of a previous pcre_get_substring() -Returns: nothing -*/ - -void -pcre_free_substring(const char *pointer) -{ -(pcre_free)((void *)pointer); -} - -/* End of get.c */ diff --git a/src/src/pcre/internal.h b/src/src/pcre/internal.h deleted file mode 100644 index 74c4dd2cd..000000000 --- a/src/src/pcre/internal.h +++ /dev/null @@ -1,754 +0,0 @@ -/* $Cambridge: exim/src/src/pcre/internal.h,v 1.2 2005/06/15 08:57:10 ph10 Exp $ */ - -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - - -/* This is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. See -the file doc/Tech.Notes for some information on the internals. - -Written by: Philip Hazel <ph10@cam.ac.uk> - - Copyright (c) 1997-2004 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - -/* This header contains definitions that are shared between the different -modules, but which are not relevant to the outside. */ - -/* Get the definitions provided by running "configure" */ - -#include "config.h" - -/* Standard C headers plus the external interface definition. The only time -setjmp and stdarg are used is when NO_RECURSE is set. */ - -#include <ctype.h> -#include <limits.h> -#include <setjmp.h> -#include <stdarg.h> -#include <stddef.h> -#include <stdio.h> -#include <stdlib.h> -#include <string.h> - -#ifndef PCRE_SPY -#define PCRE_DEFINITION /* Win32 __declspec(export) trigger for .dll */ -#endif - -/* We need to have types that specify unsigned 16-bit and 32-bit integers. We -cannot determine these outside the compilation (e.g. by running a program as -part of "configure") because PCRE is often cross-compiled for use on other -systems. Instead we make use of the maximum sizes that are available at -preprocessor time in standard C environments. */ - -#if USHRT_MAX == 65535 - typedef unsigned short pcre_uint16; -#elif UINT_MAX == 65535 - typedef unsigned int pcre_uint16; -#else - #error Cannot determine a type for 16-bit unsigned integers -#endif - -#if UINT_MAX == 4294967295 - typedef unsigned int pcre_uint32; -#elif ULONG_MAX == 4294967295 - typedef unsigned long int pcre_uint32; -#else - #error Cannot determine a type for 32-bit unsigned integers -#endif - -/* All character handling must be done as unsigned characters. Otherwise there -are problems with top-bit-set characters and functions such as isspace(). -However, we leave the interface to the outside world as char *, because that -should make things easier for callers. We define a short type for unsigned char -to save lots of typing. I tried "uchar", but it causes problems on Digital -Unix, where it is defined in sys/types, so use "uschar" instead. */ - -typedef unsigned char uschar; - -/* Include the public PCRE header */ - -#include "pcre.h" - -/* When compiling for use with the Virtual Pascal compiler, these functions -need to have their names changed. PCRE must be compiled with the -DVPCOMPAT -option on the command line. */ - -#ifdef VPCOMPAT -#define strncmp(s1,s2,m) _strncmp(s1,s2,m) -#define memcpy(d,s,n) _memcpy(d,s,n) -#define memmove(d,s,n) _memmove(d,s,n) -#define memset(s,c,n) _memset(s,c,n) -#else /* VPCOMPAT */ - -/* To cope with SunOS4 and other systems that lack memmove() but have bcopy(), -define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY -is set. Otherwise, include an emulating function for those systems that have -neither (there some non-Unix environments where this is the case). This assumes -that all calls to memmove are moving strings upwards in store, which is the -case in PCRE. */ - -#if ! HAVE_MEMMOVE -#undef memmove /* some systems may have a macro */ -#if HAVE_BCOPY -#define memmove(a, b, c) bcopy(b, a, c) -#else /* HAVE_BCOPY */ -void * -pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n) -{ -int i; -dest += n; -src += n; -for (i = 0; i < n; ++i) *(--dest) = *(--src); -} -#define memmove(a, b, c) pcre_memmove(a, b, c) -#endif /* not HAVE_BCOPY */ -#endif /* not HAVE_MEMMOVE */ -#endif /* not VPCOMPAT */ - - -/* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored -in big-endian order) by default. These are used, for example, to link from the -start of a subpattern to its alternatives and its end. The use of 2 bytes per -offset limits the size of the compiled regex to around 64K, which is big enough -for almost everybody. However, I received a request for an even bigger limit. -For this reason, and also to make the code easier to maintain, the storing and -loading of offsets from the byte string is now handled by the macros that are -defined here. - -The macros are controlled by the value of LINK_SIZE. This defaults to 2 in -the config.h file, but can be overridden by using -D on the command line. This -is automated on Unix systems via the "configure" command. */ - -#if LINK_SIZE == 2 - -#define PUT(a,n,d) \ - (a[n] = (d) >> 8), \ - (a[(n)+1] = (d) & 255) - -#define GET(a,n) \ - (((a)[n] << 8) | (a)[(n)+1]) - -#define MAX_PATTERN_SIZE (1 << 16) - - -#elif LINK_SIZE == 3 - -#define PUT(a,n,d) \ - (a[n] = (d) >> 16), \ - (a[(n)+1] = (d) >> 8), \ - (a[(n)+2] = (d) & 255) - -#define GET(a,n) \ - (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2]) - -#define MAX_PATTERN_SIZE (1 << 24) - - -#elif LINK_SIZE == 4 - -#define PUT(a,n,d) \ - (a[n] = (d) >> 24), \ - (a[(n)+1] = (d) >> 16), \ - (a[(n)+2] = (d) >> 8), \ - (a[(n)+3] = (d) & 255) - -#define GET(a,n) \ - (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3]) - -#define MAX_PATTERN_SIZE (1 << 30) /* Keep it positive */ - - -#else -#error LINK_SIZE must be either 2, 3, or 4 -#endif - - -/* Convenience macro defined in terms of the others */ - -#define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE - - -/* PCRE uses some other 2-byte quantities that do not change when the size of -offsets changes. There are used for repeat counts and for other things such as -capturing parenthesis numbers in back references. */ - -#define PUT2(a,n,d) \ - a[n] = (d) >> 8; \ - a[(n)+1] = (d) & 255 - -#define GET2(a,n) \ - (((a)[n] << 8) | (a)[(n)+1]) - -#define PUT2INC(a,n,d) PUT2(a,n,d), a += 2 - - -/* In case there is no definition of offsetof() provided - though any proper -Standard C system should have one. */ - -#ifndef offsetof -#define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field)) -#endif - - -/* These are the public options that can change during matching. */ - -#define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL) - -/* Private options flags start at the most significant end of the four bytes, -but skip the top bit so we can use ints for convenience without getting tangled -with negative values. The public options defined in pcre.h start at the least -significant end. Make sure they don't overlap, though now that we have expanded -to four bytes, there is plenty of space. */ - -#define PCRE_FIRSTSET 0x40000000 /* first_byte is set */ -#define PCRE_REQCHSET 0x20000000 /* req_byte is set */ -#define PCRE_STARTLINE 0x10000000 /* start after \n for multiline */ -#define PCRE_ICHANGED 0x08000000 /* i option changes within regex */ -#define PCRE_NOPARTIAL 0x04000000 /* can't use partial with this regex */ - -/* Options for the "extra" block produced by pcre_study(). */ - -#define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */ - -/* Masks for identifying the public options which are permitted at compile -time, run time or study time, respectively. */ - -#define PUBLIC_OPTIONS \ - (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \ - PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \ - PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT) - -#define PUBLIC_EXEC_OPTIONS \ - (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \ - PCRE_PARTIAL) - -#define PUBLIC_STUDY_OPTIONS 0 /* None defined */ - -/* Magic number to provide a small check against being handed junk. */ - -#define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */ - -/* Negative values for the firstchar and reqchar variables */ - -#define REQ_UNSET (-2) -#define REQ_NONE (-1) - -/* Flags added to firstbyte or reqbyte; a "non-literal" item is either a -variable-length repeat, or a anything other than literal characters. */ - -#define REQ_CASELESS 0x0100 /* indicates caselessness */ -#define REQ_VARY 0x0200 /* reqbyte followed non-literal item */ - -/* Miscellaneous definitions */ - -typedef int BOOL; - -#define FALSE 0 -#define TRUE 1 - -/* Escape items that are just an encoding of a particular data value. Note that -ESC_n is defined as yet another macro, which is set in config.h to either \n -(the default) or \r (which some people want). */ - -#ifndef ESC_e -#define ESC_e 27 -#endif - -#ifndef ESC_f -#define ESC_f '\f' -#endif - -#ifndef ESC_n -#define ESC_n NEWLINE -#endif - -#ifndef ESC_r -#define ESC_r '\r' -#endif - -/* We can't officially use ESC_t because it is a POSIX reserved identifier -(presumably because of all the others like size_t). */ - -#ifndef ESC_tee -#define ESC_tee '\t' -#endif - -/* These are escaped items that aren't just an encoding of a particular data -value such as \n. They must have non-zero values, as check_escape() returns -their negation. Also, they must appear in the same order as in the opcode -definitions below, up to ESC_z. There's a dummy for OP_ANY because it -corresponds to "." rather than an escape sequence. The final one must be -ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two -tests in the code for an escape greater than ESC_b and less than ESC_Z to -detect the types that may be repeated. These are the types that consume -characters. If any new escapes are put in between that don't consume a -character, that code will have to change. */ - -enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, - ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E, - ESC_Q, ESC_REF }; - -/* Flag bits and data types for the extended class (OP_XCLASS) for classes that -contain UTF-8 characters with values greater than 255. */ - -#define XCL_NOT 0x01 /* Flag: this is a negative class */ -#define XCL_MAP 0x02 /* Flag: a 32-byte map is present */ - -#define XCL_END 0 /* Marks end of individual items */ -#define XCL_SINGLE 1 /* Single item (one multibyte char) follows */ -#define XCL_RANGE 2 /* A range (two multibyte chars) follows */ -#define XCL_PROP 3 /* Unicode property (one property code) follows */ -#define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */ - - -/* Opcode table: OP_BRA must be last, as all values >= it are used for brackets -that extract substrings. Starting from 1 (i.e. after OP_END), the values up to -OP_EOD must correspond in order to the list of escapes immediately above. -Note that whenever this list is updated, the two macro definitions that follow -must also be updated to match. */ - -enum { - OP_END, /* 0 End of pattern */ - - /* Values corresponding to backslashed metacharacters */ - - OP_SOD, /* 1 Start of data: \A */ - OP_SOM, /* 2 Start of match (subject + offset): \G */ - OP_NOT_WORD_BOUNDARY, /* 3 \B */ - OP_WORD_BOUNDARY, /* 4 \b */ - OP_NOT_DIGIT, /* 5 \D */ - OP_DIGIT, /* 6 \d */ - OP_NOT_WHITESPACE, /* 7 \S */ - OP_WHITESPACE, /* 8 \s */ - OP_NOT_WORDCHAR, /* 9 \W */ - OP_WORDCHAR, /* 10 \w */ - OP_ANY, /* 11 Match any character */ - OP_ANYBYTE, /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */ - OP_NOTPROP, /* 13 \P (not Unicode property) */ - OP_PROP, /* 14 \p (Unicode property) */ - OP_EXTUNI, /* 15 \X (extended Unicode sequence */ - OP_EODN, /* 16 End of data or \n at end of data: \Z. */ - OP_EOD, /* 17 End of data: \z */ - - OP_OPT, /* 18 Set runtime options */ - OP_CIRC, /* 19 Start of line - varies with multiline switch */ - OP_DOLL, /* 20 End of line - varies with multiline switch */ - OP_CHAR, /* 21 Match one character, casefully */ - OP_CHARNC, /* 22 Match one character, caselessly */ - OP_NOT, /* 23 Match anything but the following char */ - - OP_STAR, /* 24 The maximizing and minimizing versions of */ - OP_MINSTAR, /* 25 all these opcodes must come in pairs, with */ - OP_PLUS, /* 26 the minimizing one second. */ - OP_MINPLUS, /* 27 This first set applies to single characters */ - OP_QUERY, /* 28 */ - OP_MINQUERY, /* 29 */ - OP_UPTO, /* 30 From 0 to n matches */ - OP_MINUPTO, /* 31 */ - OP_EXACT, /* 32 Exactly n matches */ - - OP_NOTSTAR, /* 33 The maximizing and minimizing versions of */ - OP_NOTMINSTAR, /* 34 all these opcodes must come in pairs, with */ - OP_NOTPLUS, /* 35 the minimizing one second. */ - OP_NOTMINPLUS, /* 36 This set applies to "not" single characters */ - OP_NOTQUERY, /* 37 */ - OP_NOTMINQUERY, /* 38 */ - OP_NOTUPTO, /* 39 From 0 to n matches */ - OP_NOTMINUPTO, /* 40 */ - OP_NOTEXACT, /* 41 Exactly n matches */ - - OP_TYPESTAR, /* 42 The maximizing and minimizing versions of */ - OP_TYPEMINSTAR, /* 43 all these opcodes must come in pairs, with */ - OP_TYPEPLUS, /* 44 the minimizing one second. These codes must */ - OP_TYPEMINPLUS, /* 45 be in exactly the same order as those above. */ - OP_TYPEQUERY, /* 46 This set applies to character types such as \d */ - OP_TYPEMINQUERY, /* 47 */ - OP_TYPEUPTO, /* 48 From 0 to n matches */ - OP_TYPEMINUPTO, /* 49 */ - OP_TYPEEXACT, /* 50 Exactly n matches */ - - OP_CRSTAR, /* 51 The maximizing and minimizing versions of */ - OP_CRMINSTAR, /* 52 all these opcodes must come in pairs, with */ - OP_CRPLUS, /* 53 the minimizing one second. These codes must */ - OP_CRMINPLUS, /* 54 be in exactly the same order as those above. */ - OP_CRQUERY, /* 55 These are for character classes and back refs */ - OP_CRMINQUERY, /* 56 */ - OP_CRRANGE, /* 57 These are different to the three sets above. */ - OP_CRMINRANGE, /* 58 */ - - OP_CLASS, /* 59 Match a character class, chars < 256 only */ - OP_NCLASS, /* 60 Same, but the bitmap was created from a negative - class - the difference is relevant only when a UTF-8 - character > 255 is encountered. */ - - OP_XCLASS, /* 61 Extended class for handling UTF-8 chars within the - class. This does both positive and negative. */ - - OP_REF, /* 62 Match a back reference */ - OP_RECURSE, /* 63 Match a numbered subpattern (possibly recursive) */ - OP_CALLOUT, /* 64 Call out to external function if provided */ - - OP_ALT, /* 65 Start of alternation */ - OP_KET, /* 66 End of group that doesn't have an unbounded repeat */ - OP_KETRMAX, /* 67 These two must remain together and in this */ - OP_KETRMIN, /* 68 order. They are for groups the repeat for ever. */ - - /* The assertions must come before ONCE and COND */ - - OP_ASSERT, /* 69 Positive lookahead */ - OP_ASSERT_NOT, /* 70 Negative lookahead */ - OP_ASSERTBACK, /* 71 Positive lookbehind */ - OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */ - OP_REVERSE, /* 73 Move pointer back - used in lookbehind assertions */ - - /* ONCE and COND must come after the assertions, with ONCE first, as there's - a test for >= ONCE for a subpattern that isn't an assertion. */ - - OP_ONCE, /* 74 Once matched, don't back up into the subpattern */ - OP_COND, /* 75 Conditional group */ - OP_CREF, /* 76 Used to hold an extraction string number (cond ref) */ - - OP_BRAZERO, /* 77 These two must remain together and in this */ - OP_BRAMINZERO, /* 78 order. */ - - OP_BRANUMBER, /* 79 Used for extracting brackets whose number is greater - than can fit into an opcode. */ - - OP_BRA /* 80 This and greater values are used for brackets that - extract substrings up to EXTRACT_BASIC_MAX. After - that, use is made of OP_BRANUMBER. */ -}; - -/* WARNING WARNING WARNING: There is an implicit assumption in pcre.c and -study.c that all opcodes are less than 128 in value. This makes handling UTF-8 -character sequences easier. */ - -/* The highest extraction number before we have to start using additional -bytes. (Originally PCRE didn't have support for extraction counts highter than -this number.) The value is limited by the number of opcodes left after OP_BRA, -i.e. 255 - OP_BRA. We actually set it a bit lower to leave room for additional -opcodes. */ - -#define EXTRACT_BASIC_MAX 100 - - -/* This macro defines textual names for all the opcodes. There are used only -for debugging, in pcre.c when DEBUG is defined, and also in pcretest.c. The -macro is referenced only in printint.c. */ - -#define OP_NAME_LIST \ - "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d", \ - "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte", \ - "notprop", "prop", "extuni", \ - "\\Z", "\\z", \ - "Opt", "^", "$", "char", "charnc", "not", \ - "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \ - "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \ - "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \ - "*", "*?", "+", "+?", "?", "??", "{", "{", \ - "class", "nclass", "xclass", "Ref", "Recurse", "Callout", \ - "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not", \ - "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\ - "Brazero", "Braminzero", "Branumber", "Bra" - - -/* This macro defines the length of fixed length operations in the compiled -regex. The lengths are used when searching for specific things, and also in the -debugging printing of a compiled regex. We use a macro so that it can be -incorporated both into pcre.c and pcretest.c without being publicly exposed. - -As things have been extended, some of these are no longer fixed lenths, but are -minima instead. For example, the length of a single-character repeat may vary -in UTF-8 mode. The code that uses this table must know about such things. */ - -#define OP_LENGTHS \ - 1, /* End */ \ - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \ - 1, 1, /* Any, Anybyte */ \ - 2, 2, 1, /* NOTPROP, PROP, EXTUNI */ \ - 1, 1, 2, 1, 1, /* \Z, \z, Opt, ^, $ */ \ - 2, /* Char - the minimum length */ \ - 2, /* Charnc - the minimum length */ \ - 2, /* not */ \ - /* Positive single-char repeats ** These are */ \ - 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \ - 4, 4, 4, /* upto, minupto, exact ** UTF-8 mode */ \ - /* Negative single-char repeats - only for chars < 256 */ \ - 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \ - 4, 4, 4, /* NOT upto, minupto, exact */ \ - /* Positive type repeats */ \ - 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \ - 4, 4, 4, /* Type upto, minupto, exact */ \ - /* Character class & ref repeats */ \ - 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \ - 5, 5, /* CRRANGE, CRMINRANGE */ \ - 33, /* CLASS */ \ - 33, /* NCLASS */ \ - 0, /* XCLASS - variable length */ \ - 3, /* REF */ \ - 1+LINK_SIZE, /* RECURSE */ \ - 2+2*LINK_SIZE, /* CALLOUT */ \ - 1+LINK_SIZE, /* Alt */ \ - 1+LINK_SIZE, /* Ket */ \ - 1+LINK_SIZE, /* KetRmax */ \ - 1+LINK_SIZE, /* KetRmin */ \ - 1+LINK_SIZE, /* Assert */ \ - 1+LINK_SIZE, /* Assert not */ \ - 1+LINK_SIZE, /* Assert behind */ \ - 1+LINK_SIZE, /* Assert behind not */ \ - 1+LINK_SIZE, /* Reverse */ \ - 1+LINK_SIZE, /* Once */ \ - 1+LINK_SIZE, /* COND */ \ - 3, /* CREF */ \ - 1, 1, /* BRAZERO, BRAMINZERO */ \ - 3, /* BRANUMBER */ \ - 1+LINK_SIZE /* BRA */ \ - - -/* A magic value for OP_CREF to indicate the "in recursion" condition. */ - -#define CREF_RECURSE 0xffff - -/* The texts of compile-time error messages are defined as macros here so that -they can be accessed by the POSIX wrapper and converted into error codes. Yes, -I could have used error codes in the first place, but didn't feel like changing -just to accommodate the POSIX wrapper. */ - -#define ERR1 "\\ at end of pattern" -#define ERR2 "\\c at end of pattern" -#define ERR3 "unrecognized character follows \\" -#define ERR4 "numbers out of order in {} quantifier" -#define ERR5 "number too big in {} quantifier" -#define ERR6 "missing terminating ] for character class" -#define ERR7 "invalid escape sequence in character class" -#define ERR8 "range out of order in character class" -#define ERR9 "nothing to repeat" -#define ERR10 "operand of unlimited repeat could match the empty string" -#define ERR11 "internal error: unexpected repeat" -#define ERR12 "unrecognized character after (?" -#define ERR13 "POSIX named classes are supported only within a class" -#define ERR14 "missing )" -#define ERR15 "reference to non-existent subpattern" -#define ERR16 "erroffset passed as NULL" -#define ERR17 "unknown option bit(s) set" -#define ERR18 "missing ) after comment" -#define ERR19 "parentheses nested too deeply" -#define ERR20 "regular expression too large" -#define ERR21 "failed to get memory" -#define ERR22 "unmatched parentheses" -#define ERR23 "internal error: code overflow" -#define ERR24 "unrecognized character after (?<" -#define ERR25 "lookbehind assertion is not fixed length" -#define ERR26 "malformed number after (?(" -#define ERR27 "conditional group contains more than two branches" -#define ERR28 "assertion expected after (?(" -#define ERR29 "(?R or (?digits must be followed by )" -#define ERR30 "unknown POSIX class name" -#define ERR31 "POSIX collating elements are not supported" -#define ERR32 "this version of PCRE is not compiled with PCRE_UTF8 support" -#define ERR33 "spare error" -#define ERR34 "character value in \\x{...} sequence is too large" -#define ERR35 "invalid condition (?(0)" -#define ERR36 "\\C not allowed in lookbehind assertion" -#define ERR37 "PCRE does not support \\L, \\l, \\N, \\U, or \\u" -#define ERR38 "number after (?C is > 255" -#define ERR39 "closing ) for (?C expected" -#define ERR40 "recursive call could loop indefinitely" -#define ERR41 "unrecognized character after (?P" -#define ERR42 "syntax error after (?P" -#define ERR43 "two named groups have the same name" -#define ERR44 "invalid UTF-8 string" -#define ERR45 "support for \\P, \\p, and \\X has not been compiled" -#define ERR46 "malformed \\P or \\p sequence" -#define ERR47 "unknown property name after \\P or \\p" - -/* The real format of the start of the pcre block; the index of names and the -code vector run on as long as necessary after the end. We store an explicit -offset to the name table so that if a regex is compiled on one host, saved, and -then run on another where the size of pointers is different, all might still -be well. For the case of compiled-on-4 and run-on-8, we include an extra -pointer that is always NULL. For future-proofing, we also include a few dummy -fields - even though you can never get this planning right! - -NOTE NOTE NOTE: -Because people can now save and re-use compiled patterns, any additions to this -structure should be made at the end, and something earlier (e.g. a new -flag in the options or one of the dummy fields) should indicate that the new -fields are present. Currently PCRE always sets the dummy fields to zero. -NOTE NOTE NOTE: -*/ - -typedef struct real_pcre { - pcre_uint32 magic_number; - pcre_uint32 size; /* Total that was malloced */ - pcre_uint32 options; - pcre_uint32 dummy1; /* For future use, maybe */ - - pcre_uint16 top_bracket; - pcre_uint16 top_backref; - pcre_uint16 first_byte; - pcre_uint16 req_byte; - pcre_uint16 name_table_offset; /* Offset to name table that follows */ - pcre_uint16 name_entry_size; /* Size of any name items */ - pcre_uint16 name_count; /* Number of name items */ - pcre_uint16 dummy2; /* For future use, maybe */ - - const unsigned char *tables; /* Pointer to tables or NULL for std */ - const unsigned char *nullpad; /* NULL padding */ -} real_pcre; - -/* The format of the block used to store data from pcre_study(). The same -remark (see NOTE above) about extending this structure applies. */ - -typedef struct pcre_study_data { - pcre_uint32 size; /* Total that was malloced */ - pcre_uint32 options; - uschar start_bits[32]; -} pcre_study_data; - -/* Structure for passing "static" information around between the functions -doing the compiling, so that they are thread-safe. */ - -typedef struct compile_data { - const uschar *lcc; /* Points to lower casing table */ - const uschar *fcc; /* Points to case-flipping table */ - const uschar *cbits; /* Points to character type table */ - const uschar *ctypes; /* Points to table of type maps */ - const uschar *start_code; /* The start of the compiled code */ - const uschar *start_pattern; /* The start of the pattern */ - uschar *name_table; /* The name/number table */ - int names_found; /* Number of entries so far */ - int name_entry_size; /* Size of each entry */ - int top_backref; /* Maximum back reference */ - unsigned int backref_map; /* Bitmap of low back refs */ - int req_varyopt; /* "After variable item" flag for reqbyte */ - BOOL nopartial; /* Set TRUE if partial won't work */ -} compile_data; - -/* Structure for maintaining a chain of pointers to the currently incomplete -branches, for testing for left recursion. */ - -typedef struct branch_chain { - struct branch_chain *outer; - uschar *current; -} branch_chain; - -/* Structure for items in a linked list that represents an explicit recursive -call within the pattern. */ - -typedef struct recursion_info { - struct recursion_info *prevrec; /* Previous recursion record (or NULL) */ - int group_num; /* Number of group that was called */ - const uschar *after_call; /* "Return value": points after the call in the expr */ - const uschar *save_start; /* Old value of md->start_match */ - int *offset_save; /* Pointer to start of saved offsets */ - int saved_max; /* Number of saved offsets */ -} recursion_info; - -/* When compiling in a mode that doesn't use recursive calls to match(), -a structure is used to remember local variables on the heap. It is defined in -pcre.c, close to the match() function, so that it is easy to keep it in step -with any changes of local variable. However, the pointer to the current frame -must be saved in some "static" place over a longjmp(). We declare the -structure here so that we can put a pointer in the match_data structure. -NOTE: This isn't used for a "normal" compilation of pcre. */ - -struct heapframe; - -/* Structure for passing "static" information around between the functions -doing the matching, so that they are thread-safe. */ - -typedef struct match_data { - unsigned long int match_call_count; /* As it says */ - unsigned long int match_limit;/* As it says */ - int *offset_vector; /* Offset vector */ - int offset_end; /* One past the end */ - int offset_max; /* The maximum usable for return data */ - const uschar *lcc; /* Points to lower casing table */ - const uschar *ctypes; /* Points to table of type maps */ - BOOL offset_overflow; /* Set if too many extractions */ - BOOL notbol; /* NOTBOL flag */ - BOOL noteol; /* NOTEOL flag */ - BOOL utf8; /* UTF8 flag */ - BOOL endonly; /* Dollar not before final \n */ - BOOL notempty; /* Empty string match not wanted */ - BOOL partial; /* PARTIAL flag */ - BOOL hitend; /* Hit the end of the subject at some point */ - const uschar *start_code; /* For use when recursing */ - const uschar *start_subject; /* Start of the subject string */ - const uschar *end_subject; /* End of the subject string */ - const uschar *start_match; /* Start of this match attempt */ - const uschar *end_match_ptr; /* Subject position at end match */ - int end_offset_top; /* Highwater mark at end of match */ - int capture_last; /* Most recent capture number */ - int start_offset; /* The start offset value */ - recursion_info *recursive; /* Linked list of recursion data */ - void *callout_data; /* To pass back to callouts */ - struct heapframe *thisframe; /* Used only when compiling for no recursion */ -} match_data; - -/* Bit definitions for entries in the pcre_ctypes table. */ - -#define ctype_space 0x01 -#define ctype_letter 0x02 -#define ctype_digit 0x04 -#define ctype_xdigit 0x08 -#define ctype_word 0x10 /* alphameric or '_' */ -#define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */ - -/* Offsets for the bitmap tables in pcre_cbits. Each table contains a set -of bits for a class map. Some classes are built by combining these tables. */ - -#define cbit_space 0 /* [:space:] or \s */ -#define cbit_xdigit 32 /* [:xdigit:] */ -#define cbit_digit 64 /* [:digit:] or \d */ -#define cbit_upper 96 /* [:upper:] */ -#define cbit_lower 128 /* [:lower:] */ -#define cbit_word 160 /* [:word:] or \w */ -#define cbit_graph 192 /* [:graph:] */ -#define cbit_print 224 /* [:print:] */ -#define cbit_punct 256 /* [:punct:] */ -#define cbit_cntrl 288 /* [:cntrl:] */ -#define cbit_length 320 /* Length of the cbits table */ - -/* Offsets of the various tables from the base tables pointer, and -total length. */ - -#define lcc_offset 0 -#define fcc_offset 256 -#define cbits_offset 512 -#define ctypes_offset (cbits_offset + cbit_length) -#define tables_length (ctypes_offset + 256) - -/* End of internal.h */ diff --git a/src/src/pcre/maketables.c b/src/src/pcre/maketables.c deleted file mode 100644 index 3ad6e70fb..000000000 --- a/src/src/pcre/maketables.c +++ /dev/null @@ -1,148 +0,0 @@ -/* $Cambridge: exim/src/src/pcre/maketables.c,v 1.2 2005/06/15 08:57:10 ph10 Exp $ */ - -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - -/* -PCRE is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. - -Written by: Philip Hazel <ph10@cam.ac.uk> - - Copyright (c) 1997-2003 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* This file is compiled on its own as part of the PCRE library. However, -it is also included in the compilation of dftables.c, in which case the macro -DFTABLES is defined. */ - -#ifndef DFTABLES -#include "internal.h" -#endif - - - -/************************************************* -* Create PCRE character tables * -*************************************************/ - -/* This function builds a set of character tables for use by PCRE and returns -a pointer to them. They are build using the ctype functions, and consequently -their contents will depend upon the current locale setting. When compiled as -part of the library, the store is obtained via pcre_malloc(), but when compiled -inside dftables, use malloc(). - -Arguments: none -Returns: pointer to the contiguous block of data -*/ - -const unsigned char * -pcre_maketables(void) -{ -unsigned char *yield, *p; -int i; - -#ifndef DFTABLES -yield = (unsigned char*)(pcre_malloc)(tables_length); -#else -yield = (unsigned char*)malloc(tables_length); -#endif - -if (yield == NULL) return NULL; -p = yield; - -/* First comes the lower casing table */ - -for (i = 0; i < 256; i++) *p++ = tolower(i); - -/* Next the case-flipping table */ - -for (i = 0; i < 256; i++) *p++ = islower(i)? toupper(i) : tolower(i); - -/* Then the character class tables. Don't try to be clever and save effort -on exclusive ones - in some locales things may be different. Note that the -table for "space" includes everything "isspace" gives, including VT in the -default locale. This makes it work for the POSIX class [:space:]. */ - -memset(p, 0, cbit_length); -for (i = 0; i < 256; i++) - { - if (isdigit(i)) - { - p[cbit_digit + i/8] |= 1 << (i&7); - p[cbit_word + i/8] |= 1 << (i&7); - } - if (isupper(i)) - { - p[cbit_upper + i/8] |= 1 << (i&7); - p[cbit_word + i/8] |= 1 << (i&7); - } - if (islower(i)) - { - p[cbit_lower + i/8] |= 1 << (i&7); - p[cbit_word + i/8] |= 1 << (i&7); - } - if (i == '_') p[cbit_word + i/8] |= 1 << (i&7); - if (isspace(i)) p[cbit_space + i/8] |= 1 << (i&7); - if (isxdigit(i))p[cbit_xdigit + i/8] |= 1 << (i&7); - if (isgraph(i)) p[cbit_graph + i/8] |= 1 << (i&7); - if (isprint(i)) p[cbit_print + i/8] |= 1 << (i&7); - if (ispunct(i)) p[cbit_punct + i/8] |= 1 << (i&7); - if (iscntrl(i)) p[cbit_cntrl + i/8] |= 1 << (i&7); - } -p += cbit_length; - -/* Finally, the character type table. In this, we exclude VT from the white -space chars, because Perl doesn't recognize it as such for \s and for comments -within regexes. */ - -for (i = 0; i < 256; i++) - { - int x = 0; - if (i != 0x0b && isspace(i)) x += ctype_space; - if (isalpha(i)) x += ctype_letter; - if (isdigit(i)) x += ctype_digit; - if (isxdigit(i)) x += ctype_xdigit; - if (isalnum(i) || i == '_') x += ctype_word; - - /* Note: strchr includes the terminating zero in the characters it considers. - In this instance, that is ok because we want binary zero to be flagged as a - meta-character, which in this sense is any character that terminates a run - of data characters. */ - - if (strchr("*+?{^.$|()[", i) != 0) x += ctype_meta; *p++ = x; } - -return yield; -} - -/* End of maketables.c */ diff --git a/src/src/pcre/pcre.c b/src/src/pcre/pcre.c deleted file mode 100644 index dbe08c23d..000000000 --- a/src/src/pcre/pcre.c +++ /dev/null @@ -1,9196 +0,0 @@ -/* $Cambridge: exim/src/src/pcre/pcre.c,v 1.2 2005/06/15 08:57:10 ph10 Exp $ */ - -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - -/* -This is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. See -the file Tech.Notes for some information on the internals. - -Written by: Philip Hazel <ph10@cam.ac.uk> - - Copyright (c) 1997-2004 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* Define DEBUG to get debugging output on stdout. */ -/* #define DEBUG */ - -/* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef -inline, and there are *still* stupid compilers about that don't like indented -pre-processor statements. I suppose it's only been 10 years... */ - -#ifdef DEBUG -#define DPRINTF(p) printf p -#else -#define DPRINTF(p) /*nothing*/ -#endif - -/* Include the internals header, which itself includes "config.h", the Standard -C headers, and the external pcre header. */ - -#include "internal.h" - -/* If Unicode Property support is wanted, include a private copy of the -function that does it, and the table that translates names to numbers. */ - -#ifdef SUPPORT_UCP -#include "ucp.c" -#include "ucptypetable.c" -#endif - -/* Maximum number of items on the nested bracket stacks at compile time. This -applies to the nesting of all kinds of parentheses. It does not limit -un-nested, non-capturing parentheses. This number can be made bigger if -necessary - it is used to dimension one int and one unsigned char vector at -compile time. */ - -#define BRASTACK_SIZE 200 - - -/* Maximum number of ints of offset to save on the stack for recursive calls. -If the offset vector is bigger, malloc is used. This should be a multiple of 3, -because the offset vector is always a multiple of 3 long. */ - -#define REC_STACK_SAVE_MAX 30 - - -/* The maximum remaining length of subject we are prepared to search for a -req_byte match. */ - -#define REQ_BYTE_MAX 1000 - - -/* Table of sizes for the fixed-length opcodes. It's defined in a macro so that -the definition is next to the definition of the opcodes in internal.h. */ - -static const uschar OP_lengths[] = { OP_LENGTHS }; - -/* Min and max values for the common repeats; for the maxima, 0 => infinity */ - -static const char rep_min[] = { 0, 0, 1, 1, 0, 0 }; -static const char rep_max[] = { 0, 0, 0, 0, 1, 1 }; - -/* Table for handling escaped characters in the range '0'-'z'. Positive returns -are simple data values; negative values are for special things like \d and so -on. Zero means further processing is needed (for things like \x), or the escape -is invalid. */ - -#if !EBCDIC /* This is the "normal" table for ASCII systems */ -static const short int escapes[] = { - 0, 0, 0, 0, 0, 0, 0, 0, /* 0 - 7 */ - 0, 0, ':', ';', '<', '=', '>', '?', /* 8 - ? */ - '@', -ESC_A, -ESC_B, -ESC_C, -ESC_D, -ESC_E, 0, -ESC_G, /* @ - G */ - 0, 0, 0, 0, 0, 0, 0, 0, /* H - O */ --ESC_P, -ESC_Q, 0, -ESC_S, 0, 0, 0, -ESC_W, /* P - W */ --ESC_X, 0, -ESC_Z, '[', '\\', ']', '^', '_', /* X - _ */ - '`', 7, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0, /* ` - g */ - 0, 0, 0, 0, 0, 0, ESC_n, 0, /* h - o */ --ESC_p, 0, ESC_r, -ESC_s, ESC_tee, 0, 0, -ESC_w, /* p - w */ - 0, 0, -ESC_z /* x - z */ -}; - -#else /* This is the "abnormal" table for EBCDIC systems */ -static const short int escapes[] = { -/* 48 */ 0, 0, 0, '.', '<', '(', '+', '|', -/* 50 */ '&', 0, 0, 0, 0, 0, 0, 0, -/* 58 */ 0, 0, '!', '$', '*', ')', ';', '~', -/* 60 */ '-', '/', 0, 0, 0, 0, 0, 0, -/* 68 */ 0, 0, '|', ',', '%', '_', '>', '?', -/* 70 */ 0, 0, 0, 0, 0, 0, 0, 0, -/* 78 */ 0, '`', ':', '#', '@', '\'', '=', '"', -/* 80 */ 0, 7, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0, -/* 88 */ 0, 0, 0, '{', 0, 0, 0, 0, -/* 90 */ 0, 0, 0, 'l', 0, ESC_n, 0, -ESC_p, -/* 98 */ 0, ESC_r, 0, '}', 0, 0, 0, 0, -/* A0 */ 0, '~', -ESC_s, ESC_tee, 0, 0, -ESC_w, 0, -/* A8 */ 0,-ESC_z, 0, 0, 0, '[', 0, 0, -/* B0 */ 0, 0, 0, 0, 0, 0, 0, 0, -/* B8 */ 0, 0, 0, 0, 0, ']', '=', '-', -/* C0 */ '{',-ESC_A, -ESC_B, -ESC_C, -ESC_D,-ESC_E, 0, -ESC_G, -/* C8 */ 0, 0, 0, 0, 0, 0, 0, 0, -/* D0 */ '}', 0, 0, 0, 0, 0, 0, -ESC_P, -/* D8 */-ESC_Q, 0, 0, 0, 0, 0, 0, 0, -/* E0 */ '\\', 0, -ESC_S, 0, 0, 0, -ESC_W, -ESC_X, -/* E8 */ 0,-ESC_Z, 0, 0, 0, 0, 0, 0, -/* F0 */ 0, 0, 0, 0, 0, 0, 0, 0, -/* F8 */ 0, 0, 0, 0, 0, 0, 0, 0 -}; -#endif - - -/* Tables of names of POSIX character classes and their lengths. The list is -terminated by a zero length entry. The first three must be alpha, upper, lower, -as this is assumed for handling case independence. */ - -static const char *const posix_names[] = { - "alpha", "lower", "upper", - "alnum", "ascii", "blank", "cntrl", "digit", "graph", - "print", "punct", "space", "word", "xdigit" }; - -static const uschar posix_name_lengths[] = { - 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 }; - -/* Table of class bit maps for each POSIX class; up to three may be combined -to form the class. The table for [:blank:] is dynamically modified to remove -the vertical space characters. */ - -static const int posix_class_maps[] = { - cbit_lower, cbit_upper, -1, /* alpha */ - cbit_lower, -1, -1, /* lower */ - cbit_upper, -1, -1, /* upper */ - cbit_digit, cbit_lower, cbit_upper, /* alnum */ - cbit_print, cbit_cntrl, -1, /* ascii */ - cbit_space, -1, -1, /* blank - a GNU extension */ - cbit_cntrl, -1, -1, /* cntrl */ - cbit_digit, -1, -1, /* digit */ - cbit_graph, -1, -1, /* graph */ - cbit_print, -1, -1, /* print */ - cbit_punct, -1, -1, /* punct */ - cbit_space, -1, -1, /* space */ - cbit_word, -1, -1, /* word - a Perl extension */ - cbit_xdigit,-1, -1 /* xdigit */ -}; - -/* Table to identify digits and hex digits. This is used when compiling -patterns. Note that the tables in chartables are dependent on the locale, and -may mark arbitrary characters as digits - but the PCRE compiling code expects -to handle only 0-9, a-z, and A-Z as digits when compiling. That is why we have -a private table here. It costs 256 bytes, but it is a lot faster than doing -character value tests (at least in some simple cases I timed), and in some -applications one wants PCRE to compile efficiently as well as match -efficiently. - -For convenience, we use the same bit definitions as in chartables: - - 0x04 decimal digit - 0x08 hexadecimal digit - -Then we can use ctype_digit and ctype_xdigit in the code. */ - -#if !EBCDIC /* This is the "normal" case, for ASCII systems */ -static const unsigned char digitab[] = - { - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - ' */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ( - / */ - 0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 */ - 0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00, /* 8 - ? */ - 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* @ - G */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H - O */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* P - W */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* X - _ */ - 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* ` - g */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h - o */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* p - w */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* x -127 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */ - -#else /* This is the "abnormal" case, for EBCDIC systems */ -static const unsigned char digitab[] = - { - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 10 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 32- 39 20 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 30 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 40 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 72- | */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 50 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 88- ¬ */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 60 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 104- ? */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 70 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */ - 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* 128- g 80 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144- p 90 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160- x A0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 B0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ - 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* { - G C0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* } - P D0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* \ - X E0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */ - 0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 F0 */ - 0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */ - -static const unsigned char ebcdic_chartab[] = { /* chartable partial dup */ - 0x80,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 0- 7 */ - 0x00,0x00,0x00,0x00,0x01,0x01,0x00,0x00, /* 8- 15 */ - 0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 16- 23 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ - 0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 32- 39 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */ - 0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 */ - 0x00,0x00,0x00,0x80,0x00,0x80,0x80,0x80, /* 72- | */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 */ - 0x00,0x00,0x00,0x80,0x80,0x80,0x00,0x00, /* 88- ¬ */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 */ - 0x00,0x00,0x00,0x00,0x00,0x10,0x00,0x80, /* 104- ? */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */ - 0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* 128- g */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */ - 0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* 144- p */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */ - 0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* 160- x */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */ - 0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 */ - 0x00,0x00,0x80,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ - 0x80,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* { - G */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */ - 0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* } - P */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */ - 0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* \ - X */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */ - 0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /* 0 - 7 */ - 0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */ -#endif - - -/* Definition to allow mutual recursion */ - -static BOOL - compile_regex(int, int, int *, uschar **, const uschar **, const char **, - BOOL, int, int *, int *, branch_chain *, compile_data *); - -/* Structure for building a chain of data that actually lives on the -stack, for holding the values of the subject pointer at the start of each -subpattern, so as to detect when an empty string has been matched by a -subpattern - to break infinite loops. When NO_RECURSE is set, these blocks -are on the heap, not on the stack. */ - -typedef struct eptrblock { - struct eptrblock *epb_prev; - const uschar *epb_saved_eptr; -} eptrblock; - -/* Flag bits for the match() function */ - -#define match_condassert 0x01 /* Called to check a condition assertion */ -#define match_isgroup 0x02 /* Set if start of bracketed group */ - -/* Non-error returns from the match() function. Error returns are externally -defined PCRE_ERROR_xxx codes, which are all negative. */ - -#define MATCH_MATCH 1 -#define MATCH_NOMATCH 0 - - - -/************************************************* -* Global variables * -*************************************************/ - -/* PCRE is thread-clean and doesn't use any global variables in the normal -sense. However, it calls memory allocation and free functions via the four -indirections below, and it can optionally do callouts. These values can be -changed by the caller, but are shared between all threads. However, when -compiling for Virtual Pascal, things are done differently (see pcre.in). */ - -#ifndef VPCOMPAT -#ifdef __cplusplus -extern "C" void *(*pcre_malloc)(size_t) = malloc; -extern "C" void (*pcre_free)(void *) = free; -extern "C" void *(*pcre_stack_malloc)(size_t) = malloc; -extern "C" void (*pcre_stack_free)(void *) = free; -extern "C" int (*pcre_callout)(pcre_callout_block *) = NULL; -#else -void *(*pcre_malloc)(size_t) = malloc; -void (*pcre_free)(void *) = free; -void *(*pcre_stack_malloc)(size_t) = malloc; -void (*pcre_stack_free)(void *) = free; -int (*pcre_callout)(pcre_callout_block *) = NULL; -#endif -#endif - - -/************************************************* -* Macros and tables for character handling * -*************************************************/ - -/* When UTF-8 encoding is being used, a character is no longer just a single -byte. The macros for character handling generate simple sequences when used in -byte-mode, and more complicated ones for UTF-8 characters. */ - -#ifndef SUPPORT_UTF8 -#define GETCHAR(c, eptr) c = *eptr; -#define GETCHARINC(c, eptr) c = *eptr++; -#define GETCHARINCTEST(c, eptr) c = *eptr++; -#define GETCHARLEN(c, eptr, len) c = *eptr; -#define BACKCHAR(eptr) - -#else /* SUPPORT_UTF8 */ - -/* Get the next UTF-8 character, not advancing the pointer. This is called when -we know we are in UTF-8 mode. */ - -#define GETCHAR(c, eptr) \ - c = *eptr; \ - if ((c & 0xc0) == 0xc0) \ - { \ - int gcii; \ - int gcaa = utf8_table4[c & 0x3f]; /* Number of additional bytes */ \ - int gcss = 6*gcaa; \ - c = (c & utf8_table3[gcaa]) << gcss; \ - for (gcii = 1; gcii <= gcaa; gcii++) \ - { \ - gcss -= 6; \ - c |= (eptr[gcii] & 0x3f) << gcss; \ - } \ - } - -/* Get the next UTF-8 character, advancing the pointer. This is called when we -know we are in UTF-8 mode. */ - -#define GETCHARINC(c, eptr) \ - c = *eptr++; \ - if ((c & 0xc0) == 0xc0) \ - { \ - int gcaa = utf8_table4[c & 0x3f]; /* Number of additional bytes */ \ - int gcss = 6*gcaa; \ - c = (c & utf8_table3[gcaa]) << gcss; \ - while (gcaa-- > 0) \ - { \ - gcss -= 6; \ - c |= (*eptr++ & 0x3f) << gcss; \ - } \ - } - -/* Get the next character, testing for UTF-8 mode, and advancing the pointer */ - -#define GETCHARINCTEST(c, eptr) \ - c = *eptr++; \ - if (md->utf8 && (c & 0xc0) == 0xc0) \ - { \ - int gcaa = utf8_table4[c & 0x3f]; /* Number of additional bytes */ \ - int gcss = 6*gcaa; \ - c = (c & utf8_table3[gcaa]) << gcss; \ - while (gcaa-- > 0) \ - { \ - gcss -= 6; \ - c |= (*eptr++ & 0x3f) << gcss; \ - } \ - } - -/* Get the next UTF-8 character, not advancing the pointer, incrementing length -if there are extra bytes. This is called when we know we are in UTF-8 mode. */ - -#define GETCHARLEN(c, eptr, len) \ - c = *eptr; \ - if ((c & 0xc0) == 0xc0) \ - { \ - int gcii; \ - int gcaa = utf8_table4[c & 0x3f]; /* Number of additional bytes */ \ - int gcss = 6*gcaa; \ - c = (c & utf8_table3[gcaa]) << gcss; \ - for (gcii = 1; gcii <= gcaa; gcii++) \ - { \ - gcss -= 6; \ - c |= (eptr[gcii] & 0x3f) << gcss; \ - } \ - len += gcaa; \ - } - -/* If the pointer is not at the start of a character, move it back until -it is. Called only in UTF-8 mode. */ - -#define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--; - -#endif - - - -/************************************************* -* Default character tables * -*************************************************/ - -/* A default set of character tables is included in the PCRE binary. Its source -is built by the maketables auxiliary program, which uses the default C ctypes -functions, and put in the file chartables.c. These tables are used by PCRE -whenever the caller of pcre_compile() does not provide an alternate set of -tables. */ - -#include "chartables.c" - - - -#ifdef SUPPORT_UTF8 -/************************************************* -* Tables for UTF-8 support * -*************************************************/ - -/* These are the breakpoints for different numbers of bytes in a UTF-8 -character. */ - -static const int utf8_table1[] = - { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff}; - -/* These are the indicator bits and the mask for the data bits to set in the -first byte of a character, indexed by the number of additional bytes. */ - -static const int utf8_table2[] = { 0, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc}; -static const int utf8_table3[] = { 0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01}; - -/* Table of the number of extra characters, indexed by the first character -masked with 0x3f. The highest number for a valid UTF-8 character is in fact -0x3d. */ - -static const uschar utf8_table4[] = { - 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, - 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, - 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, - 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5 }; - - -/************************************************* -* Convert character value to UTF-8 * -*************************************************/ - -/* This function takes an integer value in the range 0 - 0x7fffffff -and encodes it as a UTF-8 character in 0 to 6 bytes. - -Arguments: - cvalue the character value - buffer pointer to buffer for result - at least 6 bytes long - -Returns: number of characters placed in the buffer -*/ - -static int -ord2utf8(int cvalue, uschar *buffer) -{ -register int i, j; -for (i = 0; i < sizeof(utf8_table1)/sizeof(int); i++) - if (cvalue <= utf8_table1[i]) break; -buffer += i; -for (j = i; j > 0; j--) - { - *buffer-- = 0x80 | (cvalue & 0x3f); - cvalue >>= 6; - } -*buffer = utf8_table2[i] | cvalue; -return i + 1; -} -#endif - - - -/************************************************* -* Print compiled regex * -*************************************************/ - -/* The code for doing this is held in a separate file that is also included in -pcretest.c. It defines a function called print_internals(). */ - -#ifdef DEBUG -#include "printint.c" -#endif - - - -/************************************************* -* Return version string * -*************************************************/ - -#define STRING(a) # a -#define XSTRING(s) STRING(s) - -EXPORT const char * -pcre_version(void) -{ -return XSTRING(PCRE_MAJOR) "." XSTRING(PCRE_MINOR) " " XSTRING(PCRE_DATE); -} - - - - -/************************************************* -* Flip bytes in an integer * -*************************************************/ - -/* This function is called when the magic number in a regex doesn't match in -order to flip its bytes to see if we are dealing with a pattern that was -compiled on a host of different endianness. If so, this function is used to -flip other byte values. - -Arguments: - value the number to flip - n the number of bytes to flip (assumed to be 2 or 4) - -Returns: the flipped value -*/ - -static long int -byteflip(long int value, int n) -{ -if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8); -return ((value & 0x000000ff) << 24) | - ((value & 0x0000ff00) << 8) | - ((value & 0x00ff0000) >> 8) | - ((value & 0xff000000) >> 24); -} - - - -/************************************************* -* Test for a byte-flipped compiled regex * -*************************************************/ - -/* This function is called from pce_exec() and also from pcre_fullinfo(). Its -job is to test whether the regex is byte-flipped - that is, it was compiled on -a system of opposite endianness. The function is called only when the native -MAGIC_NUMBER test fails. If the regex is indeed flipped, we flip all the -relevant values into a different data block, and return it. - -Arguments: - re points to the regex - study points to study data, or NULL - internal_re points to a new regex block - internal_study points to a new study block - -Returns: the new block if is is indeed a byte-flipped regex - NULL if it is not -*/ - -static real_pcre * -try_flipped(const real_pcre *re, real_pcre *internal_re, - const pcre_study_data *study, pcre_study_data *internal_study) -{ -if (byteflip(re->magic_number, sizeof(re->magic_number)) != MAGIC_NUMBER) - return NULL; - -*internal_re = *re; /* To copy other fields */ -internal_re->size = byteflip(re->size, sizeof(re->size)); -internal_re->options = byteflip(re->options, sizeof(re->options)); -internal_re->top_bracket = byteflip(re->top_bracket, sizeof(re->top_bracket)); -internal_re->top_backref = byteflip(re->top_backref, sizeof(re->top_backref)); -internal_re->first_byte = byteflip(re->first_byte, sizeof(re->first_byte)); -internal_re->req_byte = byteflip(re->req_byte, sizeof(re->req_byte)); -internal_re->name_table_offset = byteflip(re->name_table_offset, - sizeof(re->name_table_offset)); -internal_re->name_entry_size = byteflip(re->name_entry_size, - sizeof(re->name_entry_size)); -internal_re->name_count = byteflip(re->name_count, sizeof(re->name_count)); - -if (study != NULL) - { - *internal_study = *study; /* To copy other fields */ - internal_study->size = byteflip(study->size, sizeof(study->size)); - internal_study->options = byteflip(study->options, sizeof(study->options)); - } - -return internal_re; -} - - - -/************************************************* -* (Obsolete) Return info about compiled pattern * -*************************************************/ - -/* This is the original "info" function. It picks potentially useful data out -of the private structure, but its interface was too rigid. It remains for -backwards compatibility. The public options are passed back in an int - though -the re->options field has been expanded to a long int, all the public options -at the low end of it, and so even on 16-bit systems this will still be OK. -Therefore, I haven't changed the API for pcre_info(). - -Arguments: - argument_re points to compiled code - optptr where to pass back the options - first_byte where to pass back the first character, - or -1 if multiline and all branches start ^, - or -2 otherwise - -Returns: number of capturing subpatterns - or negative values on error -*/ - -EXPORT int -pcre_info(const pcre *argument_re, int *optptr, int *first_byte) -{ -real_pcre internal_re; -const real_pcre *re = (const real_pcre *)argument_re; -if (re == NULL) return PCRE_ERROR_NULL; -if (re->magic_number != MAGIC_NUMBER) - { - re = try_flipped(re, &internal_re, NULL, NULL); - if (re == NULL) return PCRE_ERROR_BADMAGIC; - } -if (optptr != NULL) *optptr = (int)(re->options & PUBLIC_OPTIONS); -if (first_byte != NULL) - *first_byte = ((re->options & PCRE_FIRSTSET) != 0)? re->first_byte : - ((re->options & PCRE_STARTLINE) != 0)? -1 : -2; -return re->top_bracket; -} - - - -/************************************************* -* Return info about compiled pattern * -*************************************************/ - -/* This is a newer "info" function which has an extensible interface so -that additional items can be added compatibly. - -Arguments: - argument_re points to compiled code - extra_data points extra data, or NULL - what what information is required - where where to put the information - -Returns: 0 if data returned, negative on error -*/ - -EXPORT int -pcre_fullinfo(const pcre *argument_re, const pcre_extra *extra_data, int what, - void *where) -{ -real_pcre internal_re; -pcre_study_data internal_study; -const real_pcre *re = (const real_pcre *)argument_re; -const pcre_study_data *study = NULL; - -if (re == NULL || where == NULL) return PCRE_ERROR_NULL; - -if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_STUDY_DATA) != 0) - study = (const pcre_study_data *)extra_data->study_data; - -if (re->magic_number != MAGIC_NUMBER) - { - re = try_flipped(re, &internal_re, study, &internal_study); - if (re == NULL) return PCRE_ERROR_BADMAGIC; - if (study != NULL) study = &internal_study; - } - -switch (what) - { - case PCRE_INFO_OPTIONS: - *((unsigned long int *)where) = re->options & PUBLIC_OPTIONS; - break; - - case PCRE_INFO_SIZE: - *((size_t *)where) = re->size; - break; - - case PCRE_INFO_STUDYSIZE: - *((size_t *)where) = (study == NULL)? 