path: root/src/modules/m_nationalchars.cpp

/*
 * InspIRCd -- Internet Relay Chat Daemon
 *
 *   Copyright (C) 2009 Daniel De Graaf <danieldg@inspircd.org>
 *   Copyright (C) 2009 Dennis Friis <peavey@inspircd.org>
 *   Copyright (C) 2009 Craig Edwards <craigedwards@brainbox.cc>
 *   Copyright (C) 2009 Robin Burchell <robin+git@viroteck.net>
 *
 * This file is part of InspIRCd.  InspIRCd is free software: you can
 * redistribute it and/or modify it under the terms of the GNU General Public
 * License as published by the Free Software Foundation, version 2.
 *
 * This program is distributed in the hope that it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 * FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more
 * details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program.  If not, see <http://www.gnu.org/licenses/>.
 */


/* Contains a code of Unreal IRCd + Bynets patch ( http://www.unrealircd.com/ and http://www.bynets.org/ )
   Original patch is made by Dmitry "Killer{R}" Kononko. ( http://killprog.com/ )
   Changed at 2008-06-15 - 2009-02-11
   by Chernov-Phoenix Alexey (Phoenix@RusNet) mailto:phoenix /email address separator/ pravmail.ru */

#include "inspircd.h"
#include "caller.h"
#include <fstream>

/* $ModDesc: Provides an ability to have non-RFC1459 nicks & support for national CASEMAPPING */

class lwbNickHandler : public HandlerBase2<bool, const char*, size_t>
{
 public:
	lwbNickHandler() { }
	virtual ~lwbNickHandler() { }
	virtual bool Call(const char*, size_t);
};

								 /*,m_reverse_additionalUp[256];*/
static unsigned char m_reverse_additional[256],m_additionalMB[256],m_additionalUtf8[256],m_additionalUtf8range[256],m_additionalUtf8interval[256];

char utf8checkrest(unsigned char * mb, unsigned char cnt)
{
	for (unsigned char * tmp=mb; tmp<mb+cnt; tmp++)
	{
		/* & is faster! -- Phoenix (char & b11000000 == b10000000) */
		if ((*tmp & 192) != 128)
			return -1;
	}
	return cnt + 1;
}


char utf8size(unsigned char * mb)
{
	if (!*mb)
		return -1;
	if (!(*mb & 128))
		return 1;
	if ((*mb & 224) == 192)
		return utf8checkrest(mb + 1,1);
	if ((*mb & 240) == 224)
		return utf8checkrest(mb + 1,2);
	if ((*mb & 248) == 240)
		return utf8checkrest(mb + 1,3);
	return -1;
}


/* Conditions added */
bool lwbNickHandler::Call(const char* n, size_t max)
{
	if (!n || !*n)
		return false;

	unsigned int p = 0;
	for (const char* i = n; *i; i++, p++)
	{
		/* 1. Multibyte encodings support:  */
		/* 1.1. 16bit char. areas, e.g. chinese:*/

		/* if current character is the last, we DO NOT check it against multibyte table */
		/* if there are mbtable ranges, use ONLY them. No 8bit at all */
		if (i[1] && m_additionalMB[0])
		{
			/* otherwise let's take a look at the current character and the following one */
			bool found = false;
			for(unsigned char * mb = m_additionalMB; (*mb) && (mb < m_additionalMB + sizeof(m_additionalMB)); mb += 4)
			{
				if ( (i[0] >= mb[0]) && (i[0] <= mb[1]) && (i[1] >= mb[2]) && (i[1] <= mb[3]) )
				{
					/* multibyte range character found */
					i++;
					p++;
					found = true;
					break;
				}
			}
			if (found)
				/* next char! */
				continue;
			else
				/* there are ranges, but incorrect char (8bit?) given, sorry */
				return false;
		}

		/* 2. 8bit character support */
		if (((*i >= 'A') && (*i <= '}')) || m_reverse_additional[(unsigned char)*i])
			/* "A"-"}" can occur anywhere in a nickname */
			continue;

		if ((((*i >= '0') && (*i <= '9')) || (*i == '-')) && (i > n))
			/* "0"-"9", "-" can occur anywhere BUT the first char of a nickname */
			continue;

		/* 3.1. Check against a simple UTF-8 characters enumeration */
		int cursize, cursize2, ncursize = utf8size((unsigned char *)i);
		/* do check only if current multibyte character is valid UTF-8 only */
		if (ncursize != -1)
		{
			bool found = false;
			for (unsigned char * mb = m_additionalUtf8; (utf8size(mb) != -1) && (mb < m_additionalUtf8 + sizeof(m_additionalUtf8)); mb += cursize)
			{
				cursize = utf8size(mb);
				/* Size differs? Pick the next! */
				if (cursize != ncursize)
					continue;

				if (!strncmp(i, (char *)mb, cursize))
				{
					i += cursize - 1;
					p += cursize - 1;
					found = true;
					break;
				}
			}
			if (found)
				continue;

