| Age | Commit message (Collapse) | Author |
|---|
|
|
|
When !time <somenick> was being used and the user <somenick> hadn't set
his or her location, the bot would wrongly assume the argument was some
timezone.
Fix by checking against local nicks first.
|
|
Sadly, the new ipinfodb API requires an API key, and I don't know yet if
I'm allowed to redistributed it (I'm betting on no, but in case I can
it will added with an appropriate patch).
|
|
|
|
If someone joins after a challengeable W+4, he would be inserted between
the last and first player, and if there is a challenge _his_ cards would
be exposed. So prevent joining in such a situation.
|
|
After a successful challenge, the challenged player can choose to
pick&pass instead of playing one of its allowed cards. In such a case,
the must_play array would not get reset, preventing the next player to
be able to play any of its card.
Fix by resetting must_play on next_turn rather than on set_discard.
|
|
Exit as soon as we find a card that reveals that the W+4 was not a valid
move.
|
|
Since commit a993d1c358e24ab85621568c10411c5496e2dea8, the message
mappers have a more rational way of reporting failures, but the auth
module was not using it, resulting in the wrong permissions being
granted/revoked whenever allow/deny was being used.
|
|
|
|
Default mIRC colors are 0-15, and some clients don't react too well to a
color > 15, while most wrap or default. Konversation, for example, will
stop the color command parsing and interpret it as color 1, followed by
the literal digit 6.
So use the 'official' (mIRC) value of 0 for white.
|
|
|
|
|
|
Some gettext versions freeze the string they return, so dup them before
modifying them.
|
|
We allow rss handles to be of any case on creation, even though matching
is case-insentivie. However, when renaming an rss using 'rss change
handle' and only changing the case, two things prevented this from
working correctly:
* since the new downcased handle was equal to the old downcased handle,
the bot would prevent the renaming due to the existence of the new
handle
* the new handle was forcefully downcased, preventing the user
from renaming handle 'case' to 'CaSe'.
Fix by checking for this case explicitly, and handling it separately.
|
|
if 'in' was present in the message but not part of the time
specification (e.g. because the specification was with 'every' or with
'at'), the parser would fail at the first attempt (look for time after
'in') and fail unnecessarily.
Fix by trying all possibilities and only giving an error if none
succeeds.
|
|
The first_run check comparing the time delta from last success failed
when feed.last_success was nil. Prevent this from happening by checking
if we are on the first run before the rest of the checks.
|
|
|
|
Commit a306b2c68eb58c1c88057ff7ee78895401adb6a9 introduced a fix to
prevent a player from playing something like g9r9, but the fix actually
prevented any single-card play. Fix by making the \1 optional.
|
|
|
|
The regexp should make sure that the second card is equal to the first
specified card, instead of matching the card expression one or two
times. Otherwise, a play like 'pl g9r9' would be accepted when the
player had two r9 cards and could play them.
This is only a aesthetical change because the plays would still be
valid, albeit (1) surprising and (2) mismatching what the user actually
wrote.
|
|
|
|
|
|
|
|
Google changed their HTML again, breaking gcalc. Fix by using the WAP
search and get the first result if it contains an equal sign.
|
|
Some plugins (ri, fortune) wrap safe_exec in begin/rescue blocks that
did not get triggered because the exceptions from the execution were
caught in the popen block (and would never get exposed anyway). Fix this
by raising if $? does not indicate success.
|
|
The Ruby/DICT library is so tightly coupled with the rdict client that
its errors exit on initialize, instead of letting the lib user do that.
This makes those errors untrappable from the code.
Fix by monkeypatching DICTError to just act like the other exceptions in
Ruby.
|
|
|
|
Convert the (optional) first argument to #each*() methods to a string to
ensure the calls work as expected. This fixes a problem with
@registry.each not working as expected, and although it would have been
sufficient to convert the default for the first argument from nil to the
empty string, the approach implemented by this patch is more robust
overall.
|
|
|
|
Don't trap them, rerise and let the bot handle them
|
|
|
|
|
|
When timing out during a wait, return false instead of raising. This
fixes a problem with TimeoutError being raised during restarts on
ruby 1.9.x
|
|
This reverts commit 06485aeb187dde5e81204b06c8e956e7e035c323. The mutex
is necessary because of the concurrent learning and database conversion.
Some other approach should be used instead (e.g. skipping the mutex if
not running a conversion).
|
|
Having Server#to_s alias the hostname would cause problems when the
hostname was nil (e.g. after a disconnect). Fix by making sure that to_s
always returns a string instead.
|
|
|
|
|
|
When using a DB backend different from BDB, the BDB constant may not be
defined, causing a NameError during error trapping in the main loop. Fix
this by defining our own DBFatal error that maps to BDB::Fatal in the
BDB case and is defined as an (unused) Exception for TokyoCabinet.
|
|
|
|
|
|
|
|
|
|
|
|
When upgrading from 0.9.5 data, write the 0.9.9-style regisitry using
BDB still rather than trying to use TokyoCabinet already. This makes
sure that the next step (0.9.9 to modern times) works correctly.
(Moreover, the existing BDB-to-TC attempt wouldn't have worked anyway
due to wrong var names. I doubt anybody will ever hit this path and
notice though.)
|
|
|
|
|
|
|
|
|
|
|
|
|