0 : study->size; - break; - - case PCRE_INFO_CAPTURECOUNT: - *((int *)where) = re->top_bracket; - break; - - case PCRE_INFO_BACKREFMAX: - *((int *)where) = re->top_backref; - break; - - case PCRE_INFO_FIRSTBYTE: - *((int *)where) = - ((re->options & PCRE_FIRSTSET) != 0)? re->first_byte : - ((re->options & PCRE_STARTLINE) != 0)? -1 : -2; - break; - - /* Make sure we pass back the pointer to the bit vector in the external - block, not the internal copy (with flipped integer fields). */ - - case PCRE_INFO_FIRSTTABLE: - *((const uschar **)where) = - (study != NULL && (study->options & PCRE_STUDY_MAPPED) != 0)? - ((const pcre_study_data *)extra_data->study_data)->start_bits : NULL; - break; - - case PCRE_INFO_LASTLITERAL: - *((int *)where) = - ((re->options & PCRE_REQCHSET) != 0)? re->req_byte : -1; - break; - - case PCRE_INFO_NAMEENTRYSIZE: - *((int *)where) = re->name_entry_size; - break; - - case PCRE_INFO_NAMECOUNT: - *((int *)where) = re->name_count; - break; - - case PCRE_INFO_NAMETABLE: - *((const uschar **)where) = (const uschar *)re + re->name_table_offset; - break; - - case PCRE_INFO_DEFAULT_TABLES: - *((const uschar **)where) = (const uschar *)pcre_default_tables; - break; - - default: return PCRE_ERROR_BADOPTION; - } - -return 0; -} - - - -/************************************************* -* Return info about what features are configured * -*************************************************/ - -/* This is function which has an extensible interface so that additional items -can be added compatibly. - -Arguments: - what what information is required - where where to put the information - -Returns: 0 if data returned, negative on error -*/ - -EXPORT int -pcre_config(int what, void *where) -{ -switch (what) - { - case PCRE_CONFIG_UTF8: -#ifdef SUPPORT_UTF8 - *((int *)where) = 1; -#else - *((int *)where) = 0; -#endif - break; - - case PCRE_CONFIG_UNICODE_PROPERTIES: -#ifdef SUPPORT_UCP - *((int *)where) = 1; -#else - *((int *)where) = 0; -#endif - break; - - case PCRE_CONFIG_NEWLINE: - *((int *)where) = NEWLINE; - break; - - case PCRE_CONFIG_LINK_SIZE: - *((int *)where) = LINK_SIZE; - break; - - case PCRE_CONFIG_POSIX_MALLOC_THRESHOLD: - *((int *)where) = POSIX_MALLOC_THRESHOLD; - break; - - case PCRE_CONFIG_MATCH_LIMIT: - *((unsigned int *)where) = MATCH_LIMIT; - break; - - case PCRE_CONFIG_STACKRECURSE: -#ifdef NO_RECURSE - *((int *)where) = 0; -#else - *((int *)where) = 1; -#endif - break; - - default: return PCRE_ERROR_BADOPTION; - } - -return 0; -} - - - -#ifdef DEBUG -/************************************************* -* Debugging function to print chars * -*************************************************/ - -/* Print a sequence of chars in printable format, stopping at the end of the -subject if the requested. - -Arguments: - p points to characters - length number to print - is_subject TRUE if printing from within md->start_subject - md pointer to matching data block, if is_subject is TRUE - -Returns: nothing -*/ - -static void -pchars(const uschar *p, int length, BOOL is_subject, match_data *md) -{ -int c; -if (is_subject && length > md->end_subject - p) length = md->end_subject - p; -while (length-- > 0) - if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c); -} -#endif - - - - -/************************************************* -* Handle escapes * -*************************************************/ - -/* This function is called when a \ has been encountered. It either returns a -positive value for a simple escape such as \n, or a negative value which -encodes one of the more complicated things such as \d. When UTF-8 is enabled, -a positive value greater than 255 may be returned. On entry, ptr is pointing at -the \. On exit, it is on the final character of the escape sequence. - -Arguments: - ptrptr points to the pattern position pointer - errorptr points to the pointer to the error message - bracount number of previous extracting brackets - options the options bits - isclass TRUE if inside a character class - -Returns: zero or positive => a data character - negative => a special escape sequence - on error, errorptr is set -*/ - -static int -check_escape(const uschar **ptrptr, const char **errorptr, int bracount, - int options, BOOL isclass) -{ -const uschar *ptr = *ptrptr; -int c, i; - -/* If backslash is at the end of the pattern, it's an error. */ - -c = *(++ptr); -if (c == 0) *errorptr = ERR1; - -/* Non-alphamerics are literals. For digits or letters, do an initial lookup in -a table. A non-zero result is something that can be returned immediately. -Otherwise further processing may be required. */ - -#if !EBCDIC /* ASCII coding */ -else if (c < '0' || c > 'z') {} /* Not alphameric */ -else if ((i = escapes[c - '0']) != 0) c = i; - -#else /* EBCDIC coding */ -else if (c < 'a' || (ebcdic_chartab[c] & 0x0E) == 0) {} /* Not alphameric */ -else if ((i = escapes[c - 0x48]) != 0) c = i; -#endif - -/* Escapes that need further processing, or are illegal. */ - -else - { - const uschar *oldptr; - switch (c) - { - /* A number of Perl escapes are not handled by PCRE. We give an explicit - error. */ - - case 'l': - case 'L': - case 'N': - case 'u': - case 'U': - *errorptr = ERR37; - break; - - /* The handling of escape sequences consisting of a string of digits - starting with one that is not zero is not straightforward. By experiment, - the way Perl works seems to be as follows: - - Outside a character class, the digits are read as a decimal number. If the - number is less than 10, or if there are that many previous extracting - left brackets, then it is a back reference. Otherwise, up to three octal - digits are read to form an escaped byte. Thus \123 is likely to be octal - 123 (cf \0123, which is octal 012 followed by the literal 3). If the octal - value is greater than 377, the least significant 8 bits are taken. Inside a - character class, \ followed by a digit is always an octal number. */ - - case '1': case '2': case '3': case '4': case '5': - case '6': case '7': case '8': case '9': - - if (!isclass) - { - oldptr = ptr; - c -= '0'; - while ((digitab[ptr[1]] & ctype_digit) != 0) - c = c * 10 + *(++ptr) - '0'; - if (c < 10 || c <= bracount) - { - c = -(ESC_REF + c); - break; - } - ptr = oldptr; /* Put the pointer back and fall through */ - } - - /* Handle an octal number following \. If the first digit is 8 or 9, Perl - generates a binary zero byte and treats the digit as a following literal. - Thus we have to pull back the pointer by one. */ - - if ((c = *ptr) >= '8') - { - ptr--; - c = 0; - break; - } - - /* \0 always starts an octal number, but we may drop through to here with a - larger first octal digit. */ - - case '0': - c -= '0'; - while(i++ < 2 && ptr[1] >= '0' && ptr[1] <= '7') - c = c * 8 + *(++ptr) - '0'; - c &= 255; /* Take least significant 8 bits */ - break; - - /* \x is complicated when UTF-8 is enabled. \x{ddd} is a character number - which can be greater than 0xff, but only if the ddd are hex digits. */ - - case 'x': -#ifdef SUPPORT_UTF8 - if (ptr[1] == '{' && (options & PCRE_UTF8) != 0) - { - const uschar *pt = ptr + 2; - register int count = 0; - c = 0; - while ((digitab[*pt] & ctype_xdigit) != 0) - { - int cc = *pt++; - count++; -#if !EBCDIC /* ASCII coding */ - if (cc >= 'a') cc -= 32; /* Convert to upper case */ - c = c * 16 + cc - ((cc < 'A')? '0' : ('A' - 10)); -#else /* EBCDIC coding */ - if (cc >= 'a' && cc <= 'z') cc += 64; /* Convert to upper case */ - c = c * 16 + cc - ((cc >= '0')? '0' : ('A' - 10)); -#endif - } - if (*pt == '}') - { - if (c < 0 || count > 8) *errorptr = ERR34; - ptr = pt; - break; - } - /* If the sequence of hex digits does not end with '}', then we don't - recognize this construct; fall through to the normal \x handling. */ - } -#endif - - /* Read just a single hex char */ - - c = 0; - while (i++ < 2 && (digitab[ptr[1]] & ctype_xdigit) != 0) - { - int cc; /* Some compilers don't like ++ */ - cc = *(++ptr); /* in initializers */ -#if !EBCDIC /* ASCII coding */ - if (cc >= 'a') cc -= 32; /* Convert to upper case */ - c = c * 16 + cc - ((cc < 'A')? '0' : ('A' - 10)); -#else /* EBCDIC coding */ - if (cc <= 'z') cc += 64; /* Convert to upper case */ - c = c * 16 + cc - ((cc >= '0')? '0' : ('A' - 10)); -#endif - } - break; - - /* Other special escapes not starting with a digit are straightforward */ - - case 'c': - c = *(++ptr); - if (c == 0) - { - *errorptr = ERR2; - return 0; - } - - /* A letter is upper-cased; then the 0x40 bit is flipped. This coding - is ASCII-specific, but then the whole concept of \cx is ASCII-specific. - (However, an EBCDIC equivalent has now been added.) */ - -#if !EBCDIC /* ASCII coding */ - if (c >= 'a' && c <= 'z') c -= 32; - c ^= 0x40; -#else /* EBCDIC coding */ - if (c >= 'a' && c <= 'z') c += 64; - c ^= 0xC0; -#endif - break; - - /* PCRE_EXTRA enables extensions to Perl in the matter of escapes. Any - other alphameric following \ is an error if PCRE_EXTRA was set; otherwise, - for Perl compatibility, it is a literal. This code looks a bit odd, but - there used to be some cases other than the default, and there may be again - in future, so I haven't "optimized" it. */ - - default: - if ((options & PCRE_EXTRA) != 0) switch(c) - { - default: - *errorptr = ERR3; - break; - } - break; - } - } - -*ptrptr = ptr; -return c; -} - - - -#ifdef SUPPORT_UCP -/************************************************* -* Handle \P and \p * -*************************************************/ - -/* This function is called after \P or \p has been encountered, provided that -PCRE is compiled with support for Unicode properties. On entry, ptrptr is -pointing at the P or p. On exit, it is pointing at the final character of the -escape sequence. - -Argument: - ptrptr points to the pattern position pointer - negptr points to a boolean that is set TRUE for negation else FALSE - errorptr points to the pointer to the error message - -Returns: value from ucp_type_table, or -1 for an invalid type -*/ - -static int -get_ucp(const uschar **ptrptr, BOOL *negptr, const char **errorptr) -{ -int c, i, bot, top; -const uschar *ptr = *ptrptr; -char name[4]; - -c = *(++ptr); -if (c == 0) goto ERROR_RETURN; - -*negptr = FALSE; - -/* \P or \p can be followed by a one- or two-character name in {}, optionally -preceded by ^ for negation. */ - -if (c == '{') - { - if (ptr[1] == '^') - { - *negptr = TRUE; - ptr++; - } - for (i = 0; i <= 2; i++) - { - c = *(++ptr); - if (c == 0) goto ERROR_RETURN; - if (c == '}') break; - name[i] = c; - } - if (c !='}') /* Try to distinguish error cases */ - { - while (*(++ptr) != 0 && *ptr != '}'); - if (*ptr == '}') goto UNKNOWN_RETURN; else goto ERROR_RETURN; - } - name[i] = 0; - } - -/* Otherwise there is just one following character */ - -else - { - name[0] = c; - name[1] = 0; - } - -*ptrptr = ptr; - -/* Search for a recognized property name using binary chop */ - -bot = 0; -top = sizeof(utt)/sizeof(ucp_type_table); - -while (bot < top) - { - i = (bot + top)/2; - c = strcmp(name, utt[i].name); - if (c == 0) return utt[i].value; - if (c > 0) bot = i + 1; else top = i; - } - -UNKNOWN_RETURN: -*errorptr = ERR47; -*ptrptr = ptr; -return -1; - -ERROR_RETURN: -*errorptr = ERR46; -*ptrptr = ptr; -return -1; -} -#endif - - - - -/************************************************* -* Check for counted repeat * -*************************************************/ - -/* This function is called when a '{' is encountered in a place where it might -start a quantifier. It looks ahead to see if it really is a quantifier or not. -It is only a quantifier if it is one of the forms {ddd} {ddd,} or {ddd,ddd} -where the ddds are digits. - -Arguments: - p pointer to the first char after '{' - -Returns: TRUE or FALSE -*/ - -static BOOL -is_counted_repeat(const uschar *p) -{ -if ((digitab[*p++] & ctype_digit) == 0) return FALSE; -while ((digitab[*p] & ctype_digit) != 0) p++; -if (*p == '}') return TRUE; - -if (*p++ != ',') return FALSE; -if (*p == '}') return TRUE; - -if ((digitab[*p++] & ctype_digit) == 0) return FALSE; -while ((digitab[*p] & ctype_digit) != 0) p++; - -return (*p == '}'); -} - - - -/************************************************* -* Read repeat counts * -*************************************************/ - -/* Read an item of the form {n,m} and return the values. This is called only -after is_counted_repeat() has confirmed that a repeat-count quantifier exists, -so the syntax is guaranteed to be correct, but we need to check the values. - -Arguments: - p pointer to first char after '{' - minp pointer to int for min - maxp pointer to int for max - returned as -1 if no max - errorptr points to pointer to error message - -Returns: pointer to '}' on success; - current ptr on error, with errorptr set -*/ - -static const uschar * -read_repeat_counts(const uschar *p, int *minp, int *maxp, const char **errorptr) -{ -int min = 0; -int max = -1; - -while ((digitab[*p] & ctype_digit) != 0) min = min * 10 + *p++ - '0'; - -if (*p == '}') max = min; else - { - if (*(++p) != '}') - { - max = 0; - while((digitab[*p] & ctype_digit) != 0) max = max * 10 + *p++ - '0'; - if (max < min) - { - *errorptr = ERR4; - return p; - } - } - } - -/* Do paranoid checks, then fill in the required variables, and pass back the -pointer to the terminating '}'. */ - -if (min > 65535 || max > 65535) - *errorptr = ERR5; -else - { - *minp = min; - *maxp = max; - } -return p; -} - - - -/************************************************* -* Find first significant op code * -*************************************************/ - -/* This is called by several functions that scan a compiled expression looking -for a fixed first character, or an anchoring op code etc. It skips over things -that do not influence this. For some calls, a change of option is important. -For some calls, it makes sense to skip negative forward and all backward -assertions, and also the \b assertion; for others it does not. - -Arguments: - code pointer to the start of the group - options pointer to external options - optbit the option bit whose changing is significant, or - zero if none are - skipassert TRUE if certain assertions are to be skipped - -Returns: pointer to the first significant opcode -*/ - -static const uschar* -first_significant_code(const uschar *code, int *options, int optbit, - BOOL skipassert) -{ -for (;;) - { - switch ((int)*code) - { - case OP_OPT: - if (optbit > 0 && ((int)code[1] & optbit) != (*options & optbit)) - *options = (int)code[1]; - code += 2; - break; - - case OP_ASSERT_NOT: - case OP_ASSERTBACK: - case OP_ASSERTBACK_NOT: - if (!skipassert) return code; - do code += GET(code, 1); while (*code == OP_ALT); - code += OP_lengths[*code]; - break; - - case OP_WORD_BOUNDARY: - case OP_NOT_WORD_BOUNDARY: - if (!skipassert) return code; - /* Fall through */ - - case OP_CALLOUT: - case OP_CREF: - case OP_BRANUMBER: - code += OP_lengths[*code]; - break; - - default: - return code; - } - } -/* Control never reaches here */ -} - - - - -/************************************************* -* Find the fixed length of a pattern * -*************************************************/ - -/* Scan a pattern and compute the fixed length of subject that will match it, -if the length is fixed. This is needed for dealing with backward assertions. -In UTF8 mode, the result is in characters rather than bytes. - -Arguments: - code points to the start of the pattern (the bracket) - options the compiling options - -Returns: the fixed length, or -1 if there is no fixed length, - or -2 if \C was encountered -*/ - -static int -find_fixedlength(uschar *code, int options) -{ -int length = -1; - -register int branchlength = 0; -register uschar *cc = code + 1 + LINK_SIZE; - -/* Scan along the opcodes for this branch. If we get to the end of the -branch, check the length against that of the other branches. */ - -for (;;) - { - int d; - register int op = *cc; - if (op >= OP_BRA) op = OP_BRA; - - switch (op) - { - case OP_BRA: - case OP_ONCE: - case OP_COND: - d = find_fixedlength(cc, options); - if (d < 0) return d; - branchlength += d; - do cc += GET(cc, 1); while (*cc == OP_ALT); - cc += 1 + LINK_SIZE; - break; - - /* Reached end of a branch; if it's a ket it is the end of a nested - call. If it's ALT it is an alternation in a nested call. If it is - END it's the end of the outer call. All can be handled by the same code. */ - - case OP_ALT: - case OP_KET: - case OP_KETRMAX: - case OP_KETRMIN: - case OP_END: - if (length < 0) length = branchlength; - else if (length != branchlength) return -1; - if (*cc != OP_ALT) return length; - cc += 1 + LINK_SIZE; - branchlength = 0; - break; - - /* Skip over assertive subpatterns */ - - case OP_ASSERT: - case OP_ASSERT_NOT: - case OP_ASSERTBACK: - case OP_ASSERTBACK_NOT: - do cc += GET(cc, 1); while (*cc == OP_ALT); - /* Fall through */ - - /* Skip over things that don't match chars */ - - case OP_REVERSE: - case OP_BRANUMBER: - case OP_CREF: - case OP_OPT: - case OP_CALLOUT: - case OP_SOD: - case OP_SOM: - case OP_EOD: - case OP_EODN: - case OP_CIRC: - case OP_DOLL: - case OP_NOT_WORD_BOUNDARY: - case OP_WORD_BOUNDARY: - cc += OP_lengths[*cc]; - break; - - /* Handle literal characters */ - - case OP_CHAR: - case OP_CHARNC: - branchlength++; - cc += 2; -#ifdef SUPPORT_UTF8 - if ((options & PCRE_UTF8) != 0) - { - while ((*cc & 0xc0) == 0x80) cc++; - } -#endif - break; - - /* Handle exact repetitions. The count is already in characters, but we - need to skip over a multibyte character in UTF8 mode. */ - - case OP_EXACT: - branchlength += GET2(cc,1); - cc += 4; -#ifdef SUPPORT_UTF8 - if ((options & PCRE_UTF8) != 0) - { - while((*cc & 0x80) == 0x80) cc++; - } -#endif - break; - - case OP_TYPEEXACT: - branchlength += GET2(cc,1); - cc += 4; - break; - - /* Handle single-char matchers */ - - case OP_PROP: - case OP_NOTPROP: - cc++; - /* Fall through */ - - case OP_NOT_DIGIT: - case OP_DIGIT: - case OP_NOT_WHITESPACE: - case OP_WHITESPACE: - case OP_NOT_WORDCHAR: - case OP_WORDCHAR: - case OP_ANY: - branchlength++; - cc++; - break; - - /* The single-byte matcher isn't allowed */ - - case OP_ANYBYTE: - return -2; - - /* Check a class for variable quantification */ - -#ifdef SUPPORT_UTF8 - case OP_XCLASS: - cc += GET(cc, 1) - 33; - /* Fall through */ -#endif - - case OP_CLASS: - case OP_NCLASS: - cc += 33; - - switch (*cc) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - case OP_CRQUERY: - case OP_CRMINQUERY: - return -1; - - case OP_CRRANGE: - case OP_CRMINRANGE: - if (GET2(cc,1) != GET2(cc,3)) return -1; - branchlength += GET2(cc,1); - cc += 5; - break; - - default: - branchlength++; - } - break; - - /* Anything else is variable length */ - - default: - return -1; - } - } -/* Control never gets here */ -} - - - - -/************************************************* -* Scan compiled regex for numbered bracket * -*************************************************/ - -/* This little function scans through a compiled pattern until it finds a -capturing bracket with the given number. - -Arguments: - code points to start of expression - utf8 TRUE in UTF-8 mode - number the required bracket number - -Returns: pointer to the opcode for the bracket, or NULL if not found -*/ - -static const uschar * -find_bracket(const uschar *code, BOOL utf8, int number) -{ -#ifndef SUPPORT_UTF8 -utf8 = utf8; /* Stop pedantic compilers complaining */ -#endif - -for (;;) - { - register int c = *code; - if (c == OP_END) return NULL; - else if (c > OP_BRA) - { - int n = c - OP_BRA; - if (n > EXTRACT_BASIC_MAX) n = GET2(code, 2+LINK_SIZE); - if (n == number) return (uschar *)code; - code += OP_lengths[OP_BRA]; - } - else - { - code += OP_lengths[c]; - -#ifdef SUPPORT_UTF8 - - /* In UTF-8 mode, opcodes that are followed by a character may be followed - by a multi-byte character. The length in the table is a minimum, so we have - to scan along to skip the extra bytes. All opcodes are less than 128, so we - can use relatively efficient code. */ - - if (utf8) switch(c) - { - case OP_CHAR: - case OP_CHARNC: - case OP_EXACT: - case OP_UPTO: - case OP_MINUPTO: - case OP_STAR: - case OP_MINSTAR: - case OP_PLUS: - case OP_MINPLUS: - case OP_QUERY: - case OP_MINQUERY: - while ((*code & 0xc0) == 0x80) code++; - break; - - /* XCLASS is used for classes that cannot be represented just by a bit - map. This includes negated single high-valued characters. The length in - the table is zero; the actual length is stored in the compiled code. */ - - case OP_XCLASS: - code += GET(code, 1) + 1; - break; - } -#endif - } - } -} - - - -/************************************************* -* Scan compiled regex for recursion reference * -*************************************************/ - -/* This little function scans through a compiled pattern until it finds an -instance of OP_RECURSE. - -Arguments: - code points to start of expression - utf8 TRUE in UTF-8 mode - -Returns: pointer to the opcode for OP_RECURSE, or NULL if not found -*/ - -static const uschar * -find_recurse(const uschar *code, BOOL utf8) -{ -#ifndef SUPPORT_UTF8 -utf8 = utf8; /* Stop pedantic compilers complaining */ -#endif - -for (;;) - { - register int c = *code; - if (c == OP_END) return NULL; - else if (c == OP_RECURSE) return code; - else if (c > OP_BRA) - { - code += OP_lengths[OP_BRA]; - } - else - { - code += OP_lengths[c]; - -#ifdef SUPPORT_UTF8 - - /* In UTF-8 mode, opcodes that are followed by a character may be followed - by a multi-byte character. The length in the table is a minimum, so we have - to scan along to skip the extra bytes. All opcodes are less than 128, so we - can use relatively efficient code. */ - - if (utf8) switch(c) - { - case OP_CHAR: - case OP_CHARNC: - case OP_EXACT: - case OP_UPTO: - case OP_MINUPTO: - case OP_STAR: - case OP_MINSTAR: - case OP_PLUS: - case OP_MINPLUS: - case OP_QUERY: - case OP_MINQUERY: - while ((*code & 0xc0) == 0x80) code++; - break; - - /* XCLASS is used for classes that cannot be represented just by a bit - map. This includes negated single high-valued characters. The length in - the table is zero; the actual length is stored in the compiled code. */ - - case OP_XCLASS: - code += GET(code, 1) + 1; - break; - } -#endif - } - } -} - - - -/************************************************* -* Scan compiled branch for non-emptiness * -*************************************************/ - -/* This function scans through a branch of a compiled pattern to see whether it -can match the empty string or not. It is called only from could_be_empty() -below. Note that first_significant_code() skips over assertions. If we hit an -unclosed bracket, we return "empty" - this means we've struck an inner bracket -whose current branch will already have been scanned. - -Arguments: - code points to start of search - endcode points to where to stop - utf8 TRUE if in UTF8 mode - -Returns: TRUE if what is matched could be empty -*/ - -static BOOL -could_be_empty_branch(const uschar *code, const uschar *endcode, BOOL utf8) -{ -register int c; -for (code = first_significant_code(code + 1 + LINK_SIZE, NULL, 0, TRUE); - code < endcode; - code = first_significant_code(code + OP_lengths[c], NULL, 0, TRUE)) - { - const uschar *ccode; - - c = *code; - - if (c >= OP_BRA) - { - BOOL empty_branch; - if (GET(code, 1) == 0) return TRUE; /* Hit unclosed bracket */ - - /* Scan a closed bracket */ - - empty_branch = FALSE; - do - { - if (!empty_branch && could_be_empty_branch(code, endcode, utf8)) - empty_branch = TRUE; - code += GET(code, 1); - } - while (*code == OP_ALT); - if (!empty_branch) return FALSE; /* All branches are non-empty */ - code += 1 + LINK_SIZE; - c = *code; - } - - else switch (c) - { - /* Check for quantifiers after a class */ - -#ifdef SUPPORT_UTF8 - case OP_XCLASS: - ccode = code + GET(code, 1); - goto CHECK_CLASS_REPEAT; -#endif - - case OP_CLASS: - case OP_NCLASS: - ccode = code + 33; - -#ifdef SUPPORT_UTF8 - CHECK_CLASS_REPEAT: -#endif - - switch (*ccode) - { - case OP_CRSTAR: /* These could be empty; continue */ - case OP_CRMINSTAR: - case OP_CRQUERY: - case OP_CRMINQUERY: - break; - - default: /* Non-repeat => class must match */ - case OP_CRPLUS: /* These repeats aren't empty */ - case OP_CRMINPLUS: - return FALSE; - - case OP_CRRANGE: - case OP_CRMINRANGE: - if (GET2(ccode, 1) > 0) return FALSE; /* Minimum > 0 */ - break; - } - break; - - /* Opcodes that must match a character */ - - case OP_PROP: - case OP_NOTPROP: - case OP_EXTUNI: - case OP_NOT_DIGIT: - case OP_DIGIT: - case OP_NOT_WHITESPACE: - case OP_WHITESPACE: - case OP_NOT_WORDCHAR: - case OP_WORDCHAR: - case OP_ANY: - case OP_ANYBYTE: - case OP_CHAR: - case OP_CHARNC: - case OP_NOT: - case OP_PLUS: - case OP_MINPLUS: - case OP_EXACT: - case OP_NOTPLUS: - case OP_NOTMINPLUS: - case OP_NOTEXACT: - case OP_TYPEPLUS: - case OP_TYPEMINPLUS: - case OP_TYPEEXACT: - return FALSE; - - /* End of branch */ - - case OP_KET: - case OP_KETRMAX: - case OP_KETRMIN: - case OP_ALT: - return TRUE; - - /* In UTF-8 mode, STAR, MINSTAR, QUERY, MINQUERY, UPTO, and MINUPTO may be - followed by a multibyte character */ - -#ifdef SUPPORT_UTF8 - case OP_STAR: - case OP_MINSTAR: - case OP_QUERY: - case OP_MINQUERY: - case OP_UPTO: - case OP_MINUPTO: - if (utf8) while ((code[2] & 0xc0) == 0x80) code++; - break; -#endif - } - } - -return TRUE; -} - - - -/************************************************* -* Scan compiled regex for non-emptiness * -*************************************************/ - -/* This function is called to check for left recursive calls. We want to check -the current branch of the current pattern to see if it could match the empty -string. If it could, we must look outwards for branches at other levels, -stopping when we pass beyond the bracket which is the subject of the recursion. - -Arguments: - code points to start of the recursion - endcode points to where to stop (current RECURSE item) - bcptr points to the chain of current (unclosed) branch starts - utf8 TRUE if in UTF-8 mode - -Returns: TRUE if what is matched could be empty -*/ - -static BOOL -could_be_empty(const uschar *code, const uschar *endcode, branch_chain *bcptr, - BOOL utf8) -{ -while (bcptr != NULL && bcptr->current >= code) - { - if (!could_be_empty_branch(bcptr->current, endcode, utf8)) return FALSE; - bcptr = bcptr->outer; - } -return TRUE; -} - - - -/************************************************* -* Check for POSIX class syntax * -*************************************************/ - -/* This function is called when the sequence "[:" or "[." or "[=" is -encountered in a character class. It checks whether this is followed by an -optional ^ and then a sequence of letters, terminated by a matching ":]" or -".]" or "=]". - -Argument: - ptr pointer to the initial [ - endptr where to return the end pointer - cd pointer to compile data - -Returns: TRUE or FALSE -*/ - -static BOOL -check_posix_syntax(const uschar *ptr, const uschar **endptr, compile_data *cd) -{ -int terminator; /* Don't combine these lines; the Solaris cc */ -terminator = *(++ptr); /* compiler warns about "non-constant" initializer. */ -if (*(++ptr) == '^') ptr++; -while ((cd->ctypes[*ptr] & ctype_letter) != 0) ptr++; -if (*ptr == terminator && ptr[1] == ']') - { - *endptr = ptr; - return TRUE; - } -return FALSE; -} - - - - -/************************************************* -* Check POSIX class name * -*************************************************/ - -/* This function is called to check the name given in a POSIX-style class entry -such as [:alnum:]. - -Arguments: - ptr points to the first letter - len the length of the name - -Returns: a value representing the name, or -1 if unknown -*/ - -static int -check_posix_name(const uschar *ptr, int len) -{ -register int yield = 0; -while (posix_name_lengths[yield] != 0) - { - if (len == posix_name_lengths[yield] && - strncmp((const char *)ptr, posix_names[yield], len) == 0) return yield; - yield++; - } -return -1; -} - - -/************************************************* -* Adjust OP_RECURSE items in repeated group * -*************************************************/ - -/* OP_RECURSE items contain an offset from the start of the regex to the group -that is referenced. This means that groups can be replicated for fixed -repetition simply by copying (because the recursion is allowed to refer to -earlier groups that are outside the current group). However, when a group is -optional (i.e. the minimum quantifier is zero), OP_BRAZERO is inserted before -it, after it has been compiled. This means that any OP_RECURSE items within it -that refer to the group itself or any contained groups have to have their -offsets adjusted. That is the job of this function. Before it is called, the -partially compiled regex must be temporarily terminated with OP_END. - -Arguments: - group points to the start of the group - adjust the amount by which the group is to be moved - utf8 TRUE in UTF-8 mode - cd contains pointers to tables etc. - -Returns: nothing -*/ - -static void -adjust_recurse(uschar *group, int adjust, BOOL utf8, compile_data *cd) -{ -uschar *ptr = group; -while ((ptr = (uschar *)find_recurse(ptr, utf8)) != NULL) - { - int offset = GET(ptr, 1); - if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust); - ptr += 1 + LINK_SIZE; - } -} - - - -/************************************************* -* Insert an automatic callout point * -*************************************************/ - -/* This function is called when the PCRE_AUTO_CALLOUT option is set, to insert -callout points before each pattern item. - -Arguments: - code current code pointer - ptr current pattern pointer - cd pointers to tables etc - -Returns: new code pointer -*/ - -static uschar * -auto_callout(uschar *code, const uschar *ptr, compile_data *cd) -{ -*code++ = OP_CALLOUT; -*code++ = 255; -PUT(code, 0, ptr - cd->start_pattern); /* Pattern offset */ -PUT(code, LINK_SIZE, 0); /* Default length */ -return code + 2*LINK_SIZE; -} - - - -/************************************************* -* Complete a callout item * -*************************************************/ - -/* A callout item contains the length of the next item in the pattern, which -we can't fill in till after we have reached the relevant point. This is used -for both automatic and manual callouts. - -Arguments: - previous_callout points to previous callout item - ptr current pattern pointer - cd pointers to tables etc - -Returns: nothing -*/ - -static void -complete_callout(uschar *previous_callout, const uschar *ptr, compile_data *cd) -{ -int length = ptr - cd->start_pattern - GET(previous_callout, 2); -PUT(previous_callout, 2 + LINK_SIZE, length); -} - - - -#ifdef SUPPORT_UCP -/************************************************* -* Get othercase range * -*************************************************/ - -/* This function is passed the start and end of a class range, in UTF-8 mode -with UCP support. It searches up the characters, looking for internal ranges of -characters in the "other" case. Each call returns the next one, updating the -start address. - -Arguments: - cptr points to starting character value; updated - d end value - ocptr where to put start of othercase range - odptr where to put end of othercase range - -Yield: TRUE when range returned; FALSE when no more -*/ - -static BOOL -get_othercase_range(int *cptr, int d, int *ocptr, int *odptr) -{ -int c, chartype, othercase, next; - -for (c = *cptr; c <= d; c++) - { - if (ucp_findchar(c, &chartype, &othercase) == ucp_L && othercase != 0) break; - } - -if (c > d) return FALSE; - -*ocptr = othercase; -next = othercase + 1; - -for (++c; c <= d; c++) - { - if (ucp_findchar(c, &chartype, &othercase) != ucp_L || othercase != next) - break; - next++; - } - -*odptr = next - 1; -*cptr = c; - -return TRUE; -} -#endif /* SUPPORT_UCP */ - - -/************************************************* -* Compile one branch * -*************************************************/ - -/* Scan the pattern, compiling it into the code vector. If the options are -changed during the branch, the pointer is used to change the external options -bits. - -Arguments: - optionsptr pointer to the option bits - brackets points to number of extracting brackets used - codeptr points to the pointer to the current code point - ptrptr points to the current pattern pointer - errorptr points to pointer to error message - firstbyteptr set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE) - reqbyteptr set to the last literal character required, else < 0 - bcptr points to current branch chain - cd contains pointers to tables etc. - -Returns: TRUE on success - FALSE, with *errorptr set on error -*/ - -static BOOL -compile_branch(int *optionsptr, int *brackets, uschar **codeptr, - const uschar **ptrptr, const char **errorptr, int *firstbyteptr, - int *reqbyteptr, branch_chain *bcptr, compile_data *cd) -{ -int repeat_type, op_type; -int repeat_min = 0, repeat_max = 0; /* To please picky compilers */ -int bravalue = 0; -int greedy_default, greedy_non_default; -int firstbyte, reqbyte; -int zeroreqbyte, zerofirstbyte; -int req_caseopt, reqvary, tempreqvary; -int condcount = 0; -int options = *optionsptr; -int after_manual_callout = 0; -register int c; -register uschar *code = *codeptr; -uschar *tempcode; -BOOL inescq = FALSE; -BOOL groupsetfirstbyte = FALSE; -const uschar *ptr = *ptrptr; -const uschar *tempptr; -uschar *previous = NULL; -uschar *previous_callout = NULL; -uschar classbits[32]; - -#ifdef SUPPORT_UTF8 -BOOL class_utf8; -BOOL utf8 = (options & PCRE_UTF8) != 0; -uschar *class_utf8data; -uschar utf8_char[6]; -#else -BOOL utf8 = FALSE; -#endif - -/* Set up the default and non-default settings for greediness */ - -greedy_default = ((options & PCRE_UNGREEDY) != 0); -greedy_non_default = greedy_default ^ 1; - -/* Initialize no first byte, no required byte. REQ_UNSET means "no char -matching encountered yet". It gets changed to REQ_NONE if we hit something that -matches a non-fixed char first char; reqbyte just remains unset if we never -find one. - -When we hit a repeat whose minimum is zero, we may have to adjust these values -to take the zero repeat into account. This is implemented by setting them to -zerofirstbyte and zeroreqbyte when such a repeat is encountered. The individual -item types that can be repeated set these backoff variables appropriately. */ - -firstbyte = reqbyte = zerofirstbyte = zeroreqbyte = REQ_UNSET; - -/* The variable req_caseopt contains either the REQ_CASELESS value or zero, -according to the current setting of the caseless flag. REQ_CASELESS is a bit -value > 255. It is added into the firstbyte or reqbyte variables to record the -case status of the value. This is used only for ASCII characters. */ - -req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS : 0; - -/* Switch on next character until the end of the branch */ - -for (;; ptr++) - { - BOOL negate_class; - BOOL possessive_quantifier; - BOOL is_quantifier; - int class_charcount; - int class_lastchar; - int newoptions; - int recno; - int skipbytes; - int subreqbyte; - int subfirstbyte; - int mclength; - uschar mcbuffer[8]; - - /* Next byte in the pattern */ - - c = *ptr; - - /* If in \Q...\E, check for the end; if not, we have a literal */ - - if (inescq && c != 0) - { - if (c == '\\' && ptr[1] == 'E') - { - inescq = FALSE; - ptr++; - continue; - } - else - { - if (previous_callout != NULL) - { - complete_callout(previous_callout, ptr, cd); - previous_callout = NULL; - } - if ((options & PCRE_AUTO_CALLOUT) != 0) - { - previous_callout = code; - code = auto_callout(code, ptr, cd); - } - goto NORMAL_CHAR; - } - } - - /* Fill in length of a previous callout, except when the next thing is - a quantifier. */ - - is_quantifier = c == '*' || c == '+' || c == '?' || - (c == '{' && is_counted_repeat(ptr+1)); - - if (!is_quantifier && previous_callout != NULL && - after_manual_callout-- <= 0) - { - complete_callout(previous_callout, ptr, cd); - previous_callout = NULL; - } - - /* In extended mode, skip white space and comments */ - - if ((options & PCRE_EXTENDED) != 0) - { - if ((cd->ctypes[c] & ctype_space) != 0) continue; - if (c == '#') - { - /* The space before the ; is to avoid a warning on a silly compiler - on the Macintosh. */ - while ((c = *(++ptr)) != 0 && c != NEWLINE) ; - if (c != 0) continue; /* Else fall through to handle end of string */ - } - } - - /* No auto callout for quantifiers. */ - - if ((options & PCRE_AUTO_CALLOUT) != 0 && !is_quantifier) - { - previous_callout = code; - code = auto_callout(code, ptr, cd); - } - - switch(c) - { - /* The branch terminates at end of string, |, or ). */ - - case 0: - case '|': - case ')': - *firstbyteptr = firstbyte; - *reqbyteptr = reqbyte; - *codeptr = code; - *ptrptr = ptr; - return TRUE; - - /* Handle single-character metacharacters. In multiline mode, ^ disables - the setting of any following char as a first character. */ - - case '^': - if ((options & PCRE_MULTILINE) != 0) - { - if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; - } - previous = NULL; - *code++ = OP_CIRC; - break; - - case '$': - previous = NULL; - *code++ = OP_DOLL; - break; - - /* There can never be a first char if '.' is first, whatever happens about - repeats. The value of reqbyte doesn't change either. */ - - case '.': - if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; - zerofirstbyte = firstbyte; - zeroreqbyte = reqbyte; - previous = code; - *code++ = OP_ANY; - break; - - /* Character classes. If the included characters are all < 255 in value, we - build a 32-byte bitmap of the permitted characters, except in the special - case where there is only one such character. For negated classes, we build - the map as usual, then invert it at the end. However, we use a different - opcode so that data characters > 255 can be handled correctly. - - If the class contains characters outside the 0-255 range, a different - opcode is compiled. It may optionally have a bit map for characters < 256, - but those above are are explicitly listed afterwards. A flag byte tells - whether the bitmap is present, and whether this is a negated class or not. - */ - - case '[': - previous = code; - - /* PCRE supports POSIX class stuff inside a class. Perl gives an error if - they are encountered at the top level, so we'll do that too. */ - - if ((ptr[1] == ':' || ptr[1] == '.' || ptr[1] == '=') && - check_posix_syntax(ptr, &tempptr, cd)) - { - *errorptr = (ptr[1] == ':')? ERR13 : ERR31; - goto FAILED; - } - - /* If the first character is '^', set the negation flag and skip it. */ - - if ((c = *(++ptr)) == '^') - { - negate_class = TRUE; - c = *(++ptr); - } - else - { - negate_class = FALSE; - } - - /* Keep a count of chars with values < 256 so that we can optimize the case - of just a single character (as long as it's < 256). For higher valued UTF-8 - characters, we don't yet do any optimization. */ - - class_charcount = 0; - class_lastchar = -1; - -#ifdef SUPPORT_UTF8 - class_utf8 = FALSE; /* No chars >= 256 */ - class_utf8data = code + LINK_SIZE + 34; /* For UTF-8 items */ -#endif - - /* Initialize the 32-char bit map to all zeros. We have to build the - map in a temporary bit of store, in case the class contains only 1 - character (< 256), because in that case the compiled code doesn't use the - bit map. */ - - memset(classbits, 0, 32 * sizeof(uschar)); - - /* Process characters until ] is reached. By writing this as a "do" it - means that an initial ] is taken as a data character. The first pass - through the regex checked the overall syntax, so we don't need to be very - strict here. At the start of the loop, c contains the first byte of the - character. */ - - do - { -#ifdef SUPPORT_UTF8 - if (utf8 && c > 127) - { /* Braces are required because the */ - GETCHARLEN(c, ptr, ptr); /* macro generates multiple statements */ - } -#endif - - /* Inside \Q...\E everything is literal except \E */ - - if (inescq) - { - if (c == '\\' && ptr[1] == 'E') - { - inescq = FALSE; - ptr++; - continue; - } - else goto LONE_SINGLE_CHARACTER; - } - - /* Handle POSIX class names. Perl allows a negation extension of the - form [:^name:]. A square bracket that doesn't match the syntax is - treated as a literal. We also recognize the POSIX constructions - [.ch.] and [=ch=] ("collating elements") and fault them, as Perl - 5.6 and 5.8 do. */ - - if (c == '[' && - (ptr[1] == ':' || ptr[1] == '.' || ptr[1] == '=') && - check_posix_syntax(ptr, &tempptr, cd)) - { - BOOL local_negate = FALSE; - int posix_class, i; - register const uschar *cbits = cd->cbits; - - if (ptr[1] != ':') - { - *errorptr = ERR31; - goto FAILED; - } - - ptr += 2; - if (*ptr == '^') - { - local_negate = TRUE; - ptr++; - } - - posix_class = check_posix_name(ptr, tempptr - ptr); - if (posix_class < 0) - { - *errorptr = ERR30; - goto FAILED; - } - - /* If matching is caseless, upper and lower are converted to - alpha. This relies on the fact that the class table starts with - alpha, lower, upper as the first 3 entries. */ - - if ((options & PCRE_CASELESS) != 0 && posix_class <= 2) - posix_class = 0; - - /* Or into the map we are building up to 3 of the static class - tables, or their negations. The [:blank:] class sets up the same - chars as the [:space:] class (all white space). We remove the vertical - white space chars afterwards. */ - - posix_class *= 3; - for (i = 0; i < 3; i++) - { - BOOL blankclass = strncmp((char *)ptr, "blank", 5) == 0; - int taboffset = posix_class_maps[posix_class + i]; - if (taboffset < 0) break; - if (local_negate) - { - if (i == 0) - for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+taboffset]; - else - for (c = 0; c < 32; c++) classbits[c] &= ~cbits[c+taboffset]; - if (blankclass) classbits[1] |= 0x3c; - } - else - { - for (c = 0; c < 32; c++) classbits[c] |= cbits[c+taboffset]; - if (blankclass) classbits[1] &= ~0x3c; - } - } - - ptr = tempptr + 1; - class_charcount = 10; /* Set > 1; assumes more than 1 per class */ - continue; /* End of POSIX syntax handling */ - } - - /* Backslash may introduce a single character, or it may introduce one - of the specials, which just set a flag. Escaped items are checked for - validity in the pre-compiling pass. The sequence \b is a special case. - Inside a class (and only there) it is treated as backspace. Elsewhere - it marks a word boundary. Other escapes have preset maps ready to - or into the one we are building. We assume they have more than one - character in them, so set class_charcount bigger than one. */ - - if (c == '\\') - { - c = check_escape(&ptr, errorptr, *brackets, options, TRUE); - - if (-c == ESC_b) c = '\b'; /* \b is backslash in a class */ - else if (-c == ESC_X) c = 'X'; /* \X is literal X in a class */ - else if (-c == ESC_Q) /* Handle start of quoted string */ - { - if (ptr[1] == '\\' && ptr[2] == 'E') - { - ptr += 2; /* avoid empty string */ - } - else inescq = TRUE; - continue; - } - - if (c < 0) - { - register const uschar *cbits = cd->cbits; - class_charcount += 2; /* Greater than 1 is what matters */ - switch (-c) - { - case ESC_d: - for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_digit]; - continue; - - case ESC_D: - for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit]; - continue; - - case ESC_w: - for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_word]; - continue; - - case ESC_W: - for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word]; - continue; - - case ESC_s: - for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_space]; - classbits[1] &= ~0x08; /* Perl 5.004 onwards omits VT from \s */ - continue; - - case ESC_S: - for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space]; - classbits[1] |= 0x08; /* Perl 5.004 onwards omits VT from \s */ - continue; - -#ifdef SUPPORT_UCP - case ESC_p: - case ESC_P: - { - BOOL negated; - int property = get_ucp(&ptr, &negated, errorptr); - if (property < 0) goto FAILED; - class_utf8 = TRUE; - *class_utf8data++ = ((-c == ESC_p) != negated)? - XCL_PROP : XCL_NOTPROP; - *class_utf8data++ = property; - class_charcount -= 2; /* Not a < 256 character */ - } - continue; -#endif - - /* Unrecognized escapes are faulted if PCRE is running in its - strict mode. By default, for compatibility with Perl, they are - treated as literals. */ - - default: - if ((options & PCRE_EXTRA) != 0) - { - *errorptr = ERR7; - goto FAILED; - } - c = *ptr; /* The final character */ - class_charcount -= 2; /* Undo the default count from above */ - } - } - - /* Fall through if we have a single character (c >= 0). This may be - > 256 in UTF-8 mode. */ - - } /* End of backslash handling */ - - /* A single character may be followed by '-' to form a range. However, - Perl does not permit ']' to be the end of the range. A '-' character - here is treated as a literal. */ - - if (ptr[1] == '-' && ptr[2] != ']') - { - int d; - ptr += 2; - -#ifdef SUPPORT_UTF8 - if (utf8) - { /* Braces are required because the */ - GETCHARLEN(d, ptr, ptr); /* macro generates multiple statements */ - } - else -#endif - d = *ptr; /* Not UTF-8 mode */ - - /* The second part of a range can be a single-character escape, but - not any of the other escapes. Perl 5.6 treats a hyphen as a literal - in such circumstances. */ - - if (d == '\\') - { - const uschar *oldptr = ptr; - d = check_escape(&ptr, errorptr, *brackets, options, TRUE); - - /* \b is backslash; \X is literal X; any other special means the '-' - was literal */ - - if (d < 0) - { - if (d == -ESC_b) d = '\b'; - else if (d == -ESC_X) d = 'X'; else - { - ptr = oldptr - 2; - goto LONE_SINGLE_CHARACTER; /* A few lines below */ - } - } - } - - /* The check that the two values are in the correct order happens in - the pre-pass. Optimize one-character ranges */ - - if (d == c) goto LONE_SINGLE_CHARACTER; /* A few lines below */ - - /* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless - matching, we have to use an XCLASS with extra data items. Caseless - matching for characters > 127 is available only if UCP support is - available. */ - -#ifdef SUPPORT_UTF8 - if (utf8 && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127))) - { - class_utf8 = TRUE; - - /* With UCP support, we can find the other case equivalents of - the relevant characters. There may be several ranges. Optimize how - they fit with the basic range. */ - -#ifdef SUPPORT_UCP - if ((options & PCRE_CASELESS) != 0) - { - int occ, ocd; - int cc = c; - int origd = d; - while (get_othercase_range(&cc, origd, &occ, &ocd)) - { - if (occ >= c && ocd <= d) continue; /* Skip embedded ranges */ - - if (occ < c && ocd >= c - 1) /* Extend the basic range */ - { /* if there is overlap, */ - c = occ; /* noting that if occ < c */ - continue; /* we can't have ocd > d */ - } /* because a subrange is */ - if (ocd > d && occ <= d + 1) /* always shorter than */ - { /* the basic range. */ - d = ocd; - continue; - } - - if (occ == ocd) - { - *class_utf8data++ = XCL_SINGLE; - } - else - { - *class_utf8data++ = XCL_RANGE; - class_utf8data += ord2utf8(occ, class_utf8data); - } - class_utf8data += ord2utf8(ocd, class_utf8data); - } - } -#endif /* SUPPORT_UCP */ - - /* Now record the original range, possibly modified for UCP caseless - overlapping ranges. */ - - *class_utf8data++ = XCL_RANGE; - class_utf8data += ord2utf8(c, class_utf8data); - class_utf8data += ord2utf8(d, class_utf8data); - - /* With UCP support, we are done. Without UCP support, there is no - caseless matching for UTF-8 characters > 127; we can use the bit map - for the smaller ones. */ - -#ifdef SUPPORT_UCP - continue; /* With next character in the class */ -#else - if ((options & PCRE_CASELESS) == 0 || c > 127) continue; - - /* Adjust upper limit and fall through to set up the map */ - - d = 127; - -#endif /* SUPPORT_UCP */ - } -#endif /* SUPPORT_UTF8 */ - - /* We use the bit map for all cases when not in UTF-8 mode; else - ranges that lie entirely within 0-127 when there is UCP support; else - for partial ranges without UCP support. */ - - for (; c <= d; c++) - { - classbits[c/8] |= (1 << (c&7)); - if ((options & PCRE_CASELESS) != 0) - { - int uc = cd->fcc[c]; /* flip case */ - classbits[uc/8] |= (1 << (uc&7)); - } - class_charcount++; /* in case a one-char range */ - class_lastchar = c; - } - - continue; /* Go get the next char in the class */ - } - - /* Handle a lone single character - we can get here for a normal - non-escape char, or after \ that introduces a single character or for an - apparent range that isn't. */ - - LONE_SINGLE_CHARACTER: - - /* Handle a character that cannot go in the bit map */ - -#ifdef SUPPORT_UTF8 - if (utf8 && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127))) - { - class_utf8 = TRUE; - *class_utf8data++ = XCL_SINGLE; - class_utf8data += ord2utf8(c, class_utf8data); - -#ifdef SUPPORT_UCP - if ((options & PCRE_CASELESS) != 0) - { - int chartype; - int othercase; - if (ucp_findchar(c, &chartype, &othercase) >= 0 && othercase > 0) - { - *class_utf8data++ = XCL_SINGLE; - class_utf8data += ord2utf8(othercase, class_utf8data); - } - } -#endif /* SUPPORT_UCP */ - - } - else -#endif /* SUPPORT_UTF8 */ - - /* Handle a single-byte character */ - { - classbits[c/8] |= (1 << (c&7)); - if ((options & PCRE_CASELESS) != 0) - { - c = cd->fcc[c]; /* flip case */ - classbits[c/8] |= (1 << (c&7)); - } - class_charcount++; - class_lastchar = c; - } - } - - /* Loop until ']' reached; the check for end of string happens inside the - loop. This "while" is the end of the "do" above. */ - - while ((c = *(++ptr)) != ']' || inescq); - - /* If class_charcount is 1, we saw precisely one character whose value is - less than 256. In non-UTF-8 mode we can always optimize. In UTF-8 mode, we - can optimize the negative case only if there were no characters >= 128 - because OP_NOT and the related opcodes like OP_NOTSTAR operate on - single-bytes only. This is an historical hangover. Maybe one day we can - tidy these opcodes to handle multi-byte characters. - - The optimization throws away the bit map. We turn the item into a - 1-character OP_CHAR[NC] if it's positive, or OP_NOT if it's negative. Note - that OP_NOT does not support multibyte characters. In the positive case, it - can cause firstbyte to be set. Otherwise, there can be no first char if - this item is first, whatever repeat count may follow. In the case of - reqbyte, save the previous value for reinstating. */ - -#ifdef SUPPORT_UTF8 - if (class_charcount == 1 && - (!utf8 || - (!class_utf8 && (!negate_class || class_lastchar < 128)))) - -#else - if (class_charcount == 1) -#endif - { - zeroreqbyte = reqbyte; - - /* The OP_NOT opcode works on one-byte characters only. */ - - if (negate_class) - { - if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; - zerofirstbyte = firstbyte; - *code++ = OP_NOT; - *code++ = class_lastchar; - break; - } - - /* For a single, positive character, get the value into mcbuffer, and - then we can handle this with the normal one-character code. */ - -#ifdef SUPPORT_UTF8 - if (utf8 && class_lastchar > 127) - mclength = ord2utf8(class_lastchar, mcbuffer); - else -#endif - { - mcbuffer[0] = class_lastchar; - mclength = 1; - } - goto ONE_CHAR; - } /* End of 1-char optimization */ - - /* The general case - not the one-char optimization. If this is the first - thing in the branch, there can be no first char setting, whatever the - repeat count. Any reqbyte setting must remain unchanged after any kind of - repeat. */ - - if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; - zerofirstbyte = firstbyte; - zeroreqbyte = reqbyte; - - /* If there are characters with values > 255, we have to compile an - extended class, with its own opcode. If there are no characters < 256, - we can omit the bitmap. */ - -#ifdef SUPPORT_UTF8 - if (class_utf8) - { - *class_utf8data++ = XCL_END; /* Marks the end of extra data */ - *code++ = OP_XCLASS; - code += LINK_SIZE; - *code = negate_class? XCL_NOT : 0; - - /* If the map is required, install it, and move on to the end of - the extra data */ - - if (class_charcount > 0) - { - *code++ |= XCL_MAP; - memcpy(code, classbits, 32); - code = class_utf8data; - } - - /* If the map is not required, slide down the extra data. */ - - else - { - int len = class_utf8data - (code + 33); - memmove(code + 1, code + 33, len); - code += len + 1; - } - - /* Now fill in the complete length of the item */ - - PUT(previous, 1, code - previous); - break; /* End of class handling */ - } -#endif - - /* If there are no characters > 255, negate the 32-byte map if necessary, - and copy it into the code vector. If this is the first thing in the branch, - there can be no first char setting, whatever the repeat count. Any reqbyte - setting must remain unchanged after any kind of repeat. */ - - if (negate_class) - { - *code++ = OP_NCLASS; - for (c = 0; c < 32; c++) code[c] = ~classbits[c]; - } - else - { - *code++ = OP_CLASS; - memcpy(code, classbits, 32); - } - code += 32; - break; - - /* Various kinds of repeat; '{' is not necessarily a quantifier, but this - has been tested above. */ - - case '{': - if (!is_quantifier) goto NORMAL_CHAR; - ptr = read_repeat_counts(ptr+1, &repeat_min, &repeat_max, errorptr); - if (*errorptr != NULL) goto FAILED; - goto REPEAT; - - case '*': - repeat_min = 0; - repeat_max = -1; - goto REPEAT; - - case '+': - repeat_min = 1; - repeat_max = -1; - goto REPEAT; - - case '?': - repeat_min = 0; - repeat_max = 1; - - REPEAT: - if (previous == NULL) - { - *errorptr = ERR9; - goto FAILED; - } - - if (repeat_min == 0) - { - firstbyte = zerofirstbyte; /* Adjust for zero repeat */ - reqbyte = zeroreqbyte; /* Ditto */ - } - - /* Remember whether this is a variable length repeat */ - - reqvary = (repeat_min == repeat_max)? 0 : REQ_VARY; - - op_type = 0; /* Default single-char op codes */ - possessive_quantifier = FALSE; /* Default not possessive quantifier */ - - /* Save start of previous item, in case we have to move it up to make space - for an inserted OP_ONCE for the additional '+' extension. */ - - tempcode = previous; - - /* If the next character is '+', we have a possessive quantifier. This - implies greediness, whatever the setting of the PCRE_UNGREEDY option. - If the next character is '?' this is a minimizing repeat, by default, - but if PCRE_UNGREEDY is set, it works the other way round. We change the - repeat type to the non-default. */ - - if (ptr[1] == '+') - { - repeat_type = 0; /* Force greedy */ - possessive_quantifier = TRUE; - ptr++; - } - else if (ptr[1] == '?') - { - repeat_type = greedy_non_default; - ptr++; - } - else repeat_type = greedy_default; - - /* If previous was a recursion, we need to wrap it inside brackets so that - it can be replicated if necessary. */ - - if (*previous == OP_RECURSE) - { - memmove(previous + 1 + LINK_SIZE, previous, 1 + LINK_SIZE); - code += 1 + LINK_SIZE; - *previous = OP_BRA; - PUT(previous, 1, code - previous); - *code = OP_KET; - PUT(code, 1, code - previous); - code += 1 + LINK_SIZE; - } - - /* If previous was a character match, abolish the item and generate a - repeat item instead. If a char item has a minumum of more than one, ensure - that it is set in reqbyte - it might not be if a sequence such as x{3} is - the first thing in a branch because the x will have gone into firstbyte - instead. */ - - if (*previous == OP_CHAR || *previous == OP_CHARNC) - { - /* Deal with UTF-8 characters that take up more than one byte. It's - easier to write this out separately than try to macrify it. Use c to - hold the length of the character in bytes, plus 0x80 to flag that it's a - length rather than a small character. */ - -#ifdef SUPPORT_UTF8 - if (utf8 && (code[-1] & 0x80) != 0) - { - uschar *lastchar = code - 1; - while((*lastchar & 0xc0) == 0x80) lastchar--; - c = code - lastchar; /* Length of UTF-8 character */ - memcpy(utf8_char, lastchar, c); /* Save the char */ - c |= 0x80; /* Flag c as a length */ - } - else -#endif - - /* Handle the case of a single byte - either with no UTF8 support, or - with UTF-8 disabled, or for a UTF-8 character < 128. */ - - { - c = code[-1]; - if (repeat_min > 1) reqbyte = c | req_caseopt | cd->req_varyopt; - } - - goto OUTPUT_SINGLE_REPEAT; /* Code shared with single character types */ - } - - /* If previous was a single negated character ([^a] or similar), we use - one of the special opcodes, replacing it. The code is shared with single- - character repeats by setting opt_type to add a suitable offset into - repeat_type. OP_NOT is currently used only for single-byte chars. */ - - else if (*previous == OP_NOT) - { - op_type = OP_NOTSTAR - OP_STAR; /* Use "not" opcodes */ - c = previous[1]; - goto OUTPUT_SINGLE_REPEAT; - } - - /* If previous was a character type match (\d or similar), abolish it and - create a suitable repeat item. The code is shared with single-character - repeats by setting op_type to add a suitable offset into repeat_type. Note - the the Unicode property types will be present only when SUPPORT_UCP is - defined, but we don't wrap the little bits of code here because it just - makes it horribly messy. */ - - else if (*previous < OP_EODN) - { - uschar *oldcode; - int prop_type; - op_type = OP_TYPESTAR - OP_STAR; /* Use type opcodes */ - c = *previous; - - OUTPUT_SINGLE_REPEAT: - prop_type = (*previous == OP_PROP || *previous == OP_NOTPROP)? - previous[1] : -1; - - oldcode = code; - code = previous; /* Usually overwrite previous item */ - - /* If the maximum is zero then the minimum must also be zero; Perl allows - this case, so we do too - by simply omitting the item altogether. */ - - if (repeat_max == 0) goto END_REPEAT; - - /* All real repeats make it impossible to handle partial matching (maybe - one day we will be able to remove this restriction). */ - - if (repeat_max != 1) cd->nopartial = TRUE; - - /* Combine the op_type with the repeat_type */ - - repeat_type += op_type; - - /* A minimum of zero is handled either as the special case * or ?, or as - an UPTO, with the maximum given. */ - - if (repeat_min == 0) - { - if (repeat_max == -1) *code++ = OP_STAR + repeat_type; - else if (repeat_max == 1) *code++ = OP_QUERY + repeat_type; - else - { - *code++ = OP_UPTO + repeat_type; - PUT2INC(code, 0, repeat_max); - } - } - - /* A repeat minimum of 1 is optimized into some special cases. If the - maximum is unlimited, we use OP_PLUS. Otherwise, the original item it - left in place and, if the maximum is greater than 1, we use OP_UPTO with - one less than the maximum. */ - - else if (repeat_min == 1) - { - if (repeat_max == -1) - *code++ = OP_PLUS + repeat_type; - else - { - code = oldcode; /* leave previous item in place */ - if (repeat_max == 1) goto END_REPEAT; - *code++ = OP_UPTO + repeat_type; - PUT2INC(code, 0, repeat_max - 1); - } - } - - /* The case {n,n} is just an EXACT, while the general case {n,m} is - handled as an EXACT followed by an UPTO. */ - - else - { - *code++ = OP_EXACT + op_type; /* NB EXACT doesn't have repeat_type */ - PUT2INC(code, 0, repeat_min); - - /* If the maximum is unlimited, insert an OP_STAR. Before doing so, - we have to insert the character for the previous code. For a repeated - Unicode property match, there is an extra byte that defines the - required property. In UTF-8 mode, long characters have their length in - c, with the 0x80 bit as a flag. */ - - if (repeat_max < 0) - { -#ifdef SUPPORT_UTF8 - if (utf8 && c >= 128) - { - memcpy(code, utf8_char, c & 7); - code += c & 7; - } - else -#endif - { - *code++ = c; - if (prop_type >= 0) *code++ = prop_type; - } - *code++ = OP_STAR + repeat_type; - } - - /* Else insert an UPTO if the max is greater than the min, again - preceded by the character, for the previously inserted code. */ - - else if (repeat_max != repeat_min) - { -#ifdef SUPPORT_UTF8 - if (utf8 && c >= 128) - { - memcpy(code, utf8_char, c & 7); - code += c & 7; - } - else -#endif - *code++ = c; - if (prop_type >= 0) *code++ = prop_type; - repeat_max -= repeat_min; - *code++ = OP_UPTO + repeat_type; - PUT2INC(code, 0, repeat_max); - } - } - - /* The character or character type itself comes last in all cases. */ - -#ifdef SUPPORT_UTF8 - if (utf8 && c >= 128) - { - memcpy(code, utf8_char, c & 7); - code += c & 7; - } - else -#endif - *code++ = c; - - /* For a repeated Unicode property match, there is an extra byte that - defines the required property. */ - -#ifdef SUPPORT_UCP - if (prop_type >= 0) *code++ = prop_type; -#endif - } - - /* If previous was a character class or a back reference, we put the repeat - stuff after it, but just skip the item if the repeat was {0,0}. */ - - else if (*previous == OP_CLASS || - *previous == OP_NCLASS || -#ifdef SUPPORT_UTF8 - *previous == OP_XCLASS || -#endif - *previous == OP_REF) - { - if (repeat_max == 0) - { - code = previous; - goto END_REPEAT; - } - - /* All real repeats make it impossible to handle partial matching (maybe - one day we will be able to remove this restriction). */ - - if (repeat_max != 1) cd->nopartial = TRUE; - - if (repeat_min == 0 && repeat_max == -1) - *code++ = OP_CRSTAR + repeat_type; - else if (repeat_min == 1 && repeat_max == -1) - *code++ = OP_CRPLUS + repeat_type; - else if (repeat_min == 0 && repeat_max == 1) - *code++ = OP_CRQUERY + repeat_type; - else - { - *code++ = OP_CRRANGE + repeat_type; - PUT2INC(code, 0, repeat_min); - if (repeat_max == -1) repeat_max = 0; /* 2-byte encoding for max */ - PUT2INC(code, 0, repeat_max); - } - } - - /* If previous was a bracket group, we may have to replicate it in certain - cases. */ - - else if (*previous >= OP_BRA || *previous == OP_ONCE || - *previous == OP_COND) - { - register int i; - int ketoffset = 0; - int len = code - previous; - uschar *bralink = NULL; - - /* If the maximum repeat count is unlimited, find the end of the bracket - by scanning through from the start, and compute the offset back to it - from the current code pointer. There may be an OP_OPT setting following - the final KET, so we can't find the end just by going back from the code - pointer. */ - - if (repeat_max == -1) - { - register uschar *ket = previous; - do ket += GET(ket, 1); while (*ket != OP_KET); - ketoffset = code - ket; - } - - /* The case of a zero minimum is special because of the need to stick - OP_BRAZERO in front of it, and because the group appears once in the - data, whereas in other cases it appears the minimum number of times. For - this reason, it is simplest to treat this case separately, as otherwise - the code gets far too messy. There are several special subcases when the - minimum is zero. */ - - if (repeat_min == 0) - { - /* If the maximum is also zero, we just omit the group from the output - altogether. */ - - if (repeat_max == 0) - { - code = previous; - goto END_REPEAT; - } - - /* If the maximum is 1 or unlimited, we just have to stick in the - BRAZERO and do no more at this point. However, we do need to adjust - any OP_RECURSE calls inside the group that refer to the group itself or - any internal group, because the offset is from the start of the whole - regex. Temporarily terminate the pattern while doing this. */ - - if (repeat_max <= 1) - { - *code = OP_END; - adjust_recurse(previous, 1, utf8, cd); - memmove(previous+1, previous, len); - code++; - *previous++ = OP_BRAZERO + repeat_type; - } - - /* If the maximum is greater than 1 and limited, we have to replicate - in a nested fashion, sticking OP_BRAZERO before each set of brackets. - The first one has to be handled carefully because it's the original - copy, which has to be moved up. The remainder can be handled by code - that is common with the non-zero minimum case below. We have to - adjust the value or repeat_max, since one less copy is required. Once - again, we may have to adjust any OP_RECURSE calls inside the group. */ - - else - { - int offset; - *code = OP_END; - adjust_recurse(previous, 2 + LINK_SIZE, utf8, cd); - memmove(previous + 2 + LINK_SIZE, previous, len); - code += 2 + LINK_SIZE; - *previous++ = OP_BRAZERO + repeat_type; - *previous++ = OP_BRA; - - /* We chain together the bracket offset fields that have to be - filled in later when the ends of the brackets are reached. */ - - offset = (bralink == NULL)? 