			/* 3.2. Check against an UTF-8 ranges: <start character> and <length of the range>. */
			found = false;
			for (unsigned char * mb = m_additionalUtf8range; (utf8size(mb) != -1) && (mb < m_additionalUtf8range + sizeof(m_additionalUtf8range)); mb += cursize + 1)
			{
				cursize = utf8size(mb);
				/* Size differs (or lengthbyte is zero)? Pick the next! */
				if ((cursize != ncursize) || (!mb[cursize]))
					continue;

				unsigned char uright[5] = {0,0,0,0,0}, range = mb[cursize] - 1;
				strncpy((char* ) uright, (char *) mb, cursize);

				for (int temp = cursize - 1; (temp >= 0) && range; --temp)
				{
					/* all but the first char are 64-based */
					if (temp)
					{
						char part64 = range & 63; /* i.e. % 64 */
						/* handle carrying over */
						if (uright[temp] + part64 - 1 > 191)
						{
							uright[temp] -= 64;
							range += 64;
						}
						uright[temp] += part64;
						range >>= 6; /* divide it on a 64 */
					}
					/* the first char of UTF-8 doesn't follow the rule */
					else
					{
						uright[temp] += range;
					}
				}

				if ((strncmp(i, (char *) mb, cursize) >= 0) && (strncmp(i, (char *) uright, cursize) <= 0))
				{
					i += cursize - 1;
					p += cursize - 1;
					found = true;
					break;
				}
			}
			if (found)
				continue;

			/* 3.3. Check against an UTF-8 intervals: <start character> and <end character>. */
			found = false;
			for (unsigned char * mb = m_additionalUtf8interval; (utf8size(mb) != -1) && (utf8size(mb+utf8size(mb)) != -1)
				&& (mb < m_additionalUtf8interval + sizeof(m_additionalUtf8interval)); mb += (cursize+cursize2) )
			{
				cursize = utf8size(mb);
				cursize2= utf8size(mb+cursize);

				int minlen  = cursize  > ncursize ? ncursize : cursize;
				int minlen2 = cursize2 > ncursize ? ncursize : cursize2;

				unsigned char* uright = mb + cursize;

				if ((strncmp(i, (char *) mb, minlen) >= 0) && (strncmp(i, (char *) uright, minlen2) <= 0))
				{
					i += cursize - 1;
					p += cursize - 1;
					found = true;
					break;
				}
			}
			if (found)
				continue;
		}

		/* invalid character! abort */
		return false;
	}

	/* too long? or not -- pointer arithmetic rocks */
	return (p < max);
}


class ModuleNationalChars : public Module
{
 private:
	lwbNickHandler myhandler;
	std::string charset, casemapping;
	unsigned char m_additional[256], m_additionalUp[256], m_lower[256], m_upper[256];
	caller2<bool, const char*, size_t> rememberer;
	bool forcequit;
	const unsigned char * lowermap_rememberer;
	unsigned char prev_map[256];

	void CheckRehash()
	{
		// See if anything changed
		if (!memcmp(prev_map, national_case_insensitive_map, sizeof(prev_map)))
			return;

		memcpy(prev_map, national_case_insensitive_map, sizeof(prev_map));

		ServerInstance->RehashUsersAndChans();

		// The OnGarbageCollect() method in m_watch rebuilds the hashmap used by it
		Module* mod = ServerInstance->Modules->Find("m_watch.so");
		if (mod)
			mod->OnGarbageCollect();
	}

 public:
	ModuleNationalChars()
		: rememberer(ServerInstance->IsNick), lowermap_rememberer(national_case_insensitive_map)
	{
		memcpy(prev_map, national_case_insensitive_map, sizeof(prev_map));
	}

	void init()
	{
		memcpy(m_lower, rfc_case_insensitive_map, 256);
		national_case_insensitive_map = m_lower;

		ServerInstance->IsNick = &myhandler;

		Implementation eventlist[] = { I_OnRehash, I_On005Numeric };
		ServerInstance->Modules->Attach(eventlist, this, sizeof(eventlist)/sizeof(Implementation));
		OnRehash(NULL);
	}

	virtual void On005Numeric(std::string &output)
	{
		std::string tmp(casemapping);
		tmp.insert(0, "CASEMAPPING=");
		SearchAndReplace(output, std::string("CASEMAPPING=rfc1459"), tmp);
	}

	virtual void OnRehash(User* user)
	{
		ConfigTag* tag = ServerInstance->Config->ConfValue("nationalchars");
		charset = tag->getString("file");
		casemapping = tag->getString("casemapping", charset);
		if(charset[0] != '/')
			charset.insert(0, "../locales/");
		unsigned char * tables[8] = { m_additional, m_additionalMB, m_additionalUp, m_lower, m_upper, m_additionalUtf8, m_additionalUtf8range, m_additionalUtf8interval };
		loadtables(charset, tables, 8, 5);
		forcequit = tag->getBool("forcequit");
		CheckForceQuit("National character set changed");
		CheckRehash();
	}