0 : previous - bralink; - bralink = previous; - PUTINC(previous, 0, offset); - } - - repeat_max--; - } - - /* If the minimum is greater than zero, replicate the group as many - times as necessary, and adjust the maximum to the number of subsequent - copies that we need. If we set a first char from the group, and didn't - set a required char, copy the latter from the former. */ - - else - { - if (repeat_min > 1) - { - if (groupsetfirstbyte && reqbyte < 0) reqbyte = firstbyte; - for (i = 1; i < repeat_min; i++) - { - memcpy(code, previous, len); - code += len; - } - } - if (repeat_max > 0) repeat_max -= repeat_min; - } - - /* This code is common to both the zero and non-zero minimum cases. If - the maximum is limited, it replicates the group in a nested fashion, - remembering the bracket starts on a stack. In the case of a zero minimum, - the first one was set up above. In all cases the repeat_max now specifies - the number of additional copies needed. */ - - if (repeat_max >= 0) - { - for (i = repeat_max - 1; i >= 0; i--) - { - *code++ = OP_BRAZERO + repeat_type; - - /* All but the final copy start a new nesting, maintaining the - chain of brackets outstanding. */ - - if (i != 0) - { - int offset; - *code++ = OP_BRA; - offset = (bralink == NULL)? 0 : code - bralink; - bralink = code; - PUTINC(code, 0, offset); - } - - memcpy(code, previous, len); - code += len; - } - - /* Now chain through the pending brackets, and fill in their length - fields (which are holding the chain links pro tem). */ - - while (bralink != NULL) - { - int oldlinkoffset; - int offset = code - bralink + 1; - uschar *bra = code - offset; - oldlinkoffset = GET(bra, 1); - bralink = (oldlinkoffset == 0)? NULL : bralink - oldlinkoffset; - *code++ = OP_KET; - PUTINC(code, 0, offset); - PUT(bra, 1, offset); - } - } - - /* If the maximum is unlimited, set a repeater in the final copy. We - can't just offset backwards from the current code point, because we - don't know if there's been an options resetting after the ket. The - correct offset was computed above. */ - - else code[-ketoffset] = OP_KETRMAX + repeat_type; - } - - /* Else there's some kind of shambles */ - - else - { - *errorptr = ERR11; - goto FAILED; - } - - /* If the character following a repeat is '+', we wrap the entire repeated - item inside OP_ONCE brackets. This is just syntactic sugar, taken from - Sun's Java package. The repeated item starts at tempcode, not at previous, - which might be the first part of a string whose (former) last char we - repeated. However, we don't support '+' after a greediness '?'. */ - - if (possessive_quantifier) - { - int len = code - tempcode; - memmove(tempcode + 1+LINK_SIZE, tempcode, len); - code += 1 + LINK_SIZE; - len += 1 + LINK_SIZE; - tempcode[0] = OP_ONCE; - *code++ = OP_KET; - PUTINC(code, 0, len); - PUT(tempcode, 1, len); - } - - /* In all case we no longer have a previous item. We also set the - "follows varying string" flag for subsequently encountered reqbytes if - it isn't already set and we have just passed a varying length item. */ - - END_REPEAT: - previous = NULL; - cd->req_varyopt |= reqvary; - break; - - - /* Start of nested bracket sub-expression, or comment or lookahead or - lookbehind or option setting or condition. First deal with special things - that can come after a bracket; all are introduced by ?, and the appearance - of any of them means that this is not a referencing group. They were - checked for validity in the first pass over the string, so we don't have to - check for syntax errors here. */ - - case '(': - newoptions = options; - skipbytes = 0; - - if (*(++ptr) == '?') - { - int set, unset; - int *optset; - - switch (*(++ptr)) - { - case '#': /* Comment; skip to ket */ - ptr++; - while (*ptr != ')') ptr++; - continue; - - case ':': /* Non-extracting bracket */ - bravalue = OP_BRA; - ptr++; - break; - - case '(': - bravalue = OP_COND; /* Conditional group */ - - /* Condition to test for recursion */ - - if (ptr[1] == 'R') - { - code[1+LINK_SIZE] = OP_CREF; - PUT2(code, 2+LINK_SIZE, CREF_RECURSE); - skipbytes = 3; - ptr += 3; - } - - /* Condition to test for a numbered subpattern match. We know that - if a digit follows ( then there will just be digits until ) because - the syntax was checked in the first pass. */ - - else if ((digitab[ptr[1]] && ctype_digit) != 0) - { - int condref; /* Don't amalgamate; some compilers */ - condref = *(++ptr) - '0'; /* grumble at autoincrement in declaration */ - while (*(++ptr) != ')') condref = condref*10 + *ptr - '0'; - if (condref == 0) - { - *errorptr = ERR35; - goto FAILED; - } - ptr++; - code[1+LINK_SIZE] = OP_CREF; - PUT2(code, 2+LINK_SIZE, condref); - skipbytes = 3; - } - /* For conditions that are assertions, we just fall through, having - set bravalue above. */ - break; - - case '=': /* Positive lookahead */ - bravalue = OP_ASSERT; - ptr++; - break; - - case '!': /* Negative lookahead */ - bravalue = OP_ASSERT_NOT; - ptr++; - break; - - case '<': /* Lookbehinds */ - switch (*(++ptr)) - { - case '=': /* Positive lookbehind */ - bravalue = OP_ASSERTBACK; - ptr++; - break; - - case '!': /* Negative lookbehind */ - bravalue = OP_ASSERTBACK_NOT; - ptr++; - break; - } - break; - - case '>': /* One-time brackets */ - bravalue = OP_ONCE; - ptr++; - break; - - case 'C': /* Callout - may be followed by digits; */ - previous_callout = code; /* Save for later completion */ - after_manual_callout = 1; /* Skip one item before completing */ - *code++ = OP_CALLOUT; /* Already checked that the terminating */ - { /* closing parenthesis is present. */ - int n = 0; - while ((digitab[*(++ptr)] & ctype_digit) != 0) - n = n * 10 + *ptr - '0'; - if (n > 255) - { - *errorptr = ERR38; - goto FAILED; - } - *code++ = n; - PUT(code, 0, ptr - cd->start_pattern + 1); /* Pattern offset */ - PUT(code, LINK_SIZE, 0); /* Default length */ - code += 2 * LINK_SIZE; - } - previous = NULL; - continue; - - case 'P': /* Named subpattern handling */ - if (*(++ptr) == '<') /* Definition */ - { - int i, namelen; - uschar *slot = cd->name_table; - const uschar *name; /* Don't amalgamate; some compilers */ - name = ++ptr; /* grumble at autoincrement in declaration */ - - while (*ptr++ != '>'); - namelen = ptr - name - 1; - - for (i = 0; i < cd->names_found; i++) - { - int crc = memcmp(name, slot+2, namelen); - if (crc == 0) - { - if (slot[2+namelen] == 0) - { - *errorptr = ERR43; - goto FAILED; - } - crc = -1; /* Current name is substring */ - } - if (crc < 0) - { - memmove(slot + cd->name_entry_size, slot, - (cd->names_found - i) * cd->name_entry_size); - break; - } - slot += cd->name_entry_size; - } - - PUT2(slot, 0, *brackets + 1); - memcpy(slot + 2, name, namelen); - slot[2+namelen] = 0; - cd->names_found++; - goto NUMBERED_GROUP; - } - - if (*ptr == '=' || *ptr == '>') /* Reference or recursion */ - { - int i, namelen; - int type = *ptr++; - const uschar *name = ptr; - uschar *slot = cd->name_table; - - while (*ptr != ')') ptr++; - namelen = ptr - name; - - for (i = 0; i < cd->names_found; i++) - { - if (strncmp((char *)name, (char *)slot+2, namelen) == 0) break; - slot += cd->name_entry_size; - } - if (i >= cd->names_found) - { - *errorptr = ERR15; - goto FAILED; - } - - recno = GET2(slot, 0); - - if (type == '>') goto HANDLE_RECURSION; /* A few lines below */ - - /* Back reference */ - - previous = code; - *code++ = OP_REF; - PUT2INC(code, 0, recno); - cd->backref_map |= (recno < 32)? (1 << recno) : 1; - if (recno > cd->top_backref) cd->top_backref = recno; - continue; - } - - /* Should never happen */ - break; - - case 'R': /* Pattern recursion */ - ptr++; /* Same as (?0) */ - /* Fall through */ - - /* Recursion or "subroutine" call */ - - case '0': case '1': case '2': case '3': case '4': - case '5': case '6': case '7': case '8': case '9': - { - const uschar *called; - recno = 0; - while((digitab[*ptr] & ctype_digit) != 0) - recno = recno * 10 + *ptr++ - '0'; - - /* Come here from code above that handles a named recursion */ - - HANDLE_RECURSION: - - previous = code; - - /* Find the bracket that is being referenced. Temporarily end the - regex in case it doesn't exist. */ - - *code = OP_END; - called = (recno == 0)? - cd->start_code : find_bracket(cd->start_code, utf8, recno); - - if (called == NULL) - { - *errorptr = ERR15; - goto FAILED; - } - - /* If the subpattern is still open, this is a recursive call. We - check to see if this is a left recursion that could loop for ever, - and diagnose that case. */ - - if (GET(called, 1) == 0 && could_be_empty(called, code, bcptr, utf8)) - { - *errorptr = ERR40; - goto FAILED; - } - - /* Insert the recursion/subroutine item */ - - *code = OP_RECURSE; - PUT(code, 1, called - cd->start_code); - code += 1 + LINK_SIZE; - } - continue; - - /* Character after (? not specially recognized */ - - default: /* Option setting */ - set = unset = 0; - optset = &set; - - while (*ptr != ')' && *ptr != ':') - { - switch (*ptr++) - { - case '-': optset = &unset; break; - - case 'i': *optset |= PCRE_CASELESS; break; - case 'm': *optset |= PCRE_MULTILINE; break; - case 's': *optset |= PCRE_DOTALL; break; - case 'x': *optset |= PCRE_EXTENDED; break; - case 'U': *optset |= PCRE_UNGREEDY; break; - case 'X': *optset |= PCRE_EXTRA; break; - } - } - - /* Set up the changed option bits, but don't change anything yet. */ - - newoptions = (options | set) & (~unset); - - /* If the options ended with ')' this is not the start of a nested - group with option changes, so the options change at this level. Compile - code to change the ims options if this setting actually changes any of - them. We also pass the new setting back so that it can be put at the - start of any following branches, and when this group ends (if we are in - a group), a resetting item can be compiled. - - Note that if this item is right at the start of the pattern, the - options will have been abstracted and made global, so there will be no - change to compile. */ - - if (*ptr == ')') - { - if ((options & PCRE_IMS) != (newoptions & PCRE_IMS)) - { - *code++ = OP_OPT; - *code++ = newoptions & PCRE_IMS; - } - - /* Change options at this level, and pass them back for use - in subsequent branches. Reset the greedy defaults and the case - value for firstbyte and reqbyte. */ - - *optionsptr = options = newoptions; - greedy_default = ((newoptions & PCRE_UNGREEDY) != 0); - greedy_non_default = greedy_default ^ 1; - req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS : 0; - - previous = NULL; /* This item can't be repeated */ - continue; /* It is complete */ - } - - /* If the options ended with ':' we are heading into a nested group - with possible change of options. Such groups are non-capturing and are - not assertions of any kind. All we need to do is skip over the ':'; - the newoptions value is handled below. */ - - bravalue = OP_BRA; - ptr++; - } - } - - /* If PCRE_NO_AUTO_CAPTURE is set, all unadorned brackets become - non-capturing and behave like (?:...) brackets */ - - else if ((options & PCRE_NO_AUTO_CAPTURE) != 0) - { - bravalue = OP_BRA; - } - - /* Else we have a referencing group; adjust the opcode. If the bracket - number is greater than EXTRACT_BASIC_MAX, we set the opcode one higher, and - arrange for the true number to follow later, in an OP_BRANUMBER item. */ - - else - { - NUMBERED_GROUP: - if (++(*brackets) > EXTRACT_BASIC_MAX) - { - bravalue = OP_BRA + EXTRACT_BASIC_MAX + 1; - code[1+LINK_SIZE] = OP_BRANUMBER; - PUT2(code, 2+LINK_SIZE, *brackets); - skipbytes = 3; - } - else bravalue = OP_BRA + *brackets; - } - - /* Process nested bracketed re. Assertions may not be repeated, but other - kinds can be. We copy code into a non-register variable in order to be able - to pass its address because some compilers complain otherwise. Pass in a - new setting for the ims options if they have changed. */ - - previous = (bravalue >= OP_ONCE)? code : NULL; - *code = bravalue; - tempcode = code; - tempreqvary = cd->req_varyopt; /* Save value before bracket */ - - if (!compile_regex( - newoptions, /* The complete new option state */ - options & PCRE_IMS, /* The previous ims option state */ - brackets, /* Extracting bracket count */ - &tempcode, /* Where to put code (updated) */ - &ptr, /* Input pointer (updated) */ - errorptr, /* Where to put an error message */ - (bravalue == OP_ASSERTBACK || - bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */ - skipbytes, /* Skip over OP_COND/OP_BRANUMBER */ - &subfirstbyte, /* For possible first char */ - &subreqbyte, /* For possible last char */ - bcptr, /* Current branch chain */ - cd)) /* Tables block */ - goto FAILED; - - /* At the end of compiling, code is still pointing to the start of the - group, while tempcode has been updated to point past the end of the group - and any option resetting that may follow it. The pattern pointer (ptr) - is on the bracket. */ - - /* If this is a conditional bracket, check that there are no more than - two branches in the group. */ - - else if (bravalue == OP_COND) - { - uschar *tc = code; - condcount = 0; - - do { - condcount++; - tc += GET(tc,1); - } - while (*tc != OP_KET); - - if (condcount > 2) - { - *errorptr = ERR27; - goto FAILED; - } - - /* If there is just one branch, we must not make use of its firstbyte or - reqbyte, because this is equivalent to an empty second branch. */ - - if (condcount == 1) subfirstbyte = subreqbyte = REQ_NONE; - } - - /* Handle updating of the required and first characters. Update for normal - brackets of all kinds, and conditions with two branches (see code above). - If the bracket is followed by a quantifier with zero repeat, we have to - back off. Hence the definition of zeroreqbyte and zerofirstbyte outside the - main loop so that they can be accessed for the back off. */ - - zeroreqbyte = reqbyte; - zerofirstbyte = firstbyte; - groupsetfirstbyte = FALSE; - - if (bravalue >= OP_BRA || bravalue == OP_ONCE || bravalue == OP_COND) - { - /* If we have not yet set a firstbyte in this branch, take it from the - subpattern, remembering that it was set here so that a repeat of more - than one can replicate it as reqbyte if necessary. If the subpattern has - no firstbyte, set "none" for the whole branch. In both cases, a zero - repeat forces firstbyte to "none". */ - - if (firstbyte == REQ_UNSET) - { - if (subfirstbyte >= 0) - { - firstbyte = subfirstbyte; - groupsetfirstbyte = TRUE; - } - else firstbyte = REQ_NONE; - zerofirstbyte = REQ_NONE; - } - - /* If firstbyte was previously set, convert the subpattern's firstbyte - into reqbyte if there wasn't one, using the vary flag that was in - existence beforehand. */ - - else if (subfirstbyte >= 0 && subreqbyte < 0) - subreqbyte = subfirstbyte | tempreqvary; - - /* If the subpattern set a required byte (or set a first byte that isn't - really the first byte - see above), set it. */ - - if (subreqbyte >= 0) reqbyte = subreqbyte; - } - - /* For a forward assertion, we take the reqbyte, if set. This can be - helpful if the pattern that follows the assertion doesn't set a different - char. For example, it's useful for /(?=abcde).+/. We can't set firstbyte - for an assertion, however because it leads to incorrect effect for patterns - such as /(?=a)a.+/ when the "real" "a" would then become a reqbyte instead - of a firstbyte. This is overcome by a scan at the end if there's no - firstbyte, looking for an asserted first char. */ - - else if (bravalue == OP_ASSERT && subreqbyte >= 0) reqbyte = subreqbyte; - - /* Now update the main code pointer to the end of the group. */ - - code = tempcode; - - /* Error if hit end of pattern */ - - if (*ptr != ')') - { - *errorptr = ERR14; - goto FAILED; - } - break; - - /* Check \ for being a real metacharacter; if not, fall through and handle - it as a data character at the start of a string. Escape items are checked - for validity in the pre-compiling pass. */ - - case '\\': - tempptr = ptr; - c = check_escape(&ptr, errorptr, *brackets, options, FALSE); - - /* Handle metacharacters introduced by \. For ones like \d, the ESC_ values - are arranged to be the negation of the corresponding OP_values. For the - back references, the values are ESC_REF plus the reference number. Only - back references and those types that consume a character may be repeated. - We can test for values between ESC_b and ESC_Z for the latter; this may - have to change if any new ones are ever created. */ - - if (c < 0) - { - if (-c == ESC_Q) /* Handle start of quoted string */ - { - if (ptr[1] == '\\' && ptr[2] == 'E') ptr += 2; /* avoid empty string */ - else inescq = TRUE; - continue; - } - - /* For metasequences that actually match a character, we disable the - setting of a first character if it hasn't already been set. */ - - if (firstbyte == REQ_UNSET && -c > ESC_b && -c < ESC_Z) - firstbyte = REQ_NONE; - - /* Set values to reset to if this is followed by a zero repeat. */ - - zerofirstbyte = firstbyte; - zeroreqbyte = reqbyte; - - /* Back references are handled specially */ - - if (-c >= ESC_REF) - { - int number = -c - ESC_REF; - previous = code; - *code++ = OP_REF; - PUT2INC(code, 0, number); - } - - /* So are Unicode property matches, if supported. We know that get_ucp - won't fail because it was tested in the pre-pass. */ - -#ifdef SUPPORT_UCP - else if (-c == ESC_P || -c == ESC_p) - { - BOOL negated; - int value = get_ucp(&ptr, &negated, errorptr); - previous = code; - *code++ = ((-c == ESC_p) != negated)? OP_PROP : OP_NOTPROP; - *code++ = value; - } -#endif - - /* For the rest, we can obtain the OP value by negating the escape - value */ - - else - { - previous = (-c > ESC_b && -c < ESC_Z)? code : NULL; - *code++ = -c; - } - continue; - } - - /* We have a data character whose value is in c. In UTF-8 mode it may have - a value > 127. We set its representation in the length/buffer, and then - handle it as a data character. */ - -#ifdef SUPPORT_UTF8 - if (utf8 && c > 127) - mclength = ord2utf8(c, mcbuffer); - else -#endif - - { - mcbuffer[0] = c; - mclength = 1; - } - - goto ONE_CHAR; - - /* Handle a literal character. It is guaranteed not to be whitespace or # - when the extended flag is set. If we are in UTF-8 mode, it may be a - multi-byte literal character. */ - - default: - NORMAL_CHAR: - mclength = 1; - mcbuffer[0] = c; - -#ifdef SUPPORT_UTF8 - if (utf8 && (c & 0xc0) == 0xc0) - { - while ((ptr[1] & 0xc0) == 0x80) - mcbuffer[mclength++] = *(++ptr); - } -#endif - - /* At this point we have the character's bytes in mcbuffer, and the length - in mclength. When not in UTF-8 mode, the length is always 1. */ - - ONE_CHAR: - previous = code; - *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR; - for (c = 0; c < mclength; c++) *code++ = mcbuffer[c]; - - /* Set the first and required bytes appropriately. If no previous first - byte, set it from this character, but revert to none on a zero repeat. - Otherwise, leave the firstbyte value alone, and don't change it on a zero - repeat. */ - - if (firstbyte == REQ_UNSET) - { - zerofirstbyte = REQ_NONE; - zeroreqbyte = reqbyte; - - /* If the character is more than one byte long, we can set firstbyte - only if it is not to be matched caselessly. */ - - if (mclength == 1 || req_caseopt == 0) - { - firstbyte = mcbuffer[0] | req_caseopt; - if (mclength != 1) reqbyte = code[-1] | cd->req_varyopt; - } - else firstbyte = reqbyte = REQ_NONE; - } - - /* firstbyte was previously set; we can set reqbyte only the length is - 1 or the matching is caseful. */ - - else - { - zerofirstbyte = firstbyte; - zeroreqbyte = reqbyte; - if (mclength == 1 || req_caseopt == 0) - reqbyte = code[-1] | req_caseopt | cd->req_varyopt; - } - - break; /* End of literal character handling */ - } - } /* end of big loop */ - -/* Control never reaches here by falling through, only by a goto for all the -error states. Pass back the position in the pattern so that it can be displayed -to the user for diagnosing the error. */ - -FAILED: -*ptrptr = ptr; -return FALSE; -} - - - - -/************************************************* -* Compile sequence of alternatives * -*************************************************/ - -/* On entry, ptr is pointing past the bracket character, but on return -it points to the closing bracket, or vertical bar, or end of string. -The code variable is pointing at the byte into which the BRA operator has been -stored. If the ims options are changed at the start (for a (?ims: group) or -during any branch, we need to insert an OP_OPT item at the start of every -following branch to ensure they get set correctly at run time, and also pass -the new options into every subsequent branch compile. - -Argument: - options option bits, including any changes for this subpattern - oldims previous settings of ims option bits - brackets -> int containing the number of extracting brackets used - codeptr -> the address of the current code pointer - ptrptr -> the address of the current pattern pointer - errorptr -> pointer to error message - lookbehind TRUE if this is a lookbehind assertion - skipbytes skip this many bytes at start (for OP_COND, OP_BRANUMBER) - firstbyteptr place to put the first required character, or a negative number - reqbyteptr place to put the last required character, or a negative number - bcptr pointer to the chain of currently open branches - cd points to the data block with tables pointers etc. - -Returns: TRUE on success -*/ - -static BOOL -compile_regex(int options, int oldims, int *brackets, uschar **codeptr, - const uschar **ptrptr, const char **errorptr, BOOL lookbehind, int skipbytes, - int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd) -{ -const uschar *ptr = *ptrptr; -uschar *code = *codeptr; -uschar *last_branch = code; -uschar *start_bracket = code; -uschar *reverse_count = NULL; -int firstbyte, reqbyte; -int branchfirstbyte, branchreqbyte; -branch_chain bc; - -bc.outer = bcptr; -bc.current = code; - -firstbyte = reqbyte = REQ_UNSET; - -/* Offset is set zero to mark that this bracket is still open */ - -PUT(code, 1, 0); -code += 1 + LINK_SIZE + skipbytes; - -/* Loop for each alternative branch */ - -for (;;) - { - /* Handle a change of ims options at the start of the branch */ - - if ((options & PCRE_IMS) != oldims) - { - *code++ = OP_OPT; - *code++ = options & PCRE_IMS; - } - - /* Set up dummy OP_REVERSE if lookbehind assertion */ - - if (lookbehind) - { - *code++ = OP_REVERSE; - reverse_count = code; - PUTINC(code, 0, 0); - } - - /* Now compile the branch */ - - if (!compile_branch(&options, brackets, &code, &ptr, errorptr, - &branchfirstbyte, &branchreqbyte, &bc, cd)) - { - *ptrptr = ptr; - return FALSE; - } - - /* If this is the first branch, the firstbyte and reqbyte values for the - branch become the values for the regex. */ - - if (*last_branch != OP_ALT) - { - firstbyte = branchfirstbyte; - reqbyte = branchreqbyte; - } - - /* If this is not the first branch, the first char and reqbyte have to - match the values from all the previous branches, except that if the previous - value for reqbyte didn't have REQ_VARY set, it can still match, and we set - REQ_VARY for the regex. */ - - else - { - /* If we previously had a firstbyte, but it doesn't match the new branch, - we have to abandon the firstbyte for the regex, but if there was previously - no reqbyte, it takes on the value of the old firstbyte. */ - - if (firstbyte >= 0 && firstbyte != branchfirstbyte) - { - if (reqbyte < 0) reqbyte = firstbyte; - firstbyte = REQ_NONE; - } - - /* If we (now or from before) have no firstbyte, a firstbyte from the - branch becomes a reqbyte if there isn't a branch reqbyte. */ - - if (firstbyte < 0 && branchfirstbyte >= 0 && branchreqbyte < 0) - branchreqbyte = branchfirstbyte; - - /* Now ensure that the reqbytes match */ - - if ((reqbyte & ~REQ_VARY) != (branchreqbyte & ~REQ_VARY)) - reqbyte = REQ_NONE; - else reqbyte |= branchreqbyte; /* To "or" REQ_VARY */ - } - - /* If lookbehind, check that this branch matches a fixed-length string, - and put the length into the OP_REVERSE item. Temporarily mark the end of - the branch with OP_END. */ - - if (lookbehind) - { - int length; - *code = OP_END; - length = find_fixedlength(last_branch, options); - DPRINTF(("fixed length = %d\n", length)); - if (length < 0) - { - *errorptr = (length == -2)? ERR36 : ERR25; - *ptrptr = ptr; - return FALSE; - } - PUT(reverse_count, 0, length); - } - - /* Reached end of expression, either ')' or end of pattern. Go back through - the alternative branches and reverse the chain of offsets, with the field in - the BRA item now becoming an offset to the first alternative. If there are - no alternatives, it points to the end of the group. The length in the - terminating ket is always the length of the whole bracketed item. If any of - the ims options were changed inside the group, compile a resetting op-code - following, except at the very end of the pattern. Return leaving the pointer - at the terminating char. */ - - if (*ptr != '|') - { - int length = code - last_branch; - do - { - int prev_length = GET(last_branch, 1); - PUT(last_branch, 1, length); - length = prev_length; - last_branch -= length; - } - while (length > 0); - - /* Fill in the ket */ - - *code = OP_KET; - PUT(code, 1, code - start_bracket); - code += 1 + LINK_SIZE; - - /* Resetting option if needed */ - - if ((options & PCRE_IMS) != oldims && *ptr == ')') - { - *code++ = OP_OPT; - *code++ = oldims; - } - - /* Set values to pass back */ - - *codeptr = code; - *ptrptr = ptr; - *firstbyteptr = firstbyte; - *reqbyteptr = reqbyte; - return TRUE; - } - - /* Another branch follows; insert an "or" node. Its length field points back - to the previous branch while the bracket remains open. At the end the chain - is reversed. It's done like this so that the start of the bracket has a - zero offset until it is closed, making it possible to detect recursion. */ - - *code = OP_ALT; - PUT(code, 1, code - last_branch); - bc.current = last_branch = code; - code += 1 + LINK_SIZE; - ptr++; - } -/* Control never reaches here */ -} - - - - -/************************************************* -* Check for anchored expression * -*************************************************/ - -/* Try to find out if this is an anchored regular expression. Consider each -alternative branch. If they all start with OP_SOD or OP_CIRC, or with a bracket -all of whose alternatives start with OP_SOD or OP_CIRC (recurse ad lib), then -it's anchored. However, if this is a multiline pattern, then only OP_SOD -counts, since OP_CIRC can match in the middle. - -We can also consider a regex to be anchored if OP_SOM starts all its branches. -This is the code for \G, which means "match at start of match position, taking -into account the match offset". - -A branch is also implicitly anchored if it starts with .* and DOTALL is set, -because that will try the rest of the pattern at all possible matching points, -so there is no point trying again.... er .... - -.... except when the .* appears inside capturing parentheses, and there is a -subsequent back reference to those parentheses. We haven't enough information -to catch that case precisely. - -At first, the best we could do was to detect when .* was in capturing brackets -and the highest back reference was greater than or equal to that level. -However, by keeping a bitmap of the first 31 back references, we can catch some -of the more common cases more precisely. - -Arguments: - code points to start of expression (the bracket) - options points to the options setting - bracket_map a bitmap of which brackets we are inside while testing; this - handles up to substring 31; after that we just have to take - the less precise approach - backref_map the back reference bitmap - -Returns: TRUE or FALSE -*/ - -static BOOL -is_anchored(register const uschar *code, int *options, unsigned int bracket_map, - unsigned int backref_map) -{ -do { - const uschar *scode = - first_significant_code(code + 1+LINK_SIZE, options, PCRE_MULTILINE, FALSE); - register int op = *scode; - - /* Capturing brackets */ - - if (op > OP_BRA) - { - int new_map; - op -= OP_BRA; - if (op > EXTRACT_BASIC_MAX) op = GET2(scode, 2+LINK_SIZE); - new_map = bracket_map | ((op < 32)? (1 << op) : 1); - if (!is_anchored(scode, options, new_map, backref_map)) return FALSE; - } - - /* Other brackets */ - - else if (op == OP_BRA || op == OP_ASSERT || op == OP_ONCE || op == OP_COND) - { - if (!is_anchored(scode, options, bracket_map, backref_map)) return FALSE; - } - - /* .* is not anchored unless DOTALL is set and it isn't in brackets that - are or may be referenced. */ - - else if ((op == OP_TYPESTAR || op == OP_TYPEMINSTAR) && - (*options & PCRE_DOTALL) != 0) - { - if (scode[1] != OP_ANY || (bracket_map & backref_map) != 0) return FALSE; - } - - /* Check for explicit anchoring */ - - else if (op != OP_SOD && op != OP_SOM && - ((*options & PCRE_MULTILINE) != 0 || op != OP_CIRC)) - return FALSE; - code += GET(code, 1); - } -while (*code == OP_ALT); /* Loop for each alternative */ -return TRUE; -} - - - -/************************************************* -* Check for starting with ^ or .* * -*************************************************/ - -/* This is called to find out if every branch starts with ^ or .* so that -"first char" processing can be done to speed things up in multiline -matching and for non-DOTALL patterns that start with .* (which must start at -the beginning or after \n). As in the case of is_anchored() (see above), we -have to take account of back references to capturing brackets that contain .* -because in that case we can't make the assumption. - -Arguments: - code points to start of expression (the bracket) - bracket_map a bitmap of which brackets we are inside while testing; this - handles up to substring 31; after that we just have to take - the less precise approach - backref_map the back reference bitmap - -Returns: TRUE or FALSE -*/ - -static BOOL -is_startline(const uschar *code, unsigned int bracket_map, - unsigned int backref_map) -{ -do { - const uschar *scode = first_significant_code(code + 1+LINK_SIZE, NULL, 0, - FALSE); - register int op = *scode; - - /* Capturing brackets */ - - if (op > OP_BRA) - { - int new_map; - op -= OP_BRA; - if (op > EXTRACT_BASIC_MAX) op = GET2(scode, 2+LINK_SIZE); - new_map = bracket_map | ((op < 32)? (1 << op) : 1); - if (!is_startline(scode, new_map, backref_map)) return FALSE; - } - - /* Other brackets */ - - else if (op == OP_BRA || op == OP_ASSERT || op == OP_ONCE || op == OP_COND) - { if (!is_startline(scode, bracket_map, backref_map)) return FALSE; } - - /* .* means "start at start or after \n" if it isn't in brackets that - may be referenced. */ - - else if (op == OP_TYPESTAR || op == OP_TYPEMINSTAR) - { - if (scode[1] != OP_ANY || (bracket_map & backref_map) != 0) return FALSE; - } - - /* Check for explicit circumflex */ - - else if (op != OP_CIRC) return FALSE; - - /* Move on to the next alternative */ - - code += GET(code, 1); - } -while (*code == OP_ALT); /* Loop for each alternative */ -return TRUE; -} - - - -/************************************************* -* Check for asserted fixed first char * -*************************************************/ - -/* During compilation, the "first char" settings from forward assertions are -discarded, because they can cause conflicts with actual literals that follow. -However, if we end up without a first char setting for an unanchored pattern, -it is worth scanning the regex to see if there is an initial asserted first -char. If all branches start with the same asserted char, or with a bracket all -of whose alternatives start with the same asserted char (recurse ad lib), then -we return that char, otherwise -1. - -Arguments: - code points to start of expression (the bracket) - options pointer to the options (used to check casing changes) - inassert TRUE if in an assertion - -Returns: -1 or the fixed first char -*/ - -static int -find_firstassertedchar(const uschar *code, int *options, BOOL inassert) -{ -register int c = -1; -do { - int d; - const uschar *scode = - first_significant_code(code + 1+LINK_SIZE, options, PCRE_CASELESS, TRUE); - register int op = *scode; - - if (op >= OP_BRA) op = OP_BRA; - - switch(op) - { - default: - return -1; - - case OP_BRA: - case OP_ASSERT: - case OP_ONCE: - case OP_COND: - if ((d = find_firstassertedchar(scode, options, op == OP_ASSERT)) < 0) - return -1; - if (c < 0) c = d; else if (c != d) return -1; - break; - - case OP_EXACT: /* Fall through */ - scode += 2; - - case OP_CHAR: - case OP_CHARNC: - case OP_PLUS: - case OP_MINPLUS: - if (!inassert) return -1; - if (c < 0) - { - c = scode[1]; - if ((*options & PCRE_CASELESS) != 0) c |= REQ_CASELESS; - } - else if (c != scode[1]) return -1; - break; - } - - code += GET(code, 1); - } -while (*code == OP_ALT); -return c; -} - - - - -#ifdef SUPPORT_UTF8 -/************************************************* -* Validate a UTF-8 string * -*************************************************/ - -/* This function is called (optionally) at the start of compile or match, to -validate that a supposed UTF-8 string is actually valid. The early check means -that subsequent code can assume it is dealing with a valid string. The check -can be turned off for maximum performance, but then consequences of supplying -an invalid string are then undefined. - -Arguments: - string points to the string - length length of string, or -1 if the string is zero-terminated - -Returns: < 0 if the string is a valid UTF-8 string - >= 0 otherwise; the value is the offset of the bad byte -*/ - -static int -valid_utf8(const uschar *string, int length) -{ -register const uschar *p; - -if (length < 0) - { - for (p = string; *p != 0; p++); - length = p - string; - } - -for (p = string; length-- > 0; p++) - { - register int ab; - register int c = *p; - if (c < 128) continue; - if ((c & 0xc0) != 0xc0) return p - string; - ab = utf8_table4[c & 0x3f]; /* Number of additional bytes */ - if (length < ab) return p - string; - length -= ab; - - /* Check top bits in the second byte */ - if ((*(++p) & 0xc0) != 0x80) return p - string; - - /* Check for overlong sequences for each different length */ - switch (ab) - { - /* Check for xx00 000x */ - case 1: - if ((c & 0x3e) == 0) return p - string; - continue; /* We know there aren't any more bytes to check */ - - /* Check for 1110 0000, xx0x xxxx */ - case 2: - if (c == 0xe0 && (*p & 0x20) == 0) return p - string; - break; - - /* Check for 1111 0000, xx00 xxxx */ - case 3: - if (c == 0xf0 && (*p & 0x30) == 0) return p - string; - break; - - /* Check for 1111 1000, xx00 0xxx */ - case 4: - if (c == 0xf8 && (*p & 0x38) == 0) return p - string; - break; - - /* Check for leading 0xfe or 0xff, and then for 1111 1100, xx00 00xx */ - case 5: - if (c == 0xfe || c == 0xff || - (c == 0xfc && (*p & 0x3c) == 0)) return p - string; - break; - } - - /* Check for valid bytes after the 2nd, if any; all must start 10 */ - while (--ab > 0) - { - if ((*(++p) & 0xc0) != 0x80) return p - string; - } - } - -return -1; -} -#endif - - - -/************************************************* -* Compile a Regular Expression * -*************************************************/ - -/* This function takes a string and returns a pointer to a block of store -holding a compiled version of the expression. - -Arguments: - pattern the regular expression - options various option bits - errorptr pointer to pointer to error text - erroroffset ptr offset in pattern where error was detected - tables pointer to character tables or NULL - -Returns: pointer to compiled data block, or NULL on error, - with errorptr and erroroffset set -*/ - -EXPORT pcre * -pcre_compile(const char *pattern, int options, const char **errorptr, - int *erroroffset, const unsigned char *tables) -{ -real_pcre *re; -int length = 1 + LINK_SIZE; /* For initial BRA plus length */ -int c, firstbyte, reqbyte; -int bracount = 0; -int branch_extra = 0; -int branch_newextra; -int item_count = -1; -int name_count = 0; -int max_name_size = 0; -int lastitemlength = 0; -#ifdef SUPPORT_UTF8 -BOOL utf8; -BOOL class_utf8; -#endif -BOOL inescq = FALSE; -unsigned int brastackptr = 0; -size_t size; -uschar *code; -const uschar *codestart; -const uschar *ptr; -compile_data compile_block; -int brastack[BRASTACK_SIZE]; -uschar bralenstack[BRASTACK_SIZE]; - -/* We can't pass back an error message if errorptr is NULL; I guess the best we -can do is just return NULL. */ - -if (errorptr == NULL) return NULL; -*errorptr = NULL; - -/* However, we can give a message for this error */ - -if (erroroffset == NULL) - { - *errorptr = ERR16; - return NULL; - } -*erroroffset = 0; - -/* Can't support UTF8 unless PCRE has been compiled to include the code. */ - -#ifdef SUPPORT_UTF8 -utf8 = (options & PCRE_UTF8) != 0; -if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0 && - (*erroroffset = valid_utf8((uschar *)pattern, -1)) >= 0) - { - *errorptr = ERR44; - return NULL; - } -#else -if ((options & PCRE_UTF8) != 0) - { - *errorptr = ERR32; - return NULL; - } -#endif - -if ((options & ~PUBLIC_OPTIONS) != 0) - { - *errorptr = ERR17; - return NULL; - } - -/* Set up pointers to the individual character tables */ - -if (tables == NULL) tables = pcre_default_tables; -compile_block.lcc = tables + lcc_offset; -compile_block.fcc = tables + fcc_offset; -compile_block.cbits = tables + cbits_offset; -compile_block.ctypes = tables + ctypes_offset; - -/* Maximum back reference and backref bitmap. This is updated for numeric -references during the first pass, but for named references during the actual -compile pass. The bitmap records up to 31 back references to help in deciding -whether (.*) can be treated as anchored or not. */ - -compile_block.top_backref = 0; -compile_block.backref_map = 0; - -/* Reflect pattern for debugging output */ - -DPRINTF(("------------------------------------------------------------------\n")); -DPRINTF(("%s\n", pattern)); - -/* The first thing to do is to make a pass over the pattern to compute the -amount of store required to hold the compiled code. This does not have to be -perfect as long as errors are overestimates. At the same time we can detect any -flag settings right at the start, and extract them. Make an attempt to correct -for any counted white space if an "extended" flag setting appears late in the -pattern. We can't be so clever for #-comments. */ - -ptr = (const uschar *)(pattern - 1); -while ((c = *(++ptr)) != 0) - { - int min, max; - int class_optcount; - int bracket_length; - int duplength; - - /* If we are inside a \Q...\E sequence, all chars are literal */ - - if (inescq) - { - if ((options & PCRE_AUTO_CALLOUT) != 0) length += 2 + 2*LINK_SIZE; - goto NORMAL_CHAR; - } - - /* Otherwise, first check for ignored whitespace and comments */ - - if ((options & PCRE_EXTENDED) != 0) - { - if ((compile_block.ctypes[c] & ctype_space) != 0) continue; - if (c == '#') - { - /* The space before the ; is to avoid a warning on a silly compiler - on the Macintosh. */ - while ((c = *(++ptr)) != 0 && c != NEWLINE) ; - if (c == 0) break; - continue; - } - } - - item_count++; /* Is zero for the first non-comment item */ - - /* Allow space for auto callout before every item except quantifiers. */ - - if ((options & PCRE_AUTO_CALLOUT) != 0 && - c != '*' && c != '+' && c != '?' && - (c != '{' || !is_counted_repeat(ptr + 1))) - length += 2 + 2*LINK_SIZE; - - switch(c) - { - /* A backslashed item may be an escaped data character or it may be a - character type. */ - - case '\\': - c = check_escape(&ptr, errorptr, bracount, options, FALSE); - if (*errorptr != NULL) goto PCRE_ERROR_RETURN; - - lastitemlength = 1; /* Default length of last item for repeats */ - - if (c >= 0) /* Data character */ - { - length += 2; /* For a one-byte character */ - -#ifdef SUPPORT_UTF8 - if (utf8 && c > 127) - { - int i; - for (i = 0; i < sizeof(utf8_table1)/sizeof(int); i++) - if (c <= utf8_table1[i]) break; - length += i; - lastitemlength += i; - } -#endif - - continue; - } - - /* If \Q, enter "literal" mode */ - - if (-c == ESC_Q) - { - inescq = TRUE; - continue; - } - - /* \X is supported only if Unicode property support is compiled */ - -#ifndef SUPPORT_UCP - if (-c == ESC_X) - { - *errorptr = ERR45; - goto PCRE_ERROR_RETURN; - } -#endif - - /* \P and \p are for Unicode properties, but only when the support has - been compiled. Each item needs 2 bytes. */ - - else if (-c == ESC_P || -c == ESC_p) - { -#ifdef SUPPORT_UCP - BOOL negated; - length += 2; - lastitemlength = 2; - if (get_ucp(&ptr, &negated, errorptr) < 0) goto PCRE_ERROR_RETURN; - continue; -#else - *errorptr = ERR45; - goto PCRE_ERROR_RETURN; -#endif - } - - /* Other escapes need one byte */ - - length++; - - /* A back reference needs an additional 2 bytes, plus either one or 5 - bytes for a repeat. We also need to keep the value of the highest - back reference. */ - - if (c <= -ESC_REF) - { - int refnum = -c - ESC_REF; - compile_block.backref_map |= (refnum < 32)? (1 << refnum) : 1; - if (refnum > compile_block.top_backref) - compile_block.top_backref = refnum; - length += 2; /* For single back reference */ - if (ptr[1] == '{' && is_counted_repeat(ptr+2)) - { - ptr = read_repeat_counts(ptr+2, &min, &max, errorptr); - if (*errorptr != NULL) goto PCRE_ERROR_RETURN; - if ((min == 0 && (max == 1 || max == -1)) || - (min == 1 && max == -1)) - length++; - else length += 5; - if (ptr[1] == '?') ptr++; - } - } - continue; - - case '^': /* Single-byte metacharacters */ - case '.': - case '$': - length++; - lastitemlength = 1; - continue; - - case '*': /* These repeats won't be after brackets; */ - case '+': /* those are handled separately */ - case '?': - length++; - goto POSESSIVE; /* A few lines below */ - - /* This covers the cases of braced repeats after a single char, metachar, - class, or back reference. */ - - case '{': - if (!is_counted_repeat(ptr+1)) goto NORMAL_CHAR; - ptr = read_repeat_counts(ptr+1, &min, &max, errorptr); - if (*errorptr != NULL) goto PCRE_ERROR_RETURN; - - /* These special cases just insert one extra opcode */ - - if ((min == 0 && (max == 1 || max == -1)) || - (min == 1 && max == -1)) - length++; - - /* These cases might insert additional copies of a preceding character. */ - - else - { - if (min != 1) - { - length -= lastitemlength; /* Uncount the original char or metachar */ - if (min > 0) length += 3 + lastitemlength; - } - length += lastitemlength + ((max > 0)? 3 : 1); - } - - if (ptr[1] == '?') ptr++; /* Needs no extra length */ - - POSESSIVE: /* Test for possessive quantifier */ - if (ptr[1] == '+') - { - ptr++; - length += 2 + 2*LINK_SIZE; /* Allow for atomic brackets */ - } - continue; - - /* An alternation contains an offset to the next branch or ket. If any ims - options changed in the previous branch(es), and/or if we are in a - lookbehind assertion, extra space will be needed at the start of the - branch. This is handled by branch_extra. */ - - case '|': - length += 1 + LINK_SIZE + branch_extra; - continue; - - /* A character class uses 33 characters provided that all the character - values are less than 256. Otherwise, it uses a bit map for low valued - characters, and individual items for others. Don't worry about character - types that aren't allowed in classes - they'll get picked up during the - compile. A character class that contains only one single-byte character - uses 2 or 3 bytes, depending on whether it is negated or not. Notice this - where we can. (In UTF-8 mode we can do this only for chars < 128.) */ - - case '[': - if (*(++ptr) == '^') - { - class_optcount = 10; /* Greater than one */ - ptr++; - } - else class_optcount = 0; - -#ifdef SUPPORT_UTF8 - class_utf8 = FALSE; -#endif - - /* Written as a "do" so that an initial ']' is taken as data */ - - if (*ptr != 0) do - { - /* Inside \Q...\E everything is literal except \E */ - - if (inescq) - { - if (*ptr != '\\' || ptr[1] != 'E') goto GET_ONE_CHARACTER; - inescq = FALSE; - ptr += 1; - continue; - } - - /* Outside \Q...\E, check for escapes */ - - if (*ptr == '\\') - { - c = check_escape(&ptr, errorptr, bracount, options, TRUE); - if (*errorptr != NULL) goto PCRE_ERROR_RETURN; - - /* \b is backspace inside a class; \X is literal */ - - if (-c == ESC_b) c = '\b'; - else if (-c == ESC_X) c = 'X'; - - /* \Q enters quoting mode */ - - else if (-c == ESC_Q) - { - inescq = TRUE; - continue; - } - - /* Handle escapes that turn into characters */ - - if (c >= 0) goto NON_SPECIAL_CHARACTER; - - /* Escapes that are meta-things. The normal ones just affect the - bit map, but Unicode properties require an XCLASS extended item. */ - - else - { - class_optcount = 10; /* \d, \s etc; make sure > 1 */ -#ifdef SUPPORT_UTF8 - if (-c == ESC_p || -c == ESC_P) - { - if (!class_utf8) - { - class_utf8 = TRUE; - length += LINK_SIZE + 2; - } - length += 2; - } -#endif - } - } - - /* Check the syntax for POSIX stuff. The bits we actually handle are - checked during the real compile phase. */ - - else if (*ptr == '[' && check_posix_syntax(ptr, &ptr, &compile_block)) - { - ptr++; - class_optcount = 10; /* Make sure > 1 */ - } - - /* Anything else increments the possible optimization count. We have to - detect ranges here so that we can compute the number of extra ranges for - caseless wide characters when UCP support is available. If there are wide - characters, we are going to have to use an XCLASS, even for single - characters. */ - - else - { - int d; - - GET_ONE_CHARACTER: - -#ifdef SUPPORT_UTF8 - if (utf8) - { - int extra = 0; - GETCHARLEN(c, ptr, extra); - ptr += extra; - } - else c = *ptr; -#else - c = *ptr; -#endif - - /* Come here from handling \ above when it escapes to a char value */ - - NON_SPECIAL_CHARACTER: - class_optcount++; - - d = -1; - if (ptr[1] == '-') - { - uschar const *hyptr = ptr++; - if (ptr[1] == '\\') - { - ptr++; - d = check_escape(&ptr, errorptr, bracount, options, TRUE); - if (*errorptr != NULL) goto PCRE_ERROR_RETURN; - if (-d == ESC_b) d = '\b'; /* backspace */ - else if (-d == ESC_X) d = 'X'; /* literal X in a class */ - } - else if (ptr[1] != 0 && ptr[1] != ']') - { - ptr++; -#ifdef SUPPORT_UTF8 - if (utf8) - { - int extra = 0; - GETCHARLEN(d, ptr, extra); - ptr += extra; - } - else -#endif - d = *ptr; - } - if (d < 0) ptr = hyptr; /* go back to hyphen as data */ - } - - /* If d >= 0 we have a range. In UTF-8 mode, if the end is > 255, or > - 127 for caseless matching, we will need to use an XCLASS. */ - - if (d >= 0) - { - class_optcount = 10; /* Ensure > 1 */ - if (d < c) - { - *errorptr = ERR8; - goto PCRE_ERROR_RETURN; - } - -#ifdef SUPPORT_UTF8 - if (utf8 && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127))) - { - uschar buffer[6]; - if (!class_utf8) /* Allow for XCLASS overhead */ - { - class_utf8 = TRUE; - length += LINK_SIZE + 2; - } - -#ifdef SUPPORT_UCP - /* If we have UCP support, find out how many extra ranges are - needed to map the other case of characters within this range. We - have to mimic the range optimization here, because extending the - range upwards might push d over a boundary that makes is use - another byte in the UTF-8 representation. */ - - if ((options & PCRE_CASELESS) != 0) - { - int occ, ocd; - int cc = c; - int origd = d; - while (get_othercase_range(&cc, origd, &occ, &ocd)) - { - if (occ >= c && ocd <= d) continue; /* Skip embedded */ - - if (occ < c && ocd >= c - 1) /* Extend the basic range */ - { /* if there is overlap, */ - c = occ; /* noting that if occ < c */ - continue; /* we can't have ocd > d */ - } /* because a subrange is */ - if (ocd > d && occ <= d + 1) /* always shorter than */ - { /* the basic range. */ - d = ocd; - continue; - } - - /* An extra item is needed */ - - length += 1 + ord2utf8(occ, buffer) + - ((occ == ocd)? 0 : ord2utf8(ocd, buffer)); - } - } -#endif /* SUPPORT_UCP */ - - /* The length of the (possibly extended) range */ - - length += 1 + ord2utf8(c, buffer) + ord2utf8(d, buffer); - } -#endif /* SUPPORT_UTF8 */ - - } - - /* We have a single character. There is nothing to be done unless we - are in UTF-8 mode. If the char is > 255, or 127 when caseless, we must - allow for an XCL_SINGLE item, doubled for caselessness if there is UCP - support. */ - - else - { -#ifdef SUPPORT_UTF8 - if (utf8 && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127))) - { - uschar buffer[6]; - class_optcount = 10; /* Ensure > 1 */ - if (!class_utf8) /* Allow for XCLASS overhead */ - { - class_utf8 = TRUE; - length += LINK_SIZE + 2; - } -#ifdef SUPPORT_UCP - length += (((options & PCRE_CASELESS) != 0)? 2 : 1) * - (1 + ord2utf8(c, buffer)); -#else /* SUPPORT_UCP */ - length += 1 + ord2utf8(c, buffer); -#endif /* SUPPORT_UCP */ - } -#endif /* SUPPORT_UTF8 */ - } - } - } - while (*(++ptr) != 0 && (inescq || *ptr != ']')); /* Concludes "do" above */ - - if (*ptr == 0) /* Missing terminating ']' */ - { - *errorptr = ERR6; - goto PCRE_ERROR_RETURN; - } - - /* We can optimize when there was only one optimizable character. Repeats - for positive and negated single one-byte chars are handled by the general - code. Here, we handle repeats for the class opcodes. */ - - if (class_optcount == 1) length += 3; else - { - length += 33; - - /* A repeat needs either 1 or 5 bytes. If it is a possessive quantifier, - we also need extra for wrapping the whole thing in a sub-pattern. */ - - if (*ptr != 0 && ptr[1] == '{' && is_counted_repeat(ptr+2)) - { - ptr = read_repeat_counts(ptr+2, &min, &max, errorptr); - if (*errorptr != NULL) goto PCRE_ERROR_RETURN; - if ((min == 0 && (max == 1 || max == -1)) || - (min == 1 && max == -1)) - length++; - else length += 5; - if (ptr[1] == '+') - { - ptr++; - length += 2 + 2*LINK_SIZE; - } - else if (ptr[1] == '?') ptr++; - } - } - continue; - - /* Brackets may be genuine groups or special things */ - - case '(': - branch_newextra = 0; - bracket_length = 1 + LINK_SIZE; - - /* Handle special forms of bracket, which all start (? */ - - if (ptr[1] == '?') - { - int set, unset; - int *optset; - - switch (c = ptr[2]) - { - /* Skip over comments entirely */ - case '#': - ptr += 3; - while (*ptr != 0 && *ptr != ')') ptr++; - if (*ptr == 0) - { - *errorptr = ERR18; - goto PCRE_ERROR_RETURN; - } - continue; - - /* Non-referencing groups and lookaheads just move the pointer on, and - then behave like a non-special bracket, except that they don't increment - the count of extracting brackets. Ditto for the "once only" bracket, - which is in Perl from version 5.005. */ - - case ':': - case '=': - case '!': - case '>': - ptr += 2; - break; - - /* (?R) specifies a recursive call to the regex, which is an extension - to provide the facility which can be obtained by (?p{perl-code}) in - Perl 5.6. In Perl 5.8 this has become (??{perl-code}). - - From PCRE 4.00, items such as (?3) specify subroutine-like "calls" to - the appropriate numbered brackets. This includes both recursive and - non-recursive calls. (?R) is now synonymous with (?0). */ - - case 'R': - ptr++; - - case '0': case '1': case '2': case '3': case '4': - case '5': case '6': case '7': case '8': case '9': - ptr += 2; - if (c != 'R') - while ((digitab[*(++ptr)] & ctype_digit) != 0); - if (*ptr != ')') - { - *errorptr = ERR29; - goto PCRE_ERROR_RETURN; - } - length += 1 + LINK_SIZE; - - /* If this item is quantified, it will get wrapped inside brackets so - as to use the code for quantified brackets. We jump down and use the - code that handles this for real brackets. */ - - if (ptr[1] == '+' || ptr[1] == '*' || ptr[1] == '?' || ptr[1] == '{') - { - length += 2 + 2 * LINK_SIZE; /* to make bracketed */ - duplength = 5 + 3 * LINK_SIZE; - goto HANDLE_QUANTIFIED_BRACKETS; - } - continue; - - /* (?C) is an extension which provides "callout" - to provide a bit of - the functionality of the Perl (?{...}) feature. An optional number may - follow (default is zero). */ - - case 'C': - ptr += 2; - while ((digitab[*(++ptr)] & ctype_digit) != 0); - if (*ptr != ')') - { - *errorptr = ERR39; - goto PCRE_ERROR_RETURN; - } - length += 2 + 2*LINK_SIZE; - continue; - - /* Named subpatterns are an extension copied from Python */ - - case 'P': - ptr += 3; - if (*ptr == '<') - { - const uschar *p; /* Don't amalgamate; some compilers */ - p = ++ptr; /* grumble at autoincrement in declaration */ - while ((compile_block.ctypes[*ptr] & ctype_word) != 0) ptr++; - if (*ptr != '>') - { - *errorptr = ERR42; - goto PCRE_ERROR_RETURN; - } - name_count++; - if (ptr - p > max_name_size) max_name_size = (ptr - p); - break; - } - - if (*ptr == '=' || *ptr == '>') - { - while ((compile_block.ctypes[*(++ptr)] & ctype_word) != 0); - if (*ptr != ')') - { - *errorptr = ERR42; - goto PCRE_ERROR_RETURN; - } - break; - } - - /* Unknown character after (?P */ - - *errorptr = ERR41; - goto PCRE_ERROR_RETURN; - - /* Lookbehinds are in Perl from version 5.005 */ - - case '<': - ptr += 3; - if (*ptr == '=' || *ptr == '!') - { - branch_newextra = 1 + LINK_SIZE; - length += 1 + LINK_SIZE; /* For the first branch */ - break; - } - *errorptr = ERR24; - goto PCRE_ERROR_RETURN; - - /* Conditionals are in Perl from version 5.005. The bracket must either - be followed by a number (for bracket reference) or by an assertion - group, or (a PCRE extension) by 'R' for a recursion test. */ - - case '(': - if (ptr[3] == 'R' && ptr[4] == ')') - { - ptr += 4; - length += 3; - } - else if ((digitab[ptr[3]] & ctype_digit) != 0) - { - ptr += 4; - length += 3; - while ((digitab[*ptr] & ctype_digit) != 0) ptr++; - if (*ptr != ')') - { - *errorptr = ERR26; - goto PCRE_ERROR_RETURN; - } - } - else /* An assertion must follow */ - { - ptr++; /* Can treat like ':' as far as spacing is concerned */ - if (ptr[2] != '?' || - (ptr[3] != '=' && ptr[3] != '!' && ptr[3] != '<') ) - { - ptr += 2; /* To get right offset in message */ - *errorptr = ERR28; - goto PCRE_ERROR_RETURN; - } - } - break; - - /* Else loop checking valid options until ) is met. Anything else is an - error. If we are without any brackets, i.e. at top level, the settings - act as if specified in the options, so massage the options immediately. - This is for backward compatibility with Perl 5.004. */ - - default: - set = unset = 0; - optset = &set; - ptr += 2; - - for (;; ptr++) - { - c = *ptr; - switch (c) - { - case 'i': - *optset |= PCRE_CASELESS; - continue; - - case 'm': - *optset |= PCRE_MULTILINE; - continue; - - case 's': - *optset |= PCRE_DOTALL; - continue; - - case 'x': - *optset |= PCRE_EXTENDED; - continue; - - case 'X': - *optset |= PCRE_EXTRA; - continue; - - case 'U': - *optset |= PCRE_UNGREEDY; - continue; - - case '-': - optset = &unset; - continue; - - /* A termination by ')' indicates an options-setting-only item; if - this is at the very start of the pattern (indicated by item_count - being zero), we use it to set the global options. This is helpful - when analyzing the pattern for first characters, etc. Otherwise - nothing is done here and it is handled during the compiling - process. - - [Historical note: Up to Perl 5.8, options settings at top level - were always global settings, wherever they appeared in the pattern. - That is, they were equivalent to an external setting. From 5.8 - onwards, they apply only to what follows (which is what you might - expect).] */ - - case ')': - if (item_count == 0) - { - options = (options | set) & (~unset); - set = unset = 0; /* To save length */ - item_count--; /* To allow for several */ - } - - /* Fall through */ - - /* A termination by ':' indicates the start of a nested group with - the given options set. This is again handled at compile time, but - we must allow for compiled space if any of the ims options are - set. We also have to allow for resetting space at the end of - the group, which is why 4 is added to the length and not just 2. - If there are several changes of options within the same group, this - will lead to an over-estimate on the length, but this shouldn't - matter very much. We also have to allow for resetting options at - the start of any alternations, which we do by setting - branch_newextra to 2. Finally, we record whether the case-dependent - flag ever changes within the regex. This is used by the "required - character" code. */ - - case ':': - if (((set|unset) & PCRE_IMS) != 0) - { - length += 4; - branch_newextra = 2; - if (((set|unset) & PCRE_CASELESS) != 0) options |= PCRE_ICHANGED; - } - goto END_OPTIONS; - - /* Unrecognized option character */ - - default: - *errorptr = ERR12; - goto PCRE_ERROR_RETURN; - } - } - - /* If we hit a closing bracket, that's it - this is a freestanding - option-setting. We need to ensure that branch_extra is updated if - necessary. The only values branch_newextra can have here are 0 or 2. - If the value is 2, then branch_extra must either be 2 or 5, depending - on whether this is a lookbehind group or not. */ - - END_OPTIONS: - if (c == ')') - { - if (branch_newextra == 2 && - (branch_extra == 0 || branch_extra == 1+LINK_SIZE)) - branch_extra += branch_newextra; - continue; - } - - /* If options were terminated by ':' control comes here. Fall through - to handle the group below. */ - } - } - - /* Extracting brackets must be counted so we can process escapes in a - Perlish way. If the number exceeds EXTRACT_BASIC_MAX we are going to - need an additional 3 bytes of store per extracting bracket. However, if - PCRE_NO_AUTO)CAPTURE is set, unadorned brackets become non-capturing, so we - must leave the count alone (it will aways be zero). */ - - else if ((options & PCRE_NO_AUTO_CAPTURE) == 0) - { - bracount++; - if (bracount > EXTRACT_BASIC_MAX) bracket_length += 3; - } - - /* Save length for computing whole length at end if there's a repeat that - requires duplication of the group. Also save the current value of - branch_extra, and start the new group with the new value. If non-zero, this - will either be 2 for a (?imsx: group, or 3 for a lookbehind assertion. */ - - if (brastackptr >= sizeof(brastack)/sizeof(int)) - { - *errorptr = ERR19; - goto PCRE_ERROR_RETURN; - } - - bralenstack[brastackptr] = branch_extra; - branch_extra = branch_newextra; - - brastack[brastackptr++] = length; - length += bracket_length; - continue; - - /* Handle ket. Look for subsequent max/min; for certain sets of values we - have to replicate this bracket up to that many times. If brastackptr is - 0 this is an unmatched bracket which will generate an error, but take care - not to try to access brastack[-1] when computing the length and restoring - the branch_extra value. */ - - case ')': - length += 1 + LINK_SIZE; - if (brastackptr > 0) - { - duplength = length - brastack[--brastackptr]; - branch_extra = bralenstack[brastackptr]; - } - else duplength = 0; - - /* The following code is also used when a recursion such as (?3) is - followed by a quantifier, because in that case, it has to be wrapped inside - brackets so that the quantifier works. The value of duplength must be - set before arrival. */ - - HANDLE_QUANTIFIED_BRACKETS: - - /* Leave ptr at the final char; for read_repeat_counts this happens - automatically; for the others we need an increment. */ - - if ((c = ptr[1]) == '{' && is_counted_repeat(ptr+2)) - { - ptr = read_repeat_counts(ptr+2, &min, &max, errorptr); - if (*errorptr != NULL) goto PCRE_ERROR_RETURN; - } - else if (c == '*') { min = 0; max = -1; ptr++; } - else if (c == '+') { min = 1; max = -1; ptr++; } - else if (c == '?') { min = 0; max = 1; ptr++; } - else { min = 1; max = 1; } - - /* If the minimum is zero, we have to allow for an OP_BRAZERO before the - group, and if the maximum is greater than zero, we have to replicate - maxval-1 times; each replication acquires an OP_BRAZERO plus a nesting - bracket set. */ - - if (min == 0) - { - length++; - if (max > 0) length += (max - 1) * (duplength + 3 + 2*LINK_SIZE); - } - - /* When the minimum is greater than zero, we have to replicate up to - minval-1 times, with no additions required in the copies. Then, if there - is a limited maximum we have to replicate up to maxval-1 times allowing - for a BRAZERO item before each optional copy and nesting brackets for all - but one of the optional copies. */ - - else - { - length += (min - 1) * duplength; - if (max > min) /* Need this test as max=-1 means no limit */ - length += (max - min) * (duplength + 3 + 2*LINK_SIZE) - - (2 + 2*LINK_SIZE); - } - - /* Allow space for once brackets for "possessive quantifier" */ - - if (ptr[1] == '+') - { - ptr++; - length += 2 + 2*LINK_SIZE; - } - continue; - - /* Non-special character. It won't be space or # in extended mode, so it is - always a genuine character. If we are in a \Q...\E sequence, check for the - end; if not, we have a literal. */ - - default: - NORMAL_CHAR: - - if (inescq && c == '\\' && ptr[1] == 'E') - { - inescq = FALSE; - ptr++; - continue; - } - - length += 2; /* For a one-byte character */ - lastitemlength = 1; /* Default length of last item for repeats */ - - /* In UTF-8 mode, check for additional bytes. */ - -#ifdef SUPPORT_UTF8 - if (utf8 && (c & 0xc0) == 0xc0) - { - while ((ptr[1] & 0xc0) == 0x80) /* Can't flow over the end */ - { /* because the end is marked */ - lastitemlength++; /* by a zero byte. */ - length++; - ptr++; - } - } -#endif - - continue; - } - } - -length += 2 + LINK_SIZE; /* For final KET and END */ - -if ((options & PCRE_AUTO_CALLOUT) != 0) - length += 2 + 2*LINK_SIZE; /* For final callout */ - -if (length > MAX_PATTERN_SIZE) - { - *errorptr = ERR20; - return NULL; - } - -/* Compute the size of data block needed and get it, either from malloc or -externally provided function. */ - -size = length + sizeof(real_pcre) + name_count * (max_name_size + 3); -re = (real_pcre *)(pcre_malloc)(size); - -if (re == NULL) - { - *errorptr = ERR21; - return NULL; - } - -/* Put in the magic number, and save the sizes, options, and character table -pointer. NULL is used for the default character tables. The nullpad field is at -the end; it's there to help in the case when a regex compiled on a system with -4-byte pointers is run on another with 8-byte pointers. */ - -re->magic_number = MAGIC_NUMBER; -re->size = size; -re->options = options; -re->dummy1 = re->dummy2 = 0; -re->name_table_offset = sizeof(real_pcre); -re->name_entry_size = max_name_size + 3; -re->name_count = name_count; -re->tables = (tables == pcre_default_tables)? NULL : tables; -re->nullpad = NULL; - -/* The starting points of the name/number translation table and of the code are -passed around in the compile data block. */ - -compile_block.names_found = 0; -compile_block.name_entry_size = max_name_size + 3; -compile_block.name_table = (uschar *)re + re->name_table_offset; -codestart = compile_block.name_table + re->name_entry_size * re->name_count; -compile_block.start_code = codestart; -compile_block.start_pattern = (const uschar *)pattern; -compile_block.req_varyopt = 0; -compile_block.nopartial = FALSE; - -/* Set up a starting, non-extracting bracket, then compile the expression. On -error, *errorptr will be set non-NULL, so we don't need to look at the result -of the function here. */ - -ptr = (const uschar *)pattern; -code = (uschar *)codestart; -*code = OP_BRA; -bracount = 0; -(void)compile_regex(options, options & PCRE_IMS, &bracount, &code, &ptr, - errorptr, FALSE, 0, &firstbyte, &reqbyte, NULL, &compile_block); -re->top_bracket = bracount; -re->top_backref = compile_block.top_backref; - -if (compile_block.nopartial) re->options |= PCRE_NOPARTIAL; - -/* If not reached end of pattern on success, there's an excess bracket. */ - -if (*errorptr == NULL && *ptr != 0) *errorptr = ERR22; - -/* Fill in the terminating state and check for disastrous overflow, but -if debugging, leave the test till after things are printed out. */ - -*code++ = OP_END; - -#ifndef DEBUG -if (code - codestart > length) *errorptr = ERR23; -#endif - -/* Give an error if there's back reference to a non-existent capturing -subpattern. */ - -if (re->top_backref > re->top_bracket) *errorptr = ERR15; - -/* Failed to compile, or error while post-processing */ - -if (*errorptr != NULL) - { - (pcre_free)(re); - PCRE_ERROR_RETURN: - *erroroffset = ptr - (const uschar *)pattern; - return NULL; - } - -/* If the anchored option was not passed, set the flag if we can determine that -the pattern is anchored by virtue of ^ characters or \A or anything else (such -as starting with .* when DOTALL is set). - -Otherwise, if we know what the first character has to be, save it, because that -speeds up unanchored matches no end. If not, see if we can set the -PCRE_STARTLINE flag. This is helpful for multiline matches when all branches -start with ^. and also when all branches start with .* for non-DOTALL matches. -*/ - -if ((options & PCRE_ANCHORED) == 0) - { - int temp_options = options; - if (is_anchored(codestart, &temp_options, 0, compile_block.backref_map)) - re->options |= PCRE_ANCHORED; - else - { - if (firstbyte < 0) - firstbyte = find_firstassertedchar(codestart, &temp_options, FALSE); - if (firstbyte >= 0) /* Remove caseless flag for non-caseable chars */ - { - int ch = firstbyte & 255; - re->first_byte = ((firstbyte & REQ_CASELESS) != 0 && - compile_block.fcc[ch] == ch)? ch : firstbyte; - re->options |= PCRE_FIRSTSET; - } - else if (is_startline(codestart, 0, compile_block.backref_map)) - re->options |= PCRE_STARTLINE; - } - } - -/* For an anchored pattern, we use the "required byte" only if it follows a -variable length item in the regex. Remove the caseless flag for non-caseable -bytes. */ - -if (reqbyte >= 0 && - ((re->options & PCRE_ANCHORED) == 0 || (reqbyte & REQ_VARY) != 0)) - { - int ch = reqbyte & 255; - re->req_byte = ((reqbyte & REQ_CASELESS) != 0 && - compile_block.fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte; - re->options |= PCRE_REQCHSET; - } - -/* Print out the compiled data for debugging */ - -#ifdef DEBUG - -printf("Length = %d top_bracket = %d top_backref = %d\n", - length, re->top_bracket, re->top_backref); - -if (re->options != 0) - { - printf("%s%s%s%s%s%s%s%s%s%s\n", - ((re->options & PCRE_NOPARTIAL) != 0)? "nopartial " : "", - ((re->options & PCRE_ANCHORED) != 0)? "anchored " : "", - ((re->options & PCRE_CASELESS) != 0)? "caseless " : "", - ((re->options & PCRE_ICHANGED) != 0)? "case state changed " : "", - ((re->options & PCRE_EXTENDED) != 0)? "extended " : "", - ((re->options & PCRE_MULTILINE) != 0)? "multiline " : "", - ((re->options & PCRE_DOTALL) != 0)? "dotall " : "", - ((re->options & PCRE_DOLLAR_ENDONLY) != 0)? "endonly " : "", - ((re->options & PCRE_EXTRA) != 0)? "extra " : "", - ((re->options & PCRE_UNGREEDY) != 0)? "ungreedy " : ""); - } - -if ((re->options & PCRE_FIRSTSET) != 0) - { - int ch = re->first_byte & 255; - const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)? "" : " (caseless)"; - if (isprint(ch)) printf("First char = %c%s\n", ch, caseless); - else printf("First char = \\x%02x%s\n", ch, caseless); - } - -if ((re->options & PCRE_REQCHSET) != 0) - { - int ch = re->req_byte & 255; - const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)? "" : " (caseless)"; - if (isprint(ch)) printf("Req char = %c%s\n", ch, caseless); - else printf("Req char = \\x%02x%s\n", ch, caseless); - } - -print_internals(re, stdout); - -/* This check is done here in the debugging case so that the code that -was compiled can be seen. */ - -if (code - codestart > length) - { - *errorptr = ERR23; - (pcre_free)(re); - *erroroffset = ptr - (uschar *)pattern; - return NULL; - } -#endif - -return (pcre *)re; -} - - - -/************************************************* -* Match a back-reference * -*************************************************/ - -/* If a back reference hasn't been set, the length that is passed is greater -than the number of characters left in the string, so the match fails. - -Arguments: - offset index into the offset vector - eptr points into the subject - length length to be matched - md points to match data block - ims the ims flags - -Returns: TRUE if matched -*/ - -static BOOL -match_ref(int offset, register const uschar *eptr, int length, match_data *md, - unsigned long int ims) -{ -const uschar *p = md->start_subject + md->offset_vector[offset]; - -#ifdef DEBUG -if (eptr >= md->end_subject) - printf("matching subject <null>"); -else - { - printf("matching subject "); - pchars(eptr, length, TRUE, md); - } -printf(" against backref "); -pchars(p, length, FALSE, md); -printf("\n"); -#endif - -/* Always fail if not enough characters left */ - -if (length > md->end_subject - eptr) return FALSE; - -/* Separate the caselesss case for speed */ - -if ((ims & PCRE_CASELESS) != 0) - { - while (length-- > 0) - if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE; - } -else - { while (length-- > 0) if (*p++ != *eptr++) return FALSE; } - -return TRUE; -} - - -#ifdef SUPPORT_UTF8 -/************************************************* -* Match character against an XCLASS * -*************************************************/ - -/* This function is called from within the XCLASS code below, to match a -character against an extended class which might match values > 255. - -Arguments: - c the character - data points to the flag byte of the XCLASS data - -Returns: TRUE if character matches, else FALSE -*/ - -static BOOL -match_xclass(int c, const uschar *data) -{ -int t; -BOOL negated = (*data & XCL_NOT) != 0; - -/* Character values < 256 are matched against a bitmap, if one is present. If -not, we still carry on, because there may be ranges that start below 256 in the -additional data. */ - -if (c < 256) - { - if ((*data & XCL_MAP) != 0 && (data[1 + c/8] & (1 << (c&7))) != 0) - return !negated; /* char found */ - } - -/* First skip the bit map if present. Then match against the list of Unicode -properties or large chars or ranges that end with a large char. We won't ever -encounter XCL_PROP or XCL_NOTPROP when UCP support is not compiled. */ - -if ((*data++ & XCL_MAP) != 0) data += 32; - -while ((t = *data++) != XCL_END) - { - int x, y; - if (t == XCL_SINGLE) - { - GETCHARINC(x, data); - if (c == x) return !negated; - } - else if (t == XCL_RANGE) - { - GETCHARINC(x, data); - GETCHARINC(y, data); - if (c >= x && c <= y) return !negated; - } - -#ifdef SUPPORT_UCP - else /* XCL_PROP & XCL_NOTPROP */ - { - int chartype, othercase; - int rqdtype = *data++; - int category = ucp_findchar(c, &chartype, &othercase); - if (rqdtype >= 128) - { - if ((rqdtype - 128 == category) == (t == XCL_PROP)) return !negated; - } - else - { - if ((rqdtype == chartype) == (t == XCL_PROP)) return !negated; - } - } -#endif /* SUPPORT_UCP */ - } - -return negated; /* char did not match */ -} -#endif - - -/*************************************************************************** -**************************************************************************** - RECURSION IN THE match() FUNCTION - -The match() function is highly recursive. Some regular expressions can cause -it to recurse thousands of times. I was writing for Unix, so I just let it -call itself recursively. This uses the stack for saving everything that has -to be saved for a recursive call. On Unix, the stack can be large, and this -works fine. - -It turns out that on non-Unix systems there are problems with programs that -use a lot of stack. (This despite the fact that every last chip has oodles -of memory these days, and techniques for extending the stack have been known -for decades.) So.... - -There is a fudge, triggered by defining NO_RECURSE, which avoids recursive -calls by keeping local variables that need to be preserved in blocks of memory -obtained from malloc instead instead of on the stack. Macros are used to -achieve this so that the actual code doesn't look very different to what it -always used to. -**************************************************************************** -***************************************************************************/ - - -/* These versions of the macros use the stack, as normal */ - -#ifndef NO_RECURSE -#define REGISTER register -#define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) rx = match(ra,rb,rc,rd,re,rf,rg) -#define RRETURN(ra) return ra -#else - - -/* These versions of the macros manage a private stack on the heap. Note -that the rd argument of RMATCH isn't actually used. It's the md argument of -match(), which never changes. */ - -#define REGISTER - -#define RMATCH(rx,ra,rb,rc,rd,re,rf,rg)\ - {\ - heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\ - if (setjmp(frame->Xwhere) == 0)\ - {\ - newframe->Xeptr = ra;\ - newframe->Xecode = rb;\ - newframe->Xoffset_top = rc;\ - newframe->Xims = re;\ - newframe->Xeptrb = rf;\ - newframe->Xflags = rg;\ - newframe->Xprevframe = frame;\ - frame = newframe;\ - DPRINTF(("restarting from line %d\n", __LINE__));\ - goto HEAP_RECURSE;\ - }\ - else\ - {\ - DPRINTF(("longjumped back to line %d\n", __LINE__));\ - frame = md->thisframe;\ - rx = frame->Xresult;\ - }\ - } - -#define RRETURN(ra)\ - {\ - heapframe *newframe = frame;\ - frame = newframe->Xprevframe;\ - (pcre_stack_free)(newframe);\ - if (frame != NULL)\ - {\ - frame->Xresult = ra;\ - md->thisframe = frame;\ - longjmp(frame->Xwhere, 1);\ - }\ - return ra;\ - } - - -/* Structure for remembering the local variables in a private frame */ - -typedef struct heapframe { - struct heapframe *Xprevframe; - - /* Function arguments that may change */ - - const uschar *Xeptr; - const uschar *Xecode; - int Xoffset_top; - long int Xims; - eptrblock *Xeptrb; - int Xflags; - - /* Function local variables */ - - const uschar *Xcallpat; - const uschar *Xcharptr; - const uschar *Xdata; - const uschar *Xnext; - const uschar *Xpp; - const uschar *Xprev; - const uschar *Xsaved_eptr; - - recursion_info Xnew_recursive; - - BOOL Xcur_is_word; - BOOL Xcondition; - BOOL Xminimize; - BOOL Xprev_is_word; - - unsigned long int Xoriginal_ims; - -#ifdef SUPPORT_UCP - int Xprop_type; - int Xprop_fail_result; - int Xprop_category; - int Xprop_chartype; - int Xprop_othercase; - int Xprop_test_against; - int *Xprop_test_variable; -#endif - - int Xctype; - int Xfc; - int Xfi; - int Xlength; - int Xmax; - int Xmin; - int Xnumber; - int Xoffset; - int Xop; - int Xsave_capture_last; - int Xsave_offset1, Xsave_offset2, Xsave_offset3; - int Xstacksave[REC_STACK_SAVE_MAX]; - - eptrblock Xnewptrb; - - /* Place to pass back result, and where to jump back to */ - - int Xresult; - jmp_buf Xwhere; - -} heapframe; - -#endif - - -/*************************************************************************** -***************************************************************************/ - - - -/************************************************* -* Match from current position * -*************************************************/ - -/* On entry ecode points to the first opcode, and eptr to the first character -in the subject string, while eptrb holds the value of eptr at the start of the -last bracketed group - used for breaking infinite loops matching zero-length -strings. This function is called recursively in many circumstances. Whenever it -returns a negative (error) response, the outer incarnation must also return the -same response. - -Performance note: It might be tempting to extract commonly used fields from the -md structure (e.g. utf8, end_subject) into individual variables to improve -performance. Tests using gcc on a SPARC disproved this; in the first case, it -made performance worse. - -Arguments: - eptr pointer in subject - ecode position in code - offset_top current top pointer - md pointer to "static" info for the match - ims current /i, /m, and /s options - eptrb pointer to chain of blocks containing eptr at start of - brackets - for testing for empty matches - flags can contain - match_condassert - this is an assertion condition - match_isgroup - this is the start of a bracketed group - -Returns: MATCH_MATCH if matched ) these values are >= 0 - MATCH_NOMATCH if failed to match ) - a negative PCRE_ERROR_xxx value if aborted by an error condition - (e.g. stopped by recursion limit) -*/ - -static int -match(REGISTER const uschar *eptr, REGISTER const uschar *ecode, - int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb, - int flags) -{ -/* These variables do not need to be preserved over recursion in this function, -so they can be ordinary variables in all cases. Mark them with "register" -because they are used a lot in loops. */ - -register int rrc; /* Returns from recursive calls */ -register int i; /* Used for loops not involving calls to RMATCH() */ -register int c; /* Character values not kept over RMATCH() calls */ - -/* When recursion is not being used, all "local" variables that have to be -preserved over calls to RMATCH() are part of a "frame" which is obtained from -heap storage. Set up the top-level frame here; others are obtained from the -heap whenever RMATCH() does a "recursion". See the macro definitions above. */ - -#ifdef NO_RECURSE -heapframe *frame = (pcre_stack_malloc)(sizeof(heapframe)); -frame->Xprevframe = NULL; /* Marks the top level */ - -/* Copy in the original argument variables */ - -frame->Xeptr = eptr; -frame->Xecode = ecode; -frame->Xoffset_top = offset_top; -frame->Xims = ims; -frame->Xeptrb = eptrb; -frame->Xflags = flags; - -/* This is where control jumps back to to effect "recursion" */ - -HEAP_RECURSE: - -/* Macros make the argument variables come from the current frame */ - -#define eptr frame->Xeptr -#define ecode frame->Xecode -#define offset_top frame->Xoffset_top -#define ims frame->Xims -#define eptrb frame->Xeptrb -#define flags frame->Xflags - -/* Ditto for the local variables */ - -#ifdef SUPPORT_UTF8 -#define charptr frame->Xcharptr -#endif -#define callpat frame->Xcallpat -#define data frame->Xdata -#define next frame->Xnext -#define pp frame->Xpp -#define prev frame->Xprev -#define saved_eptr frame->Xsaved_eptr - -#define new_recursive frame->Xnew_recursive - -#define cur_is_word frame->Xcur_is_word -#define condition frame->Xcondition -#define minimize frame->Xminimize -#define prev_is_word frame->Xprev_is_word - -#define original_ims frame->Xoriginal_ims - -#ifdef SUPPORT_UCP -#define prop_type frame->Xprop_type -#define prop_fail_result frame->Xprop_fail_result -#define prop_category frame->Xprop_category -#define prop_chartype frame->Xprop_chartype -#define prop_othercase frame->Xprop_othercase -#define prop_test_against frame->Xprop_test_against -#define prop_test_variable frame->Xprop_test_variable -#endif - -#define ctype frame->Xctype -#define fc frame->Xfc -#define fi frame->Xfi -#define length frame->Xlength -#define max frame->Xmax -#define min frame->Xmin -#define number frame->Xnumber -#define offset frame->Xoffset -#define op frame->Xop -#define save_capture_last frame->Xsave_capture_last -#define save_offset1 frame->Xsave_offset1 -#define save_offset2 frame->Xsave_offset2 -#define save_offset3 frame->Xsave_offset3 -#define stacksave frame->Xstacksave - -#define newptrb frame->Xnewptrb - -/* When recursion is being used, local variables are allocated on the stack and -get preserved during recursion in the normal way. In this environment, fi and -i, and fc and c, can be the same variables. */ - -#else -#define fi i -#define fc c - - -#ifdef SUPPORT_UTF8 /* Many of these variables are used ony */ -const uschar *charptr; /* small blocks of the code. My normal */ -#endif /* style of coding would have declared */ -const uschar *callpat; /* them within each of those blocks. */ -const uschar *data; /* However, in order to accommodate the */ -const uschar *next; /* version of this code that uses an */ -const uschar *pp; /* external "stack" implemented on the */ -const uschar *prev; /* heap, it is easier to declare them */ -const uschar *saved_eptr; /* all here, so the declarations can */ - /* be cut out in a block. The only */ -recursion_info new_recursive; /* declarations within blocks below are */ - /* for variables that do not have to */ -BOOL cur_is_word; /* be preserved over a recursive call */ -BOOL condition; /* to RMATCH(). */ -BOOL minimize; -BOOL prev_is_word; - -unsigned long int original_ims; - -#ifdef SUPPORT_UCP -int prop_type; -int prop_fail_result; -int prop_category; -int prop_chartype; -int prop_othercase; -int prop_test_against; -int *prop_test_variable; -#endif - -int ctype; -int length; -int max; -int min; -int number; -int offset; -int op; -int save_capture_last; -int save_offset1, save_offset2, save_offset3; -int stacksave[REC_STACK_SAVE_MAX]; - -eptrblock newptrb; -#endif - -/* These statements are here to stop the compiler complaining about unitialized -variables. */ - -#ifdef SUPPORT_UCP -prop_fail_result = 0; -prop_test_against = 0; -prop_test_variable = NULL; -#endif - -/* OK, now we can get on with the real code of the function. Recursion is -specified by the macros RMATCH and RRETURN. When NO_RECURSE is *not* defined, -these just turn into a recursive call to match() and a "return", respectively. -However, RMATCH isn't like a function call because it's quite a complicated -macro. It has to be used in one particular way. This shouldn't, however, impact -performance when true recursion is being used. */ - -if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT); - -original_ims = ims; /* Save for resetting on ')' */ - -/* At the start of a bracketed group, add the current subject pointer to the -stack of such pointers, to be re-instated at the end of the group when we hit -the closing ket. When match() is called in other circumstances, we don't add to -this stack. */ - -if ((flags & match_isgroup) != 0) - { - newptrb.epb_prev = eptrb; - newptrb.epb_saved_eptr = eptr; - eptrb = &newptrb; - } - -/* Now start processing the operations. */ - -for (;;) - { - op = *ecode; - minimize = FALSE; - - /* For partial matching, remember if we ever hit the end of the subject after - matching at least one subject character. */ - - if (md->partial && - eptr >= md->end_subject && - eptr > md->start_match) - md->hitend = TRUE; - - /* Opening capturing bracket. If there is space in the offset vector, save - the current subject position in the working slot at the top of the vector. We - mustn't change the current values of the data slot, because they may be set - from a previous iteration of this group, and be referred to by a reference - inside the group. - - If the bracket fails to match, we need to restore this value and also the - values of the final offsets, in case they were set by a previous iteration of - the same bracket. - - If there isn't enough space in the offset vector, treat this as if it were a - non-capturing bracket. Don't worry about setting the flag for the error case - here; that is handled in the code for KET. */ - - if (op > OP_BRA) - { - number = op - OP_BRA; - - /* For extended extraction brackets (large number), we have to fish out the - number from a dummy opcode at the start. */ - - if (number > EXTRACT_BASIC_MAX) - number = GET2(ecode, 2+LINK_SIZE); - offset = number << 1; - -#ifdef DEBUG - printf("start bracket %d subject=", number); - pchars(eptr, 16, TRUE, md); - printf("\n"); -#endif - - if (offset < md->offset_max) - { - save_offset1 = md->offset_vector[offset]; - save_offset2 = md->offset_vector[offset+1]; - save_offset3 = md->offset_vector[md->offset_end - number]; - save_capture_last = md->capture_last; - - DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3)); - md->offset_vector[md->offset_end - number] = eptr - md->start_subject; - - do - { - RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, - match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - md->capture_last = save_capture_last; - ecode += GET(ecode, 1); - } - while (*ecode == OP_ALT); - - DPRINTF(("bracket %d failed\n", number)); - - md->offset_vector[offset] = save_offset1; - md->offset_vector[offset+1] = save_offset2; - md->offset_vector[md->offset_end - number] = save_offset3; - - RRETURN(MATCH_NOMATCH); - } - - /* Insufficient room for saving captured contents */ - - else op = OP_BRA; - } - - /* Other types of node can be handled by a switch */ - - switch(op) - { - case OP_BRA: /* Non-capturing bracket: optimized */ - DPRINTF(("start bracket 0\n")); - do - { - RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, - match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - ecode += GET(ecode, 1); - } - while (*ecode == OP_ALT); - DPRINTF(("bracket 0 failed\n")); - RRETURN(MATCH_NOMATCH); - - /* Conditional group: compilation checked that there are no more than - two branches. If the condition is false, skipping the first branch takes us - past the end if there is only one branch, but that's OK because that is - exactly what going to the ket would do. */ - - case OP_COND: - if (ecode[LINK_SIZE+1] == OP_CREF) /* Condition extract or recurse test */ - { - offset = GET2(ecode, LINK_SIZE+2) << 1; /* Doubled ref number */ - condition = (offset == CREF_RECURSE * 2)? - (md->recursive != NULL) : - (offset < offset_top && md->offset_vector[offset] >= 0); - RMATCH(rrc, eptr, ecode + (condition? - (LINK_SIZE + 4) : (LINK_SIZE + 1 + GET(ecode, 1))), - offset_top, md, ims, eptrb, match_isgroup); - RRETURN(rrc); - } - - /* The condition is an assertion. Call match() to evaluate it - setting - the final argument TRUE causes it to stop at the end of an assertion. */ - - else - { - RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, - match_condassert | match_isgroup); - if (rrc == MATCH_MATCH) - { - ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE+2); - while (*ecode == OP_ALT) ecode += GET(ecode, 1); - } - else if (rrc != MATCH_NOMATCH) - { - RRETURN(rrc); /* Need braces because of following else */ - } - else ecode += GET(ecode, 1); - RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, - match_isgroup); - RRETURN(rrc); - } - /* Control never reaches here */ - - /* Skip over conditional reference or large extraction number data if - encountered. */ - - case OP_CREF: - case OP_BRANUMBER: - ecode += 3; - break; - - /* End of the pattern. If we are in a recursion, we should restore the - offsets appropriately and continue from after the call. */ - - case OP_END: - if (md->recursive != NULL && md->recursive->group_num == 0) - { - recursion_info *rec = md->recursive; - DPRINTF(("Hit the end in a (?0) recursion\n")); - md->recursive = rec->prevrec; - memmove(md->offset_vector, rec->offset_save, - rec->saved_max * sizeof(int)); - md->start_match = rec->save_start; - ims = original_ims; - ecode = rec->after_call; - break; - } - - /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty - string - backtracking will then try other alternatives, if any. */ - - if (md->notempty && eptr == md->start_match) RRETURN(MATCH_NOMATCH); - md->end_match_ptr = eptr; /* Record where we ended */ - md->end_offset_top = offset_top; /* and how many extracts were taken */ - RRETURN(MATCH_MATCH); - - /* Change option settings */ - - case OP_OPT: - ims = ecode[1]; - ecode += 2; - DPRINTF(("ims set to %02lx\n", ims)); - break; - - /* Assertion brackets. Check the alternative branches in turn - the - matching won't pass the KET for an assertion. If any one branch matches, - the assertion is true. Lookbehind assertions have an OP_REVERSE item at the - start of each branch to move the current point backwards, so the code at - this level is identical to the lookahead case. */ - - case OP_ASSERT: - case OP_ASSERTBACK: - do - { - RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, - match_isgroup); - if (rrc == MATCH_MATCH) break; - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - ecode += GET(ecode, 1); - } - while (*ecode == OP_ALT); - if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH); - - /* If checking an assertion for a condition, return MATCH_MATCH. */ - - if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH); - - /* Continue from after the assertion, updating the offsets high water - mark, since extracts may have been taken during the assertion. */ - - do ecode += GET(ecode,1); while (*ecode == OP_ALT); - ecode += 1 + LINK_SIZE; - offset_top = md->end_offset_top; - continue; - - /* Negative assertion: all branches must fail to match */ - - case OP_ASSERT_NOT: - case OP_ASSERTBACK_NOT: - do - { - RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, - match_isgroup); - if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - ecode += GET(ecode,1); - } - while (*ecode == OP_ALT); - - if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH); - - ecode += 1 + LINK_SIZE; - continue; - - /* Move the subject pointer back. This occurs only at the start of - each branch of a lookbehind assertion. If we are too close to the start to - move back, this match function fails. When working with UTF-8 we move - back a number of characters, not bytes. */ - - case OP_REVERSE: -#ifdef SUPPORT_UTF8 - if (md->utf8) - { - c = GET(ecode,1); - for (i = 0; i < c; i++) - { - eptr--; - if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH); - BACKCHAR(eptr) - } - } - else -#endif - - /* No UTF-8 support, or not in UTF-8 mode: count is byte count */ - - { - eptr -= GET(ecode,1); - if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH); - } - - /* Skip to next op code */ - - ecode += 1 + LINK_SIZE; - break; - - /* The callout item calls an external function, if one is provided, passing - details of the match so far. This is mainly for debugging, though the - function is able to force a failure. */ - - case OP_CALLOUT: - if (pcre_callout != NULL) - { - pcre_callout_block cb; - cb.version = 1; /* Version 1 of the callout block */ - cb.callout_number = ecode[1]; - cb.offset_vector = md->offset_vector; - cb.subject = (const char *)md->start_subject; - cb.subject_length = md->end_subject - md->start_subject; - cb.start_match = md->start_match - md->start_subject; - cb.current_position = eptr - md->start_subject; - cb.pattern_position = GET(ecode, 2); - cb.next_item_length = GET(ecode, 2 + LINK_SIZE); - cb.capture_top = offset_top/2; - cb.capture_last = md->capture_last; - cb.callout_data = md->callout_data; - if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH); - if (rrc < 0) RRETURN(rrc); - } - ecode += 2 + 2*LINK_SIZE; - break; - - /* Recursion either matches the current regex, or some subexpression. The - offset data is the offset to the starting bracket from the start of the - whole pattern. (This is so that it works from duplicated subpatterns.) - - If there are any capturing brackets started but not finished, we have to - save their starting points and reinstate them after the recursion. However, - we don't know how many such there are (offset_top records the completed - total) so we just have to save all the potential data. There may be up to - 65535 such values, which is too large to put on the stack, but using malloc - for small numbers seems expensive. As a compromise, the stack is used when - there are no more than REC_STACK_SAVE_MAX values to store; otherwise malloc - is used. A problem is what to do if the malloc fails ... there is no way of - returning to the top level with an error. Save the top REC_STACK_SAVE_MAX - values on the stack, and accept that the rest may be wrong. - - There are also other values that have to be saved. We use a chained - sequence of blocks that actually live on the stack. Thanks to Robin Houston - for the original version of this logic. */ - - case OP_RECURSE: - { - callpat = md->start_code + GET(ecode, 1); - new_recursive.group_num = *callpat - OP_BRA; - - /* For extended extraction brackets (large number), we have to fish out - the number from a dummy opcode at the start. */ - - if (new_recursive.group_num > EXTRACT_BASIC_MAX) - new_recursive.group_num = GET2(callpat, 2+LINK_SIZE); - - /* Add to "recursing stack" */ - - new_recursive.prevrec = md->recursive; - md->recursive = &new_recursive; - - /* Find where to continue from afterwards */ - - ecode += 1 + LINK_SIZE; - new_recursive.after_call = ecode; - - /* Now save the offset data. */ - - new_recursive.saved_max = md->offset_end; - if (new_recursive.saved_max <= REC_STACK_SAVE_MAX) - new_recursive.offset_save = stacksave; - else - { - new_recursive.offset_save = - (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int)); - if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY); - } - - memcpy(new_recursive.offset_save, md->offset_vector, - new_recursive.saved_max * sizeof(int)); - new_recursive.save_start = md->start_match; - md->start_match = eptr; - - /* OK, now we can do the recursion. For each top-level alternative we - restore the offset and recursion data. */ - - DPRINTF(("Recursing into group %d\n", new_recursive.group_num)); - do - { - RMATCH(rrc, eptr, callpat + 1 + LINK_SIZE, offset_top, md, ims, - eptrb, match_isgroup); - if (rrc == MATCH_MATCH) - { - md->recursive = new_recursive.prevrec; - if (new_recursive.offset_save != stacksave) - (pcre_free)(new_recursive.offset_save); - RRETURN(MATCH_MATCH); - } - else if (rrc != MATCH_NOMATCH) RRETURN(rrc); - - md->recursive = &new_recursive; - memcpy(md->offset_vector, new_recursive.offset_save, - new_recursive.saved_max * sizeof(int)); - callpat += GET(callpat, 1); - } - while (*callpat == OP_ALT); - - DPRINTF(("Recursion didn't match\n")); - md->recursive = new_recursive.prevrec; - if (new_recursive.offset_save != stacksave) - (pcre_free)(new_recursive.offset_save); - RRETURN(MATCH_NOMATCH); - } - /* Control never reaches here */ - - /* "Once" brackets are like assertion brackets except that after a match, - the point in the subject string is not moved back. Thus there can never be - a move back into the brackets. Friedl calls these "atomic" subpatterns. - Check the alternative branches in turn - the matching won't pass the KET - for this kind of subpattern. If any one branch matches, we carry on as at - the end of a normal bracket, leaving the subject pointer. */ - - case OP_ONCE: - { - prev = ecode; - saved_eptr = eptr; - - do - { - RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, - eptrb, match_isgroup); - if (rrc == MATCH_MATCH) break; - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - ecode += GET(ecode,1); - } - while (*ecode == OP_ALT); - - /* If hit the end of the group (which could be repeated), fail */ - - if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH); - - /* Continue as from after the assertion, updating the offsets high water - mark, since extracts may have been taken. */ - - do ecode += GET(ecode,1); while (*ecode == OP_ALT); - - offset_top = md->end_offset_top; - eptr = md->end_match_ptr; - - /* For a non-repeating ket, just continue at this level. This also - happens for a repeating ket if no characters were matched in the group. - This is the forcible breaking of infinite loops as implemented in Perl - 5.005. If there is an options reset, it will get obeyed in the normal - course of events. */ - - if (*ecode == OP_KET || eptr == saved_eptr) - { - ecode += 1+LINK_SIZE; - break; - } - - /* The repeating kets try the rest of the pattern or restart from the - preceding bracket, in the appropriate order. We need to reset any options - that changed within the bracket before re-running it, so check the next - opcode. */ - - if (ecode[1+LINK_SIZE] == OP_OPT) - { - ims = (ims & ~PCRE_IMS) | ecode[4]; - DPRINTF(("ims set to %02lx at group repeat\n", ims)); - } - - if (*ecode == OP_KETRMIN) - { - RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - else /* OP_KETRMAX */ - { - RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - } - RRETURN(MATCH_NOMATCH); - - /* An alternation is the end of a branch; scan along to find the end of the - bracketed group and go to there. */ - - case OP_ALT: - do ecode += GET(ecode,1); while (*ecode == OP_ALT); - break; - - /* BRAZERO and BRAMINZERO occur just before a bracket group, indicating - that it may occur zero times. It may repeat infinitely, or not at all - - i.e. it could be ()* or ()? in the pattern. Brackets with fixed upper - repeat limits are compiled as a number of copies, with the optional ones - preceded by BRAZERO or BRAMINZERO. */ - - case OP_BRAZERO: - { - next = ecode+1; - RMATCH(rrc, eptr, next, offset_top, md, ims, eptrb, match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - do next += GET(next,1); while (*next == OP_ALT); - ecode = next + 1+LINK_SIZE; - } - break; - - case OP_BRAMINZERO: - { - next = ecode+1; - do next += GET(next,1); while (*next == OP_ALT); - RMATCH(rrc, eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, - match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - ecode++; - } - break; - - /* End of a group, repeated or non-repeating. If we are at the end of - an assertion "group", stop matching and return MATCH_MATCH, but record the - current high water mark for use by positive assertions. Do this also - for the "once" (not-backup up) groups. */ - - case OP_KET: - case OP_KETRMIN: - case OP_KETRMAX: - { - prev = ecode - GET(ecode, 1); - saved_eptr = eptrb->epb_saved_eptr; - - /* Back up the stack of bracket start pointers. */ - - eptrb = eptrb->epb_prev; - - if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT || - *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT || - *prev == OP_ONCE) - { - md->end_match_ptr = eptr; /* For ONCE */ - md->end_offset_top = offset_top; - RRETURN(MATCH_MATCH); - } - - /* In all other cases except a conditional group we have to check the - group number back at the start and if necessary complete handling an - extraction by setting the offsets and bumping the high water mark. */ - - if (*prev != OP_COND) - { - number = *prev - OP_BRA; - - /* For extended extraction brackets (large number), we have to fish out - the number from a dummy opcode at the start. */ - - if (number > EXTRACT_BASIC_MAX) number = GET2(prev, 2+LINK_SIZE); - offset = number << 1; - -#ifdef DEBUG - printf("end bracket %d", number); - printf("\n"); -#endif - - /* Test for a numbered group. This includes groups called as a result - of recursion. Note that whole-pattern recursion is coded as a recurse - into group 0, so it won't be picked up here. Instead, we catch it when - the OP_END is reached. */ - - if (number > 0) - { - md->capture_last = number; - if (offset >= md->offset_max) md->offset_overflow = TRUE; else - { - md->offset_vector[offset] = - md->offset_vector[md->offset_end - number]; - md->offset_vector[offset+1] = eptr - md->start_subject; - if (offset_top <= offset) offset_top = offset + 2; - } - - /* Handle a recursively called group. Restore the offsets - appropriately and continue from after the call. */ - - if (md->recursive != NULL && md->recursive->group_num == number) - { - recursion_info *rec = md->recursive; - DPRINTF(("Recursion (%d) succeeded - continuing\n", number)); - md->recursive = rec->prevrec; - md->start_match = rec->save_start; - memcpy(md->offset_vector, rec->offset_save, - rec->saved_max * sizeof(int)); - ecode = rec->after_call; - ims = original_ims; - break; - } - } - } - - /* Reset the value of the ims flags, in case they got changed during - the group. */ - - ims = original_ims; - DPRINTF(("ims reset to %02lx\n", ims)); - - /* For a non-repeating ket, just continue at this level. This also - happens for a repeating ket if no characters were matched in the group. - This is the forcible breaking of infinite loops as implemented in Perl - 5.005. If there is an options reset, it will get obeyed in the normal - course of events. */ - - if (*ecode == OP_KET || eptr == saved_eptr) - { - ecode += 1 + LINK_SIZE; - break; - } - - /* The repeating kets try the rest of the pattern or restart from the - preceding bracket, in the appropriate order. */ - - if (*ecode == OP_KETRMIN) - { - RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - else /* OP_KETRMAX */ - { - RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - } - - RRETURN(MATCH_NOMATCH); - - /* Start of subject unless notbol, or after internal newline if multiline */ - - case OP_CIRC: - if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH); - if ((ims & PCRE_MULTILINE) != 0) - { - if (eptr != md->start_subject && eptr[-1] != NEWLINE) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - } - /* ... else fall through */ - - /* Start of subject assertion */ - - case OP_SOD: - if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH); - ecode++; - break; - - /* Start of match assertion */ - - case OP_SOM: - if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH); - ecode++; - break; - - /* Assert before internal newline if multiline, or before a terminating - newline unless endonly is set, else end of subject unless noteol is set. */ - - case OP_DOLL: - if ((ims & PCRE_MULTILINE) != 0) - { - if (eptr < md->end_subject) - { if (*eptr != NEWLINE) RRETURN(MATCH_NOMATCH); } - else - { if (md->noteol) RRETURN(MATCH_NOMATCH); } - ecode++; - break; - } - else - { - if (md->noteol) RRETURN(MATCH_NOMATCH); - if (!md->endonly) - { - if (eptr < md->end_subject - 1 || - (eptr == md->end_subject - 1 && *eptr != NEWLINE)) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - } - } - /* ... else fall through */ - - /* End of subject assertion (\z) */ - - case OP_EOD: - if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH); - ecode++; - break; - - /* End of subject or ending \n assertion (\Z) */ - - case OP_EODN: - if (eptr < md->end_subject - 1 || - (eptr == md->end_subject - 1 && *eptr != NEWLINE)) RRETURN(MATCH_NOMATCH); - ecode++; - break; - - /* Word boundary assertions */ - - case OP_NOT_WORD_BOUNDARY: - case OP_WORD_BOUNDARY: - { - - /* Find out if the previous and current characters are "word" characters. - It takes a bit more work in UTF-8 mode. Characters > 255 are assumed to - be "non-word" characters. */ - -#ifdef SUPPORT_UTF8 - if (md->utf8) - { - if (eptr == md->start_subject) prev_is_word = FALSE; else - { - const uschar *lastptr = eptr - 1; - while((*lastptr & 0xc0) == 0x80) lastptr--; - GETCHAR(c, lastptr); - prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0; - } - if (eptr >= md->end_subject) cur_is_word = FALSE; else - { - GETCHAR(c, eptr); - cur_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0; - } - } - else -#endif - - /* More streamlined when not in UTF-8 mode */ - - { - prev_is_word = (eptr != md->start_subject) && - ((md->ctypes[eptr[-1]] & ctype_word) != 0); - cur_is_word = (eptr < md->end_subject) && - ((md->ctypes[*eptr] & ctype_word) != 0); - } - - /* Now see if the situation is what we want */ - - if ((*ecode++ == OP_WORD_BOUNDARY)? - cur_is_word == prev_is_word : cur_is_word != prev_is_word) - RRETURN(MATCH_NOMATCH); - } - break; - - /* Match a single character type; inline for speed */ - - case OP_ANY: - if ((ims & PCRE_DOTALL) == 0 && eptr < md->end_subject && *eptr == NEWLINE) - RRETURN(MATCH_NOMATCH); - if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH); -#ifdef SUPPORT_UTF8 - if (md->utf8) - while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; -#endif - ecode++; - break; - - /* Match a single byte, even in UTF-8 mode. This opcode really does match - any byte, even newline, independent of the setting of PCRE_DOTALL. */ - - case OP_ANYBYTE: - if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_NOT_DIGIT: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c < 256 && -#endif - (md->ctypes[c] & ctype_digit) != 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_DIGIT: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c >= 256 || -#endif - (md->ctypes[c] & ctype_digit) == 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_NOT_WHITESPACE: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c < 256 && -#endif - (md->ctypes[c] & ctype_space) != 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_WHITESPACE: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c >= 256 || -#endif - (md->ctypes[c] & ctype_space) == 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_NOT_WORDCHAR: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c < 256 && -#endif - (md->ctypes[c] & ctype_word) != 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_WORDCHAR: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c >= 256 || -#endif - (md->ctypes[c] & ctype_word) == 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - -#ifdef SUPPORT_UCP - /* Check the next character by Unicode property. We will get here only - if the support is in the binary; otherwise a compile-time error occurs. */ - - case OP_PROP: - case OP_NOTPROP: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - { - int chartype, rqdtype; - int othercase; - int category = ucp_findchar(c, &chartype, &othercase); - - rqdtype = *(++ecode); - ecode++; - - if (rqdtype >= 128) - { - if ((rqdtype - 128 != category) == (op == OP_PROP)) - RRETURN(MATCH_NOMATCH); - } - else - { - if ((rqdtype != chartype) == (op == OP_PROP)) - RRETURN(MATCH_NOMATCH); - } - } - break; - - /* Match an extended Unicode sequence. We will get here only if the support - is in the binary; otherwise a compile-time error occurs. */ - - case OP_EXTUNI: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - { - int chartype; - int othercase; - int category = ucp_findchar(c, &chartype, &othercase); - if (category == ucp_M) RRETURN(MATCH_NOMATCH); - while (eptr < md->end_subject) - { - int len = 1; - if (!md->utf8) c = *eptr; else - { - GETCHARLEN(c, eptr, len); - } - category = ucp_findchar(c, &chartype, &othercase); - if (category != ucp_M) break; - eptr += len; - } - } - ecode++; - break; -#endif - - - /* Match a back reference, possibly repeatedly. Look past the end of the - item to see if there is repeat information following. The code is similar - to that for character classes, but repeated for efficiency. Then obey - similar code to character type repeats - written out again for speed. - However, if the referenced string is the empty string, always treat - it as matched, any number of times (otherwise there could be infinite - loops). */ - - case OP_REF: - { - offset = GET2(ecode, 1) << 1; /* Doubled ref number */ - ecode += 3; /* Advance past item */ - - /* If the reference is unset, set the length to be longer than the amount - of subject left; this ensures that every attempt at a match fails. We - can't just fail here, because of the possibility of quantifiers with zero - minima. */ - - length = (offset >= offset_top || md->offset_vector[offset] < 0)? - md->end_subject - eptr + 1 : - md->offset_vector[offset+1] - md->offset_vector[offset]; - - /* Set up for repetition, or handle the non-repeated case */ - - switch (*ecode) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - case OP_CRPLUS: - case OP_CRMINPLUS: - case OP_CRQUERY: - case OP_CRMINQUERY: - c = *ecode++ - OP_CRSTAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - break; - - case OP_CRRANGE: - case OP_CRMINRANGE: - minimize = (*ecode == OP_CRMINRANGE); - min = GET2(ecode, 1); - max = GET2(ecode, 3); - if (max == 0) max = INT_MAX; - ecode += 5; - break; - - default: /* No repeat follows */ - if (!match_ref(offset, eptr, length, md, ims)) RRETURN(MATCH_NOMATCH); - eptr += length; - continue; /* With the main loop */ - } - - /* If the length of the reference is zero, just continue with the - main loop. */ - - if (length == 0) continue; - - /* First, ensure the minimum number of matches are present. We get back - the length of the reference string explicitly rather than passing the - address of eptr, so that eptr can be a register variable. */ - - for (i = 1; i <= min; i++) - { - if (!match_ref(offset, eptr, length, md, ims)) RRETURN(MATCH_NOMATCH); - eptr += length; - } - - /* If min = max, continue at the same level without recursion. - They are not both allowed to be zero. */ - - if (min == max) continue; - - /* If minimizing, keep trying and advancing the pointer */ - - if (minimize) - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || !match_ref(offset, eptr, length, md, ims)) - RRETURN(MATCH_NOMATCH); - eptr += length; - } - /* Control never gets here */ - } - - /* If maximizing, find the longest string and work backwards */ - - else - { - pp = eptr; - for (i = min; i < max; i++) - { - if (!match_ref(offset, eptr, length, md, ims)) break; - eptr += length; - } - while (eptr >= pp) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - eptr -= length; - } - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - - - - /* Match a bit-mapped character class, possibly repeatedly. This op code is - used when all the characters in the class have values in the range 0-255, - and either the matching is caseful, or the characters are in the range - 0-127 when UTF-8 processing is enabled. The only difference between - OP_CLASS and OP_NCLASS occurs when a data character outside the range is - encountered. - - First, look past the end of the item to see if there is repeat information - following. Then obey similar code to character type repeats - written out - again for speed. */ - - case OP_NCLASS: - case OP_CLASS: - { - data = ecode + 1; /* Save for matching */ - ecode += 33; /* Advance past the item */ - - switch (*ecode) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - case OP_CRPLUS: - case OP_CRMINPLUS: - case OP_CRQUERY: - case OP_CRMINQUERY: - c = *ecode++ - OP_CRSTAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - break; - - case OP_CRRANGE: - case OP_CRMINRANGE: - minimize = (*ecode == OP_CRMINRANGE); - min = GET2(ecode, 1); - max = GET2(ecode, 3); - if (max == 0) max = INT_MAX; - ecode += 5; - break; - - default: /* No repeat follows */ - min = max = 1; - break; - } - - /* First, ensure the minimum number of matches are present. */ - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (md->utf8) - { - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - if (c > 255) - { - if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); - } - else - { - if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); - } - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - c = *eptr++; - if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); - } - } - - /* If max == min we can continue with the main loop without the - need to recurse. */ - - if (min == max) continue; - - /* If minimizing, keep testing the rest of the expression and advancing - the pointer while it matches the class. */ - - if (minimize) - { -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (md->utf8) - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - if (c > 