	void CheckForceQuit(const char * message)
	{
		if (!forcequit)
			return;

		for (LocalUserList::const_iterator iter = ServerInstance->Users->local_users.begin(); iter != ServerInstance->Users->local_users.end(); ++iter)
		{
			/* Fix by Brain: Dont quit UID users */
			User* n = *iter;
			if (!isdigit(n->nick[0]) && !ServerInstance->IsNick(n->nick.c_str(), ServerInstance->Config->Limits.NickMax))
				ServerInstance->Users->QuitUser(n, message);
		}
	}

	virtual ~ModuleNationalChars()
	{
		ServerInstance->IsNick = rememberer;
		national_case_insensitive_map = lowermap_rememberer;
		CheckForceQuit("National characters module unloaded");
		CheckRehash();
	}

	virtual Version GetVersion()
	{
		return Version("Provides an ability to have non-RFC1459 nicks & support for national CASEMAPPING", VF_VENDOR | VF_COMMON, charset);
	}

	/*make an array to check against it 8bit characters a bit faster. Whether allowed or uppercase (for your needs).*/
	void makereverse(unsigned char * from, unsigned  char * to, unsigned int cnt)
	{
		memset(to, 0, cnt);
		for(unsigned char * n=from; (*n) && ((*n)<cnt) && (n<from+cnt); n++)
			to[*n] = 1;
	}

	/*so Bynets Unreal distribution stuff*/
	void loadtables(std::string filename, unsigned char ** tables, unsigned char cnt, char faillimit)
	{
		std::ifstream ifs(filename.c_str());
		if (ifs.fail())
		{
			ServerInstance->Logs->Log("m_nationalchars",DEFAULT,"loadtables() called for missing file: %s", filename.c_str());
			return;
		}

		for (unsigned char n=0; n< cnt; n++)
		{
			memset(tables[n], 0, 256);
		}

		memcpy(m_lower, rfc_case_insensitive_map, 256);

		for (unsigned char n = 0; n < cnt; n++)
		{
			if (loadtable(ifs, tables[n], 255) && (n < faillimit))
			{
				ServerInstance->Logs->Log("m_nationalchars",DEFAULT,"loadtables() called for illegal file: %s (line %d)", filename.c_str(), n+1);
				return;
			}
		}

		makereverse(m_additional, m_reverse_additional, sizeof(m_additional));
	}

	unsigned char symtoi(const char *t,unsigned char base)
	/* base = 16 for hexadecimal, 10 for decimal, 8 for octal ;) */
	{
		unsigned char tmp = 0, current;
		while ((*t) && (*t !=' ') && (*t != 13) && (*t != 10) && (*t != ','))
		{
			tmp *= base;
			current = ascii_case_insensitive_map[(unsigned char)*t];
			if (current >= 'a')
				current = current - 'a' + 10;
			else
				current = current - '0';
			tmp+=current;
			t++;
		}
		return tmp;
	}

	int loadtable(std::ifstream &ifs , unsigned char *chartable, unsigned int maxindex)
	{
		std::string buf;
		getline(ifs, buf);

		unsigned int i = 0;
		int fail = 0;

		buf.erase(buf.find_last_not_of("\n") + 1);

		if (buf[0] == '.')	/* simple plain-text string after dot */
		{
			i = buf.size() - 1;

			if (i > (maxindex + 1))
				i = maxindex + 1;

			memcpy(chartable, buf.c_str() + 1, i);
		}
		else
		{
			const char * p = buf.c_str();
			while (*p)
			{
				if (i > maxindex)
				{
					fail = 1;
					break;
				}

				if (*p != '\'')		/* decimal or hexadecimal char code */
				{
					if (*p == '0')
					{
						if (p[1] == 'x')
							 /* hex with the leading "0x" */
							chartable[i] = symtoi(p + 2, 16);
						else
							chartable[i] = symtoi(p + 1, 8);
					}
					/* hex form */
					else if (*p == 'x')
					{
						chartable[i] = symtoi(p + 1, 16);
					}else	 /* decimal form */
					{
						chartable[i] = symtoi(p, 10);
					}
				}
				else		 /* plain-text char between '' */
				{
					if (*(p + 1) == '\\')
					{
						chartable[i] = *(p + 2);
						p += 3;
					}else
					{
						chartable[i] = *(p + 1);
						p += 2;
					}
				}
				while (*p && (*p != ',') && (*p != ' ') && (*p != 13) && (*p != 10))
					p++;
				while (*p && ((*p == ',') || (*p == ' ') || (*p == 13) || (*p == 10)))
					p++;
				i++;
			}
		}
		return fail;
	}
};

MODULE_INIT(ModuleNationalChars)