255) - { - if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); - } - else - { - if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); - } - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - c = *eptr++; - if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - } - - /* If maximizing, find the longest possible run, then work backwards. */ - - else - { - pp = eptr; - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (md->utf8) - { - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c > 255) - { - if (op == OP_CLASS) break; - } - else - { - if ((data[c/8] & (1 << (c&7))) == 0) break; - } - eptr += len; - } - for (;;) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - BACKCHAR(eptr); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject) break; - c = *eptr; - if ((data[c/8] & (1 << (c&7))) == 0) break; - eptr++; - } - while (eptr >= pp) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - eptr--; - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - } - - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - - - /* Match an extended character class. This opcode is encountered only - in UTF-8 mode, because that's the only time it is compiled. */ - -#ifdef SUPPORT_UTF8 - case OP_XCLASS: - { - data = ecode + 1 + LINK_SIZE; /* Save for matching */ - ecode += GET(ecode, 1); /* Advance past the item */ - - switch (*ecode) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - case OP_CRPLUS: - case OP_CRMINPLUS: - case OP_CRQUERY: - case OP_CRMINQUERY: - c = *ecode++ - OP_CRSTAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - break; - - case OP_CRRANGE: - case OP_CRMINRANGE: - minimize = (*ecode == OP_CRMINRANGE); - min = GET2(ecode, 1); - max = GET2(ecode, 3); - if (max == 0) max = INT_MAX; - ecode += 5; - break; - - default: /* No repeat follows */ - min = max = 1; - break; - } - - /* First, ensure the minimum number of matches are present. */ - - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - if (!match_xclass(c, data)) RRETURN(MATCH_NOMATCH); - } - - /* If max == min we can continue with the main loop without the - need to recurse. */ - - if (min == max) continue; - - /* If minimizing, keep testing the rest of the expression and advancing - the pointer while it matches the class. */ - - if (minimize) - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - if (!match_xclass(c, data)) RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - } - - /* If maximizing, find the longest possible run, then work backwards. */ - - else - { - pp = eptr; - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (!match_xclass(c, data)) break; - eptr += len; - } - for(;;) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - BACKCHAR(eptr) - } - RRETURN(MATCH_NOMATCH); - } - - /* Control never gets here */ - } -#endif /* End of XCLASS */ - - /* Match a single character, casefully */ - - case OP_CHAR: -#ifdef SUPPORT_UTF8 - if (md->utf8) - { - length = 1; - ecode++; - GETCHARLEN(fc, ecode, length); - if (length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH); - } - else -#endif - - /* Non-UTF-8 mode */ - { - if (md->end_subject - eptr < 1) RRETURN(MATCH_NOMATCH); - if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH); - ecode += 2; - } - break; - - /* Match a single character, caselessly */ - - case OP_CHARNC: -#ifdef SUPPORT_UTF8 - if (md->utf8) - { - length = 1; - ecode++; - GETCHARLEN(fc, ecode, length); - - if (length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - - /* If the pattern character's value is < 128, we have only one byte, and - can use the fast lookup table. */ - - if (fc < 128) - { - if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH); - } - - /* Otherwise we must pick up the subject character */ - - else - { - int dc; - GETCHARINC(dc, eptr); - ecode += length; - - /* If we have Unicode property support, we can use it to test the other - case of the character, if there is one. The result of ucp_findchar() is - < 0 if the char isn't found, and othercase is returned as zero if there - isn't one. */ - - if (fc != dc) - { -#ifdef SUPPORT_UCP - int chartype; - int othercase; - if (ucp_findchar(fc, &chartype, &othercase) < 0 || dc != othercase) -#endif - RRETURN(MATCH_NOMATCH); - } - } - } - else -#endif /* SUPPORT_UTF8 */ - - /* Non-UTF-8 mode */ - { - if (md->end_subject - eptr < 1) RRETURN(MATCH_NOMATCH); - if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH); - ecode += 2; - } - break; - - /* Match a single character repeatedly; different opcodes share code. */ - - case OP_EXACT: - min = max = GET2(ecode, 1); - ecode += 3; - goto REPEATCHAR; - - case OP_UPTO: - case OP_MINUPTO: - min = 0; - max = GET2(ecode, 1); - minimize = *ecode == OP_MINUPTO; - ecode += 3; - goto REPEATCHAR; - - case OP_STAR: - case OP_MINSTAR: - case OP_PLUS: - case OP_MINPLUS: - case OP_QUERY: - case OP_MINQUERY: - c = *ecode++ - OP_STAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - - /* Common code for all repeated single-character matches. We can give - up quickly if there are fewer than the minimum number of characters left in - the subject. */ - - REPEATCHAR: -#ifdef SUPPORT_UTF8 - if (md->utf8) - { - length = 1; - charptr = ecode; - GETCHARLEN(fc, ecode, length); - if (min * length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - ecode += length; - - /* Handle multibyte character matching specially here. There is - support for caseless matching if UCP support is present. */ - - if (length > 1) - { - int oclength = 0; - uschar occhars[8]; - -#ifdef SUPPORT_UCP - int othercase; - int chartype; - if ((ims & PCRE_CASELESS) != 0 && - ucp_findchar(fc, &chartype, &othercase) >= 0 && - othercase > 0) - oclength = ord2utf8(othercase, occhars); -#endif /* SUPPORT_UCP */ - - for (i = 1; i <= min; i++) - { - if (memcmp(eptr, charptr, length) == 0) eptr += length; - /* Need braces because of following else */ - else if (oclength == 0) { RRETURN(MATCH_NOMATCH); } - else - { - if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH); - eptr += oclength; - } - } - - if (min == max) continue; - - if (minimize) - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - if (memcmp(eptr, charptr, length) == 0) eptr += length; - /* Need braces because of following else */ - else if (oclength == 0) { RRETURN(MATCH_NOMATCH); } - else - { - if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH); - eptr += oclength; - } - } - /* Control never gets here */ - } - else - { - pp = eptr; - for (i = min; i < max; i++) - { - if (eptr > md->end_subject - length) break; - if (memcmp(eptr, charptr, length) == 0) eptr += length; - else if (oclength == 0) break; - else - { - if (memcmp(eptr, occhars, oclength) != 0) break; - eptr += oclength; - } - } - while (eptr >= pp) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - eptr -= length; - } - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - } - - /* If the length of a UTF-8 character is 1, we fall through here, and - obey the code as for non-UTF-8 characters below, though in this case the - value of fc will always be < 128. */ - } - else -#endif /* SUPPORT_UTF8 */ - - /* When not in UTF-8 mode, load a single-byte character. */ - { - if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - fc = *ecode++; - } - - /* The value of fc at this point is always less than 256, though we may or - may not be in UTF-8 mode. The code is duplicated for the caseless and - caseful cases, for speed, since matching characters is likely to be quite - common. First, ensure the minimum number of matches are present. If min = - max, continue at the same level without recursing. Otherwise, if - minimizing, keep trying the rest of the expression and advancing one - matching character if failing, up to the maximum. Alternatively, if - maximizing, find the maximum number of characters and work backwards. */ - - DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max, - max, eptr)); - - if ((ims & PCRE_CASELESS) != 0) - { - fc = md->lcc[fc]; - for (i = 1; i <= min; i++) - if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH); - if (min == max) continue; - if (minimize) - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject || - fc != md->lcc[*eptr++]) - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - } - else - { - pp = eptr; - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break; - eptr++; - } - while (eptr >= pp) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - eptr--; - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - } - - /* Caseful comparisons (includes all multi-byte characters) */ - - else - { - for (i = 1; i <= min; i++) if (fc != *eptr++) RRETURN(MATCH_NOMATCH); - if (min == max) continue; - if (minimize) - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject || fc != *eptr++) - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - } - else - { - pp = eptr; - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || fc != *eptr) break; - eptr++; - } - while (eptr >= pp) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - eptr--; - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - - /* Match a negated single one-byte character. The character we are - checking can be multibyte. */ - - case OP_NOT: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - ecode++; - GETCHARINCTEST(c, eptr); - if ((ims & PCRE_CASELESS) != 0) - { -#ifdef SUPPORT_UTF8 - if (c < 256) -#endif - c = md->lcc[c]; - if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH); - } - else - { - if (*ecode++ == c) RRETURN(MATCH_NOMATCH); - } - break; - - /* Match a negated single one-byte character repeatedly. This is almost a - repeat of the code for a repeated single character, but I haven't found a - nice way of commoning these up that doesn't require a test of the - positive/negative option for each character match. Maybe that wouldn't add - very much to the time taken, but character matching *is* what this is all - about... */ - - case OP_NOTEXACT: - min = max = GET2(ecode, 1); - ecode += 3; - goto REPEATNOTCHAR; - - case OP_NOTUPTO: - case OP_NOTMINUPTO: - min = 0; - max = GET2(ecode, 1); - minimize = *ecode == OP_NOTMINUPTO; - ecode += 3; - goto REPEATNOTCHAR; - - case OP_NOTSTAR: - case OP_NOTMINSTAR: - case OP_NOTPLUS: - case OP_NOTMINPLUS: - case OP_NOTQUERY: - case OP_NOTMINQUERY: - c = *ecode++ - OP_NOTSTAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - - /* Common code for all repeated single-byte matches. We can give up quickly - if there are fewer than the minimum number of bytes left in the - subject. */ - - REPEATNOTCHAR: - if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - fc = *ecode++; - - /* The code is duplicated for the caseless and caseful cases, for speed, - since matching characters is likely to be quite common. First, ensure the - minimum number of matches are present. If min = max, continue at the same - level without recursing. Otherwise, if minimizing, keep trying the rest of - the expression and advancing one matching character if failing, up to the - maximum. Alternatively, if maximizing, find the maximum number of - characters and work backwards. */ - - DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max, - max, eptr)); - - if ((ims & PCRE_CASELESS) != 0) - { - fc = md->lcc[fc]; - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (md->utf8) - { - register int d; - for (i = 1; i <= min; i++) - { - GETCHARINC(d, eptr); - if (d < 256) d = md->lcc[d]; - if (fc == d) RRETURN(MATCH_NOMATCH); - } - } - else -#endif - - /* Not UTF-8 mode */ - { - for (i = 1; i <= min; i++) - if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH); - } - - if (min == max) continue; - - if (minimize) - { -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (md->utf8) - { - register int d; - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - GETCHARINC(d, eptr); - if (d < 256) d = md->lcc[d]; - if (fi >= max || eptr >= md->end_subject || fc == d) - RRETURN(MATCH_NOMATCH); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++]) - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - } - - /* Maximize case */ - - else - { - pp = eptr; - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (md->utf8) - { - register int d; - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(d, eptr, len); - if (d < 256) d = md->lcc[d]; - if (fc == d) break; - eptr += len; - } - for(;;) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - BACKCHAR(eptr); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break; - eptr++; - } - while (eptr >= pp) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - eptr--; - } - } - - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - } - - /* Caseful comparisons */ - - else - { -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (md->utf8) - { - register int d; - for (i = 1; i <= min; i++) - { - GETCHARINC(d, eptr); - if (fc == d) RRETURN(MATCH_NOMATCH); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (i = 1; i <= min; i++) - if (fc == *eptr++) RRETURN(MATCH_NOMATCH); - } - - if (min == max) continue; - - if (minimize) - { -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (md->utf8) - { - register int d; - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - GETCHARINC(d, eptr); - if (fi >= max || eptr >= md->end_subject || fc == d) - RRETURN(MATCH_NOMATCH); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject || fc == *eptr++) - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - } - - /* Maximize case */ - - else - { - pp = eptr; - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (md->utf8) - { - register int d; - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(d, eptr, len); - if (fc == d) break; - eptr += len; - } - for(;;) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - BACKCHAR(eptr); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || fc == *eptr) break; - eptr++; - } - while (eptr >= pp) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - eptr--; - } - } - - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - - /* Match a single character type repeatedly; several different opcodes - share code. This is very similar to the code for single characters, but we - repeat it in the interests of efficiency. */ - - case OP_TYPEEXACT: - min = max = GET2(ecode, 1); - minimize = TRUE; - ecode += 3; - goto REPEATTYPE; - - case OP_TYPEUPTO: - case OP_TYPEMINUPTO: - min = 0; - max = GET2(ecode, 1); - minimize = *ecode == OP_TYPEMINUPTO; - ecode += 3; - goto REPEATTYPE; - - case OP_TYPESTAR: - case OP_TYPEMINSTAR: - case OP_TYPEPLUS: - case OP_TYPEMINPLUS: - case OP_TYPEQUERY: - case OP_TYPEMINQUERY: - c = *ecode++ - OP_TYPESTAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - - /* Common code for all repeated single character type matches. Note that - in UTF-8 mode, '.' matches a character of any length, but for the other - character types, the valid characters are all one-byte long. */ - - REPEATTYPE: - ctype = *ecode++; /* Code for the character type */ - -#ifdef SUPPORT_UCP - if (ctype == OP_PROP || ctype == OP_NOTPROP) - { - prop_fail_result = ctype == OP_NOTPROP; - prop_type = *ecode++; - if (prop_type >= 128) - { - prop_test_against = prop_type - 128; - prop_test_variable = &prop_category; - } - else - { - prop_test_against = prop_type; - prop_test_variable = &prop_chartype; - } - } - else prop_type = -1; -#endif - - /* First, ensure the minimum number of matches are present. Use inline - code for maximizing the speed, and do the type test once at the start - (i.e. keep it out of the loop). Also we can test that there are at least - the minimum number of bytes before we start. This isn't as effective in - UTF-8 mode, but it does no harm. Separate the UTF-8 code completely as that - is tidier. Also separate the UCP code, which can be the same for both UTF-8 - and single-bytes. */ - - if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - if (min > 0) - { -#ifdef SUPPORT_UCP - if (prop_type > 0) - { - for (i = 1; i <= min; i++) - { - GETCHARINC(c, eptr); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if ((*prop_test_variable == prop_test_against) == prop_fail_result) - RRETURN(MATCH_NOMATCH); - } - } - - /* Match extended Unicode sequences. We will get here only if the - support is in the binary; otherwise a compile-time error occurs. */ - - else if (ctype == OP_EXTUNI) - { - for (i = 1; i <= min; i++) - { - GETCHARINCTEST(c, eptr); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH); - while (eptr < md->end_subject) - { - int len = 1; - if (!md->utf8) c = *eptr; else - { - GETCHARLEN(c, eptr, len); - } - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if (prop_category != ucp_M) break; - eptr += len; - } - } - } - - else -#endif /* SUPPORT_UCP */ - -/* Handle all other cases when the coding is UTF-8 */ - -#ifdef SUPPORT_UTF8 - if (md->utf8) switch(ctype) - { - case OP_ANY: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || - (*eptr++ == NEWLINE && (ims & PCRE_DOTALL) == 0)) - RRETURN(MATCH_NOMATCH); - while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; - } - break; - - case OP_ANYBYTE: - eptr += min; - break; - - case OP_NOT_DIGIT: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - if (c < 128 && (md->ctypes[c] & ctype_digit) != 0) - RRETURN(MATCH_NOMATCH); - } - break; - - case OP_DIGIT: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || - *eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0) - RRETURN(MATCH_NOMATCH); - /* No need to skip more bytes - we know it's a 1-byte character */ - } - break; - - case OP_NOT_WHITESPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || - (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0)) - RRETURN(MATCH_NOMATCH); - while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; - } - break; - - case OP_WHITESPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || - *eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0) - RRETURN(MATCH_NOMATCH); - /* No need to skip more bytes - we know it's a 1-byte character */ - } - break; - - case OP_NOT_WORDCHAR: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || - (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0)) - RRETURN(MATCH_NOMATCH); - while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; - } - break; - - case OP_WORDCHAR: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || - *eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0) - RRETURN(MATCH_NOMATCH); - /* No need to skip more bytes - we know it's a 1-byte character */ - } - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } /* End switch(ctype) */ - - else -#endif /* SUPPORT_UTF8 */ - - /* Code for the non-UTF-8 case for minimum matching of operators other - than OP_PROP and OP_NOTPROP. */ - - switch(ctype) - { - case OP_ANY: - if ((ims & PCRE_DOTALL) == 0) - { - for (i = 1; i <= min; i++) - if (*eptr++ == NEWLINE) RRETURN(MATCH_NOMATCH); - } - else eptr += min; - break; - - case OP_ANYBYTE: - eptr += min; - break; - - case OP_NOT_DIGIT: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_DIGIT: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WHITESPACE: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_WHITESPACE: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WORDCHAR: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_word) != 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_WORDCHAR: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_word) == 0) - RRETURN(MATCH_NOMATCH); - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - } - - /* If min = max, continue at the same level without recursing */ - - if (min == max) continue; - - /* If minimizing, we have to test the rest of the pattern before each - subsequent match. Again, separate the UTF-8 case for speed, and also - separate the UCP cases. */ - - if (minimize) - { -#ifdef SUPPORT_UCP - if (prop_type > 0) - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if ((*prop_test_variable == prop_test_against) == prop_fail_result) - RRETURN(MATCH_NOMATCH); - } - } - - /* Match extended Unicode sequences. We will get here only if the - support is in the binary; otherwise a compile-time error occurs. */ - - else if (ctype == OP_EXTUNI) - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH); - while (eptr < md->end_subject) - { - int len = 1; - if (!md->utf8) c = *eptr; else - { - GETCHARLEN(c, eptr, len); - } - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if (prop_category != ucp_M) break; - eptr += len; - } - } - } - - else -#endif /* SUPPORT_UCP */ - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (md->utf8) - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - - GETCHARINC(c, eptr); - switch(ctype) - { - case OP_ANY: - if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH); - break; - - case OP_ANYBYTE: - break; - - case OP_NOT_DIGIT: - if (c < 256 && (md->ctypes[c] & ctype_digit) != 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_DIGIT: - if (c >= 256 || (md->ctypes[c] & ctype_digit) == 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WHITESPACE: - if (c < 256 && (md->ctypes[c] & ctype_space) != 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_WHITESPACE: - if (c >= 256 || (md->ctypes[c] & ctype_space) == 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WORDCHAR: - if (c < 256 && (md->ctypes[c] & ctype_word) != 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_WORDCHAR: - if (c >= 256 && (md->ctypes[c] & ctype_word) == 0) - RRETURN(MATCH_NOMATCH); - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (fi = min;; fi++) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - c = *eptr++; - switch(ctype) - { - case OP_ANY: - if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH); - break; - - case OP_ANYBYTE: - break; - - case OP_NOT_DIGIT: - if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_DIGIT: - if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WHITESPACE: - if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_WHITESPACE: - if ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WORDCHAR: - if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_WORDCHAR: - if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH); - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - } - } - /* Control never gets here */ - } - - /* If maximizing it is worth using inline code for speed, doing the type - test once at the start (i.e. keep it out of the loop). Again, keep the - UTF-8 and UCP stuff separate. */ - - else - { - pp = eptr; /* Remember where we started */ - -#ifdef SUPPORT_UCP - if (prop_type > 0) - { - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if ((*prop_test_variable == prop_test_against) == prop_fail_result) - break; - eptr+= len; - } - - /* eptr is now past the end of the maximum run */ - - for(;;) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - BACKCHAR(eptr); - } - } - - /* Match extended Unicode sequences. We will get here only if the - support is in the binary; otherwise a compile-time error occurs. */ - - else if (ctype == OP_EXTUNI) - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject) break; - GETCHARINCTEST(c, eptr); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if (prop_category == ucp_M) break; - while (eptr < md->end_subject) - { - int len = 1; - if (!md->utf8) c = *eptr; else - { - GETCHARLEN(c, eptr, len); - } - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if (prop_category != ucp_M) break; - eptr += len; - } - } - - /* eptr is now past the end of the maximum run */ - - for(;;) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - for (;;) /* Move back over one extended */ - { - int len = 1; - BACKCHAR(eptr); - if (!md->utf8) c = *eptr; else - { - GETCHARLEN(c, eptr, len); - } - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if (prop_category != ucp_M) break; - eptr--; - } - } - } - - else -#endif /* SUPPORT_UCP */ - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - - if (md->utf8) - { - switch(ctype) - { - case OP_ANY: - - /* Special code is required for UTF8, but when the maximum is unlimited - we don't need it, so we repeat the non-UTF8 code. This is probably - worth it, because .* is quite a common idiom. */ - - if (max < INT_MAX) - { - if ((ims & PCRE_DOTALL) == 0) - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || *eptr == NEWLINE) break; - eptr++; - while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; - } - } - else - { - for (i = min; i < max; i++) - { - eptr++; - while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; - } - } - } - - /* Handle unlimited UTF-8 repeat */ - - else - { - if ((ims & PCRE_DOTALL) == 0) - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || *eptr == NEWLINE) break; - eptr++; - } - break; - } - else - { - c = max - min; - if (c > md->end_subject - eptr) c = md->end_subject - eptr; - eptr += c; - } - } - break; - - /* The byte case is the same as non-UTF8 */ - - case OP_ANYBYTE: - c = max - min; - if (c > md->end_subject - eptr) c = md->end_subject - eptr; - eptr += c; - break; - - case OP_NOT_DIGIT: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c < 256 && (md->ctypes[c] & ctype_digit) != 0) break; - eptr+= len; - } - break; - - case OP_DIGIT: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c >= 256 ||(md->ctypes[c] & ctype_digit) == 0) break; - eptr+= len; - } - break; - - case OP_NOT_WHITESPACE: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c < 256 && (md->ctypes[c] & ctype_space) != 0) break; - eptr+= len; - } - break; - - case OP_WHITESPACE: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c >= 256 ||(md->ctypes[c] & ctype_space) == 0) break; - eptr+= len; - } - break; - - case OP_NOT_WORDCHAR: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c < 256 && (md->ctypes[c] & ctype_word) != 0) break; - eptr+= len; - } - break; - - case OP_WORDCHAR: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c >= 256 || (md->ctypes[c] & ctype_word) == 0) break; - eptr+= len; - } - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - - /* eptr is now past the end of the maximum run */ - - for(;;) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - BACKCHAR(eptr); - } - } - else -#endif - - /* Not UTF-8 mode */ - { - switch(ctype) - { - case OP_ANY: - if ((ims & PCRE_DOTALL) == 0) - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || *eptr == NEWLINE) break; - eptr++; - } - break; - } - /* For DOTALL case, fall through and treat as \C */ - - case OP_ANYBYTE: - c = max - min; - if (c > md->end_subject - eptr) c = md->end_subject - eptr; - eptr += c; - break; - - case OP_NOT_DIGIT: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_digit) != 0) - break; - eptr++; - } - break; - - case OP_DIGIT: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_digit) == 0) - break; - eptr++; - } - break; - - case OP_NOT_WHITESPACE: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_space) != 0) - break; - eptr++; - } - break; - - case OP_WHITESPACE: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_space) == 0) - break; - eptr++; - } - break; - - case OP_NOT_WORDCHAR: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_word) != 0) - break; - eptr++; - } - break; - - case OP_WORDCHAR: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_word) == 0) - break; - eptr++; - } - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - - /* eptr is now past the end of the maximum run */ - - while (eptr >= pp) - { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - eptr--; - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - } - - /* Get here if we can't make it match with any permitted repetitions */ - - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - - /* There's been some horrible disaster. Since all codes > OP_BRA are - for capturing brackets, and there shouldn't be any gaps between 0 and - OP_BRA, arrival here can only mean there is something seriously wrong - in the code above or the OP_xxx definitions. */ - - default: - DPRINTF(("Unknown opcode %d\n", *ecode)); - RRETURN(PCRE_ERROR_UNKNOWN_NODE); - } - - /* Do not stick any code in here without much thought; it is assumed - that "continue" in the code above comes out to here to repeat the main - loop. */ - - } /* End of main loop */ -/* Control never reaches here */ -} - - -/*************************************************************************** -**************************************************************************** - RECURSION IN THE match() FUNCTION - -Undefine all the macros that were defined above to handle this. */ - -#ifdef NO_RECURSE -#undef eptr -#undef ecode -#undef offset_top -#undef ims -#undef eptrb -#undef flags - -#undef callpat -#undef charptr -#undef data -#undef next -#undef pp -#undef prev -#undef saved_eptr - -#undef new_recursive - -#undef cur_is_word -#undef condition -#undef minimize -#undef prev_is_word - -#undef original_ims - -#undef ctype -#undef length -#undef max -#undef min -#undef number -#undef offset -#undef op -#undef save_capture_last -#undef save_offset1 -#undef save_offset2 -#undef save_offset3 -#undef stacksave - -#undef newptrb - -#endif - -/* These two are defined as macros in both cases */ - -#undef fc -#undef fi - -/*************************************************************************** -***************************************************************************/ - - - -/************************************************* -* Execute a Regular Expression * -*************************************************/ - -/* This function applies a compiled re to a subject string and picks out -portions of the string if it matches. Two elements in the vector are set for -each substring: the offsets to the start and end of the substring. - -Arguments: - argument_re points to the compiled expression - extra_data points to extra data or is NULL - subject points to the subject string - length length of subject string (may contain binary zeros) - start_offset where to start in the subject string - options option bits - offsets points to a vector of ints to be filled in with offsets - offsetcount the number of elements in the vector - -Returns: > 0 => success; value is the number of elements filled in - = 0 => success, but offsets is not big enough - -1 => failed to match - < -1 => some kind of unexpected problem -*/ - -EXPORT int -pcre_exec(const pcre *argument_re, const pcre_extra *extra_data, - const char *subject, int length, int start_offset, int options, int *offsets, - int offsetcount) -{ -int rc, resetcount, ocount; -int first_byte = -1; -int req_byte = -1; -int req_byte2 = -1; -unsigned long int ims = 0; -BOOL using_temporary_offsets = FALSE; -BOOL anchored; -BOOL startline; -BOOL first_byte_caseless = FALSE; -BOOL req_byte_caseless = FALSE; -match_data match_block; -const uschar *tables; -const uschar *start_bits = NULL; -const uschar *start_match = (const uschar *)subject + start_offset; -const uschar *end_subject; -const uschar *req_byte_ptr = start_match - 1; - -pcre_study_data internal_study; -const pcre_study_data *study; - -real_pcre internal_re; -const real_pcre *external_re = (const real_pcre *)argument_re; -const real_pcre *re = external_re; - -/* Plausibility checks */ - -if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION; -if (re == NULL || subject == NULL || - (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL; -if (offsetcount < 0) return PCRE_ERROR_BADCOUNT; - -/* Fish out the optional data from the extra_data structure, first setting -the default values. */ - -study = NULL; -match_block.match_limit = MATCH_LIMIT; -match_block.callout_data = NULL; - -/* The table pointer is always in native byte order. */ - -tables = external_re->tables; - -if (extra_data != NULL) - { - register unsigned int flags = extra_data->flags; - if ((flags & PCRE_EXTRA_STUDY_DATA) != 0) - study = (const pcre_study_data *)extra_data->study_data; - if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0) - match_block.match_limit = extra_data->match_limit; - if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0) - match_block.callout_data = extra_data->callout_data; - if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables; - } - -/* If the exec call supplied NULL for tables, use the inbuilt ones. This -is a feature that makes it possible to save compiled regex and re-use them -in other programs later. */ - -if (tables == NULL) tables = pcre_default_tables; - -/* Check that the first field in the block is the magic number. If it is not, -test for a regex that was compiled on a host of opposite endianness. If this is -the case, flipped values are put in internal_re and internal_study if there was -study data too. */ - -if (re->magic_number != MAGIC_NUMBER) - { - re = try_flipped(re, &internal_re, study, &internal_study); - if (re == NULL) return PCRE_ERROR_BADMAGIC; - if (study != NULL) study = &internal_study; - } - -/* Set up other data */ - -anchored = ((re->options | options) & PCRE_ANCHORED) != 0; -startline = (re->options & PCRE_STARTLINE) != 0; - -/* The code starts after the real_pcre block and the capture name table. */ - -match_block.start_code = (const uschar *)external_re + re->name_table_offset + - re->name_count * re->name_entry_size; - -match_block.start_subject = (const uschar *)subject; -match_block.start_offset = start_offset; -match_block.end_subject = match_block.start_subject + length; -end_subject = match_block.end_subject; - -match_block.endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0; -match_block.utf8 = (re->options & PCRE_UTF8) != 0; - -match_block.notbol = (options & PCRE_NOTBOL) != 0; -match_block.noteol = (options & PCRE_NOTEOL) != 0; -match_block.notempty = (options & PCRE_NOTEMPTY) != 0; -match_block.partial = (options & PCRE_PARTIAL) != 0; -match_block.hitend = FALSE; - -match_block.recursive = NULL; /* No recursion at top level */ - -match_block.lcc = tables + lcc_offset; -match_block.ctypes = tables + ctypes_offset; - -/* Partial matching is supported only for a restricted set of regexes at the -moment. */ - -if (match_block.partial && (re->options & PCRE_NOPARTIAL) != 0) - return PCRE_ERROR_BADPARTIAL; - -/* Check a UTF-8 string if required. Unfortunately there's no way of passing -back the character offset. */ - -#ifdef SUPPORT_UTF8 -if (match_block.utf8 && (options & PCRE_NO_UTF8_CHECK) == 0) - { - if (valid_utf8((uschar *)subject, length) >= 0) - return PCRE_ERROR_BADUTF8; - if (start_offset > 0 && start_offset < length) - { - int tb = ((uschar *)subject)[start_offset]; - if (tb > 127) - { - tb &= 0xc0; - if (tb != 0 && tb != 0xc0) return PCRE_ERROR_BADUTF8_OFFSET; - } - } - } -#endif - -/* The ims options can vary during the matching as a result of the presence -of (?ims) items in the pattern. They are kept in a local variable so that -restoring at the exit of a group is easy. */ - -ims = re->options & (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL); - -/* If the expression has got more back references than the offsets supplied can -hold, we get a temporary chunk of working store to use during the matching. -Otherwise, we can use the vector supplied, rounding down its size to a multiple -of 3. */ - -ocount = offsetcount - (offsetcount % 3); - -if (re->top_backref > 0 && re->top_backref >= ocount/3) - { - ocount = re->top_backref * 3 + 3; - match_block.offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int)); - if (match_block.offset_vector == NULL) return PCRE_ERROR_NOMEMORY; - using_temporary_offsets = TRUE; - DPRINTF(("Got memory to hold back references\n")); - } -else match_block.offset_vector = offsets; - -match_block.offset_end = ocount; -match_block.offset_max = (2*ocount)/3; -match_block.offset_overflow = FALSE; -match_block.capture_last = -1; - -/* Compute the minimum number of offsets that we need to reset each time. Doing -this makes a huge difference to execution time when there aren't many brackets -in the pattern. */ - -resetcount = 2 + re->top_bracket * 2; -if (resetcount > offsetcount) resetcount = ocount; - -/* Reset the working variable associated with each extraction. These should -never be used unless previously set, but they get saved and restored, and so we -initialize them to avoid reading uninitialized locations. */ - -if (match_block.offset_vector != NULL) - { - register int *iptr = match_block.offset_vector + ocount; - register int *iend = iptr - resetcount/2 + 1; - while (--iptr >= iend) *iptr = -1; - } - -/* Set up the first character to match, if available. The first_byte value is -never set for an anchored regular expression, but the anchoring may be forced -at run time, so we have to test for anchoring. The first char may be unset for -an unanchored pattern, of course. If there's no first char and the pattern was -studied, there may be a bitmap of possible first characters. */ - -if (!anchored) - { - if ((re->options & PCRE_FIRSTSET) != 0) - { - first_byte = re->first_byte & 255; - if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE) - first_byte = match_block.lcc[first_byte]; - } - else - if (!startline && study != NULL && - (study->options & PCRE_STUDY_MAPPED) != 0) - start_bits = study->start_bits; - } - -/* For anchored or unanchored matches, there may be a "last known required -character" set. */ - -if ((re->options & PCRE_REQCHSET) != 0) - { - req_byte = re->req_byte & 255; - req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0; - req_byte2 = (tables + fcc_offset)[req_byte]; /* case flipped */ - } - -/* Loop for handling unanchored repeated matching attempts; for anchored regexs -the loop runs just once. */ - -do - { - /* Reset the maximum number of extractions we might see. */ - - if (match_block.offset_vector != NULL) - { - register int *iptr = match_block.offset_vector; - register int *iend = iptr + resetcount; - while (iptr < iend) *iptr++ = -1; - } - - /* Advance to a unique first char if possible */ - - if (first_byte >= 0) - { - if (first_byte_caseless) - while (start_match < end_subject && - match_block.lcc[*start_match] != first_byte) - start_match++; - else - while (start_match < end_subject && *start_match != first_byte) - start_match++; - } - - /* Or to just after \n for a multiline match if possible */ - - else if (startline) - { - if (start_match > match_block.start_subject + start_offset) - { - while (start_match < end_subject && start_match[-1] != NEWLINE) - start_match++; - } - } - - /* Or to a non-unique first char after study */ - - else if (start_bits != NULL) - { - while (start_match < end_subject) - { - register unsigned int c = *start_match; - if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++; else break; - } - } - -#ifdef DEBUG /* Sigh. Some compilers never learn. */ - printf(">>>> Match against: "); - pchars(start_match, end_subject - start_match, TRUE, &match_block); - printf("\n"); -#endif - - /* If req_byte is set, we know that that character must appear in the subject - for the match to succeed. If the first character is set, req_byte must be - later in the subject; otherwise the test starts at the match point. This - optimization can save a huge amount of backtracking in patterns with nested - unlimited repeats that aren't going to match. Writing separate code for - cased/caseless versions makes it go faster, as does using an autoincrement - and backing off on a match. - - HOWEVER: when the subject string is very, very long, searching to its end can - take a long time, and give bad performance on quite ordinary patterns. This - showed up when somebody was matching /^C/ on a 32-megabyte string... so we - don't do this when the string is sufficiently long. - - ALSO: this processing is disabled when partial matching is requested. - */ - - if (req_byte >= 0 && - end_subject - start_match < REQ_BYTE_MAX && - !match_block.partial) - { - register const uschar *p = start_match + ((first_byte >= 0)? 1 : 0); - - /* We don't need to repeat the search if we haven't yet reached the - place we found it at last time. */ - - if (p > req_byte_ptr) - { - if (req_byte_caseless) - { - while (p < end_subject) - { - register int pp = *p++; - if (pp == req_byte || pp == req_byte2) { p--; break; } - } - } - else - { - while (p < end_subject) - { - if (*p++ == req_byte) { p--; break; } - } - } - - /* If we can't find the required character, break the matching loop */ - - if (p >= end_subject) break; - - /* If we have found the required character, save the point where we - found it, so that we don't search again next time round the loop if - the start hasn't passed this character yet. */ - - req_byte_ptr = p; - } - } - - /* When a match occurs, substrings will be set for all internal extractions; - we just need to set up the whole thing as substring 0 before returning. If - there were too many extractions, set the return code to zero. In the case - where we had to get some local store to hold offsets for backreferences, copy - those back references that we can. In this case there need not be overflow - if certain parts of the pattern were not used. */ - - match_block.start_match = start_match; - match_block.match_call_count = 0; - - rc = match(start_match, match_block.start_code, 2, &match_block, ims, NULL, - match_isgroup); - - if (rc == MATCH_NOMATCH) - { - start_match++; -#ifdef SUPPORT_UTF8 - if (match_block.utf8) - while(start_match < end_subject && (*start_match & 0xc0) == 0x80) - start_match++; -#endif - continue; - } - - if (rc != MATCH_MATCH) - { - DPRINTF((">>>> error: returning %d\n", rc)); - return rc; - } - - /* We have a match! Copy the offset information from temporary store if - necessary */ - - if (using_temporary_offsets) - { - if (offsetcount >= 4) - { - memcpy(offsets + 2, match_block.offset_vector + 2, - (offsetcount - 2) * sizeof(int)); - DPRINTF(("Copied offsets from temporary memory\n")); - } - if (match_block.end_offset_top > offsetcount) - match_block.offset_overflow = TRUE; - - DPRINTF(("Freeing temporary memory\n")); - (pcre_free)(match_block.offset_vector); - } - - rc = match_block.offset_overflow? 0 : match_block.end_offset_top/2; - - if (offsetcount < 2) rc = 0; else - { - offsets[0] = start_match - match_block.start_subject; - offsets[1] = match_block.end_match_ptr - match_block.start_subject; - } - - DPRINTF((">>>> returning %d\n", rc)); - return rc; - } - -/* This "while" is the end of the "do" above */ - -while (!anchored && start_match <= end_subject); - -if (using_temporary_offsets) - { - DPRINTF(("Freeing temporary memory\n")); - (pcre_free)(match_block.offset_vector); - } - -if (match_block.partial && match_block.hitend) - { - DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n")); - return PCRE_ERROR_PARTIAL; - } -else - { - DPRINTF((">>>> returning PCRE_ERROR_NOMATCH\n")); - return PCRE_ERROR_NOMATCH; - } -} - -/* End of pcre.c */ diff --git a/src/src/pcre/pcre.h b/src/src/pcre/pcre.h index 1a1e9b87d..46be6ed7e 100644 --- a/src/src/pcre/pcre.h +++ b/src/src/pcre/pcre.h @@ -1,9 +1,11 @@ +/* $Cambridge: exim/src/src/pcre/pcre.h,v 1.4 2006/11/07 16:50:36 ph10 Exp $ */ + /************************************************* * Perl-Compatible Regular Expressions * *************************************************/ -/* In its original form, this is the .in file that is transformed by -"configure" into pcre.h. +/* This is the public header file for the PCRE library, to be #included by +applications that call the PCRE functions. Copyright (c) 1997-2005 University of Cambridge @@ -39,14 +41,26 @@ POSSIBILITY OF SUCH DAMAGE. #ifndef _PCRE_H #define _PCRE_H -/* The file pcre.h is build by "configure". Do not edit it; instead -make changes to pcre.in. */ +/* The current PCRE version information. */ + +/* NOTES FOR FUTURE MAINTAINERS: Do not use numbers with leading zeros, because +they may be treated as octal constants. The PCRE_PRERELEASE feature is for +identifying release candidates. It might be defined as -RC2, for example. In +real releases, it should be defined empty. Do not change the alignment of these +statments. The code in ./configure greps out the version numbers by using "cut" +to get values from column 29 onwards. These are substituted into pcre-config +and libpcre.pc. The values are not put into configure.ac and substituted here +(which would simplify this issue) because that makes life harder for those who +cannot run ./configure. As it now stands, this file need not be edited in that +circumstance. */ #define PCRE_MAJOR 6 -#define PCRE_MINOR 2 -#define PCRE_DATE 01-Aug-2005 +#define PCRE_MINOR 7 +#define PCRE_PRERELEASE +#define PCRE_DATE 04-Jul-2006 -/* Win32 uses DLL by default; it needs special stuff for exported functions. */ +/* Win32 uses DLL by default; it needs special stuff for exported functions +when building PCRE. */ #ifdef _WIN32 # ifdef PCRE_DEFINITION @@ -60,7 +74,7 @@ make changes to pcre.in. */ # endif #endif -/* For other operating systems, we use the standard "extern". */ +/* Otherwise, we use the standard "extern". */ #ifndef PCRE_DATA_SCOPE # ifdef __cplusplus @@ -102,6 +116,10 @@ extern "C" { #define PCRE_DFA_SHORTEST 0x00010000 #define PCRE_DFA_RESTART 0x00020000 #define PCRE_FIRSTLINE 0x00040000 +#define PCRE_DUPNAMES 0x00080000 +#define PCRE_NEWLINE_CR 0x00100000 +#define PCRE_NEWLINE_LF 0x00200000 +#define PCRE_NEWLINE_CRLF 0x00300000 /* Exec-time and get/set-time error codes */ @@ -125,6 +143,7 @@ extern "C" { #define PCRE_ERROR_DFA_UMLIMIT (-18) #define PCRE_ERROR_DFA_WSSIZE (-19) #define PCRE_ERROR_DFA_RECURSE (-20) +#define PCRE_ERROR_RECURSIONLIMIT (-21) /* Request types for pcre_fullinfo() */ @@ -142,7 +161,8 @@ extern "C" { #define PCRE_INFO_STUDYSIZE 10 #define PCRE_INFO_DEFAULT_TABLES 11 -/* Request types for pcre_config() */ +/* Request types for pcre_config(). Do not re-arrange, in order to remain +compatible. */ #define PCRE_CONFIG_UTF8 0 #define PCRE_CONFIG_NEWLINE 1 @@ -151,19 +171,30 @@ extern "C" { #define PCRE_CONFIG_MATCH_LIMIT 4 #define PCRE_CONFIG_STACKRECURSE 5 #define PCRE_CONFIG_UNICODE_PROPERTIES 6 +#define PCRE_CONFIG_MATCH_LIMIT_RECURSION 7 -/* Bit flags for the pcre_extra structure */ +/* Bit flags for the pcre_extra structure. Do not re-arrange or redefine +these bits, just add new ones on the end, in order to remain compatible. */ -#define PCRE_EXTRA_STUDY_DATA 0x0001 -#define PCRE_EXTRA_MATCH_LIMIT 0x0002 -#define PCRE_EXTRA_CALLOUT_DATA 0x0004 -#define PCRE_EXTRA_TABLES 0x0008 +#define PCRE_EXTRA_STUDY_DATA 0x0001 +#define PCRE_EXTRA_MATCH_LIMIT 0x0002 +#define PCRE_EXTRA_CALLOUT_DATA 0x0004 +#define PCRE_EXTRA_TABLES 0x0008 +#define PCRE_EXTRA_MATCH_LIMIT_RECURSION 0x0010 /* Types */ struct real_pcre; /* declaration; the definition is private */ typedef struct real_pcre pcre; +/* When PCRE is compiled as a C++ library, the subject pointer type can be +replaced with a custom type. For conventional use, the public interface is a +const char *. */ + +#ifndef PCRE_SPTR +#define PCRE_SPTR const char * +#endif + /* The structure for passing additional data to pcre_exec(). This is defined in such as way as to be extensible. Always add new fields at the end, in order to remain compatible. */ @@ -174,6 +205,7 @@ typedef struct pcre_extra { unsigned long int match_limit; /* Maximum number of calls to match() */ void *callout_data; /* Data passed back in callouts */ const unsigned char *tables; /* Pointer to character tables */ + unsigned long int match_limit_recursion; /* Max recursive calls to match() */ } pcre_extra; /* The structure for passing out data via the pcre_callout_function. We use a @@ -186,7 +218,7 @@ typedef struct pcre_callout_block { /* ------------------------ Version 0 ------------------------------- */ int callout_number; /* Number compiled into pattern */ int *offset_vector; /* The offset vector */ - const char *subject; /* The subject being matched */ + PCRE_SPTR subject; /* The subject being matched */ int subject_length; /* The length of the subject */ int start_match; /* Offset to start of this match attempt */ int current_position; /* Where we currently are in the subject */ @@ -232,7 +264,7 @@ PCRE_DATA_SCOPE int pcre_copy_substring(const char *, int *, int, int, char *, int); PCRE_DATA_SCOPE int pcre_dfa_exec(const pcre *, const pcre_extra *, const char *, int, int, int, int *, int , int *, int); -PCRE_DATA_SCOPE int pcre_exec(const pcre *, const pcre_extra *, const char *, +PCRE_DATA_SCOPE int pcre_exec(const pcre *, const pcre_extra *, PCRE_SPTR, int, int, int, int *, int); PCRE_DATA_SCOPE void pcre_free_substring(const char *); PCRE_DATA_SCOPE void pcre_free_substring_list(const char **); @@ -241,6 +273,8 @@ PCRE_DATA_SCOPE int pcre_fullinfo(const pcre *, const pcre_extra *, int, PCRE_DATA_SCOPE int pcre_get_named_substring(const pcre *, const char *, int *, int, const char *, const char **); PCRE_DATA_SCOPE int pcre_get_stringnumber(const pcre *, const char *); +PCRE_DATA_SCOPE int pcre_get_stringtable_entries(const pcre *, const char *, + char **, char **); PCRE_DATA_SCOPE int pcre_get_substring(const char *, int *, int, int, const char **); PCRE_DATA_SCOPE int pcre_get_substring_list(const char *, int *, int, diff --git a/src/src/pcre/pcre_compile.c b/src/src/pcre/pcre_compile.c index 7306a278c..eaa9550a7 100644 --- a/src/src/pcre/pcre_compile.c +++ b/src/src/pcre/pcre_compile.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_compile.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_compile.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -44,9 +44,19 @@ POSSIBILITY OF SUCH DAMAGE. supporting internal functions that are not used by other modules. */ +#define NLBLOCK cd /* The block containing newline information */ #include "pcre_internal.h" +/* When DEBUG is defined, we need the pcre_printint() function, which is also +used by pcretest. DEBUG is not defined when building a production library. */ + +#ifdef DEBUG +#include "pcre_printint.src" +#endif + + + /************************************************* * Code parameters and static tables * *************************************************/ @@ -109,7 +119,7 @@ static const short int escapes[] = { /* Tables of names of POSIX character classes and their lengths. The list is -terminated by a zero length entry. The first three must be alpha, upper, lower, +terminated by a zero length entry. The first three must be alpha, lower, upper, as this is assumed for handling case independence. */ static const char *const posix_names[] = { @@ -120,25 +130,31 @@ static const char *const posix_names[] = { static const uschar posix_name_lengths[] = { 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 }; -/* Table of class bit maps for each POSIX class; up to three may be combined -to form the class. The table for [:blank:] is dynamically modified to remove -the vertical space characters. */ +/* Table of class bit maps for each POSIX class. Each class is formed from a +base map, with an optional addition or removal of another map. Then, for some +classes, there is some additional tweaking: for [:blank:] the vertical space +characters are removed, and for [:alpha:] and [:alnum:] the underscore +character is removed. The triples in the table consist of the base map offset, +second map offset or -1 if no second map, and a non-negative value for map +addition or a negative value for map subtraction (if there are two maps). The +absolute value of the third field has these meanings: 0 => no tweaking, 1 => +remove vertical space characters, 2 => remove underscore. */ static const int posix_class_maps[] = { - cbit_lower, cbit_upper, -1, /* alpha */ - cbit_lower, -1, -1, /* lower */ - cbit_upper, -1, -1, /* upper */ - cbit_digit, cbit_lower, cbit_upper, /* alnum */ - cbit_print, cbit_cntrl, -1, /* ascii */ - cbit_space, -1, -1, /* blank - a GNU extension */ - cbit_cntrl, -1, -1, /* cntrl */ - cbit_digit, -1, -1, /* digit */ - cbit_graph, -1, -1, /* graph */ - cbit_print, -1, -1, /* print */ - cbit_punct, -1, -1, /* punct */ - cbit_space, -1, -1, /* space */ - cbit_word, -1, -1, /* word - a Perl extension */ - cbit_xdigit,-1, -1 /* xdigit */ + cbit_word, cbit_digit, -2, /* alpha */ + cbit_lower, -1, 0, /* lower */ + cbit_upper, -1, 0, /* upper */ + cbit_word, -1, 2, /* alnum - word without underscore */ + cbit_print, cbit_cntrl, 0, /* ascii */ + cbit_space, -1, 1, /* blank - a GNU extension */ + cbit_cntrl, -1, 0, /* cntrl */ + cbit_digit, -1, 0, /* digit */ + cbit_graph, -1, 0, /* graph */ + cbit_print, -1, 0, /* print */ + cbit_punct, -1, 0, /* punct */ + cbit_space, -1, 0, /* space */ + cbit_word, -1, 0, /* word - a Perl extension */ + cbit_xdigit,-1, 0 /* xdigit */ }; @@ -177,7 +193,7 @@ static const char *error_texts[] = { "unrecognized character after (?<", /* 25 */ "lookbehind assertion is not fixed length", - "malformed number after (?(", + "malformed number or name after (?(", "conditional group contains more than two branches", "assertion expected after (?(", "(?R or (?digits must be followed by )", @@ -197,12 +213,17 @@ static const char *error_texts[] = { "recursive call could loop indefinitely", "unrecognized character after (?P", "syntax error after (?P", - "two named groups have the same name", + "two named subpatterns have the same name", "invalid UTF-8 string", /* 45 */ "support for \\P, \\p, and \\X has not been compiled", "malformed \\P or \\p sequence", - "unknown property name after \\P or \\p" + "unknown property name after \\P or \\p", + "subpattern name is too long (maximum 32 characters)", + "too many named subpatterns (maximum 10,000)", + /* 50 */ + "repeated subpattern is too long", + "octal value is greater than \\377 (not in UTF-8 mode)" }; @@ -364,12 +385,15 @@ static int check_escape(const uschar **ptrptr, int *errorcodeptr, int bracount, int options, BOOL isclass) { -const uschar *ptr = *ptrptr; +BOOL utf8 = (options & PCRE_UTF8) != 0; +const uschar *ptr = *ptrptr + 1; int c, i; +GETCHARINCTEST(c, ptr); /* Get character value, increment pointer */ +ptr--; /* Set pointer back to the last byte */ + /* If backslash is at the end of the pattern, it's an error. */ -c = *(++ptr); if (c == 0) *errorcodeptr = ERR1; /* Non-alphamerics are literals. For digits or letters, do an initial lookup in @@ -444,49 +468,56 @@ else } /* \0 always starts an octal number, but we may drop through to here with a - larger first octal digit. */ + larger first octal digit. The original code used just to take the least + significant 8 bits of octal numbers (I think this is what early Perls used + to do). Nowadays we allow for larger numbers in UTF-8 mode, but no more + than 3 octal digits. */ case '0': c -= '0'; while(i++ < 2 && ptr[1] >= '0' && ptr[1] <= '7') c = c * 8 + *(++ptr) - '0'; - c &= 255; /* Take least significant 8 bits */ + if (!utf8 && c > 255) *errorcodeptr = ERR51; break; - /* \x is complicated when UTF-8 is enabled. \x{ddd} is a character number - which can be greater than 0xff, but only if the ddd are hex digits. */ + /* \x is complicated. \x{ddd} is a character number which can be greater + than 0xff in utf8 mode, but only if the ddd are hex digits. If not, { is + treated as a data character. */ case 'x': -#ifdef SUPPORT_UTF8 - if (ptr[1] == '{' && (options & PCRE_UTF8) != 0) + if (ptr[1] == '{') { const uschar *pt = ptr + 2; - register int count = 0; + int count = 0; + c = 0; while ((digitab[*pt] & ctype_xdigit) != 0) { - int cc = *pt++; + register int cc = *pt++; + if (c == 0 && cc == '0') continue; /* Leading zeroes */ count++; + #if !EBCDIC /* ASCII coding */ if (cc >= 'a') cc -= 32; /* Convert to upper case */ - c = c * 16 + cc - ((cc < 'A')? '0' : ('A' - 10)); + c = (c << 4) + cc - ((cc < 'A')? '0' : ('A' - 10)); #else /* EBCDIC coding */ if (cc >= 'a' && cc <= 'z') cc += 64; /* Convert to upper case */ - c = c * 16 + cc - ((cc >= '0')? '0' : ('A' - 10)); + c = (c << 4) + cc - ((cc >= '0')? '0' : ('A' - 10)); #endif } + if (*pt == '}') { - if (c < 0 || count > 8) *errorcodeptr = ERR34; + if (c < 0 || count > (utf8? 8 : 2)) *errorcodeptr = ERR34; ptr = pt; break; } + /* If the sequence of hex digits does not end with '}', then we don't recognize this construct; fall through to the normal \x handling. */ } -#endif - /* Read just a single hex char */ + /* Read just a single-byte hex-defined char */ c = 0; while (i++ < 2 && (digitab[ptr[1]] & ctype_xdigit) != 0) @@ -562,25 +593,26 @@ escape sequence. Argument: ptrptr points to the pattern position pointer negptr points to a boolean that is set TRUE for negation else FALSE + dptr points to an int that is set to the detailed property value errorcodeptr points to the error code variable -Returns: value from ucp_type_table, or -1 for an invalid type +Returns: type value from ucp_type_table, or -1 for an invalid type */ static int -get_ucp(const uschar **ptrptr, BOOL *negptr, int *errorcodeptr) +get_ucp(const uschar **ptrptr, BOOL *negptr, int *dptr, int *errorcodeptr) { int c, i, bot, top; const uschar *ptr = *ptrptr; -char name[4]; +char name[32]; c = *(++ptr); if (c == 0) goto ERROR_RETURN; *negptr = FALSE; -/* \P or \p can be followed by a one- or two-character name in {}, optionally -preceded by ^ for negation. */ +/* \P or \p can be followed by a name in {}, optionally preceded by ^ for +negation. */ if (c == '{') { @@ -589,18 +621,14 @@ if (c == '{') *negptr = TRUE; ptr++; } - for (i = 0; i <= 2; i++) + for (i = 0; i < sizeof(name) - 1; i++) { c = *(++ptr); if (c == 0) goto ERROR_RETURN; if (c == '}') break; name[i] = c; } - if (c !='}') /* Try to distinguish error cases */ - { - while (*(++ptr) != 0 && *ptr != '}'); - if (*ptr == '}') goto UNKNOWN_RETURN; else goto ERROR_RETURN; - } + if (c !='}') goto ERROR_RETURN; name[i] = 0; } @@ -621,13 +649,16 @@ top = _pcre_utt_size; while (bot < top) { - i = (bot + top)/2; + i = (bot + top) >> 1; c = strcmp(name, _pcre_utt[i].name); - if (c == 0) return _pcre_utt[i].value; + if (c == 0) + { + *dptr = _pcre_utt[i].value; + return _pcre_utt[i].type; + } if (c > 0) bot = i + 1; else top = i; } -UNKNOWN_RETURN: *errorcodeptr = ERR47; *ptrptr = ptr; return -1; @@ -743,6 +774,49 @@ return p; /************************************************* +* Find forward referenced named subpattern * +*************************************************/ + +/* This function scans along a pattern looking for capturing subpatterns, and +counting them. If it finds a named pattern that matches the name it is given, +it returns its number. This is used for forward references to named +subpatterns. We know that if (?P< is encountered, the name will be terminated +by '>' because that is checked in the first pass. + +Arguments: + pointer current position in the pattern + count current count of capturing parens + name name to seek + namelen name length + +Returns: the number of the named subpattern, or -1 if not found +*/ + +static int +find_named_parens(const uschar *ptr, int count, const uschar *name, int namelen) +{ +const uschar *thisname; +for (; *ptr != 0; ptr++) + { + if (*ptr == '\\' && ptr[1] != 0) { ptr++; continue; } + if (*ptr != '(') continue; + if (ptr[1] != '?') { count++; continue; } + if (ptr[2] == '(') { ptr += 2; continue; } + if (ptr[2] != 'P' || ptr[3] != '<') continue; + count++; + ptr += 4; + thisname = ptr; + while (*ptr != '>') ptr++; + if (namelen == ptr - thisname && + strncmp((char *)name, (char*)thisname, namelen) == 0) + return count; + } +return -1; +} + + + +/************************************************* * Find first significant op code * *************************************************/ @@ -897,6 +971,7 @@ for (;;) case OP_CHAR: case OP_CHARNC: + case OP_NOT: branchlength++; cc += 2; #ifdef SUPPORT_UTF8 @@ -930,7 +1005,7 @@ for (;;) case OP_PROP: case OP_NOTPROP: - cc++; + cc += 2; /* Fall through */ case OP_NOT_DIGIT: @@ -1011,14 +1086,19 @@ Returns: pointer to the opcode for the bracket, or NULL if not found static const uschar * find_bracket(const uschar *code, BOOL utf8, int number) { -#ifndef SUPPORT_UTF8 -utf8 = utf8; /* Stop pedantic compilers complaining */ -#endif - for (;;) { register int c = *code; if (c == OP_END) return NULL; + + /* XCLASS is used for classes that cannot be represented just by a bit + map. This includes negated single high-valued characters. The length in + the table is zero; the actual length is stored in the compiled code. */ + + if (c == OP_XCLASS) code += GET(code, 1); + + /* Handle bracketed group */ + else if (c > OP_BRA) { int n = c - OP_BRA; @@ -1026,17 +1106,16 @@ for (;;) if (n == number) return (uschar *)code; code += _pcre_OP_lengths[OP_BRA]; } + + /* Otherwise, we get the item's length from the table. In UTF-8 mode, opcodes + that are followed by a character may be followed by a multi-byte character. + The length in the table is a minimum, so we have to scan along to skip the + extra bytes. All opcodes are less than 128, so we can use relatively + efficient code. */ + else { code += _pcre_OP_lengths[c]; - -#ifdef SUPPORT_UTF8 - - /* In UTF-8 mode, opcodes that are followed by a character may be followed - by a multi-byte character. The length in the table is a minimum, so we have - to scan along to skip the extra bytes. All opcodes are less than 128, so we - can use relatively efficient code. */ - if (utf8) switch(c) { case OP_CHAR: @@ -1052,16 +1131,7 @@ for (;;) case OP_MINQUERY: while ((*code & 0xc0) == 0x80) code++; break; - - /* XCLASS is used for classes that cannot be represented just by a bit - map. This includes negated single high-valued characters. The length in - the table is zero; the actual length is stored in the compiled code. */ - - case OP_XCLASS: - code += GET(code, 1) + 1; - break; } -#endif } } } @@ -1085,30 +1155,34 @@ Returns: pointer to the opcode for OP_RECURSE, or NULL if not found static const uschar * find_recurse(const uschar *code, BOOL utf8) { -#ifndef SUPPORT_UTF8 -utf8 = utf8; /* Stop pedantic compilers complaining */ -#endif - for (;;) { register int c = *code; if (c == OP_END) return NULL; - else if (c == OP_RECURSE) return code; + if (c == OP_RECURSE) return code; + + /* XCLASS is used for classes that cannot be represented just by a bit + map. This includes negated single high-valued characters. The length in + the table is zero; the actual length is stored in the compiled code. */ + + if (c == OP_XCLASS) code += GET(code, 1); + + /* All bracketed groups have the same length. */ + else if (c > OP_BRA) { code += _pcre_OP_lengths[OP_BRA]; } + + /* Otherwise, we get the item's length from the table. In UTF-8 mode, opcodes + that are followed by a character may be followed by a multi-byte character. + The length in the table is a minimum, so we have to scan along to skip the + extra bytes. All opcodes are less than 128, so we can use relatively + efficient code. */ + else { code += _pcre_OP_lengths[c]; - -#ifdef SUPPORT_UTF8 - - /* In UTF-8 mode, opcodes that are followed by a character may be followed - by a multi-byte character. The length in the table is a minimum, so we have - to scan along to skip the extra bytes. All opcodes are less than 128, so we - can use relatively efficient code. */ - if (utf8) switch(c) { case OP_CHAR: @@ -1124,16 +1198,7 @@ for (;;) case OP_MINQUERY: while ((*code & 0xc0) == 0x80) code++; break; - - /* XCLASS is used for classes that cannot be represented just by a bit - map. This includes negated single high-valued characters. The length in - the table is zero; the actual length is stored in the compiled code. */ - - case OP_XCLASS: - code += GET(code, 1) + 1; - break; } -#endif } } } @@ -1490,13 +1555,10 @@ Yield: TRUE when range returned; FALSE when no more static BOOL get_othercase_range(int *cptr, int d, int *ocptr, int *odptr) { -int c, chartype, othercase, next; +int c, othercase, next; for (c = *cptr; c <= d; c++) - { - if (_pcre_ucp_findchar(c, &chartype, &othercase) == ucp_L && othercase != 0) - break; - } + { if ((othercase = _pcre_ucp_othercase(c)) >= 0) break; } if (c > d) return FALSE; @@ -1505,9 +1567,7 @@ next = othercase + 1; for (++c; c <= d; c++) { - if (_pcre_ucp_findchar(c, &chartype, &othercase) != ucp_L || - othercase != next) - break; + if (_pcre_ucp_othercase(c) != next) break; next++; } @@ -1554,7 +1614,6 @@ int greedy_default, greedy_non_default; int firstbyte, reqbyte; int zeroreqbyte, zerofirstbyte; int req_caseopt, reqvary, tempreqvary; -int condcount = 0; int options = *optionsptr; int after_manual_callout = 0; register int c; @@ -1668,10 +1727,14 @@ for (;; ptr++) if ((cd->ctypes[c] & ctype_space) != 0) continue; if (c == '#') { - /* The space before the ; is to avoid a warning on a silly compiler - on the Macintosh. */ - while ((c = *(++ptr)) != 0 && c != NEWLINE) ; - if (c != 0) continue; /* Else fall through to handle end of string */ + while (*(++ptr) != 0) if (IS_NEWLINE(ptr)) break; + if (*ptr != 0) + { + ptr += cd->nllen - 1; + continue; + } + /* Else fall through to handle end of string */ + c = 0; } } @@ -1724,11 +1787,11 @@ for (;; ptr++) *code++ = OP_ANY; break; - /* Character classes. If the included characters are all < 255 in value, we - build a 32-byte bitmap of the permitted characters, except in the special - case where there is only one such character. For negated classes, we build - the map as usual, then invert it at the end. However, we use a different - opcode so that data characters > 255 can be handled correctly. + /* Character classes. If the included characters are all < 256, we build a + 32-byte bitmap of the permitted characters, except in the special case + where there is only one such character. For negated classes, we build the + map as usual, then invert it at the end. However, we use a different opcode + so that data characters > 255 can be handled correctly. If the class contains characters outside the 0-255 range, a different opcode is compiled. It may optionally have a bit map for characters < 256, @@ -1819,8 +1882,9 @@ for (;; ptr++) check_posix_syntax(ptr, &tempptr, cd)) { BOOL local_negate = FALSE; - int posix_class, i; + int posix_class, taboffset, tabopt; register const uschar *cbits = cd->cbits; + uschar pbits[32]; if (ptr[1] != ':') { @@ -1849,32 +1913,46 @@ for (;; ptr++) if ((options & PCRE_CASELESS) != 0 && posix_class <= 2) posix_class = 0; - /* Or into the map we are building up to 3 of the static class - tables, or their negations. The [:blank:] class sets up the same - chars as the [:space:] class (all white space). We remove the vertical - white space chars afterwards. */ + /* We build the bit map for the POSIX class in a chunk of local store + because we may be adding and subtracting from it, and we don't want to + subtract bits that may be in the main map already. At the end we or the + result into the bit map that is being built. */ posix_class *= 3; - for (i = 0; i < 3; i++) + + /* Copy in the first table (always present) */ + + memcpy(pbits, cbits + posix_class_maps[posix_class], + 32 * sizeof(uschar)); + + /* If there is a second table, add or remove it as required. */ + + taboffset = posix_class_maps[posix_class + 1]; + tabopt = posix_class_maps[posix_class + 2]; + + if (taboffset >= 0) { - BOOL blankclass = strncmp((char *)ptr, "blank", 5) == 0; - int taboffset = posix_class_maps[posix_class + i]; - if (taboffset < 0) break; - if (local_negate) - { - if (i == 0) - for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+taboffset]; - else - for (c = 0; c < 32; c++) classbits[c] &= ~cbits[c+taboffset]; - if (blankclass) classbits[1] |= 0x3c; - } + if (tabopt >= 0) + for (c = 0; c < 32; c++) pbits[c] |= cbits[c + taboffset]; else - { - for (c = 0; c < 32; c++) classbits[c] |= cbits[c+taboffset]; - if (blankclass) classbits[1] &= ~0x3c; - } + for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset]; } + /* Not see if we need to remove any special characters. An option + value of 1 removes vertical space and 2 removes underscore. */ + + if (tabopt < 0) tabopt = -tabopt; + if (tabopt == 1) pbits[1] &= ~0x3c; + else if (tabopt == 2) pbits[11] &= 0x7f; + + /* Add the POSIX table or its complement into the main table that is + being built and we are done. */ + + if (local_negate) + for (c = 0; c < 32; c++) classbits[c] |= ~pbits[c]; + else + for (c = 0; c < 32; c++) classbits[c] |= pbits[c]; + ptr = tempptr + 1; class_charcount = 10; /* Set > 1; assumes more than 1 per class */ continue; /* End of POSIX syntax handling */ @@ -1941,12 +2019,14 @@ for (;; ptr++) case ESC_P: { BOOL negated; - int property = get_ucp(&ptr, &negated, errorcodeptr); - if (property < 0) goto FAILED; + int pdata; + int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr); + if (ptype < 0) goto FAILED; class_utf8 = TRUE; *class_utf8data++ = ((-c == ESC_p) != negated)? XCL_PROP : XCL_NOTPROP; - *class_utf8data++ = property; + *class_utf8data++ = ptype; + *class_utf8data++ = pdata; class_charcount -= 2; /* Not a < 256 character */ } continue; @@ -2128,10 +2208,8 @@ for (;; ptr++) #ifdef SUPPORT_UCP if ((options & PCRE_CASELESS) != 0) { - int chartype; int othercase; - if (_pcre_ucp_findchar(c, &chartype, &othercase) >= 0 && - othercase > 0) + if ((othercase = _pcre_ucp_othercase(c)) >= 0) { *class_utf8data++ = XCL_SINGLE; class_utf8data += _pcre_ord2utf8(othercase, class_utf8data); @@ -2416,13 +2494,17 @@ for (;; ptr++) else if (*previous < OP_EODN) { uschar *oldcode; - int prop_type; + int prop_type, prop_value; op_type = OP_TYPESTAR - OP_STAR; /* Use type opcodes */ c = *previous; OUTPUT_SINGLE_REPEAT: - prop_type = (*previous == OP_PROP || *previous == OP_NOTPROP)? - previous[1] : -1; + if (*previous == OP_PROP || *previous == OP_NOTPROP) + { + prop_type = previous[1]; + prop_value = previous[2]; + } + else prop_type = prop_value = -1; oldcode = code; code = previous; /* Usually overwrite previous item */ @@ -2483,7 +2565,7 @@ for (;; ptr++) /* If the maximum is unlimited, insert an OP_STAR. Before doing so, we have to insert the character for the previous code. For a repeated - Unicode property match, there is an extra byte that defines the + Unicode property match, there are two extra bytes that define the required property. In UTF-8 mode, long characters have their length in c, with the 0x80 bit as a flag. */ @@ -2499,7 +2581,11 @@ for (;; ptr++) #endif { *code++ = c; - if (prop_type >= 0) *code++ = prop_type; + if (prop_type >= 0) + { + *code++ = prop_type; + *code++ = prop_value; + } } *code++ = OP_STAR + repeat_type; } @@ -2518,7 +2604,11 @@ for (;; ptr++) else #endif *code++ = c; - if (prop_type >= 0) *code++ = prop_type; + if (prop_type >= 0) + { + *code++ = prop_type; + *code++ = prop_value; + } repeat_max -= repeat_min; *code++ = OP_UPTO + repeat_type; PUT2INC(code, 0, repeat_max); @@ -2537,11 +2627,15 @@ for (;; ptr++) #endif *code++ = c; - /* For a repeated Unicode property match, there is an extra byte that - defines the required property. */ + /* For a repeated Unicode property match, there are two extra bytes that + define the required property. */ #ifdef SUPPORT_UCP - if (prop_type >= 0) *code++ = prop_type; + if (prop_type >= 0) + { + *code++ = prop_type; + *code++ = prop_value; + } #endif } @@ -2805,37 +2899,91 @@ for (;; ptr++) case '(': bravalue = OP_COND; /* Conditional group */ - /* Condition to test for recursion */ - - if (ptr[1] == 'R') + /* A condition can be a number, referring to a numbered group, a name, + referring to a named group, 'R', referring to recursion, or an + assertion. There are two unfortunate ambiguities, caused by history. + (a) 'R' can be the recursive thing or the name 'R', and (b) a number + could be a name that consists of digits. In both cases, we look for a + name first; if not found, we try the other cases. If the first + character after (?( is a word character, we know the rest up to ) will + also be word characters because the syntax was checked in the first + pass. */ + + if ((cd->ctypes[ptr[1]] & ctype_word) != 0) { - code[1+LINK_SIZE] = OP_CREF; - PUT2(code, 2+LINK_SIZE, CREF_RECURSE); + int i, namelen; + int condref = 0; + const uschar *name; + uschar *slot = cd->name_table; + + /* This is needed for all successful cases. */ + skipbytes = 3; - ptr += 3; - } - /* Condition to test for a numbered subpattern match. We know that - if a digit follows ( then there will just be digits until ) because - the syntax was checked in the first pass. */ + /* Read the name, but also get it as a number if it's all digits */ - else if ((digitab[ptr[1]] && ctype_digit) != 0) - { - int condref; /* Don't amalgamate; some compilers */ - condref = *(++ptr) - '0'; /* grumble at autoincrement in declaration */ - while (*(++ptr) != ')') condref = condref*10 + *ptr - '0'; - if (condref == 0) + name = ++ptr; + while (*ptr != ')') { - *errorcodeptr = ERR35; - goto FAILED; + if (condref >= 0) + condref = ((digitab[*ptr] & ctype_digit) != 0)? + condref * 10 + *ptr - '0' : -1; + ptr++; } + namelen = ptr - name; ptr++; - code[1+LINK_SIZE] = OP_CREF; - PUT2(code, 2+LINK_SIZE, condref); - skipbytes = 3; + + for (i = 0; i < cd->names_found; i++) + { + if (strncmp((char *)name, (char *)slot+2, namelen) == 0) break; + slot += cd->name_entry_size; + } + + /* Found a previous named subpattern */ + + if (i < cd->names_found) + { + condref = GET2(slot, 0); + code[1+LINK_SIZE] = OP_CREF; + PUT2(code, 2+LINK_SIZE, condref); + } + + /* Search the pattern for a forward reference */ + + else if ((i = find_named_parens(ptr, *brackets, name, namelen)) > 0) + { + code[1+LINK_SIZE] = OP_CREF; + PUT2(code, 2+LINK_SIZE, i); + } + + /* Check for 'R' for recursion */ + + else if (namelen == 1 && *name == 'R') + { + code[1+LINK_SIZE] = OP_CREF; + PUT2(code, 2+LINK_SIZE, CREF_RECURSE); + } + + /* Check for a subpattern number */ + + else if (condref > 0) + { + code[1+LINK_SIZE] = OP_CREF; + PUT2(code, 2+LINK_SIZE, condref); + } + + /* Either an unidentified subpattern, or a reference to (?(0) */ + + else + { + *errorcodeptr = (condref == 0)? ERR35: ERR15; + goto FAILED; + } } + /* For conditions that are assertions, we just fall through, having set bravalue above. */ + break; case '=': /* Positive lookahead */ @@ -2907,10 +3055,13 @@ for (;; ptr++) { if (slot[2+namelen] == 0) { - *errorcodeptr = ERR43; - goto FAILED; + if ((options & PCRE_DUPNAMES) == 0) + { + *errorcodeptr = ERR43; + goto FAILED; + } } - crc = -1; /* Current name is substring */ + else crc = -1; /* Current name is substring */ } if (crc < 0) { @@ -2943,14 +3094,18 @@ for (;; ptr++) if (strncmp((char *)name, (char *)slot+2, namelen) == 0) break; slot += cd->name_entry_size; } - if (i >= cd->names_found) + + if (i < cd->names_found) /* Back reference */ + { + recno = GET2(slot, 0); + } + else if ((recno = /* Forward back reference */ + find_named_parens(ptr, *brackets, name, namelen)) <= 0) { *errorcodeptr = ERR15; goto FAILED; } - recno = GET2(slot, 0); - if (type == '>') goto HANDLE_RECURSION; /* A few lines below */ /* Back reference */ @@ -2990,9 +3145,8 @@ for (;; ptr++) regex in case it doesn't exist. */ *code = OP_END; - called = (recno == 0)? - cd->start_code : find_bracket(cd->start_code, utf8, recno); - + called = (recno == 0)? cd->start_code : + find_bracket(cd->start_code, utf8, recno); if (called == NULL) { *errorcodeptr = ERR15; @@ -3009,11 +3163,20 @@ for (;; ptr++) goto FAILED; } - /* Insert the recursion/subroutine item */ + /* Insert the recursion/subroutine item, automatically wrapped inside + "once" brackets. */ + + *code = OP_ONCE; + PUT(code, 1, 2 + 2*LINK_SIZE); + code += 1 + LINK_SIZE; *code = OP_RECURSE; PUT(code, 1, called - cd->start_code); code += 1 + LINK_SIZE; + + *code = OP_KET; + PUT(code, 1, 2 + 2*LINK_SIZE); + code += 1 + LINK_SIZE; } continue; @@ -3030,6 +3193,7 @@ for (;; ptr++) case '-': optset = &unset; break; case 'i': *optset |= PCRE_CASELESS; break; + case 'J': *optset |= PCRE_DUPNAMES; break; case 'm': *optset |= PCRE_MULTILINE; break; case 's': *optset |= PCRE_DOTALL; break; case 'x': *optset |= PCRE_EXTENDED; break; @@ -3146,7 +3310,7 @@ for (;; ptr++) else if (bravalue == OP_COND) { uschar *tc = code; - condcount = 0; + int condcount = 0; do { condcount++; @@ -3283,10 +3447,12 @@ for (;; ptr++) else if (-c == ESC_P || -c == ESC_p) { BOOL negated; - int value = get_ucp(&ptr, &negated, errorcodeptr); + int pdata; + int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr); previous = code; *code++ = ((-c == ESC_p) != negated)? OP_PROP : OP_NOTPROP; - *code++ = value; + *code++ = ptype; + *code++ = pdata; } #endif @@ -3841,7 +4007,7 @@ Returns: pointer to compiled data block, or NULL on error, with errorptr and erroroffset set */ -EXPORT pcre * +PCRE_DATA_SCOPE pcre * pcre_compile(const char *pattern, int options, const char **errorptr, int *erroroffset, const unsigned char *tables) { @@ -3849,13 +4015,14 @@ return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables); } -EXPORT pcre * + +PCRE_DATA_SCOPE pcre * pcre_compile2(const char *pattern, int options, int *errorcodeptr, const char **errorptr, int *erroroffset, const unsigned char *tables) { real_pcre *re; int length = 1 + LINK_SIZE; /* For initial BRA plus length */ -int c, firstbyte, reqbyte; +int c, firstbyte, reqbyte, newline; int bracount = 0; int branch_extra = 0; int branch_newextra; @@ -3876,6 +4043,7 @@ uschar *code; const uschar *codestart; const uschar *ptr; compile_data compile_block; +compile_data *cd = &compile_block; int brastack[BRASTACK_SIZE]; uschar bralenstack[BRASTACK_SIZE]; @@ -3929,18 +4097,42 @@ if ((options & ~PUBLIC_OPTIONS) != 0) /* Set up pointers to the individual character tables */ if (tables == NULL) tables = _pcre_default_tables; -compile_block.lcc = tables + lcc_offset; -compile_block.fcc = tables + fcc_offset; -compile_block.cbits = tables + cbits_offset; -compile_block.ctypes = tables + ctypes_offset; +cd->lcc = tables + lcc_offset; +cd->fcc = tables + fcc_offset; +cd->cbits = tables + cbits_offset; +cd->ctypes = tables + ctypes_offset; + +/* Handle different types of newline. The two bits give four cases. The current +code allows for one- or two-byte sequences. */ + +switch (options & PCRE_NEWLINE_CRLF) + { + default: newline = NEWLINE; break; /* Compile-time default */ + case PCRE_NEWLINE_CR: newline = '\r'; break; + case PCRE_NEWLINE_LF: newline = '\n'; break; + case PCRE_NEWLINE_CR+ + PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break; + } + +if (newline > 255) + { + cd->nllen = 2; + cd->nl[0] = (newline >> 8) & 255; + cd->nl[1] = newline & 255; + } +else + { + cd->nllen = 1; + cd->nl[0] = newline; + } /* Maximum back reference and backref bitmap. This is updated for numeric references during the first pass, but for named references during the actual compile pass. The bitmap records up to 31 back references to help in deciding whether (.*) can be treated as anchored or not. */ -compile_block.top_backref = 0; -compile_block.backref_map = 0; +cd->top_backref = 0; +cd->backref_map = 0; /* Reflect pattern for debugging output */ @@ -3974,14 +4166,16 @@ while ((c = *(++ptr)) != 0) if ((options & PCRE_EXTENDED) != 0) { - if ((compile_block.ctypes[c] & ctype_space) != 0) continue; + if ((cd->ctypes[c] & ctype_space) != 0) continue; if (c == '#') { - /* The space before the ; is to avoid a warning on a silly compiler - on the Macintosh. */ - while ((c = *(++ptr)) != 0 && c != NEWLINE) ; - if (c == 0) break; - continue; + while (*(++ptr) != 0) if (IS_NEWLINE(ptr)) break; + if (*ptr != 0) + { + ptr += cd->nllen - 1; + continue; + } + break; /* End loop at end of pattern */ } } @@ -4042,15 +4236,17 @@ while ((c = *(++ptr)) != 0) #endif /* \P and \p are for Unicode properties, but only when the support has - been compiled. Each item needs 2 bytes. */ + been compiled. Each item needs 3 bytes. */ else if (-c == ESC_P || -c == ESC_p) { #ifdef SUPPORT_UCP BOOL negated; - length += 2; - lastitemlength = 2; - if (get_ucp(&ptr, &negated, &errorcode) < 0) goto PCRE_ERROR_RETURN; + BOOL pdata; + length += 3; + lastitemlength = 3; + if (get_ucp(&ptr, &negated, &pdata, &errorcode) < 0) + goto PCRE_ERROR_RETURN; continue; #else errorcode = ERR45; @@ -4069,9 +4265,9 @@ while ((c = *(++ptr)) != 0) if (c <= -ESC_REF) { int refnum = -c - ESC_REF; - compile_block.backref_map |= (refnum < 32)? (1 << refnum) : 1; - if (refnum > compile_block.top_backref) - compile_block.top_backref = refnum; + cd->backref_map |= (refnum < 32)? (1 << refnum) : 1; + if (refnum > cd->top_backref) + cd->top_backref = refnum; length += 2; /* For single back reference */ if (ptr[1] == '{' && is_counted_repeat(ptr+2)) { @@ -4216,7 +4412,7 @@ while ((c = *(++ptr)) != 0) class_utf8 = TRUE; length += LINK_SIZE + 2; } - length += 2; + length += 3; } #endif } @@ -4225,7 +4421,9 @@ while ((c = *(++ptr)) != 0) /* Check the syntax for POSIX stuff. The bits we actually handle are checked during the real compile phase. */ - else if (*ptr == '[' && check_posix_syntax(ptr, &ptr, &compile_block)) + else if (*ptr == '[' && + (ptr[1] == ':' || ptr[1] == '.' || ptr[1] == '=') && + check_posix_syntax(ptr, &ptr, cd)) { ptr++; class_optcount = 10; /* Make sure > 1 */ @@ -4458,6 +4656,61 @@ while ((c = *(++ptr)) != 0) ptr += 2; break; + /* Named subpatterns are an extension copied from Python */ + + case 'P': + ptr += 3; + + /* Handle the definition of a named subpattern */ + + if (*ptr == '<') + { + const uschar *p; /* Don't amalgamate; some compilers */ + p = ++ptr; /* grumble at autoincrement in declaration */ + while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++; + if (*ptr != '>') + { + errorcode = ERR42; + goto PCRE_ERROR_RETURN; + } + name_count++; + if (name_count > MAX_NAME_COUNT) + { + errorcode = ERR49; + goto PCRE_ERROR_RETURN; + } + if (ptr - p > max_name_size) + { + max_name_size = (ptr - p); + if (max_name_size > MAX_NAME_SIZE) + { + errorcode = ERR48; + goto PCRE_ERROR_RETURN; + } + } + capturing = TRUE; /* Named parentheses are always capturing */ + break; /* Go handle capturing parentheses */ + } + + /* Handle back references and recursive calls to named subpatterns */ + + if (*ptr == '=' || *ptr == '>') + { + length += 3 + 3*LINK_SIZE; /* Allow for the automatic "once" */ + while ((cd->ctypes[*(++ptr)] & ctype_word) != 0); + if (*ptr != ')') + { + errorcode = ERR42; + goto PCRE_ERROR_RETURN; + } + goto RECURSE_CHECK_QUANTIFIED; + } + + /* Unknown character after (?P */ + + errorcode = ERR41; + goto PCRE_ERROR_RETURN; + /* (?R) specifies a recursive call to the regex, which is an extension to provide the facility which can be obtained by (?p{perl-code}) in Perl 5.6. In Perl 5.8 this has become (??{perl-code}). @@ -4479,12 +4732,14 @@ while ((c = *(++ptr)) != 0) errorcode = ERR29; goto PCRE_ERROR_RETURN; } - length += 1 + LINK_SIZE; + length += 3 + 3*LINK_SIZE; /* Allows for the automatic "once" */ /* If this item is quantified, it will get wrapped inside brackets so as to use the code for quantified brackets. We jump down and use the - code that handles this for real brackets. */ + code that handles this for real brackets. Come here from code for + named recursions/subroutines. */ + RECURSE_CHECK_QUANTIFIED: if (ptr[1] == '+' || ptr[1] == '*' || ptr[1] == '?' || ptr[1] == '{') { length += 2 + 2 * LINK_SIZE; /* to make bracketed */ @@ -4508,47 +4763,6 @@ while ((c = *(++ptr)) != 0) length += 2 + 2*LINK_SIZE; continue; - /* Named subpatterns are an extension copied from Python */ - - case 'P': - ptr += 3; - - /* Handle the definition of a named subpattern */ - - if (*ptr == '<') - { - const uschar *p; /* Don't amalgamate; some compilers */ - p = ++ptr; /* grumble at autoincrement in declaration */ - while ((compile_block.ctypes[*ptr] & ctype_word) != 0) ptr++; - if (*ptr != '>') - { - errorcode = ERR42; - goto PCRE_ERROR_RETURN; - } - name_count++; - if (ptr - p > max_name_size) max_name_size = (ptr - p); - capturing = TRUE; /* Named parentheses are always capturing */ - break; - } - - /* Handle back references and recursive calls to named subpatterns */ - - if (*ptr == '=' || *ptr == '>') - { - while ((compile_block.ctypes[*(++ptr)] & ctype_word) != 0); - if (*ptr != ')') - { - errorcode = ERR42; - goto PCRE_ERROR_RETURN; - } - break; - } - - /* Unknown character after (?P */ - - errorcode = ERR41; - goto PCRE_ERROR_RETURN; - /* Lookbehinds are in Perl from version 5.005 */ case '<': @@ -4564,19 +4778,17 @@ while ((c = *(++ptr)) != 0) /* Conditionals are in Perl from version 5.005. The bracket must either be followed by a number (for bracket reference) or by an assertion - group, or (a PCRE extension) by 'R' for a recursion test. */ + group. PCRE extends this by allowing a name to reference a named group; + unfortunately, previously 'R' was implemented for a recursion test. + When this is compiled, we look for the named group 'R' first. At this + point we just do a basic syntax check. */ case '(': - if (ptr[3] == 'R' && ptr[4] == ')') - { - ptr += 4; - length += 3; - } - else if ((digitab[ptr[3]] & ctype_digit) != 0) + if ((cd->ctypes[ptr[3]] & ctype_word) != 0) { ptr += 4; length += 3; - while ((digitab[*ptr] & ctype_digit) != 0) ptr++; + while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++; if (*ptr != ')') { errorcode = ERR26; @@ -4615,6 +4827,11 @@ while ((c = *(++ptr)) != 0) *optset |= PCRE_CASELESS; continue; + case 'J': + *optset |= PCRE_DUPNAMES; + options |= PCRE_JCHANGED; /* Record that it changed */ + continue; + case 'm': *optset |= PCRE_MULTILINE; continue; @@ -4680,16 +4897,13 @@ while ((c = *(++ptr)) != 0) will lead to an over-estimate on the length, but this shouldn't matter very much. We also have to allow for resetting options at the start of any alternations, which we do by setting - branch_newextra to 2. Finally, we record whether the case-dependent - flag ever changes within the regex. This is used by the "required - character" code. */ + branch_newextra to 2. */ case ':': if (((set|unset) & PCRE_IMS) != 0) { length += 4; branch_newextra = 2; - if (((set|unset) & PCRE_CASELESS) != 0) options |= PCRE_ICHANGED; } goto END_OPTIONS; @@ -4769,6 +4983,12 @@ while ((c = *(++ptr)) != 0) { duplength = length - brastack[--brastackptr]; branch_extra = bralenstack[brastackptr]; + /* This is a paranoid check to stop integer overflow later on */ + if (duplength > MAX_DUPLENGTH) + { + errorcode = ERR50; + goto PCRE_ERROR_RETURN; + } } else duplength = 0; @@ -4873,7 +5093,8 @@ if (length > MAX_PATTERN_SIZE) } /* Compute the size of data block needed and get it, either from malloc or -externally provided function. */ +externally provided function. Integer overflow should no longer be possible +because nowadays we limit the maximum value of name_count and max_name size. */ size = length + sizeof(real_pcre) + name_count * (max_name_size + 3); re = (real_pcre *)(pcre_malloc)(size); @@ -4903,14 +5124,14 @@ re->nullpad = NULL; /* The starting points of the name/number translation table and of the code are passed around in the compile data block. */ -compile_block.names_found = 0; -compile_block.name_entry_size = max_name_size + 3; -compile_block.name_table = (uschar *)re + re->name_table_offset; -codestart = compile_block.name_table + re->name_entry_size * re->name_count; -compile_block.start_code = codestart; -compile_block.start_pattern = (const uschar *)pattern; -compile_block.req_varyopt = 0; -compile_block.nopartial = FALSE; +cd->names_found = 0; +cd->name_entry_size = max_name_size + 3; +cd->name_table = (uschar *)re + re->name_table_offset; +codestart = cd->name_table + re->name_entry_size * re->name_count; +cd->start_code = codestart; +cd->start_pattern = (const uschar *)pattern; +cd->req_varyopt = 0; +cd->nopartial = FALSE; /* Set up a starting, non-extracting bracket, then compile the expression. On error, errorcode will be set non-zero, so we don't need to look at the result @@ -4921,11 +5142,11 @@ code = (uschar *)codestart; *code = OP_BRA; bracount = 0; (void)compile_regex(options, options & PCRE_IMS, &bracount, &code, &ptr, - &errorcode, FALSE, 0, &firstbyte, &reqbyte, NULL, &compile_block); + &errorcode, FALSE, 0, &firstbyte, &reqbyte, NULL, cd); re->top_bracket = bracount; -re->top_backref = compile_block.top_backref; +re->top_backref = cd->top_backref; -if (compile_block.nopartial) re->options |= PCRE_NOPARTIAL; +if (cd->nopartial) re->options |= PCRE_NOPARTIAL; /* If not reached end of pattern on success, there's an excess bracket. */ @@ -4971,7 +5192,7 @@ start with ^. and also when all branches start with .* for non-DOTALL matches. if ((options & PCRE_ANCHORED) == 0) { int temp_options = options; - if (is_anchored(codestart, &temp_options, 0, compile_block.backref_map)) + if (is_anchored(codestart, &temp_options, 0, cd->backref_map)) re->options |= PCRE_ANCHORED; else { @@ -4981,10 +5202,10 @@ if ((options & PCRE_ANCHORED) == 0) { int ch = firstbyte & 255; re->first_byte = ((firstbyte & REQ_CASELESS) != 0 && - compile_block.fcc[ch] == ch)? ch : firstbyte; + cd->fcc[ch] == ch)? ch : firstbyte; re->options |= PCRE_FIRSTSET; } - else if (is_startline(codestart, 0, compile_block.backref_map)) + else if (is_startline(codestart, 0, cd->backref_map)) re->options |= PCRE_STARTLINE; } } @@ -4998,11 +5219,12 @@ if (reqbyte >= 0 && { int ch = reqbyte & 255; re->req_byte = ((reqbyte & REQ_CASELESS) != 0 && - compile_block.fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte; + cd->fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte; re->options |= PCRE_REQCHSET; } -/* Print out the compiled data for debugging */ +/* Print out the compiled data if debugging is enabled. This is never the +case when building a production library. */ #ifdef DEBUG @@ -5011,11 +5233,10 @@ printf("Length = %d top_bracket = %d top_backref = %d\n", if (re->options != 0) { - printf("%s%s%s%s%s%s%s%s%s%s\n", + printf("%s%s%s%s%s%s%s%s%s\n", ((re->options & PCRE_NOPARTIAL) != 0)? "nopartial " : "", ((re->options & PCRE_ANCHORED) != 0)? "anchored " : "", ((re->options & PCRE_CASELESS) != 0)? "caseless " : "", - ((re->options & PCRE_ICHANGED) != 0)? "case state changed " : "", ((re->options & PCRE_EXTENDED) != 0)? "extended " : "", ((re->options & PCRE_MULTILINE) != 0)? "multiline " : "", ((re->options & PCRE_DOTALL) != 0)? "dotall " : "", @@ -5027,7 +5248,8 @@ if (re->options != 0) if ((re->options & PCRE_FIRSTSET) != 0) { int ch = re->first_byte & 255; - const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)? "" : " (caseless)"; + const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)? + "" : " (caseless)"; if (isprint(ch)) printf("First char = %c%s\n", ch, caseless); else printf("First char = \\x%02x%s\n", ch, caseless); } @@ -5035,12 +5257,13 @@ if ((re->options & PCRE_FIRSTSET) != 0) if ((re->options & PCRE_REQCHSET) != 0) { int ch = re->req_byte & 255; - const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)? "" : " (caseless)"; + const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)? + "" : " (caseless)"; if (isprint(ch)) printf("Req char = %c%s\n", ch, caseless); else printf("Req char = \\x%02x%s\n", ch, caseless); } -_pcre_printint(re, stdout); +pcre_printint(re, stdout); /* This check is done here in the debugging case so that the code that was compiled can be seen. */ diff --git a/src/src/pcre/pcre_config.c b/src/src/pcre/pcre_config.c index ce02c878b..e20450eea 100644 --- a/src/src/pcre/pcre_config.c +++ b/src/src/pcre/pcre_config.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_config.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_config.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -60,7 +60,7 @@ Arguments: Returns: 0 if data returned, negative on error */ -EXPORT int +PCRE_DATA_SCOPE int pcre_config(int what, void *where) { switch (what) @@ -97,6 +97,10 @@ switch (what) *((unsigned int *)where) = MATCH_LIMIT; break; + case PCRE_CONFIG_MATCH_LIMIT_RECURSION: + *((unsigned int *)where) = MATCH_LIMIT_RECURSION; + break; + case PCRE_CONFIG_STACKRECURSE: #ifdef NO_RECURSE *((int *)where) = 0; diff --git a/src/src/pcre/pcre_exec.c b/src/src/pcre/pcre_exec.c index 85feb45e8..4484eea20 100644 --- a/src/src/pcre/pcre_exec.c +++ b/src/src/pcre/pcre_exec.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_exec.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_exec.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -44,7 +44,7 @@ POSSIBILITY OF SUCH DAMAGE. pattern matching using an NFA algorithm, trying to mimic Perl as closely as possible. There are also some static supporting functions. */ - +#define NLBLOCK md /* The block containing newline information */ #include "pcre_internal.h" @@ -56,7 +56,7 @@ are on the heap, not on the stack. */ typedef struct eptrblock { struct eptrblock *epb_prev; - const uschar *epb_saved_eptr; + USPTR epb_saved_eptr; } eptrblock; /* Flag bits for the match() function */ @@ -130,10 +130,10 @@ Returns: TRUE if matched */ static BOOL -match_ref(int offset, register const uschar *eptr, int length, match_data *md, +match_ref(int offset, register USPTR eptr, int length, match_data *md, unsigned long int ims) { -const uschar *p = md->start_subject + md->offset_vector[offset]; +USPTR p = md->start_subject + md->offset_vector[offset]; #ifdef DEBUG if (eptr >= md->end_subject) @@ -171,32 +171,50 @@ return TRUE; **************************************************************************** RECURSION IN THE match() FUNCTION -The match() function is highly recursive. Some regular expressions can cause -it to recurse thousands of times. I was writing for Unix, so I just let it -call itself recursively. This uses the stack for saving everything that has -to be saved for a recursive call. On Unix, the stack can be large, and this -works fine. +The match() function is highly recursive, though not every recursive call +increases the recursive depth. Nevertheless, some regular expressions can cause +it to recurse to a great depth. I was writing for Unix, so I just let it call +itself recursively. This uses the stack for saving everything that has to be +saved for a recursive call. On Unix, the stack can be large, and this works +fine. -It turns out that on non-Unix systems there are problems with programs that -use a lot of stack. (This despite the fact that every last chip has oodles -of memory these days, and techniques for extending the stack have been known -for decades.) So.... +It turns out that on some non-Unix-like systems there are problems with +programs that use a lot of stack. (This despite the fact that every last chip +has oodles of memory these days, and techniques for extending the stack have +been known for decades.) So.... There is a fudge, triggered by defining NO_RECURSE, which avoids recursive calls by keeping local variables that need to be preserved in blocks of memory -obtained from malloc instead instead of on the stack. Macros are used to +obtained from malloc() instead instead of on the stack. Macros are used to achieve this so that the actual code doesn't look very different to what it always used to. **************************************************************************** ***************************************************************************/ -/* These versions of the macros use the stack, as normal */ +/* These versions of the macros use the stack, as normal. There are debugging +versions and production versions. */ #ifndef NO_RECURSE #define REGISTER register -#define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) rx = match(ra,rb,rc,rd,re,rf,rg) +#ifdef DEBUG +#define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) \ + { \ + printf("match() called in line %d\n", __LINE__); \ + rx = match(ra,rb,rc,rd,re,rf,rg,rdepth+1); \ + printf("to line %d\n", __LINE__); \ + } +#define RRETURN(ra) \ + { \ + printf("match() returned %d from line %d ", ra, __LINE__); \ + return ra; \ + } +#else +#define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) \ + rx = match(ra,rb,rc,rd,re,rf,rg,rdepth+1) #define RRETURN(ra) return ra +#endif + #else @@ -217,6 +235,7 @@ match(), which never changes. */ newframe->Xims = re;\ newframe->Xeptrb = rf;\ newframe->Xflags = rg;\ + newframe->Xrdepth = frame->Xrdepth + 1;\ newframe->Xprevframe = frame;\ frame = newframe;\ DPRINTF(("restarting from line %d\n", __LINE__));\ @@ -258,6 +277,7 @@ typedef struct heapframe { long int Xims; eptrblock *Xeptrb; int Xflags; + unsigned int Xrdepth; /* Function local variables */ @@ -280,11 +300,11 @@ typedef struct heapframe { #ifdef SUPPORT_UCP int Xprop_type; + int Xprop_value; int Xprop_fail_result; int Xprop_category; int Xprop_chartype; - int Xprop_othercase; - int Xprop_test_against; + int Xprop_script; int *Xprop_test_variable; #endif @@ -345,26 +365,27 @@ Arguments: flags can contain match_condassert - this is an assertion condition match_isgroup - this is the start of a bracketed group + rdepth the recursion depth Returns: MATCH_MATCH if matched ) these values are >= 0 MATCH_NOMATCH if failed to match ) a negative PCRE_ERROR_xxx value if aborted by an error condition - (e.g. stopped by recursion limit) + (e.g. stopped by repeated call or recursion limit) */ static int -match(REGISTER const uschar *eptr, REGISTER const uschar *ecode, +match(REGISTER USPTR eptr, REGISTER const uschar *ecode, int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb, - int flags) + int flags, unsigned int rdepth) { /* These variables do not need to be preserved over recursion in this function, so they can be ordinary variables in all cases. Mark them with "register" because they are used a lot in loops. */ -register int rrc; /* Returns from recursive calls */ -register int i; /* Used for loops not involving calls to RMATCH() */ -register int c; /* Character values not kept over RMATCH() calls */ -register BOOL utf8; /* Local copy of UTF-8 flag for speed */ +register int rrc; /* Returns from recursive calls */ +register int i; /* Used for loops not involving calls to RMATCH() */ +register unsigned int c; /* Character values not kept over RMATCH() calls */ +register BOOL utf8; /* Local copy of UTF-8 flag for speed */ /* When recursion is not being used, all "local" variables that have to be preserved over calls to RMATCH() are part of a "frame" which is obtained from @@ -383,6 +404,7 @@ frame->Xoffset_top = offset_top; frame->Xims = ims; frame->Xeptrb = eptrb; frame->Xflags = flags; +frame->Xrdepth = rdepth; /* This is where control jumps back to to effect "recursion" */ @@ -396,6 +418,7 @@ HEAP_RECURSE: #define ims frame->Xims #define eptrb frame->Xeptrb #define flags frame->Xflags +#define rdepth frame->Xrdepth /* Ditto for the local variables */ @@ -420,11 +443,11 @@ HEAP_RECURSE: #ifdef SUPPORT_UCP #define prop_type frame->Xprop_type +#define prop_value frame->Xprop_value #define prop_fail_result frame->Xprop_fail_result #define prop_category frame->Xprop_category #define prop_chartype frame->Xprop_chartype -#define prop_othercase frame->Xprop_othercase -#define prop_test_against frame->Xprop_test_against +#define prop_script frame->Xprop_script #define prop_test_variable frame->Xprop_test_variable #endif @@ -454,20 +477,20 @@ i, and fc and c, can be the same variables. */ #define fc c -#ifdef SUPPORT_UTF8 /* Many of these variables are used ony */ -const uschar *charptr; /* small blocks of the code. My normal */ -#endif /* style of coding would have declared */ -const uschar *callpat; /* them within each of those blocks. */ -const uschar *data; /* However, in order to accommodate the */ -const uschar *next; /* version of this code that uses an */ -const uschar *pp; /* external "stack" implemented on the */ -const uschar *prev; /* heap, it is easier to declare them */ -const uschar *saved_eptr; /* all here, so the declarations can */ - /* be cut out in a block. The only */ -recursion_info new_recursive; /* declarations within blocks below are */ - /* for variables that do not have to */ -BOOL cur_is_word; /* be preserved over a recursive call */ -BOOL condition; /* to RMATCH(). */ +#ifdef SUPPORT_UTF8 /* Many of these variables are used only */ +const uschar *charptr; /* in small blocks of the code. My normal */ +#endif /* style of coding would have declared */ +const uschar *callpat; /* them within each of those blocks. */ +const uschar *data; /* However, in order to accommodate the */ +const uschar *next; /* version of this code that uses an */ +USPTR pp; /* external "stack" implemented on the */ +const uschar *prev; /* heap, it is easier to declare them all */ +USPTR saved_eptr; /* here, so the declarations can be cut */ + /* out in a block. The only declarations */ +recursion_info new_recursive; /* within blocks below are for variables */ + /* that do not have to be preserved over */ +BOOL cur_is_word; /* a recursive call to RMATCH(). */ +BOOL condition; BOOL minimize; BOOL prev_is_word; @@ -475,11 +498,11 @@ unsigned long int original_ims; #ifdef SUPPORT_UCP int prop_type; +int prop_value; int prop_fail_result; int prop_category; int prop_chartype; -int prop_othercase; -int prop_test_against; +int prop_script; int *prop_test_variable; #endif @@ -501,22 +524,39 @@ eptrblock newptrb; variables. */ #ifdef SUPPORT_UCP +prop_value = 0; prop_fail_result = 0; -prop_test_against = 0; prop_test_variable = NULL; #endif -/* OK, now we can get on with the real code of the function. Recursion is -specified by the macros RMATCH and RRETURN. When NO_RECURSE is *not* defined, -these just turn into a recursive call to match() and a "return", respectively. -However, RMATCH isn't like a function call because it's quite a complicated -macro. It has to be used in one particular way. This shouldn't, however, impact -performance when true recursion is being used. */ +/* This label is used for tail recursion, which is used in a few cases even +when NO_RECURSE is not defined, in order to reduce the amount of stack that is +used. Thanks to Ian Taylor for noticing this possibility and sending the +original patch. */ + +TAIL_RECURSE: + +/* OK, now we can get on with the real code of the function. Recursive calls +are specified by the macro RMATCH and RRETURN is used to return. When +NO_RECURSE is *not* defined, these just turn into a recursive call to match() +and a "return", respectively (possibly with some debugging if DEBUG is +defined). However, RMATCH isn't like a function call because it's quite a +complicated macro. It has to be used in one particular way. This shouldn't, +however, impact performance when true recursion is being used. */ + +/* First check that we haven't called match() too many times, or that we +haven't exceeded the recursive call limit. */ if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT); +if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT); original_ims = ims; /* Save for resetting on ')' */ + +#ifdef SUPPORT_UTF8 utf8 = md->utf8; /* Local copy of the flag */ +#else +utf8 = FALSE; +#endif /* At the start of a bracketed group, add the current subject pointer to the stack of such pointers, to be re-instated at the end of the group when we hit @@ -616,21 +656,38 @@ for (;;) { case OP_BRA: /* Non-capturing bracket: optimized */ DPRINTF(("start bracket 0\n")); - do + + /* Loop for all the alternatives */ + + for (;;) { + /* When we get to the final alternative within the brackets, we would + return the result of a recursive call to match() whatever happened. We + can reduce stack usage by turning this into a tail recursion. */ + + if (ecode[GET(ecode, 1)] != OP_ALT) + { + ecode += 1 + LINK_SIZE; + flags = match_isgroup; + DPRINTF(("bracket 0 tail recursion\n")); + goto TAIL_RECURSE; + } + + /* For non-final alternatives, continue the loop for a NOMATCH result; + otherwise return. */ + RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, match_isgroup); if (rrc != MATCH_NOMATCH) RRETURN(rrc); ecode += GET(ecode, 1); } - while (*ecode == OP_ALT); - DPRINTF(("bracket 0 failed\n")); - RRETURN(MATCH_NOMATCH); + /* Control never reaches here. */ /* Conditional group: compilation checked that there are no more than two branches. If the condition is false, skipping the first branch takes us past the end if there is only one branch, but that's OK because that is - exactly what going to the ket would do. */ + exactly what going to the ket would do. As there is only one branch to be + obeyed, we can use tail recursion to avoid using another stack frame. */ case OP_COND: if (ecode[LINK_SIZE+1] == OP_CREF) /* Condition extract or recurse test */ @@ -639,10 +696,9 @@ for (;;) condition = (offset == CREF_RECURSE * 2)? (md->recursive != NULL) : (offset < offset_top && md->offset_vector[offset] >= 0); - RMATCH(rrc, eptr, ecode + (condition? - (LINK_SIZE + 4) : (LINK_SIZE + 1 + GET(ecode, 1))), - offset_top, md, ims, eptrb, match_isgroup); - RRETURN(rrc); + ecode += condition? (LINK_SIZE + 4) : (LINK_SIZE + 1 + GET(ecode, 1)); + flags = match_isgroup; + goto TAIL_RECURSE; } /* The condition is an assertion. Call match() to evaluate it - setting @@ -662,9 +718,13 @@ for (;;) RRETURN(rrc); /* Need braces because of following else */ } else ecode += GET(ecode, 1); - RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, - match_isgroup); - RRETURN(rrc); + + /* We are now at the branch that is to be obeyed. As there is only one, + we can use tail recursion to avoid using another stack frame. */ + + ecode += 1 + LINK_SIZE; + flags = match_isgroup; + goto TAIL_RECURSE; } /* Control never reaches here */ @@ -683,7 +743,7 @@ for (;;) if (md->recursive != NULL && md->recursive->group_num == 0) { recursion_info *rec = md->recursive; - DPRINTF(("Hit the end in a (?0) recursion\n")); + DPRINTF(("End of pattern in a (?0) recursion\n")); md->recursive = rec->prevrec; memmove(md->offset_vector, rec->offset_save, rec->saved_max * sizeof(int)); @@ -802,7 +862,7 @@ for (;;) cb.version = 1; /* Version 1 of the callout block */ cb.callout_number = ecode[1]; cb.offset_vector = md->offset_vector; - cb.subject = (const char *)md->start_subject; + cb.subject = (PCRE_SPTR)md->start_subject; cb.subject_length = md->end_subject - md->start_subject; cb.start_match = md->start_match - md->start_subject; cb.current_position = eptr - md->start_subject; @@ -884,12 +944,17 @@ for (;;) eptrb, match_isgroup); if (rrc == MATCH_MATCH) { + DPRINTF(("Recursion matched\n")); md->recursive = new_recursive.prevrec; if (new_recursive.offset_save != stacksave) (pcre_free)(new_recursive.offset_save); RRETURN(MATCH_MATCH); } - else if (rrc != MATCH_NOMATCH) RRETURN(rrc); + else if (rrc != MATCH_NOMATCH) + { + DPRINTF(("Recursion gave error %d\n", rrc)); + RRETURN(rrc); + } md->recursive = &new_recursive; memcpy(md->offset_vector, new_recursive.offset_save, @@ -914,71 +979,72 @@ for (;;) the end of a normal bracket, leaving the subject pointer. */ case OP_ONCE: - { - prev = ecode; - saved_eptr = eptr; + prev = ecode; + saved_eptr = eptr; - do - { - RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, - eptrb, match_isgroup); - if (rrc == MATCH_MATCH) break; - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - ecode += GET(ecode,1); - } - while (*ecode == OP_ALT); + do + { + RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, + eptrb, match_isgroup); + if (rrc == MATCH_MATCH) break; + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode += GET(ecode,1); + } + while (*ecode == OP_ALT); - /* If hit the end of the group (which could be repeated), fail */ + /* If hit the end of the group (which could be repeated), fail */ - if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH); + if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH); - /* Continue as from after the assertion, updating the offsets high water - mark, since extracts may have been taken. */ + /* Continue as from after the assertion, updating the offsets high water + mark, since extracts may have been taken. */ - do ecode += GET(ecode,1); while (*ecode == OP_ALT); + do ecode += GET(ecode,1); while (*ecode == OP_ALT); - offset_top = md->end_offset_top; - eptr = md->end_match_ptr; + offset_top = md->end_offset_top; + eptr = md->end_match_ptr; - /* For a non-repeating ket, just continue at this level. This also - happens for a repeating ket if no characters were matched in the group. - This is the forcible breaking of infinite loops as implemented in Perl - 5.005. If there is an options reset, it will get obeyed in the normal - course of events. */ + /* For a non-repeating ket, just continue at this level. This also + happens for a repeating ket if no characters were matched in the group. + This is the forcible breaking of infinite loops as implemented in Perl + 5.005. If there is an options reset, it will get obeyed in the normal + course of events. */ - if (*ecode == OP_KET || eptr == saved_eptr) - { - ecode += 1+LINK_SIZE; - break; - } + if (*ecode == OP_KET || eptr == saved_eptr) + { + ecode += 1+LINK_SIZE; + break; + } - /* The repeating kets try the rest of the pattern or restart from the - preceding bracket, in the appropriate order. We need to reset any options - that changed within the bracket before re-running it, so check the next - opcode. */ + /* The repeating kets try the rest of the pattern or restart from the + preceding bracket, in the appropriate order. The second "call" of match() + uses tail recursion, to avoid using another stack frame. We need to reset + any options that changed within the bracket before re-running it, so + check the next opcode. */ - if (ecode[1+LINK_SIZE] == OP_OPT) - { - ims = (ims & ~PCRE_IMS) | ecode[4]; - DPRINTF(("ims set to %02lx at group repeat\n", ims)); - } + if (ecode[1+LINK_SIZE] == OP_OPT) + { + ims = (ims & ~PCRE_IMS) | ecode[4]; + DPRINTF(("ims set to %02lx at group repeat\n", ims)); + } - if (*ecode == OP_KETRMIN) - { - RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - else /* OP_KETRMAX */ - { - RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } + if (*ecode == OP_KETRMIN) + { + RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode = prev; + flags = match_isgroup; + goto TAIL_RECURSE; + } + else /* OP_KETRMAX */ + { + RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode += 1 + LINK_SIZE; + flags = 0; + goto TAIL_RECURSE; } - RRETURN(MATCH_NOMATCH); + /* Control never gets here */ /* An alternation is the end of a branch; scan along to find the end of the bracketed group and go to there. */ @@ -1022,114 +1088,114 @@ for (;;) case OP_KET: case OP_KETRMIN: case OP_KETRMAX: - { - prev = ecode - GET(ecode, 1); - saved_eptr = eptrb->epb_saved_eptr; + prev = ecode - GET(ecode, 1); + saved_eptr = eptrb->epb_saved_eptr; - /* Back up the stack of bracket start pointers. */ + /* Back up the stack of bracket start pointers. */ - eptrb = eptrb->epb_prev; + eptrb = eptrb->epb_prev; - if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT || - *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT || - *prev == OP_ONCE) - { - md->end_match_ptr = eptr; /* For ONCE */ - md->end_offset_top = offset_top; - RRETURN(MATCH_MATCH); - } + if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT || + *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT || + *prev == OP_ONCE) + { + md->end_match_ptr = eptr; /* For ONCE */ + md->end_offset_top = offset_top; + RRETURN(MATCH_MATCH); + } - /* In all other cases except a conditional group we have to check the - group number back at the start and if necessary complete handling an - extraction by setting the offsets and bumping the high water mark. */ + /* In all other cases except a conditional group we have to check the + group number back at the start and if necessary complete handling an + extraction by setting the offsets and bumping the high water mark. */ - if (*prev != OP_COND) - { - number = *prev - OP_BRA; + if (*prev != OP_COND) + { + number = *prev - OP_BRA; - /* For extended extraction brackets (large number), we have to fish out - the number from a dummy opcode at the start. */ + /* For extended extraction brackets (large number), we have to fish out + the number from a dummy opcode at the start. */ - if (number > EXTRACT_BASIC_MAX) number = GET2(prev, 2+LINK_SIZE); - offset = number << 1; + if (number > EXTRACT_BASIC_MAX) number = GET2(prev, 2+LINK_SIZE); + offset = number << 1; #ifdef DEBUG - printf("end bracket %d", number); - printf("\n"); + printf("end bracket %d", number); + printf("\n"); #endif - /* Test for a numbered group. This includes groups called as a result - of recursion. Note that whole-pattern recursion is coded as a recurse - into group 0, so it won't be picked up here. Instead, we catch it when - the OP_END is reached. */ + /* Test for a numbered group. This includes groups called as a result + of recursion. Note that whole-pattern recursion is coded as a recurse + into group 0, so it won't be picked up here. Instead, we catch it when + the OP_END is reached. */ - if (number > 0) + if (number > 0) + { + md->capture_last = number; + if (offset >= md->offset_max) md->offset_overflow = TRUE; else { - md->capture_last = number; - if (offset >= md->offset_max) md->offset_overflow = TRUE; else - { - md->offset_vector[offset] = - md->offset_vector[md->offset_end - number]; - md->offset_vector[offset+1] = eptr - md->start_subject; - if (offset_top <= offset) offset_top = offset + 2; - } + md->offset_vector[offset] = + md->offset_vector[md->offset_end - number]; + md->offset_vector[offset+1] = eptr - md->start_subject; + if (offset_top <= offset) offset_top = offset + 2; + } - /* Handle a recursively called group. Restore the offsets - appropriately and continue from after the call. */ + /* Handle a recursively called group. Restore the offsets + appropriately and continue from after the call. */ - if (md->recursive != NULL && md->recursive->group_num == number) - { - recursion_info *rec = md->recursive; - DPRINTF(("Recursion (%d) succeeded - continuing\n", number)); - md->recursive = rec->prevrec; - md->start_match = rec->save_start; - memcpy(md->offset_vector, rec->offset_save, - rec->saved_max * sizeof(int)); - ecode = rec->after_call; - ims = original_ims; - break; - } + if (md->recursive != NULL && md->recursive->group_num == number) + { + recursion_info *rec = md->recursive; + DPRINTF(("Recursion (%d) succeeded - continuing\n", number)); + md->recursive = rec->prevrec; + md->start_match = rec->save_start; + memcpy(md->offset_vector, rec->offset_save, + rec->saved_max * sizeof(int)); + ecode = rec->after_call; + ims = original_ims; + break; } } + } - /* Reset the value of the ims flags, in case they got changed during - the group. */ + /* Reset the value of the ims flags, in case they got changed during + the group. */ - ims = original_ims; - DPRINTF(("ims reset to %02lx\n", ims)); + ims = original_ims; + DPRINTF(("ims reset to %02lx\n", ims)); - /* For a non-repeating ket, just continue at this level. This also - happens for a repeating ket if no characters were matched in the group. - This is the forcible breaking of infinite loops as implemented in Perl - 5.005. If there is an options reset, it will get obeyed in the normal - course of events. */ + /* For a non-repeating ket, just continue at this level. This also + happens for a repeating ket if no characters were matched in the group. + This is the forcible breaking of infinite loops as implemented in Perl + 5.005. If there is an options reset, it will get obeyed in the normal + course of events. */ - if (*ecode == OP_KET || eptr == saved_eptr) - { - ecode += 1 + LINK_SIZE; - break; - } + if (*ecode == OP_KET || eptr == saved_eptr) + { + ecode += 1 + LINK_SIZE; + break; + } - /* The repeating kets try the rest of the pattern or restart from the - preceding bracket, in the appropriate order. */ + /* The repeating kets try the rest of the pattern or restart from the + preceding bracket, in the appropriate order. In the second case, we can use + tail recursion to avoid using another stack frame. */ - if (*ecode == OP_KETRMIN) - { - RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - else /* OP_KETRMAX */ - { - RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } + if (*ecode == OP_KETRMIN) + { + RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode = prev; + flags = match_isgroup; + goto TAIL_RECURSE; } - - RRETURN(MATCH_NOMATCH); + else /* OP_KETRMAX */ + { + RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode += 1 + LINK_SIZE; + flags = 0; + goto TAIL_RECURSE; + } + /* Control never gets here */ /* Start of subject unless notbol, or after internal newline if multiline */ @@ -1137,7 +1203,10 @@ for (;;) if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH); if ((ims & PCRE_MULTILINE) != 0) { - if (eptr != md->start_subject && eptr[-1] != NEWLINE) + if (eptr != md->start_subject && + (eptr == md->end_subject || + eptr < md->start_subject + md->nllen || + !IS_NEWLINE(eptr - md->nllen))) RRETURN(MATCH_NOMATCH); ecode++; break; @@ -1165,7 +1234,7 @@ for (;;) if ((ims & PCRE_MULTILINE) != 0) { if (eptr < md->end_subject) - { if (*eptr != NEWLINE) RRETURN(MATCH_NOMATCH); } + { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); } else { if (md->noteol) RRETURN(MATCH_NOMATCH); } ecode++; @@ -1176,14 +1245,14 @@ for (;;) if (md->noteol) RRETURN(MATCH_NOMATCH); if (!md->endonly) { - if (eptr < md->end_subject - 1 || - (eptr == md->end_subject - 1 && *eptr != NEWLINE)) + if (eptr != md->end_subject && + (eptr != md->end_subject - md->nllen || !IS_NEWLINE(eptr))) RRETURN(MATCH_NOMATCH); ecode++; break; } } - /* ... else fall through */ + /* ... else fall through for endonly */ /* End of subject assertion (\z) */ @@ -1195,8 +1264,9 @@ for (;;) /* End of subject or ending \n assertion (\Z) */ case OP_EODN: - if (eptr < md->end_subject - 1 || - (eptr == md->end_subject - 1 && *eptr != NEWLINE)) RRETURN(MATCH_NOMATCH); + if (eptr != md->end_subject && + (eptr != md->end_subject - md->nllen || !IS_NEWLINE(eptr))) + RRETURN(MATCH_NOMATCH); ecode++; break; @@ -1249,13 +1319,14 @@ for (;;) /* Match a single character type; inline for speed */ case OP_ANY: - if ((ims & PCRE_DOTALL) == 0 && eptr < md->end_subject && *eptr == NEWLINE) - RRETURN(MATCH_NOMATCH); + if ((ims & PCRE_DOTALL) == 0) + { + if (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)) + RRETURN(MATCH_NOMATCH); + } if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH); -#ifdef SUPPORT_UTF8 if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; -#endif ecode++; break; @@ -1354,23 +1425,43 @@ for (;;) if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); GETCHARINCTEST(c, eptr); { - int chartype, rqdtype; - int othercase; - int category = ucp_findchar(c, &chartype, &othercase); - - rqdtype = *(++ecode); - ecode++; + int chartype, script; + int category = _pcre_ucp_findprop(c, &chartype, &script); - if (rqdtype >= 128) + switch(ecode[1]) { - if ((rqdtype - 128 != category) == (op == OP_PROP)) + case PT_ANY: + if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH); + break; + + case PT_LAMP: + if ((chartype == ucp_Lu || + chartype == ucp_Ll || + chartype == ucp_Lt) == (op == OP_NOTPROP)) RRETURN(MATCH_NOMATCH); - } - else - { - if ((rqdtype != chartype) == (op == OP_PROP)) + break; + + case PT_GC: + if ((ecode[2] != category) == (op == OP_PROP)) + RRETURN(MATCH_NOMATCH); + break; + + case PT_PC: + if ((ecode[2] != chartype) == (op == OP_PROP)) RRETURN(MATCH_NOMATCH); + break; + + case PT_SC: + if ((ecode[2] != script) == (op == OP_PROP)) + RRETURN(MATCH_NOMATCH); + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + break; } + + ecode += 3; } break; @@ -1381,9 +1472,8 @@ for (;;) if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); GETCHARINCTEST(c, eptr); { - int chartype; - int othercase; - int category = ucp_findchar(c, &chartype, &othercase); + int chartype, script; + int category = _pcre_ucp_findprop(c, &chartype, &script); if (category == ucp_M) RRETURN(MATCH_NOMATCH); while (eptr < md->end_subject) { @@ -1392,7 +1482,7 @@ for (;;) { GETCHARLEN(c, eptr, len); } - category = ucp_findchar(c, &chartype, &othercase); + category = _pcre_ucp_findprop(c, &chartype, &script); if (category != ucp_M) break; eptr += len; } @@ -1685,8 +1775,8 @@ for (;;) while (eptr >= pp) { RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - eptr--; if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; } } @@ -1843,16 +1933,12 @@ for (;;) ecode += length; /* If we have Unicode property support, we can use it to test the other - case of the character, if there is one. The result of ucp_findchar() is - < 0 if the char isn't found, and othercase is returned as zero if there - isn't one. */ + case of the character, if there is one. */ if (fc != dc) { #ifdef SUPPORT_UCP - int chartype; - int othercase; - if (ucp_findchar(fc, &chartype, &othercase) < 0 || dc != othercase) + if (dc != _pcre_ucp_othercase(fc)) #endif RRETURN(MATCH_NOMATCH); } @@ -1920,10 +2006,9 @@ for (;;) #ifdef SUPPORT_UCP int othercase; - int chartype; if ((ims & PCRE_CASELESS) != 0 && - ucp_findchar(fc, &chartype, &othercase) >= 0 && - othercase > 0) + (othercase = _pcre_ucp_othercase(fc)) >= 0 && + othercase >= 0) oclength = _pcre_ord2utf8(othercase, occhars); #endif /* SUPPORT_UCP */ @@ -2410,16 +2495,7 @@ for (;;) { prop_fail_result = ctype == OP_NOTPROP; prop_type = *ecode++; - if (prop_type >= 128) - { - prop_test_against = prop_type - 128; - prop_test_variable = &prop_category; - } - else - { - prop_test_against = prop_type; - prop_test_variable = &prop_chartype; - } + prop_value = *ecode++; } else prop_type = -1; #endif @@ -2436,14 +2512,68 @@ for (;;) if (min > 0) { #ifdef SUPPORT_UCP - if (prop_type > 0) + if (prop_type >= 0) { - for (i = 1; i <= min; i++) + switch(prop_type) { - GETCHARINC(c, eptr); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if ((*prop_test_variable == prop_test_against) == prop_fail_result) - RRETURN(MATCH_NOMATCH); + case PT_ANY: + if (prop_fail_result) RRETURN(MATCH_NOMATCH); + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + } + break; + + case PT_LAMP: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_chartype == ucp_Lu || + prop_chartype == ucp_Ll || + prop_chartype == ucp_Lt) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_GC: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_category == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_PC: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_chartype == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_SC: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_script == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + break; } } @@ -2455,7 +2585,7 @@ for (;;) for (i = 1; i <= min; i++) { GETCHARINCTEST(c, eptr); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH); while (eptr < md->end_subject) { @@ -2464,7 +2594,7 @@ for (;;) { GETCHARLEN(c, eptr, len); } - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); if (prop_category != ucp_M) break; eptr += len; } @@ -2483,8 +2613,11 @@ for (;;) for (i = 1; i <= min; i++) { if (eptr >= md->end_subject || - (*eptr++ == NEWLINE && (ims & PCRE_DOTALL) == 0)) + ((ims & PCRE_DOTALL) == 0 && + eptr <= md->end_subject - md->nllen && + IS_NEWLINE(eptr))) RRETURN(MATCH_NOMATCH); + eptr++; while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; } break; @@ -2569,7 +2702,11 @@ for (;;) if ((ims & PCRE_DOTALL) == 0) { for (i = 1; i <= min; i++) - if (*eptr++ == NEWLINE) RRETURN(MATCH_NOMATCH); + { + if (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)) + RRETURN(MATCH_NOMATCH); + eptr++; + } } else eptr += min; break; @@ -2626,17 +2763,78 @@ for (;;) if (minimize) { #ifdef SUPPORT_UCP - if (prop_type > 0) + if (prop_type >= 0) { - for (fi = min;; fi++) + switch(prop_type) { - RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if ((*prop_test_variable == prop_test_against) == prop_fail_result) - RRETURN(MATCH_NOMATCH); + case PT_ANY: + for (fi = min;; fi++) + { + RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + if (prop_fail_result) RRETURN(MATCH_NOMATCH); + } + break; + + case PT_LAMP: + for (fi = min;; fi++) + { + RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_chartype == ucp_Lu || + prop_chartype == ucp_Ll || + prop_chartype == ucp_Lt) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_GC: + for (fi = min;; fi++) + { + RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_category == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_PC: + for (fi = min;; fi++) + { + RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_chartype == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_SC: + for (fi = min;; fi++) + { + RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_script == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + break; } } @@ -2651,7 +2849,7 @@ for (;;) if (rrc != MATCH_NOMATCH) RRETURN(rrc); if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); GETCHARINCTEST(c, eptr); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH); while (eptr < md->end_subject) { @@ -2660,7 +2858,7 @@ for (;;) { GETCHARLEN(c, eptr, len); } - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); if (prop_category != ucp_M) break; eptr += len; } @@ -2678,13 +2876,15 @@ for (;;) { RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + if (fi >= max || eptr >= md->end_subject || + (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 && + eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr))) + RRETURN(MATCH_NOMATCH); GETCHARINC(c, eptr); switch(ctype) { - case OP_ANY: - if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH); + case OP_ANY: /* This is the DOTALL case */ break; case OP_ANYBYTE: @@ -2733,12 +2933,15 @@ for (;;) { RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0); if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); + if (fi >= max || eptr >= md->end_subject || + ((ims & PCRE_DOTALL) == 0 && + eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr))) + RRETURN(MATCH_NOMATCH); + c = *eptr++; switch(ctype) { - case OP_ANY: - if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH); + case OP_ANY: /* This is the DOTALL case */ break; case OP_ANYBYTE: @@ -2785,17 +2988,74 @@ for (;;) pp = eptr; /* Remember where we started */ #ifdef SUPPORT_UCP - if (prop_type > 0) + if (prop_type >= 0) { - for (i = min; i < max; i++) + switch(prop_type) { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); - if ((*prop_test_variable == prop_test_against) == prop_fail_result) - break; - eptr+= len; + case PT_ANY: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) break; + GETCHARLEN(c, eptr, len); + if (prop_fail_result) break; + eptr+= len; + } + break; + + case PT_LAMP: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) break; + GETCHARLEN(c, eptr, len); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_chartype == ucp_Lu || + prop_chartype == ucp_Ll || + prop_chartype == ucp_Lt) == prop_fail_result) + break; + eptr+= len; + } + break; + + case PT_GC: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) break; + GETCHARLEN(c, eptr, len); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_category == prop_value) == prop_fail_result) + break; + eptr+= len; + } + break; + + case PT_PC: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) break; + GETCHARLEN(c, eptr, len); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_chartype == prop_value) == prop_fail_result) + break; + eptr+= len; + } + break; + + case PT_SC: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) break; + GETCHARLEN(c, eptr, len); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); + if ((prop_script == prop_value) == prop_fail_result) + break; + eptr+= len; + } + break; } /* eptr is now past the end of the maximum run */ @@ -2818,7 +3078,7 @@ for (;;) { if (eptr >= md->end_subject) break; GETCHARINCTEST(c, eptr); - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); if (prop_category == ucp_M) break; while (eptr < md->end_subject) { @@ -2827,7 +3087,7 @@ for (;;) { GETCHARLEN(c, eptr, len); } - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); if (prop_category != ucp_M) break; eptr += len; } @@ -2848,7 +3108,7 @@ for (;;) { GETCHARLEN(c, eptr, len); } - prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase); + prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); if (prop_category != ucp_M) break; eptr--; } @@ -2867,9 +3127,9 @@ for (;;) { case OP_ANY: - /* Special code is required for UTF8, but when the maximum is unlimited - we don't need it, so we repeat the non-UTF8 code. This is probably - worth it, because .* is quite a common idiom. */ + /* Special code is required for UTF8, but when the maximum is + unlimited we don't need it, so we repeat the non-UTF8 code. This is + probably worth it, because .* is quite a common idiom. */ if (max < INT_MAX) { @@ -2877,7 +3137,9 @@ for (;;) { for (i = min; i < max; i++) { - if (eptr >= md->end_subject || *eptr == NEWLINE) break; + if (eptr >= md->end_subject || + (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr))) + break; eptr++; while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; } @@ -2886,6 +3148,7 @@ for (;;) { for (i = min; i < max; i++) { + if (eptr >= md->end_subject) break; eptr++; while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; } @@ -2900,7 +3163,9 @@ for (;;) { for (i = min; i < max; i++) { - if (eptr >= md->end_subject || *eptr == NEWLINE) break; + if (eptr >= md->end_subject || + (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr))) + break; eptr++; } break; @@ -3014,7 +3279,9 @@ for (;;) { for (i = min; i < max; i++) { - if (eptr >= md->end_subject || *eptr == NEWLINE) break; + if (eptr >= md->end_subject || + (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr))) + break; eptr++; } break; @@ -3202,16 +3469,17 @@ Returns: > 0 => success; value is the number of elements filled in < -1 => some kind of unexpected problem */ -EXPORT int +PCRE_DATA_SCOPE int pcre_exec(const pcre *argument_re, const pcre_extra *extra_data, - const char *subject, int length, int start_offset, int options, int *offsets, + PCRE_SPTR subject, int length, int start_offset, int options, int *offsets, int offsetcount) { int rc, resetcount, ocount; int first_byte = -1; int req_byte = -1; int req_byte2 = -1; -unsigned long int ims = 0; +int newline; +unsigned long int ims; BOOL using_temporary_offsets = FALSE; BOOL anchored; BOOL startline; @@ -3219,11 +3487,12 @@ BOOL firstline; BOOL first_byte_caseless = FALSE; BOOL req_byte_caseless = FALSE; match_data match_block; +match_data *md = &match_block; const uschar *tables; const uschar *start_bits = NULL; -const uschar *start_match = (const uschar *)subject + start_offset; -const uschar *end_subject; -const uschar *req_byte_ptr = start_match - 1; +USPTR start_match = (USPTR)subject + start_offset; +USPTR end_subject; +USPTR req_byte_ptr = start_match - 1; pcre_study_data internal_study; const pcre_study_data *study; @@ -3243,8 +3512,9 @@ if (offsetcount < 0) return PCRE_ERROR_BADCOUNT; the default values. */ study = NULL; -match_block.match_limit = MATCH_LIMIT; -match_block.callout_data = NULL; +md->match_limit = MATCH_LIMIT; +md->match_limit_recursion = MATCH_LIMIT_RECURSION; +md->callout_data = NULL; /* The table pointer is always in native byte order. */ @@ -3256,9 +3526,11 @@ if (extra_data != NULL) if ((flags & PCRE_EXTRA_STUDY_DATA) != 0) study = (const pcre_study_data *)extra_data->study_data; if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0) - match_block.match_limit = extra_data->match_limit; + md->match_limit = extra_data->match_limit; + if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0) + md->match_limit_recursion = extra_data->match_limit_recursion; if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0) - match_block.callout_data = extra_data->callout_data; + md->callout_data = extra_data->callout_data; if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables; } @@ -3288,39 +3560,64 @@ firstline = (re->options & PCRE_FIRSTLINE) != 0; /* The code starts after the real_pcre block and the capture name table. */ -match_block.start_code = (const uschar *)external_re + re->name_table_offset + +md->start_code = (const uschar *)external_re + re->name_table_offset + re->name_count * re->name_entry_size; -match_block.start_subject = (const uschar *)subject; -match_block.start_offset = start_offset; -match_block.end_subject = match_block.start_subject + length; -end_subject = match_block.end_subject; +md->start_subject = (USPTR)subject; +md->start_offset = start_offset; +md->end_subject = md->start_subject + length; +end_subject = md->end_subject; + +md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0; +md->utf8 = (re->options & PCRE_UTF8) != 0; -match_block.endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0; -match_block.utf8 = (re->options & PCRE_UTF8) != 0; +md->notbol = (options & PCRE_NOTBOL) != 0; +md->noteol = (options & PCRE_NOTEOL) != 0; +md->notempty = (options & PCRE_NOTEMPTY) != 0; +md->partial = (options & PCRE_PARTIAL) != 0; +md->hitend = FALSE; -match_block.notbol = (options & PCRE_NOTBOL) != 0; -match_block.noteol = (options & PCRE_NOTEOL) != 0; -match_block.notempty = (options & PCRE_NOTEMPTY) != 0; -match_block.partial = (options & PCRE_PARTIAL) != 0; -match_block.hitend = FALSE; +md->recursive = NULL; /* No recursion at top level */ -match_block.recursive = NULL; /* No recursion at top level */ +md->lcc = tables + lcc_offset; +md->ctypes = tables + ctypes_offset; -match_block.lcc = tables + lcc_offset; -match_block.ctypes = tables + ctypes_offset; +/* Handle different types of newline. The two bits give four cases. If nothing +is set at run time, whatever was used at compile time applies. */ + +switch ((((options & PCRE_NEWLINE_CRLF) == 0)? re->options : options) & + PCRE_NEWLINE_CRLF) + { + default: newline = NEWLINE; break; /* Compile-time default */ + case PCRE_NEWLINE_CR: newline = '\r'; break; + case PCRE_NEWLINE_LF: newline = '\n'; break; + case PCRE_NEWLINE_CR+ + PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break; + } + +if (newline > 255) + { + md->nllen = 2; + md->nl[0] = (newline >> 8) & 255; + md->nl[1] = newline & 255; + } +else + { + md->nllen = 1; + md->nl[0] = newline; + } /* Partial matching is supported only for a restricted set of regexes at the moment. */ -if (match_block.partial && (re->options & PCRE_NOPARTIAL) != 0) +if (md->partial && (re->options & PCRE_NOPARTIAL) != 0) return PCRE_ERROR_BADPARTIAL; /* Check a UTF-8 string if required. Unfortunately there's no way of passing back the character offset. */ #ifdef SUPPORT_UTF8 -if (match_block.utf8 && (options & PCRE_NO_UTF8_CHECK) == 0) +if (md->utf8 && (options & PCRE_NO_UTF8_CHECK) == 0) { if (_pcre_valid_utf8((uschar *)subject, length) >= 0) return PCRE_ERROR_BADUTF8; @@ -3352,17 +3649,17 @@ ocount = offsetcount - (offsetcount % 3); if (re->top_backref > 0 && re->top_backref >= ocount/3) { ocount = re->top_backref * 3 + 3; - match_block.offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int)); - if (match_block.offset_vector == NULL) return PCRE_ERROR_NOMEMORY; + md->offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int)); + if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY; using_temporary_offsets = TRUE; DPRINTF(("Got memory to hold back references\n")); } -else match_block.offset_vector = offsets; +else md->offset_vector = offsets; -match_block.offset_end = ocount; -match_block.offset_max = (2*ocount)/3; -match_block.offset_overflow = FALSE; -match_block.capture_last = -1; +md->offset_end = ocount; +md->offset_max = (2*ocount)/3; +md->offset_overflow = FALSE; +md->capture_last = -1; /* Compute the minimum number of offsets that we need to reset each time. Doing this makes a huge difference to execution time when there aren't many brackets @@ -3375,9 +3672,9 @@ if (resetcount > offsetcount) resetcount = ocount; never be used unless previously set, but they get saved and restored, and so we initialize them to avoid reading uninitialized locations. */ -if (match_block.offset_vector != NULL) +if (md->offset_vector != NULL) { - register int *iptr = match_block.offset_vector + ocount; + register int *iptr = md->offset_vector + ocount; register int *iend = iptr - resetcount/2 + 1; while (--iptr >= iend) *iptr = -1; } @@ -3394,7 +3691,7 @@ if (!anchored) { first_byte = re->first_byte & 255; if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE) - first_byte = match_block.lcc[first_byte]; + first_byte = md->lcc[first_byte]; } else if (!startline && study != NULL && @@ -3417,13 +3714,13 @@ the loop runs just once. */ do { - const uschar *save_end_subject = end_subject; + USPTR save_end_subject = end_subject; /* Reset the maximum number of extractions we might see. */ - if (match_block.offset_vector != NULL) + if (md->offset_vector != NULL) { - register int *iptr = match_block.offset_vector; + register int *iptr = md->offset_vector; register int *iend = iptr + resetcount; while (iptr < iend) *iptr++ = -1; } @@ -3436,8 +3733,8 @@ do if (firstline) { - const uschar *t = start_match; - while (t < save_end_subject && *t != '\n') t++; + USPTR t = start_match; + while (t <= save_end_subject - md->nllen && !IS_NEWLINE(t)) t++; end_subject = t; } @@ -3447,20 +3744,22 @@ do { if (first_byte_caseless) while (start_match < end_subject && - match_block.lcc[*start_match] != first_byte) + md->lcc[*start_match] != first_byte) start_match++; else while (start_match < end_subject && *start_match != first_byte) start_match++; } - /* Or to just after \n for a multiline match if possible */ + /* Or to just after a linebreak for a multiline match if possible */ else if (startline) { - if (start_match > match_block.start_subject + start_offset) + if (start_match >= md->start_subject + md->nllen + + start_offset) { - while (start_match < end_subject && start_match[-1] != NEWLINE) + while (start_match <= end_subject && + !IS_NEWLINE(start_match - md->nllen)) start_match++; } } @@ -3482,7 +3781,7 @@ do #ifdef DEBUG /* Sigh. Some compilers never learn. */ printf(">>>> Match against: "); - pchars(start_match, end_subject - start_match, TRUE, &match_block); + pchars(start_match, end_subject - start_match, TRUE, md); printf("\n"); #endif @@ -3504,9 +3803,9 @@ do if (req_byte >= 0 && end_subject - start_match < REQ_BYTE_MAX && - !match_block.partial) + !md->partial) { - register const uschar *p = start_match + ((first_byte >= 0)? 1 : 0); + register USPTR p = start_match + ((first_byte >= 0)? 1 : 0); /* We don't need to repeat the search if we haven't yet reached the place we found it at last time. */ @@ -3548,11 +3847,10 @@ do those back references that we can. In this case there need not be overflow if certain parts of the pattern were not used. */ - match_block.start_match = start_match; - match_block.match_call_count = 0; + md->start_match = start_match; + md->match_call_count = 0; - rc = match(start_match, match_block.start_code, 2, &match_block, ims, NULL, - match_isgroup); + rc = match(start_match, md->start_code, 2, md, ims, NULL, match_isgroup, 0); /* When the result is no match, if the subject's first character was a newline and the PCRE_FIRSTLINE option is set, break (which will return @@ -3563,10 +3861,13 @@ do if (rc == MATCH_NOMATCH) { - if (firstline && *start_match == NEWLINE) break; + if (firstline && + start_match <= md->end_subject - md->nllen && + IS_NEWLINE(start_match)) + break; start_match++; #ifdef SUPPORT_UTF8 - if (match_block.utf8) + if (md->utf8) while(start_match < end_subject && (*start_match & 0xc0) == 0x80) start_match++; #endif @@ -3586,23 +3887,23 @@ do { if (offsetcount >= 4) { - memcpy(offsets + 2, match_block.offset_vector + 2, + memcpy(offsets + 2, md->offset_vector + 2, (offsetcount - 2) * sizeof(int)); DPRINTF(("Copied offsets from temporary memory\n")); } - if (match_block.end_offset_top > offsetcount) - match_block.offset_overflow = TRUE; + if (md->end_offset_top > offsetcount) + md->offset_overflow = TRUE; DPRINTF(("Freeing temporary memory\n")); - (pcre_free)(match_block.offset_vector); + (pcre_free)(md->offset_vector); } - rc = match_block.offset_overflow? 0 : match_block.end_offset_top/2; + rc = md->offset_overflow? 0 : md->end_offset_top/2; if (offsetcount < 2) rc = 0; else { - offsets[0] = start_match - match_block.start_subject; - offsets[1] = match_block.end_match_ptr - match_block.start_subject; + offsets[0] = start_match - md->start_subject; + offsets[1] = md->end_match_ptr - md->start_subject; } DPRINTF((">>>> returning %d\n", rc)); @@ -3616,10 +3917,10 @@ while (!anchored && start_match <= end_subject); if (using_temporary_offsets) { DPRINTF(("Freeing temporary memory\n")); - (pcre_free)(match_block.offset_vector); + (pcre_free)(md->offset_vector); } -if (match_block.partial && match_block.hitend) +if (md->partial && md->hitend) { DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n")); return PCRE_ERROR_PARTIAL; diff --git a/src/src/pcre/pcre_fullinfo.c b/src/src/pcre/pcre_fullinfo.c index 645a58280..244afd9c7 100644 --- a/src/src/pcre/pcre_fullinfo.c +++ b/src/src/pcre/pcre_fullinfo.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_fullinfo.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_fullinfo.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -63,7 +63,7 @@ Arguments: Returns: 0 if data returned, negative on error */ -EXPORT int +PCRE_DATA_SCOPE int pcre_fullinfo(const pcre *argument_re, const pcre_extra *extra_data, int what, void *where) { diff --git a/src/src/pcre/pcre_get.c b/src/src/pcre/pcre_get.c index e0b00f171..835ce0d97 100644 --- a/src/src/pcre/pcre_get.c +++ b/src/src/pcre/pcre_get.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_get.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_get.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -52,8 +52,8 @@ for these functions came from Scott Wimer. */ * Find number for named string * *************************************************/ -/* This function is used by the two extraction functions below, as well -as being generally available. +/* This function is used by the get_first_set() function below, as well +as being generally available. It assumes that names are unique. Arguments: code the compiled regex @@ -96,6 +96,113 @@ return PCRE_ERROR_NOSUBSTRING; /************************************************* +* Find (multiple) entries for named string * +*************************************************/ + +/* This is used by the get_first_set() function below, as well as being +generally available. It is used when duplicated names are permitted. + +Arguments: + code the compiled regex + stringname the name whose entries required + firstptr where to put the pointer to the first entry + lastptr where to put the pointer to the last entry + +Returns: the length of each entry, or a negative number + (PCRE_ERROR_NOSUBSTRING) if not found +*/ + +int +pcre_get_stringtable_entries(const pcre *code, const char *stringname, + char **firstptr, char **lastptr) +{ +int rc; +int entrysize; +int top, bot; +uschar *nametable, *lastentry; + +if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMECOUNT, &top)) != 0) + return rc; +if (top <= 0) return PCRE_ERROR_NOSUBSTRING; + +if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMEENTRYSIZE, &entrysize)) != 0) + return rc; +if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMETABLE, &nametable)) != 0) + return rc; + +lastentry = nametable + entrysize * (top - 1); +bot = 0; +while (top > bot) + { + int mid = (top + bot) / 2; + uschar *entry = nametable + entrysize*mid; + int c = strcmp(stringname, (char *)(entry + 2)); + if (c == 0) + { + uschar *first = entry; + uschar *last = entry; + while (first > nametable) + { + if (strcmp(stringname, (char *)(first - entrysize + 2)) != 0) break; + first -= entrysize; + } + while (last < lastentry) + { + if (strcmp(stringname, (char *)(last + entrysize + 2)) != 0) break; + last += entrysize; + } + *firstptr = (char *)first; + *lastptr = (char *)last; + return entrysize; + } + if (c > 0) bot = mid + 1; else top = mid; + } + +return PCRE_ERROR_NOSUBSTRING; +} + + + +/************************************************* +* Find first set of multiple named strings * +*************************************************/ + +/* This function allows for duplicate names in the table of named substrings. +It returns the number of the first one that was set in a pattern match. + +Arguments: + code the compiled regex + stringname the name of the capturing substring + ovector the vector of matched substrings + +Returns: the number of the first that is set, + or the number of the last one if none are set, + or a negative number on error +*/ + +static int +get_first_set(const pcre *code, const char *stringname, int *ovector) +{ +const real_pcre *re = (const real_pcre *)code; +int entrysize; +char *first, *last; +uschar *entry; +if ((re->options & (PCRE_DUPNAMES | PCRE_JCHANGED)) == 0) + return pcre_get_stringnumber(code, stringname); +entrysize = pcre_get_stringtable_entries(code, stringname, &first, &last); +if (entrysize <= 0) return entrysize; +for (entry = (uschar *)first; entry <= (uschar *)last; entry += entrysize) + { + int n = (entry[0] << 8) + entry[1]; + if (ovector[n*2] >= 0) return n; + } +return (first[0] << 8) + first[1]; +} + + + + +/************************************************* * Copy captured string to given buffer * *************************************************/ @@ -144,7 +251,8 @@ return yield; *************************************************/ /* This function copies a single captured substring into a given buffer, -identifying it by name. +identifying it by name. If the regex permits duplicate names, the first +substring that is set is chosen. Arguments: code the compiled regex @@ -170,7 +278,7 @@ int pcre_copy_named_substring(const pcre *code, const char *subject, int *ovector, int stringcount, const char *stringname, char *buffer, int size) { -int n = pcre_get_stringnumber(code, stringname); +int n = get_first_set(code, stringname, ovector); if (n <= 0) return n; return pcre_copy_substring(subject, ovector, stringcount, n, buffer, size); } @@ -301,7 +409,8 @@ return yield; *************************************************/ /* This function copies a single captured substring, identified by name, into -new store. +new store. If the regex permits duplicate names, the first substring that is +set is chosen. Arguments: code the compiled regex @@ -326,9 +435,10 @@ int pcre_get_named_substring(const pcre *code, const char *subject, int *ovector, int stringcount, const char *stringname, const char **stringptr) { -int n = pcre_get_stringnumber(code, stringname); +int n = get_first_set(code, stringname, ovector); if (n <= 0) return n; return pcre_get_substring(subject, ovector, stringcount, n, stringptr); + } diff --git a/src/src/pcre/pcre_globals.c b/src/src/pcre/pcre_globals.c index 434a57a24..90ff487da 100644 --- a/src/src/pcre/pcre_globals.c +++ b/src/src/pcre/pcre_globals.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_globals.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_globals.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without diff --git a/src/src/pcre/pcre_internal.h b/src/src/pcre/pcre_internal.h index 4b35bcb9b..0eec49e4c 100644 --- a/src/src/pcre/pcre_internal.h +++ b/src/src/pcre/pcre_internal.h @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_internal.h,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_internal.h,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -9,7 +9,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -44,12 +44,14 @@ POSSIBILITY OF SUCH DAMAGE. modules, but which are not relevant to the exported API. This includes some functions whose names all begin with "_pcre_". */ +#ifndef PCRE_INTERNAL_H +#define PCRE_INTERNAL_H /* Define DEBUG to get debugging output on stdout. */ -/**** +#if 0 #define DEBUG -****/ +#endif /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef inline, and there are *still* stupid compilers about that don't like indented @@ -114,14 +116,35 @@ Unix, where it is defined in sys/types, so use "uschar" instead. */ typedef unsigned char uschar; -/* Include the public PCRE header */ - -#include "pcre.h" +/* PCRE is able to support 3 different kinds of newline (CR, LF, CRLF). The +following macro is used to package up testing for newlines. NLBLOCK is defined +in the various modules to indicate in which datablock the parameters exist. */ + +#define IS_NEWLINE(p) \ + ((p)[0] == NLBLOCK->nl[0] && \ + (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1])) + +/* When PCRE is compiled as a C++ library, the subject pointer can be replaced +with a custom type. This makes it possible, for example, to allow pcre_exec() +to process subject strings that are discontinuous by using a smart pointer +class. It must always be possible to inspect all of the subject string in +pcre_exec() because of the way it backtracks. Two macros are required in the +normal case, for sign-unspecified and unsigned char pointers. The former is +used for the external interface and appears in pcre.h, which is why its name +must begin with PCRE_. */ + +#ifdef CUSTOM_SUBJECT_PTR +#define PCRE_SPTR CUSTOM_SUBJECT_PTR +#define USPTR CUSTOM_SUBJECT_PTR +#else +#define PCRE_SPTR const char * +#define USPTR const unsigned char * +#endif -/* Include the (copy of) the public ucp header, changing the external name into -a private one. This does no harm, even if we aren't compiling UCP support. */ +/* Include the public PCRE header and the definitions of UCP character property +values. */ -#define ucp_findchar _pcre_ucp_findchar +#include "pcre.h" #include "ucp.h" /* When compiling for use with the Virtual Pascal compiler, these functions @@ -147,13 +170,14 @@ case in PCRE. */ #if HAVE_BCOPY #define memmove(a, b, c) bcopy(b, a, c) #else /* HAVE_BCOPY */ -void * +static void * pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n) { -int i; +size_t i; dest += n; src += n; for (i = 0; i < n; ++i) *(--dest) = *(--src); +return dest; } #define memmove(a, b, c) pcre_memmove(a, b, c) #endif /* not HAVE_BCOPY */ @@ -359,16 +383,17 @@ Standard C system should have one. */ #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL) -/* Private options flags start at the most significant end of the four bytes, -but skip the top bit so we can use ints for convenience without getting tangled -with negative values. The public options defined in pcre.h start at the least -significant end. Make sure they don't overlap! */ +/* Private options flags start at the most significant end of the four bytes. +The public options defined in pcre.h start at the least significant end. Make +sure they don't overlap! The bits are getting a bit scarce now -- when we run +out, there is a dummy word in the structure that could be used for the private +bits. */ +#define PCRE_NOPARTIAL 0x80000000 /* can't use partial with this regex */ #define PCRE_FIRSTSET 0x40000000 /* first_byte is set */ #define PCRE_REQCHSET 0x20000000 /* req_byte is set */ #define PCRE_STARTLINE 0x10000000 /* start after \n for multiline */ -#define PCRE_ICHANGED 0x08000000 /* i option changes within regex */ -#define PCRE_NOPARTIAL 0x04000000 /* can't use partial with this regex */ +#define PCRE_JCHANGED 0x08000000 /* j option changes within regex */ /* Options for the "extra" block produced by pcre_study(). */ @@ -380,15 +405,17 @@ time, run time, or study time, respectively. */ #define PUBLIC_OPTIONS \ (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \ PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \ - PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE) + PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \ + PCRE_DUPNAMES|PCRE_NEWLINE_CR|PCRE_NEWLINE_LF) #define PUBLIC_EXEC_OPTIONS \ (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \ - PCRE_PARTIAL) + PCRE_PARTIAL|PCRE_NEWLINE_CR|PCRE_NEWLINE_LF) #define PUBLIC_DFA_EXEC_OPTIONS \ (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \ - PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART) + PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_CR| \ + PCRE_NEWLINE_LF) #define PUBLIC_STUDY_OPTIONS 0 /* None defined */ @@ -447,6 +474,26 @@ ESC_n is defined as yet another macro, which is set in config.h to either \n #define ESC_tee '\t' #endif +/* Codes for different types of Unicode property */ + +#define PT_ANY 0 /* Any property - matches all chars */ +#define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */ +#define PT_GC 2 /* General characteristic (e.g. L) */ +#define PT_PC 3 /* Particular characteristic (e.g. Lu) */ +#define PT_SC 4 /* Script (e.g. Han) */ + +/* Flag bits and data types for the extended class (OP_XCLASS) for classes that +contain UTF-8 characters with values greater than 255. */ + +#define XCL_NOT 0x01 /* Flag: this is a negative class */ +#define XCL_MAP 0x02 /* Flag: a 32-byte map is present */ + +#define XCL_END 0 /* Marks end of individual items */ +#define XCL_SINGLE 1 /* Single item (one multibyte char) follows */ +#define XCL_RANGE 2 /* A range (two multibyte chars) follows */ +#define XCL_PROP 3 /* Unicode property (2-byte property code follows) */ +#define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */ + /* These are escaped items that aren't just an encoding of a particular data value such as \n. They must have non-zero values, as check_escape() returns their negation. Also, they must appear in the same order as in the opcode @@ -462,19 +509,6 @@ enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_REF }; -/* Flag bits and data types for the extended class (OP_XCLASS) for classes that -contain UTF-8 characters with values greater than 255. */ - -#define XCL_NOT 0x01 /* Flag: this is a negative class */ -#define XCL_MAP 0x02 /* Flag: a 32-byte map is present */ - -#define XCL_END 0 /* Marks end of individual items */ -#define XCL_SINGLE 1 /* Single item (one multibyte char) follows */ -#define XCL_RANGE 2 /* A range (two multibyte chars) follows */ -#define XCL_PROP 3 /* Unicode property (one property code) follows */ -#define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */ - - /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets that extract substrings. Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in order to the list of escapes immediately above. @@ -509,7 +543,7 @@ enum { OP_DOLL, /* 20 End of line - varies with multiline switch */ OP_CHAR, /* 21 Match one character, casefully */ OP_CHARNC, /* 22 Match one character, caselessly */ - OP_NOT, /* 23 Match anything but the following char */ + OP_NOT, /* 23 Match one character, not the following one */ OP_STAR, /* 24 The maximizing and minimizing versions of */ OP_MINSTAR, /* 25 all these opcodes must come in pairs, with */ @@ -638,7 +672,7 @@ in UTF-8 mode. The code that uses this table must know about such things. */ 1, /* End */ \ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \ 1, 1, /* Any, Anybyte */ \ - 2, 2, 1, /* NOTPROP, PROP, EXTUNI */ \ + 3, 3, 1, /* NOTPROP, PROP, EXTUNI */ \ 1, 1, 2, 1, 1, /* \Z, \z, Opt, ^, $ */ \ 2, /* Char - the minimum length */ \ 2, /* Charnc - the minimum length */ \ @@ -689,7 +723,8 @@ enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9, ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19, ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29, ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39, - ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 }; + ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49, + ERR50, ERR51 }; /* The real format of the start of the pcre block; the index of names and the code vector run on as long as necessary after the end. We store an explicit @@ -753,6 +788,8 @@ typedef struct compile_data { unsigned int backref_map; /* Bitmap of low back refs */ int req_varyopt; /* "After variable item" flag for reqbyte */ BOOL nopartial; /* Set TRUE if partial won't work */ + int nllen; /* 1 or 2 for newline string length */ + uschar nl[4]; /* Newline string */ } compile_data; /* Structure for maintaining a chain of pointers to the currently incomplete @@ -770,18 +807,18 @@ typedef struct recursion_info { struct recursion_info *prevrec; /* Previous recursion record (or NULL) */ int group_num; /* Number of group that was called */ const uschar *after_call; /* "Return value": points after the call in the expr */ - const uschar *save_start; /* Old value of md->start_match */ + USPTR save_start; /* Old value of md->start_match */ int *offset_save; /* Pointer to start of saved offsets */ int saved_max; /* Number of saved offsets */ } recursion_info; /* When compiling in a mode that doesn't use recursive calls to match(), a structure is used to remember local variables on the heap. It is defined in -pcre.c, close to the match() function, so that it is easy to keep it in step -with any changes of local variable. However, the pointer to the current frame -must be saved in some "static" place over a longjmp(). We declare the -structure here so that we can put a pointer in the match_data structure. -NOTE: This isn't used for a "normal" compilation of pcre. */ +pcre_exec.c, close to the match() function, so that it is easy to keep it in +step with any changes of local variable. However, the pointer to the current +frame must be saved in some "static" place over a longjmp(). We declare the +structure here so that we can put a pointer in the match_data structure. NOTE: +This isn't used for a "normal" compilation of pcre. */ struct heapframe; @@ -789,11 +826,14 @@ struct heapframe; doing traditional NFA matching, so that they are thread-safe. */ typedef struct match_data { - unsigned long int match_call_count; /* As it says */ - unsigned long int match_limit;/* As it says */ + unsigned long int match_call_count; /* As it says */ + unsigned long int match_limit; /* As it says */ + unsigned long int match_limit_recursion; /* As it says */ int *offset_vector; /* Offset vector */ int offset_end; /* One past the end */ int offset_max; /* The maximum usable for return data */ + int nllen; /* 1 or 2 for newline string length */ + uschar nl[4]; /* Newline string */ const uschar *lcc; /* Points to lower casing table */ const uschar *ctypes; /* Points to table of type maps */ BOOL offset_overflow; /* Set if too many extractions */ @@ -805,10 +845,10 @@ typedef struct match_data { BOOL partial; /* PARTIAL flag */ BOOL hitend; /* Hit the end of the subject at some point */ const uschar *start_code; /* For use when recursing */ - const uschar *start_subject; /* Start of the subject string */ - const uschar *end_subject; /* End of the subject string */ - const uschar *start_match; /* Start of this match attempt */ - const uschar *end_match_ptr; /* Subject position at end match */ + USPTR start_subject; /* Start of the subject string */ + USPTR end_subject; /* End of the subject string */ + USPTR start_match; /* Start of this match attempt */ + USPTR end_match_ptr; /* Subject position at end match */ int end_offset_top; /* Highwater mark at end of match */ int capture_last; /* Most recent capture number */ int start_offset; /* The start offset value */ @@ -827,6 +867,8 @@ typedef struct dfa_match_data { const uschar *tables; /* Character tables */ int moptions; /* Match options */ int poptions; /* Pattern options */ + int nllen; /* 1 or 2 for newline string length */ + uschar nl[4]; /* Newline string */ void *callout_data; /* To pass back to callouts */ } dfa_match_data; @@ -863,12 +905,13 @@ total length. */ #define ctypes_offset (cbits_offset + cbit_length) #define tables_length (ctypes_offset + 256) -/* Layout of the UCP type table that translates property names into codes for -ucp_findchar(). */ +/* Layout of the UCP type table that translates property names into types and +codes. */ typedef struct { const char *name; - int value; + pcre_uint16 type; + pcre_uint16 value; } ucp_type_table; @@ -897,11 +940,13 @@ one of the exported public functions. They have to be "external" in the C sense, but are not part of the PCRE public API. */ extern int _pcre_ord2utf8(int, uschar *); -extern void _pcre_printint(pcre *, FILE *); extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *, const pcre_study_data *, pcre_study_data *); -extern int _pcre_ucp_findchar(const int, int *, int *); +extern int _pcre_ucp_findprop(const unsigned int, int *, int *); +extern int _pcre_ucp_othercase(const int); extern int _pcre_valid_utf8(const uschar *, int); extern BOOL _pcre_xclass(int, const uschar *); +#endif + /* End of pcre_internal.h */ diff --git a/src/src/pcre/pcre_maketables.c b/src/src/pcre/pcre_maketables.c index d686383ac..3f6ec8415 100644 --- a/src/src/pcre/pcre_maketables.c +++ b/src/src/pcre/pcre_maketables.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_maketables.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_maketables.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -88,29 +88,22 @@ for (i = 0; i < 256; i++) *p++ = tolower(i); for (i = 0; i < 256; i++) *p++ = islower(i)? toupper(i) : tolower(i); -/* Then the character class tables. Don't try to be clever and save effort -on exclusive ones - in some locales things may be different. Note that the -table for "space" includes everything "isspace" gives, including VT in the -default locale. This makes it work for the POSIX class [:space:]. */ +/* Then the character class tables. Don't try to be clever and save effort on +exclusive ones - in some locales things may be different. Note that the table +for "space" includes everything "isspace" gives, including VT in the default +locale. This makes it work for the POSIX class [:space:]. Note also that it is +possible for a character to be alnum or alpha without being lower or upper, +such as "male and female ordinals" (\xAA and \xBA) in the fr_FR locale (at +least under Debian Linux's locales as of 12/2005). So we must test for alnum +specially. */ memset(p, 0, cbit_length); for (i = 0; i < 256; i++) { - if (isdigit(i)) - { - p[cbit_digit + i/8] |= 1 << (i&7); - p[cbit_word + i/8] |= 1 << (i&7); - } - if (isupper(i)) - { - p[cbit_upper + i/8] |= 1 << (i&7); - p[cbit_word + i/8] |= 1 << (i&7); - } - if (islower(i)) - { - p[cbit_lower + i/8] |= 1 << (i&7); - p[cbit_word + i/8] |= 1 << (i&7); - } + if (isdigit(i)) p[cbit_digit + i/8] |= 1 << (i&7); + if (isupper(i)) p[cbit_upper + i/8] |= 1 << (i&7); + if (islower(i)) p[cbit_lower + i/8] |= 1 << (i&7); + if (isalnum(i)) p[cbit_word + i/8] |= 1 << (i&7); if (i == '_') p[cbit_word + i/8] |= 1 << (i&7); if (isspace(i)) p[cbit_space + i/8] |= 1 << (i&7); if (isxdigit(i))p[cbit_xdigit + i/8] |= 1 << (i&7); @@ -139,7 +132,9 @@ for (i = 0; i < 256; i++) meta-character, which in this sense is any character that terminates a run of data characters. */ - if (strchr("*+?{^.$|()[", i) != 0) x += ctype_meta; *p++ = x; } + if (strchr("*+?{^.$|()[", i) != 0) x += ctype_meta; + *p++ = x; + } return yield; } diff --git a/src/src/pcre/pcre_printint.c b/src/src/pcre/pcre_printint.c index aef437380..913d294ca 100644 --- a/src/src/pcre/pcre_printint.c +++ b/src/src/pcre/pcre_printint.c @@ -1,453 +1,2 @@ -/* $Cambridge: exim/src/src/pcre/pcre_printint.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_printint.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - -/* PCRE is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. - - Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* This module contains an PCRE private debugging function for printing out the -internal form of a compiled regular expression, along with some supporting -local functions. */ - - -#include "pcre_internal.h" - - -static const char *OP_names[] = { OP_NAME_LIST }; - - -/************************************************* -* Print single- or multi-byte character * -*************************************************/ - -static int -print_char(FILE *f, uschar *ptr, BOOL utf8) -{ -int c = *ptr; - -if (!utf8 || (c & 0xc0) != 0xc0) - { - if (isprint(c)) fprintf(f, "%c", c); else fprintf(f, "\\x%02x", c); - return 0; - } -else - { - int i; - int a = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ - int s = 6*a; - c = (c & _pcre_utf8_table3[a]) << s; - for (i = 1; i <= a; i++) - { - /* This is a check for malformed UTF-8; it should only occur if the sanity - check has been turned off. Rather than swallow random bytes, just stop if - we hit a bad one. Print it with \X instead of \x as an indication. */ - - if ((ptr[i] & 0xc0) != 0x80) - { - fprintf(f, "\\X{%x}", c); - return i - 1; - } - - /* The byte is OK */ - - s -= 6; - c |= (ptr[i] & 0x3f) << s; - } - if (c < 128) fprintf(f, "\\x%02x", c); else fprintf(f, "\\x{%x}", c); - return a; - } -} - - - -/************************************************* -* Find Unicode property name * -*************************************************/ - -static const char * -get_ucpname(int property) -{ -#ifdef SUPPORT_UCP -int i; -for (i = _pcre_utt_size; i >= 0; i--) - { - if (property == _pcre_utt[i].value) break; - } -return (i >= 0)? _pcre_utt[i].name : "??"; -#else -return "??"; -#endif -} - - - -/************************************************* -* Print compiled regex * -*************************************************/ - -/* Make this function work for a regex with integers either byte order. -However, we assume that what we are passed is a compiled regex. */ - -EXPORT void -_pcre_printint(pcre *external_re, FILE *f) -{ -real_pcre *re = (real_pcre *)external_re; -uschar *codestart, *code; -BOOL utf8; - -unsigned int options = re->options; -int offset = re->name_table_offset; -int count = re->name_count; -int size = re->name_entry_size; - -if (re->magic_number != MAGIC_NUMBER) - { - offset = ((offset << 8) & 0xff00) | ((offset >> 8) & 0xff); - count = ((count << 8) & 0xff00) | ((count >> 8) & 0xff); - size = ((size << 8) & 0xff00) | ((size >> 8) & 0xff); - options = ((options << 24) & 0xff000000) | - ((options << 8) & 0x00ff0000) | - ((options >> 8) & 0x0000ff00) | - ((options >> 24) & 0x000000ff); - } - -code = codestart = (uschar *)re + offset + count * size; -utf8 = (options & PCRE_UTF8) != 0; - -for(;;) - { - uschar *ccode; - int c; - int extra = 0; - - fprintf(f, "%3d ", (int)(code - codestart)); - - if (*code >= OP_BRA) - { - if (*code - OP_BRA > EXTRACT_BASIC_MAX) - fprintf(f, "%3d Bra extra\n", GET(code, 1)); - else - fprintf(f, "%3d Bra %d\n", GET(code, 1), *code - OP_BRA); - code += _pcre_OP_lengths[OP_BRA]; - continue; - } - - switch(*code) - { - case OP_END: - fprintf(f, " %s\n", OP_names[*code]); - fprintf(f, "------------------------------------------------------------------\n"); - return; - - case OP_OPT: - fprintf(f, " %.2x %s", code[1], OP_names[*code]); - break; - - case OP_CHAR: - { - fprintf(f, " "); - do - { - code++; - code += 1 + print_char(f, code, utf8); - } - while (*code == OP_CHAR); - fprintf(f, "\n"); - continue; - } - break; - - case OP_CHARNC: - { - fprintf(f, " NC "); - do - { - code++; - code += 1 + print_char(f, code, utf8); - } - while (*code == OP_CHARNC); - fprintf(f, "\n"); - continue; - } - break; - - case OP_KETRMAX: - case OP_KETRMIN: - case OP_ALT: - case OP_KET: - case OP_ASSERT: - case OP_ASSERT_NOT: - case OP_ASSERTBACK: - case OP_ASSERTBACK_NOT: - case OP_ONCE: - case OP_COND: - case OP_REVERSE: - fprintf(f, "%3d %s", GET(code, 1), OP_names[*code]); - break; - - case OP_BRANUMBER: - printf("%3d %s", GET2(code, 1), OP_names[*code]); - break; - - case OP_CREF: - if (GET2(code, 1) == CREF_RECURSE) - fprintf(f, " Cond recurse"); - else - fprintf(f, "%3d %s", GET2(code,1), OP_names[*code]); - break; - - case OP_STAR: - case OP_MINSTAR: - case OP_PLUS: - case OP_MINPLUS: - case OP_QUERY: - case OP_MINQUERY: - case OP_TYPESTAR: - case OP_TYPEMINSTAR: - case OP_TYPEPLUS: - case OP_TYPEMINPLUS: - case OP_TYPEQUERY: - case OP_TYPEMINQUERY: - fprintf(f, " "); - if (*code >= OP_TYPESTAR) - { - fprintf(f, "%s", OP_names[code[1]]); - if (code[1] == OP_PROP || code[1] == OP_NOTPROP) - { - fprintf(f, " %s ", get_ucpname(code[2])); - extra = 1; - } - } - else extra = print_char(f, code+1, utf8); - fprintf(f, "%s", OP_names[*code]); - break; - - case OP_EXACT: - case OP_UPTO: - case OP_MINUPTO: - fprintf(f, " "); - extra = print_char(f, code+3, utf8); - fprintf(f, "{"); - if (*code != OP_EXACT) fprintf(f, ","); - fprintf(f, "%d}", GET2(code,1)); - if (*code == OP_MINUPTO) fprintf(f, "?"); - break; - - case OP_TYPEEXACT: - case OP_TYPEUPTO: - case OP_TYPEMINUPTO: - fprintf(f, " %s", OP_names[code[3]]); - if (code[3] == OP_PROP || code[3] == OP_NOTPROP) - { - fprintf(f, " %s ", get_ucpname(code[4])); - extra = 1; - } - fprintf(f, "{"); - if (*code != OP_TYPEEXACT) fprintf(f, "0,"); - fprintf(f, "%d}", GET2(code,1)); - if (*code == OP_TYPEMINUPTO) fprintf(f, "?"); - break; - - case OP_NOT: - if (isprint(c = code[1])) fprintf(f, " [^%c]", c); - else fprintf(f, " [^\\x%02x]", c); - break; - - case OP_NOTSTAR: - case OP_NOTMINSTAR: - case OP_NOTPLUS: - case OP_NOTMINPLUS: - case OP_NOTQUERY: - case OP_NOTMINQUERY: - if (isprint(c = code[1])) fprintf(f, " [^%c]", c); - else fprintf(f, " [^\\x%02x]", c); - fprintf(f, "%s", OP_names[*code]); - break; - - case OP_NOTEXACT: - case OP_NOTUPTO: - case OP_NOTMINUPTO: - if (isprint(c = code[3])) fprintf(f, " [^%c]{", c); - else fprintf(f, " [^\\x%02x]{", c); - if (*code != OP_NOTEXACT) fprintf(f, "0,"); - fprintf(f, "%d}", GET2(code,1)); - if (*code == OP_NOTMINUPTO) fprintf(f, "?"); - break; - - case OP_RECURSE: - fprintf(f, "%3d %s", GET(code, 1), OP_names[*code]); - break; - - case OP_REF: - fprintf(f, " \\%d", GET2(code,1)); - ccode = code + _pcre_OP_lengths[*code]; - goto CLASS_REF_REPEAT; - - case OP_CALLOUT: - fprintf(f, " %s %d %d %d", OP_names[*code], code[1], GET(code,2), - GET(code, 2 + LINK_SIZE)); - break; - - case OP_PROP: - case OP_NOTPROP: - fprintf(f, " %s %s", OP_names[*code], get_ucpname(code[1])); - break; - - /* OP_XCLASS can only occur in UTF-8 mode. However, there's no harm in - having this code always here, and it makes it less messy without all those - #ifdefs. */ - - case OP_CLASS: - case OP_NCLASS: - case OP_XCLASS: - { - int i, min, max; - BOOL printmap; - - fprintf(f, " ["); - - if (*code == OP_XCLASS) - { - extra = GET(code, 1); - ccode = code + LINK_SIZE + 1; - printmap = (*ccode & XCL_MAP) != 0; - if ((*ccode++ & XCL_NOT) != 0) fprintf(f, "^"); - } - else - { - printmap = TRUE; - ccode = code + 1; - } - - /* Print a bit map */ - - if (printmap) - { - for (i = 0; i < 256; i++) - { - if ((ccode[i/8] & (1 << (i&7))) != 0) - { - int j; - for (j = i+1; j < 256; j++) - if ((ccode[j/8] & (1 << (j&7))) == 0) break; - if (i == '-' || i == ']') fprintf(f, "\\"); - if (isprint(i)) fprintf(f, "%c", i); else fprintf(f, "\\x%02x", i); - if (--j > i) - { - if (j != i + 1) fprintf(f, "-"); - if (j == '-' || j == ']') fprintf(f, "\\"); - if (isprint(j)) fprintf(f, "%c", j); else fprintf(f, "\\x%02x", j); - } - i = j; - } - } - ccode += 32; - } - - /* For an XCLASS there is always some additional data */ - - if (*code == OP_XCLASS) - { - int ch; - while ((ch = *ccode++) != XCL_END) - { - if (ch == XCL_PROP) - { - fprintf(f, "\\p{%s}", get_ucpname(*ccode++)); - } - else if (ch == XCL_NOTPROP) - { - fprintf(f, "\\P{%s}", get_ucpname(*ccode++)); - } - else - { - ccode += 1 + print_char(f, ccode, TRUE); - if (ch == XCL_RANGE) - { - fprintf(f, "-"); - ccode += 1 + print_char(f, ccode, TRUE); - } - } - } - } - - /* Indicate a non-UTF8 class which was created by negation */ - - fprintf(f, "]%s", (*code == OP_NCLASS)? " (neg)" : ""); - - /* Handle repeats after a class or a back reference */ - - CLASS_REF_REPEAT: - switch(*ccode) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - case OP_CRPLUS: - case OP_CRMINPLUS: - case OP_CRQUERY: - case OP_CRMINQUERY: - fprintf(f, "%s", OP_names[*ccode]); - extra += _pcre_OP_lengths[*ccode]; - break; - - case OP_CRRANGE: - case OP_CRMINRANGE: - min = GET2(ccode,1); - max = GET2(ccode,3); - if (max == 0) fprintf(f, "{%d,}", min); - else fprintf(f, "{%d,%d}", min, max); - if (*ccode == OP_CRMINRANGE) fprintf(f, "?"); - extra += _pcre_OP_lengths[*ccode]; - break; - } - } - break; - - /* Anything else is just an item with no data*/ - - default: - fprintf(f, " %s", OP_names[*code]); - break; - } - - code += _pcre_OP_lengths[*code] + extra; - fprintf(f, "\n"); - } -} - -/* End of pcre_printint.c */ diff --git a/src/src/pcre/printint.c b/src/src/pcre/pcre_printint.src index ed3742f65..4144adfb2 100644 --- a/src/src/pcre/printint.c +++ b/src/src/pcre/pcre_printint.src @@ -1,17 +1,14 @@ -/* $Cambridge: exim/src/src/pcre/printint.c,v 1.2 2005/06/15 08:57:10 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_printint.src,v 1.1 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * *************************************************/ -/* -This is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. See -the file Tech.Notes for some information on the internals. +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. -Written by: Philip Hazel <ph10@cam.ac.uk> - - Copyright (c) 1997-2004 University of Cambridge + Written by Philip Hazel + Copyright (c) 1997-2005 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -43,10 +40,15 @@ POSSIBILITY OF SUCH DAMAGE. */ -/* This module contains a debugging function for printing out the internal form -of a compiled regular expression. It is kept in a separate file so that it can -be #included both in the pcretest program, and in the library itself when -compiled with the debugging switch. */ +/* This module contains a PCRE private debugging function for printing out the +internal form of a compiled regular expression, along with some supporting +local functions. This source file is used in two places: + +(1) It is #included by pcre_compile.c when it is compiled in debugging mode +(DEBUG defined in pcre_internal.h). It is not included in production compiles. + +(2) It is always #included by pcretest.c, which can be asked to print out a +compiled regex for debugging purposes. */ static const char *OP_names[] = { OP_NAME_LIST }; @@ -56,18 +58,6 @@ static const char *OP_names[] = { OP_NAME_LIST }; * Print single- or multi-byte character * *************************************************/ -/* These tables are actually copies of ones in pcre.c. If we compile the -library with debugging, they are included twice, but that isn't really a -problem - compiling with debugging is pretty rare and these are very small. */ - -static const int utf8_t3[] = { 0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01}; - -static const uschar utf8_t4[] = { - 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, - 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, - 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, - 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5 }; - static int print_char(FILE *f, uschar *ptr, BOOL utf8) { @@ -81,9 +71,9 @@ if (!utf8 || (c & 0xc0) != 0xc0) else { int i; - int a = utf8_t4[c & 0x3f]; /* Number of additional bytes */ + int a = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ int s = 6*a; - c = (c & utf8_t3[a]) << s; + c = (c & _pcre_utf8_table3[a]) << s; for (i = 1; i <= a; i++) { /* This is a check for malformed UTF-8; it should only occur if the sanity @@ -108,23 +98,24 @@ else - /************************************************* * Find Unicode property name * *************************************************/ static const char * -get_ucpname(int property) +get_ucpname(int ptype, int pvalue) { #ifdef SUPPORT_UCP int i; -for (i = sizeof(utt)/sizeof(ucp_type_table); i >= 0; i--) +for (i = _pcre_utt_size; i >= 0; i--) { - if (property == utt[i].value) break; + if (ptype == _pcre_utt[i].type && pvalue == _pcre_utt[i].value) break; } -return (i >= 0)? utt[i].name : "??"; +return (i >= 0)? _pcre_utt[i].name : "??"; #else -return "??"; +/* It gets harder and harder to shut off unwanted compiler warnings. */ +ptype = ptype * pvalue; +return (ptype == pvalue)? "??" : "??"; #endif } @@ -138,7 +129,7 @@ return "??"; However, we assume that what we are passed is a compiled regex. */ static void -print_internals(pcre *external_re, FILE *f) +pcre_printint(pcre *external_re, FILE *f) { real_pcre *re = (real_pcre *)external_re; uschar *codestart, *code; @@ -177,7 +168,7 @@ for(;;) fprintf(f, "%3d Bra extra\n", GET(code, 1)); else fprintf(f, "%3d Bra %d\n", GET(code, 1), *code - OP_BRA); - code += OP_lengths[OP_BRA]; + code += _pcre_OP_lengths[OP_BRA]; continue; } @@ -193,32 +184,26 @@ for(;;) break; case OP_CHAR: + fprintf(f, " "); + do { - fprintf(f, " "); - do - { - code++; - code += 1 + print_char(f, code, utf8); - } - while (*code == OP_CHAR); - fprintf(f, "\n"); - continue; + code++; + code += 1 + print_char(f, code, utf8); } - break; + while (*code == OP_CHAR); + fprintf(f, "\n"); + continue; case OP_CHARNC: + fprintf(f, " NC "); + do { - fprintf(f, " NC "); - do - { - code++; - code += 1 + print_char(f, code, utf8); - } - while (*code == OP_CHARNC); - fprintf(f, "\n"); - continue; + code++; + code += 1 + print_char(f, code, utf8); } - break; + while (*code == OP_CHARNC); + fprintf(f, "\n"); + continue; case OP_KETRMAX: case OP_KETRMIN: @@ -263,8 +248,8 @@ for(;;) fprintf(f, "%s", OP_names[code[1]]); if (code[1] == OP_PROP || code[1] == OP_NOTPROP) { - fprintf(f, " %s ", get_ucpname(code[2])); - extra = 1; + fprintf(f, " %s ", get_ucpname(code[2], code[3])); + extra = 2; } } else extra = print_char(f, code+1, utf8); @@ -288,8 +273,8 @@ for(;;) fprintf(f, " %s", OP_names[code[3]]); if (code[3] == OP_PROP || code[3] == OP_NOTPROP) { - fprintf(f, " %s ", get_ucpname(code[4])); - extra = 1; + fprintf(f, " %s ", get_ucpname(code[4], code[5])); + extra = 2; } fprintf(f, "{"); if (*code != OP_TYPEEXACT) fprintf(f, "0,"); @@ -318,7 +303,7 @@ for(;;) case OP_NOTMINUPTO: if (isprint(c = code[3])) fprintf(f, " [^%c]{", c); else fprintf(f, " [^\\x%02x]{", c); - if (*code != OP_NOTEXACT) fprintf(f, ","); + if (*code != OP_NOTEXACT) fprintf(f, "0,"); fprintf(f, "%d}", GET2(code,1)); if (*code == OP_NOTMINUPTO) fprintf(f, "?"); break; @@ -329,7 +314,7 @@ for(;;) case OP_REF: fprintf(f, " \\%d", GET2(code,1)); - ccode = code + OP_lengths[*code]; + ccode = code + _pcre_OP_lengths[*code]; goto CLASS_REF_REPEAT; case OP_CALLOUT: @@ -339,7 +324,7 @@ for(;;) case OP_PROP: case OP_NOTPROP: - fprintf(f, " %s %s", OP_names[*code], get_ucpname(code[1])); + fprintf(f, " %s %s", OP_names[*code], get_ucpname(code[1], code[2])); break; /* OP_XCLASS can only occur in UTF-8 mode. However, there's no harm in @@ -402,11 +387,15 @@ for(;;) { if (ch == XCL_PROP) { - fprintf(f, "\\p{%s}", get_ucpname(*ccode++)); + int ptype = *ccode++; + int pvalue = *ccode++; + fprintf(f, "\\p{%s}", get_ucpname(ptype, pvalue)); } else if (ch == XCL_NOTPROP) { - fprintf(f, "\\P{%s}", get_ucpname(*ccode++)); + int ptype = *ccode++; + int pvalue = *ccode++; + fprintf(f, "\\P{%s}", get_ucpname(ptype, pvalue)); } else { @@ -436,7 +425,7 @@ for(;;) case OP_CRQUERY: case OP_CRMINQUERY: fprintf(f, "%s", OP_names[*ccode]); - extra += OP_lengths[*ccode]; + extra += _pcre_OP_lengths[*ccode]; break; case OP_CRRANGE: @@ -446,7 +435,13 @@ for(;;) if (max == 0) fprintf(f, "{%d,}", min); else fprintf(f, "{%d,%d}", min, max); if (*ccode == OP_CRMINRANGE) fprintf(f, "?"); - extra += OP_lengths[*ccode]; + extra += _pcre_OP_lengths[*ccode]; + break; + + /* Do nothing if it's not a repeat; this code stops picky compilers + warning about the lack of a default code path. */ + + default: break; } } @@ -459,9 +454,9 @@ for(;;) break; } - code += OP_lengths[*code] + extra; + code += _pcre_OP_lengths[*code] + extra; fprintf(f, "\n"); } } -/* End of printint.c */ +/* End of pcre_printint.src */ diff --git a/src/src/pcre/pcre_study.c b/src/src/pcre/pcre_study.c index 18b5b680a..62990940a 100644 --- a/src/src/pcre/pcre_study.c +++ b/src/src/pcre/pcre_study.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_study.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_study.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -97,6 +97,13 @@ set_start_bits(const uschar *code, uschar *start_bits, BOOL caseless, { register int c; +#if 0 +/* ========================================================================= */ +/* The following comment and code was inserted in January 1999. In May 2006, +when it was observed to cause compiler warnings about unused values, I took it +out again. If anybody is still using OS/2, they will have to put it back +manually. */ + /* This next statement and the later reference to dummy are here in order to trick the optimizer of the IBM C compiler for OS/2 into generating correct code. Apparently IBM isn't going to fix the problem, and we would rather not @@ -104,6 +111,8 @@ disable optimization (in this module it actually makes a big difference, and the pcre module can use all the optimization it can get). */ volatile int dummy; +/* ========================================================================= */ +#endif do { @@ -161,7 +170,11 @@ do case OP_BRAMINZERO: if (!set_start_bits(++tcode, start_bits, caseless, utf8, cd)) return FALSE; +/* ========================================================================= + See the comment at the head of this function concerning the next line, + which was an old fudge for the benefit of OS/2. dummy = 1; + ========================================================================= */ do tcode += GET(tcode,1); while (*tcode == OP_ALT); tcode += 1+LINK_SIZE; break; @@ -217,15 +230,29 @@ do try_next = FALSE; break; + /* The cbit_space table has vertical tab as whitespace; we have to + discard it. */ + case OP_NOT_WHITESPACE: for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_space]; + { + int d = cd->cbits[c+cbit_space]; + if (c == 1) d &= ~0x08; + start_bits[c] |= ~d; + } try_next = FALSE; break; + /* The cbit_space table has vertical tab as whitespace; we have to + discard it. */ + case OP_WHITESPACE: for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_space]; + { + int d = cd->cbits[c+cbit_space]; + if (c == 1) d &= ~0x08; + start_bits[c] |= d; + } try_next = FALSE; break; @@ -279,14 +306,28 @@ do start_bits[c] |= cd->cbits[c+cbit_digit]; break; + /* The cbit_space table has vertical tab as whitespace; we have to + discard it. */ + case OP_NOT_WHITESPACE: for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_space]; + { + int d = cd->cbits[c+cbit_space]; + if (c == 1) d &= ~0x08; + start_bits[c] |= ~d; + } break; + /* The cbit_space table has vertical tab as whitespace; we have to + discard it. */ + case OP_WHITESPACE: for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_space]; + { + int d = cd->cbits[c+cbit_space]; + if (c == 1) d &= ~0x08; + start_bits[c] |= d; + } break; case OP_NOT_WORDCHAR: @@ -403,17 +444,16 @@ Returns: pointer to a pcre_extra block, with study_data filled in and the NULL on error or if no optimization possible */ -EXPORT pcre_extra * +PCRE_DATA_SCOPE pcre_extra * pcre_study(const pcre *external_re, int options, const char **errorptr) { uschar start_bits[32]; pcre_extra *extra; pcre_study_data *study; const uschar *tables; -const real_pcre *re = (const real_pcre *)external_re; -uschar *code = (uschar *)re + re->name_table_offset + - (re->name_count * re->name_entry_size); +uschar *code; compile_data compile_block; +const real_pcre *re = (const real_pcre *)external_re; *errorptr = NULL; @@ -429,6 +469,9 @@ if ((options & ~PUBLIC_STUDY_OPTIONS) != 0) return NULL; } +code = (uschar *)re + re->name_table_offset + + (re->name_count * re->name_entry_size); + /* For an anchored pattern, or an unanchored pattern that has a first char, or a multiline pattern that matches only at "line starts", no further processing at present. */ diff --git a/src/src/pcre/pcre_tables.c b/src/src/pcre/pcre_tables.c index e8120ccc5..0cb530a0b 100644 --- a/src/src/pcre/pcre_tables.c +++ b/src/src/pcre/pcre_tables.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_tables.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_tables.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -41,14 +41,16 @@ POSSIBILITY OF SUCH DAMAGE. /* This module contains some fixed tables that are used by more than one of the -PCRE code modules. */ +PCRE code modules. The tables are also #included by the pcretest program, which +uses macros to change their names from _pcre_xxx to xxxx, thereby avoiding name +clashes with the library. */ #include "pcre_internal.h" /* Table of sizes for the fixed-length opcodes. It's defined in a macro so that -the definition is next to the definition of the opcodes in internal.h. */ +the definition is next to the definition of the opcodes in pcre_internal.h. */ const uschar _pcre_OP_lengths[] = { OP_LENGTHS }; @@ -82,48 +84,110 @@ const uschar _pcre_utf8_table4[] = { 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5 }; -/* This table translates Unicode property names into code values for the -ucp_findchar() function. It is used by pcretest as well as by the library -functions. */ +/* This table translates Unicode property names into type and code values. It +is searched by binary chop, so must be in collating sequence of name. */ const ucp_type_table _pcre_utt[] = { - { "C", 128 + ucp_C }, - { "Cc", ucp_Cc }, - { "Cf", ucp_Cf }, - { "Cn", ucp_Cn }, - { "Co", ucp_Co }, - { "Cs", ucp_Cs }, - { "L", 128 + ucp_L }, - { "Ll", ucp_Ll }, - { "Lm", ucp_Lm }, - { "Lo", ucp_Lo }, - { "Lt", ucp_Lt }, - { "Lu", ucp_Lu }, - { "M", 128 + ucp_M }, - { "Mc", ucp_Mc }, - { "Me", ucp_Me }, - { "Mn", ucp_Mn }, - { "N", 128 + ucp_N }, - { "Nd", ucp_Nd }, - { "Nl", ucp_Nl }, - { "No", ucp_No }, - { "P", 128 + ucp_P }, - { "Pc", ucp_Pc }, - { "Pd", ucp_Pd }, - { "Pe", ucp_Pe }, - { "Pf", ucp_Pf }, - { "Pi", ucp_Pi }, - { "Po", ucp_Po }, - { "Ps", ucp_Ps }, - { "S", 128 + ucp_S }, - { "Sc", ucp_Sc }, - { "Sk", ucp_Sk }, - { "Sm", ucp_Sm }, - { "So", ucp_So }, - { "Z", 128 + ucp_Z }, - { "Zl", ucp_Zl }, - { "Zp", ucp_Zp }, - { "Zs", ucp_Zs } + { "Any", PT_ANY, 0 }, + { "Arabic", PT_SC, ucp_Arabic }, + { "Armenian", PT_SC, ucp_Armenian }, + { "Bengali", PT_SC, ucp_Bengali }, + { "Bopomofo", PT_SC, ucp_Bopomofo }, + { "Braille", PT_SC, ucp_Braille }, + { "Buginese", PT_SC, ucp_Buginese }, + { "Buhid", PT_SC, ucp_Buhid }, + { "C", PT_GC, ucp_C }, + { "Canadian_Aboriginal", PT_SC, ucp_Canadian_Aboriginal }, + { "Cc", PT_PC, ucp_Cc }, + { "Cf", PT_PC, ucp_Cf }, + { "Cherokee", PT_SC, ucp_Cherokee }, + { "Cn", PT_PC, ucp_Cn }, + { "Co", PT_PC, ucp_Co }, + { "Common", PT_SC, ucp_Common }, + { "Coptic", PT_SC, ucp_Coptic }, + { "Cs", PT_PC, ucp_Cs }, + { "Cypriot", PT_SC, ucp_Cypriot }, + { "Cyrillic", PT_SC, ucp_Cyrillic }, + { "Deseret", PT_SC, ucp_Deseret }, + { "Devanagari", PT_SC, ucp_Devanagari }, + { "Ethiopic", PT_SC, ucp_Ethiopic }, + { "Georgian", PT_SC, ucp_Georgian }, + { "Glagolitic", PT_SC, ucp_Glagolitic }, + { "Gothic", PT_SC, ucp_Gothic }, + { "Greek", PT_SC, ucp_Greek }, + { "Gujarati", PT_SC, ucp_Gujarati }, + { "Gurmukhi", PT_SC, ucp_Gurmukhi }, + { "Han", PT_SC, ucp_Han }, + { "Hangul", PT_SC, ucp_Hangul }, + { "Hanunoo", PT_SC, ucp_Hanunoo }, + { "Hebrew", PT_SC, ucp_Hebrew }, + { "Hiragana", PT_SC, ucp_Hiragana }, + { "Inherited", PT_SC, ucp_Inherited }, + { "Kannada", PT_SC, ucp_Kannada }, + { "Katakana", PT_SC, ucp_Katakana }, + { "Kharoshthi", PT_SC, ucp_Kharoshthi }, + { "Khmer", PT_SC, ucp_Khmer }, + { "L", PT_GC, ucp_L }, + { "L&", PT_LAMP, 0 }, + { "Lao", PT_SC, ucp_Lao }, + { "Latin", PT_SC, ucp_Latin }, + { "Limbu", PT_SC, ucp_Limbu }, + { "Linear_B", PT_SC, ucp_Linear_B }, + { "Ll", PT_PC, ucp_Ll }, + { "Lm", PT_PC, ucp_Lm }, + { "Lo", PT_PC, ucp_Lo }, + { "Lt", PT_PC, ucp_Lt }, + { "Lu", PT_PC, ucp_Lu }, + { "M", PT_GC, ucp_M }, + { "Malayalam", PT_SC, ucp_Malayalam }, + { "Mc", PT_PC, ucp_Mc }, + { "Me", PT_PC, ucp_Me }, + { "Mn", PT_PC, ucp_Mn }, + { "Mongolian", PT_SC, ucp_Mongolian }, + { "Myanmar", PT_SC, ucp_Myanmar }, + { "N", PT_GC, ucp_N }, + { "Nd", PT_PC, ucp_Nd }, + { "New_Tai_Lue", PT_SC, ucp_New_Tai_Lue }, + { "Nl", PT_PC, ucp_Nl }, + { "No", PT_PC, ucp_No }, + { "Ogham", PT_SC, ucp_Ogham }, + { "Old_Italic", PT_SC, ucp_Old_Italic }, + { "Old_Persian", PT_SC, ucp_Old_Persian }, + { "Oriya", PT_SC, ucp_Oriya }, + { "Osmanya", PT_SC, ucp_Osmanya }, + { "P", PT_GC, ucp_P }, + { "Pc", PT_PC, ucp_Pc }, + { "Pd", PT_PC, ucp_Pd }, + { "Pe", PT_PC, ucp_Pe }, + { "Pf", PT_PC, ucp_Pf }, + { "Pi", PT_PC, ucp_Pi }, + { "Po", PT_PC, ucp_Po }, + { "Ps", PT_PC, ucp_Ps }, + { "Runic", PT_SC, ucp_Runic }, + { "S", PT_GC, ucp_S }, + { "Sc", PT_PC, ucp_Sc }, + { "Shavian", PT_SC, ucp_Shavian }, + { "Sinhala", PT_SC, ucp_Sinhala }, + { "Sk", PT_PC, ucp_Sk }, + { "Sm", PT_PC, ucp_Sm }, + { "So", PT_PC, ucp_So }, + { "Syloti_Nagri", PT_SC, ucp_Syloti_Nagri }, + { "Syriac", PT_SC, ucp_Syriac }, + { "Tagalog", PT_SC, ucp_Tagalog }, + { "Tagbanwa", PT_SC, ucp_Tagbanwa }, + { "Tai_Le", PT_SC, ucp_Tai_Le }, + { "Tamil", PT_SC, ucp_Tamil }, + { "Telugu", PT_SC, ucp_Telugu }, + { "Thaana", PT_SC, ucp_Thaana }, + { "Thai", PT_SC, ucp_Thai }, + { "Tibetan", PT_SC, ucp_Tibetan }, + { "Tifinagh", PT_SC, ucp_Tifinagh }, + { "Ugaritic", PT_SC, ucp_Ugaritic }, + { "Yi", PT_SC, ucp_Yi }, + { "Z", PT_GC, ucp_Z }, + { "Zl", PT_PC, ucp_Zl }, + { "Zp", PT_PC, ucp_Zp }, + { "Zs", PT_PC, ucp_Zs } }; const int _pcre_utt_size = sizeof(_pcre_utt)/sizeof(ucp_type_table); diff --git a/src/src/pcre/pcre_try_flipped.c b/src/src/pcre/pcre_try_flipped.c index 3069d1fb0..6466f23e1 100644 --- a/src/src/pcre/pcre_try_flipped.c +++ b/src/src/pcre/pcre_try_flipped.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_try_flipped.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_try_flipped.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -64,8 +64,8 @@ Arguments: Returns: the flipped value */ -static long int -byteflip(long int value, int n) +static unsigned long int +byteflip(unsigned long int value, int n) { if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8); return ((value & 0x000000ff) << 24) | @@ -96,7 +96,7 @@ Returns: the new block if is is indeed a byte-flipped regex NULL if it is not */ -EXPORT real_pcre * +real_pcre * _pcre_try_flipped(const real_pcre *re, real_pcre *internal_re, const pcre_study_data *study, pcre_study_data *internal_study) { diff --git a/src/src/pcre/pcre_version.c b/src/src/pcre/pcre_version.c index 871d19fed..47f889f1d 100644 --- a/src/src/pcre/pcre_version.c +++ b/src/src/pcre/pcre_version.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcre_version.c,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcre_version.c,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * Perl-Compatible Regular Expressions * @@ -8,7 +8,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2005 University of Cambridge + Copyright (c) 1997-2006 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -54,10 +54,13 @@ string that identifies the PCRE version that is in use. */ #define STRING(a) # a #define XSTRING(s) STRING(s) -EXPORT const char * +PCRE_DATA_SCOPE const char * pcre_version(void) { -return XSTRING(PCRE_MAJOR) "." XSTRING(PCRE_MINOR) " " XSTRING(PCRE_DATE); +return XSTRING(PCRE_MAJOR) + "." XSTRING(PCRE_MINOR) + XSTRING(PCRE_PRERELEASE) + " " XSTRING(PCRE_DATE); } /* End of pcre_version.c */ diff --git a/src/src/pcre/pcretest.c b/src/src/pcre/pcretest.c index 17e646044..fda541a0c 100644 --- a/src/src/pcre/pcretest.c +++ b/src/src/pcre/pcretest.c @@ -1,4 +1,4 @@ -/* $Cambridge: exim/src/src/pcre/pcretest.c,v 1.3 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/pcretest.c,v 1.4 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* * PCRE testing program * @@ -46,15 +46,40 @@ POSSIBILITY OF SUCH DAMAGE. #include <locale.h> #include <errno.h> +#ifndef _WIN32 +#include <sys/resource.h> +#endif + #define PCRE_SPY /* For Win32 build, import data, not export */ -/* We need the internal info for displaying the results of pcre_study() and -other internal data; pcretest also uses some of the fixed tables, and generally -has "inside information" compared to a program that strictly follows the PCRE -API. */ +/* We include pcre_internal.h because we need the internal info for displaying +the results of pcre_study() and we also need to know about the internal +macros, structures, and other internal data values; pcretest has "inside +information" compared to a program that strictly follows the PCRE API. */ #include "pcre_internal.h" +/* We need access to the data tables that PCRE uses. So as not to have to keep +two copies, we include the source file here, changing the names of the external +symbols to prevent clashes. */ + +#define _pcre_utf8_table1 utf8_table1 +#define _pcre_utf8_table1_size utf8_table1_size +#define _pcre_utf8_table2 utf8_table2 +#define _pcre_utf8_table3 utf8_table3 +#define _pcre_utf8_table4 utf8_table4 +#define _pcre_utt utt +#define _pcre_utt_size utt_size +#define _pcre_OP_lengths OP_lengths + +#include "pcre_tables.c" + +/* We also need the pcre_printint() function for printing out compiled +patterns. This function is in a separate file so that it can be included in +pcre_compile.c when that module is compiled with debugging enabled. */ + +#include "pcre_printint.src" + /* It is possible to compile this test program without including support for testing the POSIX interface, though this is not available via the standard @@ -70,6 +95,8 @@ to the DFA matcher (NODFA), and without the doublecheck of the old "info" function (define NOINFOCHECK). */ +/* Other parameters */ + #ifndef CLOCKS_PER_SEC #ifdef CLK_TCK #define CLOCKS_PER_SEC CLK_TCK @@ -80,10 +107,7 @@ function (define NOINFOCHECK). */ #define LOOPREPEAT 500000 -#define BUFFER_SIZE 30000 -#define PBUFFER_SIZE BUFFER_SIZE -#define DBUFFER_SIZE BUFFER_SIZE - +/* Static variables */ static FILE *outfile; static int log_store = 0; @@ -96,11 +120,96 @@ static int show_malloc; static int use_utf8; static size_t gotten_store; +/* The buffers grow automatically if very long input lines are encountered. */ + +static int buffer_size = 50000; +static uschar *buffer = NULL; +static uschar *dbuffer = NULL; static uschar *pbuffer = NULL; /************************************************* +* Read or extend an input line * +*************************************************/ + +/* Input lines are read into buffer, but both patterns and data lines can be +continued over multiple input lines. In addition, if the buffer fills up, we +want to automatically expand it so as to be able to handle extremely large +lines that are needed for certain stress tests. When the input buffer is +expanded, the other two buffers must also be expanded likewise, and the +contents of pbuffer, which are a copy of the input for callouts, must be +preserved (for when expansion happens for a data line). This is not the most +optimal way of handling this, but hey, this is just a test program! + +Arguments: + f the file to read + start where in buffer to start (this *must* be within buffer) + +Returns: pointer to the start of new data + could be a copy of start, or could be moved + NULL if no data read and EOF reached +*/ + +static uschar * +extend_inputline(FILE *f, uschar *start) +{ +uschar *here = start; + +for (;;) + { + int rlen = buffer_size - (here - buffer); + if (rlen > 1000) + { + int dlen; + if (fgets((char *)here, rlen, f) == NULL) + return (here == start)? NULL : start; + dlen = (int)strlen((char *)here); + if (dlen > 0 && here[dlen - 1] == '\n') return start; + here += dlen; + } + + else + { + int new_buffer_size = 2*buffer_size; + uschar *new_buffer = (unsigned char *)malloc(new_buffer_size); + uschar *new_dbuffer = (unsigned char *)malloc(new_buffer_size); + uschar *new_pbuffer = (unsigned char *)malloc(new_buffer_size); + + if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL) + { + fprintf(stderr, "pcretest: malloc(%d) failed\n", new_buffer_size); + exit(1); + } + + memcpy(new_buffer, buffer, buffer_size); + memcpy(new_pbuffer, pbuffer, buffer_size); + + buffer_size = new_buffer_size; + + start = new_buffer + (start - buffer); + here = new_buffer + (here - buffer); + + free(buffer); + free(dbuffer); + free(pbuffer); + + buffer = new_buffer; + dbuffer = new_dbuffer; + pbuffer = new_pbuffer; + } + } + +return NULL; /* Control never gets here */ +} + + + + + + + +/************************************************* * Read number from string * *************************************************/ @@ -136,19 +245,19 @@ return(result); and returns the value of the character. Argument: - buffer a pointer to the byte vector - vptr a pointer to an int to receive the value + utf8bytes a pointer to the byte vector + vptr a pointer to an int to receive the value -Returns: > 0 => the number of bytes consumed - -6 to 0 => malformed UTF-8 character at offset = (-return) +Returns: > 0 => the number of bytes consumed + -6 to 0 => malformed UTF-8 character at offset = (-return) */ #if !defined NOUTF8 static int -utf82ord(unsigned char *buffer, int *vptr) +utf82ord(unsigned char *utf8bytes, int *vptr) { -int c = *buffer++; +int c = *utf8bytes++; int d = c; int i, j, s; @@ -164,11 +273,11 @@ if (i == 0 || i == 6) return 0; /* invalid UTF-8 */ /* i now has a value in the range 1-5 */ s = 6*i; -d = (c & _pcre_utf8_table3[i]) << s; +d = (c & utf8_table3[i]) << s; for (j = 0; j < i; j++) { - c = *buffer++; + c = *utf8bytes++; if ((c & 0xc0) != 0x80) return -(j+1); s -= 6; d |= (c & 0x3f) << s; @@ -176,8 +285,8 @@ for (j = 0; j < i; j++) /* Check that encoding was the correct unique one */ -for (j = 0; j < _pcre_utf8_table1_size; j++) - if (d <= _pcre_utf8_table1[j]) break; +for (j = 0; j < utf8_table1_size; j++) + if (d <= utf8_table1[j]) break; if (j != i) return -(i+1); /* Valid value */ @@ -191,6 +300,42 @@ return i+1; /************************************************* +* Convert character value to UTF-8 * +*************************************************/ + +/* This function takes an integer value in the range 0 - 0x7fffffff +and encodes it as a UTF-8 character in 0 to 6 bytes. + +Arguments: + cvalue the character value + utf8bytes pointer to buffer for result - at least 6 bytes long + +Returns: number of characters placed in the buffer +*/ + +#if !defined NOUTF8 + +static int +ord2utf8(int cvalue, uschar *utf8bytes) +{ +register int i, j; +for (i = 0; i < utf8_table1_size; i++) + if (cvalue <= utf8_table1[i]) break; +utf8bytes += i; +for (j = i; j > 0; j--) + { + *utf8bytes-- = 0x80 | (cvalue & 0x3f); + cvalue >>= 6; + } +*utf8bytes = utf8_table2[i] | cvalue; +return i + 1; +} + +#endif + + + +/************************************************* * Print character string * *************************************************/ @@ -200,7 +345,7 @@ chars without printing. */ static int pchars(unsigned char *p, int length, FILE *f) { -int c; +int c = 0; int yield = 0; while (length-- > 0) @@ -406,8 +551,8 @@ if ((rc = pcre_fullinfo(re, study, option, ptr)) < 0) * Byte flipping function * *************************************************/ -static long int -byteflip(long int value, int n) +static unsigned long int +byteflip(unsigned long int value, int n) { if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8); return ((value & 0x000000ff) << 24) | @@ -420,6 +565,83 @@ return ((value & 0x000000ff) << 24) | /************************************************* +* Check match or recursion limit * +*************************************************/ + +static int +check_match_limit(pcre *re, pcre_extra *extra, uschar *bptr, int len, + int start_offset, int options, int *use_offsets, int use_size_offsets, + int flag, unsigned long int *limit, int errnumber, const char *msg) +{ +int count; +int min = 0; +int mid = 64; +int max = -1; + +extra->flags |= flag; + +for (;;) + { + *limit = mid; + + count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options, + use_offsets, use_size_offsets); + + if (count == errnumber) + { + /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */ + min = mid; + mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2; + } + + else if (count >= 0 || count == PCRE_ERROR_NOMATCH || + count == PCRE_ERROR_PARTIAL) + { + if (mid == min + 1) + { + fprintf(outfile, "Minimum %s limit = %d\n", msg, mid); + break; + } + /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */ + max = mid; + mid = (min + mid)/2; + } + else break; /* Some other error */ + } + +extra->flags &= ~flag; +return count; +} + + + +/************************************************* +* Check newline indicator * +*************************************************/ + +/* This is used both at compile and run-time to check for <xxx> escapes, where +xxx is LF, CR, or CRLF. Print a message and return 0 if there is no match. + +Arguments: + p points after the leading '<' + f file for error message + +Returns: appropriate PCRE_NEWLINE_xxx flags, or 0 +*/ + +static int +check_newline(uschar *p, FILE *f) +{ +if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR; +if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF; +if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF; +fprintf(f, "Unknown newline type at: <%s\n", p); +return 0; +} + + + +/************************************************* * Main Program * *************************************************/ @@ -436,6 +658,7 @@ int op = 1; int timeit = 0; int showinfo = 0; int showstore = 0; +int quiet = 0; int size_offsets = 45; int size_offsets_max; int *offsets = NULL; @@ -446,16 +669,23 @@ int debug = 0; int done = 0; int all_use_dfa = 0; int yield = 0; +int stack_size; + +/* These vectors store, end-to-end, a list of captured substring names. Assume +that 1024 is plenty long enough for the few names we'll be testing. */ + +uschar copynames[1024]; +uschar getnames[1024]; -unsigned char *buffer; -unsigned char *dbuffer; +uschar *copynamesptr; +uschar *getnamesptr; /* Get buffers from malloc() so that Electric Fence will check their misuse -when I am debugging. */ +when I am debugging. They grow automatically when very long lines are read. */ -buffer = (unsigned char *)malloc(BUFFER_SIZE); -dbuffer = (unsigned char *)malloc(DBUFFER_SIZE); -pbuffer = (unsigned char *)malloc(PBUFFER_SIZE); +buffer = (unsigned char *)malloc(buffer_size); +dbuffer = (unsigned char *)malloc(buffer_size); +pbuffer = (unsigned char *)malloc(buffer_size); /* The outfile variable is static so that new_malloc can use it. The _setmode() stuff is some magic that I don't understand, but which apparently does good @@ -476,6 +706,7 @@ while (argc > 1 && argv[op][0] == '-') if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0) showstore = 1; else if (strcmp(argv[op], "-t") == 0) timeit = 1; + else if (strcmp(argv[op], "-q") == 0) quiet = 1; else if (strcmp(argv[op], "-i") == 0) showinfo = 1; else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1; #if !defined NODFA @@ -488,6 +719,28 @@ while (argc > 1 && argv[op][0] == '-') op++; argc--; } + else if (strcmp(argv[op], "-S") == 0 && argc > 2 && + ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)), + *endptr == 0)) + { +#ifdef _WIN32 + printf("PCRE: -S not supported on this OS\n"); + exit(1); +#else + int rc; + struct rlimit rlim; + getrlimit(RLIMIT_STACK, &rlim); + rlim.rlim_cur = stack_size * 1024 * 1024; + rc = setrlimit(RLIMIT_STACK, &rlim); + if (rc != 0) + { + printf("PCRE: setrlimit() failed with error %d\n", rc); + exit(1); + } + op++; + argc--; +#endif + } #if !defined NOPOSIX else if (strcmp(argv[op], "-p") == 0) posix = 1; #endif @@ -501,13 +754,16 @@ while (argc > 1 && argv[op][0] == '-') (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc); printf(" %sUnicode properties support\n", rc? "" : "No "); (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc); - printf(" Newline character is %s\n", (rc == '\r')? "CR" : "LF"); + printf(" Newline sequence is %s\n", (rc == '\r')? "CR" : + (rc == '\n')? "LF" : "CRLF"); (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc); printf(" Internal link size = %d\n", rc); (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc); printf(" POSIX malloc threshold = %d\n", rc); (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc); printf(" Default match limit = %d\n", rc); + (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc); + printf(" Default recursion depth limit = %d\n", rc); (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc); printf(" Match recursion uses %s\n", rc? "stack" : "heap"); exit(0); @@ -515,7 +771,7 @@ while (argc > 1 && argv[op][0] == '-') else { printf("** Unknown or malformed option %s\n", argv[op]); - printf("Usage: pcretest [-d] [-i] [-o <n>] [-p] [-s] [-t] [<input> [<output>]]\n"); + printf("Usage: pcretest [options] [<input> [<output>]]\n"); printf(" -C show PCRE compile-time options and exit\n"); printf(" -d debug: show compiled code; implies -i\n"); #if !defined NODFA @@ -527,6 +783,7 @@ while (argc > 1 && argv[op][0] == '-') #if !defined NOPOSIX printf(" -p use POSIX interface\n"); #endif + printf(" -S <n> set stack size to <n> megabytes\n"); printf(" -s output store (memory) used information\n" " -t time compilation and execution\n"); yield = 1; @@ -579,9 +836,9 @@ pcre_free = new_free; pcre_stack_malloc = stack_malloc; pcre_stack_free = stack_free; -/* Heading line, then prompt for first regex if stdin */ +/* Heading line unless quiet, then prompt for first regex if stdin */ -fprintf(outfile, "PCRE version %s\n\n", pcre_version()); +if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version()); /* Main loop */ @@ -613,7 +870,7 @@ while (!done) use_utf8 = 0; if (infile == stdin) printf(" re> "); - if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL) break; + if (extend_inputline(infile, buffer) == NULL) break; if (infile != stdin) fprintf(outfile, "%s", (char *)buffer); fflush(outfile); @@ -625,7 +882,7 @@ while (!done) if (*p == '<' && strchr((char *)(p+1), '<') == NULL) { - unsigned long int magic; + unsigned long int magic, get_options; uschar sbuf[8]; FILE *f; @@ -673,8 +930,8 @@ while (!done) /* Need to know if UTF-8 for printing data strings */ - new_info(re, NULL, PCRE_INFO_OPTIONS, &options); - use_utf8 = (options & PCRE_UTF8) != 0; + new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options); + use_utf8 = (get_options & PCRE_UTF8) != 0; /* Now see if there is any following study data */ @@ -728,16 +985,8 @@ while (!done) pp++; } if (*pp != 0) break; - - len = BUFFER_SIZE - (pp - buffer); - if (len < 256) - { - fprintf(outfile, "** Expression too long - missing delimiter?\n"); - goto SKIP_DATA; - } - if (infile == stdin) printf(" > "); - if (fgets((char *)pp, len, infile) == NULL) + if ((pp = extend_inputline(infile, pp)) == NULL) { fprintf(outfile, "** Unexpected EOF\n"); done = 1; @@ -783,6 +1032,7 @@ while (!done) case 'F': do_flip = 1; break; case 'G': do_G = 1; break; case 'I': do_showinfo = 1; break; + case 'J': options |= PCRE_DUPNAMES; break; case 'M': log_store = 1; break; case 'N': options |= PCRE_NO_AUTO_CAPTURE; break; @@ -817,6 +1067,15 @@ while (!done) *pp = 0; break; + case '<': + { + int x = check_newline(pp, outfile); + if (x == 0) goto SKIP_DATA; + options |= x; + while (*pp++ != '>'); + } + break; + case '\r': /* So that it works in Windows */ case '\n': case ' ': @@ -841,6 +1100,9 @@ while (!done) if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE; if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE; if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL; + if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB; + if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8; + rc = regcomp(&preg, (char *)p, cflags); /* Compilation failed; go back for another re, skipping to blank line @@ -848,7 +1110,7 @@ while (!done) if (rc != 0) { - (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE); + (void)regerror(rc, &preg, (char *)buffer, buffer_size); fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer); goto SKIP_DATA; } @@ -889,7 +1151,7 @@ while (!done) { for (;;) { - if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL) + if (extend_inputline(infile, buffer) == NULL) { done = 1; goto CONTINUE; @@ -990,7 +1252,7 @@ while (!done) if (do_debug) { fprintf(outfile, "------------------------------------------------------------------\n"); - _pcre_printint(re, outfile); + pcre_printint(re, outfile); } new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options); @@ -1050,13 +1312,13 @@ while (!done) if (do_flip) { all_options = byteflip(all_options, sizeof(all_options)); - } + } if ((all_options & PCRE_NOPARTIAL) != 0) fprintf(outfile, "Partial matching not supported\n"); if (get_options == 0) fprintf(outfile, "No options\n"); - else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s\n", + else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s\n", ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "", ((get_options & PCRE_CASELESS) != 0)? " caseless" : "", ((get_options & PCRE_EXTENDED) != 0)? " extended" : "", @@ -1066,15 +1328,32 @@ while (!done) ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "", ((get_options & PCRE_EXTRA) != 0)? " extra" : "", ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "", + ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "", ((get_options & PCRE_UTF8) != 0)? " utf8" : "", - ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : ""); + ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "", + ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : ""); + + switch (get_options & PCRE_NEWLINE_CRLF) + { + case PCRE_NEWLINE_CR: + fprintf(outfile, "Forced newline sequence: CR\n"); + break; + + case PCRE_NEWLINE_LF: + fprintf(outfile, "Forced newline sequence: LF\n"); + break; - if (((((real_pcre *)re)->options) & PCRE_ICHANGED) != 0) - fprintf(outfile, "Case state changes\n"); + case PCRE_NEWLINE_CRLF: + fprintf(outfile, "Forced newline sequence: CRLF\n"); + break; + + default: + break; + } if (first_char == -1) { - fprintf(outfile, "First char at start or follows \\n\n"); + fprintf(outfile, "First char at start or follows newline\n"); } else if (first_char < 0) { @@ -1211,8 +1490,8 @@ while (!done) for (;;) { - unsigned char *q; - unsigned char *bptr = dbuffer; + uschar *q; + uschar *bptr = dbuffer; int *use_offsets = offsets; int use_size_offsets = size_offsets; int callout_data = 0; @@ -1229,6 +1508,12 @@ while (!done) options = 0; + *copynames = 0; + *getnames = 0; + + copynamesptr = copynames; + getnamesptr = getnames; + pcre_callout = callout; first_callout = 1; callout_extra = 0; @@ -1237,15 +1522,24 @@ while (!done) callout_fail_id = -1; show_malloc = 0; - if (infile == stdin) printf("data> "); - if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL) + if (extra != NULL) extra->flags &= + ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION); + + len = 0; + for (;;) { - done = 1; - goto CONTINUE; + if (infile == stdin) printf("data> "); + if (extend_inputline(infile, buffer + len) == NULL) + { + if (len > 0) break; + done = 1; + goto CONTINUE; + } + if (infile != stdin) fprintf(outfile, "%s", (char *)buffer); + len = (int)strlen((char *)buffer); + if (buffer[len-1] == '\n') break; } - if (infile != stdin) fprintf(outfile, "%s", (char *)buffer); - len = (int)strlen((char *)buffer); while (len > 0 && isspace(buffer[len-1])) len--; buffer[len] = 0; if (len == 0) break; @@ -1275,6 +1569,17 @@ while (!done) c -= '0'; while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9') c = c * 8 + *p++ - '0'; + +#if !defined NOUTF8 + if (use_utf8 && c > 255) + { + unsigned char buff8[8]; + int ii, utn; + utn = ord2utf8(c, buff8); + for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii]; + c = buff8[ii]; /* Last byte */ + } +#endif break; case 'x': @@ -1292,7 +1597,7 @@ while (!done) { unsigned char buff8[8]; int ii, utn; - utn = _pcre_ord2utf8(c, buff8); + utn = ord2utf8(c, buff8); for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii]; c = buff8[ii]; /* Last byte */ p = pt + 1; @@ -1336,14 +1641,14 @@ while (!done) } else if (isalnum(*p)) { - uschar name[256]; - uschar *npp = name; + uschar *npp = copynamesptr; while (isalnum(*p)) *npp++ = *p++; + *npp++ = 0; *npp = 0; - n = pcre_get_stringnumber(re, (char *)name); + n = pcre_get_stringnumber(re, (char *)copynamesptr); if (n < 0) - fprintf(outfile, "no parentheses with name \"%s\"\n", name); - else copystrings |= 1 << n; + fprintf(outfile, "no parentheses with name \"%s\"\n", copynamesptr); + copynamesptr = npp; } else if (*p == '+') { @@ -1404,14 +1709,14 @@ while (!done) } else if (isalnum(*p)) { - uschar name[256]; - uschar *npp = name; + uschar *npp = getnamesptr; while (isalnum(*p)) *npp++ = *p++; + *npp++ = 0; *npp = 0; - n = pcre_get_stringnumber(re, (char *)name); + n = pcre_get_stringnumber(re, (char *)getnamesptr); if (n < 0) - fprintf(outfile, "no parentheses with name \"%s\"\n", name); - else getstrings |= 1 << n; + fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr); + getnamesptr = npp; } continue; @@ -1450,6 +1755,28 @@ while (!done) options |= PCRE_PARTIAL; continue; + case 'Q': + while(isdigit(*p)) n = n * 10 + *p++ - '0'; + if (extra == NULL) + { + extra = (pcre_extra *)malloc(sizeof(pcre_extra)); + extra->flags = 0; + } + extra->flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION; + extra->match_limit_recursion = n; + continue; + + case 'q': + while(isdigit(*p)) n = n * 10 + *p++ - '0'; + if (extra == NULL) + { + extra = (pcre_extra *)malloc(sizeof(pcre_extra)); + extra->flags = 0; + } + extra->flags |= PCRE_EXTRA_MATCH_LIMIT; + extra->match_limit = n; + continue; + #if !defined NODFA case 'R': options |= PCRE_DFA_RESTART; @@ -1467,6 +1794,15 @@ while (!done) case '?': options |= PCRE_NO_UTF8_CHECK; continue; + + case '<': + { + int x = check_newline(p, outfile); + if (x == 0) goto NEXT_DATA; + options |= x; + while (*p++ != '>'); + } + continue; } *q++ = c; } @@ -1497,9 +1833,14 @@ while (!done) if (rc != 0) { - (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE); + (void)regerror(rc, &preg, (char *)buffer, buffer_size); fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer); } + else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE) + != 0) + { + fprintf(outfile, "Matched with REG_NOSUB\n"); + } else { size_t i; @@ -1560,48 +1901,26 @@ while (!done) } /* If find_match_limit is set, we want to do repeated matches with - varying limits in order to find the minimum value. */ + varying limits in order to find the minimum value for the match limit and + for the recursion limit. */ if (find_match_limit) { - int min = 0; - int mid = 64; - int max = -1; - if (extra == NULL) { extra = (pcre_extra *)malloc(sizeof(pcre_extra)); extra->flags = 0; } - extra->flags |= PCRE_EXTRA_MATCH_LIMIT; - for (;;) - { - extra->match_limit = mid; - count = pcre_exec(re, extra, (char *)bptr, len, start_offset, - options | g_notempty, use_offsets, use_size_offsets); - if (count == PCRE_ERROR_MATCHLIMIT) - { - /* fprintf(outfile, "Testing match limit = %d\n", mid); */ - min = mid; - mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2; - } - else if (count >= 0 || count == PCRE_ERROR_NOMATCH || - count == PCRE_ERROR_PARTIAL) - { - if (mid == min + 1) - { - fprintf(outfile, "Minimum match limit = %d\n", mid); - break; - } - /* fprintf(outfile, "Testing match limit = %d\n", mid); */ - max = mid; - mid = (min + mid)/2; - } - else break; /* Some other error */ - } + (void)check_match_limit(re, extra, bptr, len, start_offset, + options|g_notempty, use_offsets, use_size_offsets, + PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit), + PCRE_ERROR_MATCHLIMIT, "match()"); - extra->flags &= ~PCRE_EXTRA_MATCH_LIMIT; + count = check_match_limit(re, extra, bptr, len, start_offset, + options|g_notempty, use_offsets, use_size_offsets, + PCRE_EXTRA_MATCH_LIMIT_RECURSION, &(extra->match_limit_recursion), + PCRE_ERROR_RECURSIONLIMIT, "match() recursion"); } /* If callout_data is set, use the interface with additional data */ @@ -1681,7 +2000,7 @@ while (!done) { if ((copystrings & (1 << i)) != 0) { - char copybuffer[16]; + char copybuffer[256]; int rc = pcre_copy_substring((char *)bptr, use_offsets, count, i, copybuffer, sizeof(copybuffer)); if (rc < 0) @@ -1691,6 +2010,19 @@ while (!done) } } + for (copynamesptr = copynames; + *copynamesptr != 0; + copynamesptr += (int)strlen((char*)copynamesptr) + 1) + { + char copybuffer[256]; + int rc = pcre_copy_named_substring(re, (char *)bptr, use_offsets, + count, (char *)copynamesptr, copybuffer, sizeof(copybuffer)); + if (rc < 0) + fprintf(outfile, "copy substring %s failed %d\n", copynamesptr, rc); + else + fprintf(outfile, " C %s (%d) %s\n", copybuffer, rc, copynamesptr); + } + for (i = 0; i < 32; i++) { if ((getstrings & (1 << i)) != 0) @@ -1703,12 +2035,27 @@ while (!done) else { fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc); - /* free((void *)substring); */ pcre_free_substring(substring); } } } + for (getnamesptr = getnames; + *getnamesptr != 0; + getnamesptr += (int)strlen((char*)getnamesptr) + 1) + { + const char *substring; + int rc = pcre_get_named_substring(re, (char *)bptr, use_offsets, + count, (char *)getnamesptr, &substring); + if (rc < 0) + fprintf(outfile, "copy substring %s failed %d\n", getnamesptr, rc); + else + { + fprintf(outfile, " G %s (%d) %s\n", substring, rc, getnamesptr); + pcre_free_substring(substring); + } + } + if (getlist) { const char **stringlist; @@ -1808,6 +2155,8 @@ while (!done) len -= use_offsets[1]; } } /* End of loop for /g and /G */ + + NEXT_DATA: continue; } /* End of loop for data lines */ CONTINUE: diff --git a/src/src/pcre/study.c b/src/src/pcre/study.c deleted file mode 100644 index 861d1a93b..000000000 --- a/src/src/pcre/study.c +++ /dev/null @@ -1,486 +0,0 @@ -/* $Cambridge: exim/src/src/pcre/study.c,v 1.2 2005/06/15 08:57:10 ph10 Exp $ */ - -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - -/* -This is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. See -the file Tech.Notes for some information on the internals. - -Written by: Philip Hazel <ph10@cam.ac.uk> - - Copyright (c) 1997-2004 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* Include the internals header, which itself includes Standard C headers plus -the external pcre header. */ - -#include "internal.h" - - - -/************************************************* -* Set a bit and maybe its alternate case * -*************************************************/ - -/* Given a character, set its bit in the table, and also the bit for the other -version of a letter if we are caseless. - -Arguments: - start_bits points to the bit map - c is the character - caseless the caseless flag - cd the block with char table pointers - -Returns: nothing -*/ - -static void -set_bit(uschar *start_bits, unsigned int c, BOOL caseless, compile_data *cd) -{ -start_bits[c/8] |= (1 << (c&7)); -if (caseless && (cd->ctypes[c] & ctype_letter) != 0) - start_bits[cd->fcc[c]/8] |= (1 << (cd->fcc[c]&7)); -} - - - -/************************************************* -* Create bitmap of starting chars * -*************************************************/ - -/* This function scans a compiled unanchored expression and attempts to build a -bitmap of the set of initial characters. If it can't, it returns FALSE. As time -goes by, we may be able to get more clever at doing this. - -Arguments: - code points to an expression - start_bits points to a 32-byte table, initialized to 0 - caseless the current state of the caseless flag - utf8 TRUE if in UTF-8 mode - cd the block with char table pointers - -Returns: TRUE if table built, FALSE otherwise -*/ - -static BOOL -set_start_bits(const uschar *code, uschar *start_bits, BOOL caseless, - BOOL utf8, compile_data *cd) -{ -register int c; - -/* This next statement and the later reference to dummy are here in order to -trick the optimizer of the IBM C compiler for OS/2 into generating correct -code. Apparently IBM isn't going to fix the problem, and we would rather not -disable optimization (in this module it actually makes a big difference, and -the pcre module can use all the optimization it can get). */ - -volatile int dummy; - -do - { - const uschar *tcode = code + 1 + LINK_SIZE; - BOOL try_next = TRUE; - - while (try_next) - { - /* If a branch starts with a bracket or a positive lookahead assertion, - recurse to set bits from within them. That's all for this branch. */ - - if ((int)*tcode >= OP_BRA || *tcode == OP_ASSERT) - { - if (!set_start_bits(tcode, start_bits, caseless, utf8, cd)) - return FALSE; - try_next = FALSE; - } - - else switch(*tcode) - { - default: - return FALSE; - - /* Skip over callout */ - - case OP_CALLOUT: - tcode += 2 + 2*LINK_SIZE; - break; - - /* Skip over extended extraction bracket number */ - - case OP_BRANUMBER: - tcode += 3; - break; - - /* Skip over lookbehind and negative lookahead assertions */ - - case OP_ASSERT_NOT: - case OP_ASSERTBACK: - case OP_ASSERTBACK_NOT: - do tcode += GET(tcode, 1); while (*tcode == OP_ALT); - tcode += 1+LINK_SIZE; - break; - - /* Skip over an option setting, changing the caseless flag */ - - case OP_OPT: - caseless = (tcode[1] & PCRE_CASELESS) != 0; - tcode += 2; - break; - - /* BRAZERO does the bracket, but carries on. */ - - case OP_BRAZERO: - case OP_BRAMINZERO: - if (!set_start_bits(++tcode, start_bits, caseless, utf8, cd)) - return FALSE; - dummy = 1; - do tcode += GET(tcode,1); while (*tcode == OP_ALT); - tcode += 1+LINK_SIZE; - break; - - /* Single-char * or ? sets the bit and tries the next item */ - - case OP_STAR: - case OP_MINSTAR: - case OP_QUERY: - case OP_MINQUERY: - set_bit(start_bits, tcode[1], caseless, cd); - tcode += 2; -#ifdef SUPPORT_UTF8 - if (utf8) while ((*tcode & 0xc0) == 0x80) tcode++; -#endif - break; - - /* Single-char upto sets the bit and tries the next */ - - case OP_UPTO: - case OP_MINUPTO: - set_bit(start_bits, tcode[3], caseless, cd); - tcode += 4; -#ifdef SUPPORT_UTF8 - if (utf8) while ((*tcode & 0xc0) == 0x80) tcode++; -#endif - break; - - /* At least one single char sets the bit and stops */ - - case OP_EXACT: /* Fall through */ - tcode += 2; - - case OP_CHAR: - case OP_CHARNC: - case OP_PLUS: - case OP_MINPLUS: - set_bit(start_bits, tcode[1], caseless, cd); - try_next = FALSE; - break; - - /* Single character type sets the bits and stops */ - - case OP_NOT_DIGIT: - for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_digit]; - try_next = FALSE; - break; - - case OP_DIGIT: - for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_digit]; - try_next = FALSE; - break; - - case OP_NOT_WHITESPACE: - for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_space]; - try_next = FALSE; - break; - - case OP_WHITESPACE: - for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_space]; - try_next = FALSE; - break; - - case OP_NOT_WORDCHAR: - for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_word]; - try_next = FALSE; - break; - - case OP_WORDCHAR: - for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_word]; - try_next = FALSE; - break; - - /* One or more character type fudges the pointer and restarts, knowing - it will hit a single character type and stop there. */ - - case OP_TYPEPLUS: - case OP_TYPEMINPLUS: - tcode++; - break; - - case OP_TYPEEXACT: - tcode += 3; - break; - - /* Zero or more repeats of character types set the bits and then - try again. */ - - case OP_TYPEUPTO: - case OP_TYPEMINUPTO: - tcode += 2; /* Fall through */ - - case OP_TYPESTAR: - case OP_TYPEMINSTAR: - case OP_TYPEQUERY: - case OP_TYPEMINQUERY: - switch(tcode[1]) - { - case OP_ANY: - return FALSE; - - case OP_NOT_DIGIT: - for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_digit]; - break; - - case OP_DIGIT: - for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_digit]; - break; - - case OP_NOT_WHITESPACE: - for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_space]; - break; - - case OP_WHITESPACE: - for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_space]; - break; - - case OP_NOT_WORDCHAR: - for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_word]; - break; - - case OP_WORDCHAR: - for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_word]; - break; - } - - tcode += 2; - break; - - /* Character class where all the information is in a bit map: set the - bits and either carry on or not, according to the repeat count. If it was - a negative class, and we are operating with UTF-8 characters, any byte - with a value >= 0xc4 is a potentially valid starter because it starts a - character with a value > 255. */ - - case OP_NCLASS: - if (utf8) - { - start_bits[24] |= 0xf0; /* Bits for 0xc4 - 0xc8 */ - memset(start_bits+25, 0xff, 7); /* Bits for 0xc9 - 0xff */ - } - /* Fall through */ - - case OP_CLASS: - { - tcode++; - - /* In UTF-8 mode, the bits in a bit map correspond to character - values, not to byte values. However, the bit map we are constructing is - for byte values. So we have to do a conversion for characters whose - value is > 127. In fact, there are only two possible starting bytes for - characters in the range 128 - 255. */ - - if (utf8) - { - for (c = 0; c < 16; c++) start_bits[c] |= tcode[c]; - for (c = 128; c < 256; c++) - { - if ((tcode[c/8] && (1 << (c&7))) != 0) - { - int d = (c >> 6) | 0xc0; /* Set bit for this starter */ - start_bits[d/8] |= (1 << (d&7)); /* and then skip on to the */ - c = (c & 0xc0) + 0x40 - 1; /* next relevant character. */ - } - } - } - - /* In non-UTF-8 mode, the two bit maps are completely compatible. */ - - else - { - for (c = 0; c < 32; c++) start_bits[c] |= tcode[c]; - } - - /* Advance past the bit map, and act on what follows */ - - tcode += 32; - switch (*tcode) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - case OP_CRQUERY: - case OP_CRMINQUERY: - tcode++; - break; - - case OP_CRRANGE: - case OP_CRMINRANGE: - if (((tcode[1] << 8) + tcode[2]) == 0) tcode += 5; - else try_next = FALSE; - break; - - default: - try_next = FALSE; - break; - } - } - break; /* End of bitmap class handling */ - - } /* End of switch */ - } /* End of try_next loop */ - - code += GET(code, 1); /* Advance to next branch */ - } -while (*code == OP_ALT); -return TRUE; -} - - - -/************************************************* -* Study a compiled expression * -*************************************************/ - -/* This function is handed a compiled expression that it must study to produce -information that will speed up the matching. It returns a pcre_extra block -which then gets handed back to pcre_exec(). - -Arguments: - re points to the compiled expression - options contains option bits - errorptr points to where to place error messages; - set NULL unless error - -Returns: pointer to a pcre_extra block, with study_data filled in and the - appropriate flag set; - NULL on error or if no optimization possible -*/ - -EXPORT pcre_extra * -pcre_study(const pcre *external_re, int options, const char **errorptr) -{ -uschar start_bits[32]; -pcre_extra *extra; -pcre_study_data *study; -const uschar *tables; -const real_pcre *re = (const real_pcre *)external_re; -uschar *code = (uschar *)re + re->name_table_offset + - (re->name_count * re->name_entry_size); -compile_data compile_block; - -*errorptr = NULL; - -if (re == NULL || re->magic_number != MAGIC_NUMBER) - { - *errorptr = "argument is not a compiled regular expression"; - return NULL; - } - -if ((options & ~PUBLIC_STUDY_OPTIONS) != 0) - { - *errorptr = "unknown or incorrect option bit(s) set"; - return NULL; - } - -/* For an anchored pattern, or an unanchored pattern that has a first char, or -a multiline pattern that matches only at "line starts", no further processing -at present. */ - -if ((re->options & (PCRE_ANCHORED|PCRE_FIRSTSET|PCRE_STARTLINE)) != 0) - return NULL; - -/* Set the character tables in the block that is passed around */ - -tables = re->tables; -if (tables == NULL) - (void)pcre_fullinfo(external_re, NULL, PCRE_INFO_DEFAULT_TABLES, &tables); - -compile_block.lcc = tables + lcc_offset; -compile_block.fcc = tables + fcc_offset; -compile_block.cbits = tables + cbits_offset; -compile_block.ctypes = tables + ctypes_offset; - -/* See if we can find a fixed set of initial characters for the pattern. */ - -memset(start_bits, 0, 32 * sizeof(uschar)); -if (!set_start_bits(code, start_bits, (re->options & PCRE_CASELESS) != 0, - (re->options & PCRE_UTF8) != 0, &compile_block)) return NULL; - -/* Get a pcre_extra block and a pcre_study_data block. The study data is put in -the latter, which is pointed to by the former, which may also get additional -data set later by the calling program. At the moment, the size of -pcre_study_data is fixed. We nevertheless save it in a field for returning via -the pcre_fullinfo() function so that if it becomes variable in the future, we -don't have to change that code. */ - -extra = (pcre_extra *)(pcre_malloc) - (sizeof(pcre_extra) + sizeof(pcre_study_data)); - -if (extra == NULL) - { - *errorptr = "failed to get memory"; - return NULL; - } - -study = (pcre_study_data *)((char *)extra + sizeof(pcre_extra)); -extra->flags = PCRE_EXTRA_STUDY_DATA; -extra->study_data = study; - -study->size = sizeof(pcre_study_data); -study->options = PCRE_STUDY_MAPPED; -memcpy(study->start_bits, start_bits, sizeof(start_bits)); - -return extra; -} - -/* End of study.c */ diff --git a/src/src/pcre/ucp.h b/src/src/pcre/ucp.h index fab202438..20fcff52c 100644 --- a/src/src/pcre/ucp.h +++ b/src/src/pcre/ucp.h @@ -1,14 +1,16 @@ -/* $Cambridge: exim/src/src/pcre/ucp.h,v 1.2 2005/08/08 10:22:14 ph10 Exp $ */ +/* $Cambridge: exim/src/src/pcre/ucp.h,v 1.3 2006/11/07 16:50:36 ph10 Exp $ */ /************************************************* -* libucp - Unicode Property Table handler * +* Unicode Property Table handler * *************************************************/ - #ifndef _UCP_H #define _UCP_H -/* These are the character categories that are returned by ucp_findchar */ +/* This file contains definitions of the property values that are returned by +the function _pcre_ucp_findprop(). */ + +/* These are the general character categories. */ enum { ucp_C, /* Other */ @@ -20,7 +22,7 @@ enum { ucp_Z /* Separator */ }; -/* These are the detailed character types that are returned by ucp_findchar */ +/* These are the particular character types. */ enum { ucp_Cc, /* Control */ @@ -55,7 +57,71 @@ enum { ucp_Zs /* Space separator */ }; -extern int ucp_findchar(const int, int *, int *); +/* These are the script identifications. */ + +enum { + ucp_Arabic, + ucp_Armenian, + ucp_Bengali, + ucp_Bopomofo, + ucp_Braille, + ucp_Buginese, + ucp_Buhid, + ucp_Canadian_Aboriginal, + ucp_Cherokee, + ucp_Common, + ucp_Coptic, + ucp_Cypriot, + ucp_Cyrillic, + ucp_Deseret, + ucp_Devanagari, + ucp_Ethiopic, + ucp_Georgian, + ucp_Glagolitic, + ucp_Gothic, + ucp_Greek, + ucp_Gujarati, + ucp_Gurmukhi, + ucp_Han, + ucp_Hangul, + ucp_Hanunoo, + ucp_Hebrew, + ucp_Hiragana, + ucp_Inherited, + ucp_Kannada, + ucp_Katakana, + ucp_Kharoshthi, + ucp_Khmer, + ucp_Lao, + ucp_Latin, + ucp_Limbu, + ucp_Linear_B, + ucp_Malayalam, + ucp_Mongolian, + ucp_Myanmar, + ucp_New_Tai_Lue, + ucp_Ogham, + ucp_Old_Italic, + ucp_Old_Persian, + ucp_Oriya, + ucp_Osmanya, + ucp_Runic, + ucp_Shavian, + ucp_Sinhala, + ucp_Syloti_Nagri, + ucp_Syriac, + ucp_Tagalog, + ucp_Tagbanwa, + ucp_Tai_Le, + ucp_Tamil, + ucp_Telugu, + ucp_Thaana, + ucp_Thai, + ucp_Tibetan, + ucp_Tifinagh, + ucp_Ugaritic, + ucp_Yi +}; #endif |