src/modules/m_nationalchars.cpp

   1 /*
   2  * InspIRCd -- Internet Relay Chat Daemon
   3  *
   4  *   Copyright (C) 2009 Daniel De Graaf <danieldg@inspircd.org>
   5  *   Copyright (C) 2009 Dennis Friis <peavey@inspircd.org>
   6  *   Copyright (C) 2009 Craig Edwards <craigedwards@brainbox.cc>
   7  *   Copyright (C) 2009 Robin Burchell <robin+git@viroteck.net>
   8  *
   9  * This file is part of InspIRCd.  InspIRCd is free software: you can
  10  * redistribute it and/or modify it under the terms of the GNU General Public
  11  * License as published by the Free Software Foundation, version 2.
  12  *
  13  * This program is distributed in the hope that it will be useful, but WITHOUT
  14  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
  15  * FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more
  16  * details.
  17  *
  18  * You should have received a copy of the GNU General Public License
  19  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  20  */
  21
  22
  23 /* Contains a code of Unreal IRCd + Bynets patch ( http://www.unrealircd.com/ and http://www.bynets.org/ )
  24    Original patch is made by Dmitry "Killer{R}" Kononko. ( http://killprog.com/ )
  25    Changed at 2008-06-15 - 2009-02-11
  26    by Chernov-Phoenix Alexey (Phoenix@RusNet) mailto:phoenix /email address separator/ pravmail.ru */
  27
  28 #include "inspircd.h"
  29 #include "caller.h"
  30 #include <fstream>
  31
  32 /* $ModDesc: Provides an ability to have non-RFC1459 nicks & support for national CASEMAPPING */
  33
  34 class lwbNickHandler : public HandlerBase2<bool, const char*, size_t>
  35 {
  36  public:
  37         lwbNickHandler() { }
  38         virtual ~lwbNickHandler() { }
  39         virtual bool Call(const char*, size_t);
  40 };
  41
  42                                                                  /*,m_reverse_additionalUp[256];*/
  43 static unsigned char m_reverse_additional[256],m_additionalMB[256],m_additionalUtf8[256],m_additionalUtf8range[256],m_additionalUtf8interval[256];
  44
  45 char utf8checkrest(unsigned char * mb, unsigned char cnt)
  46 {
  47         for (unsigned char * tmp=mb; tmp<mb+cnt; tmp++)
  48         {
  49                 /* & is faster! -- Phoenix (char & b11000000 == b10000000) */
  50                 if ((*tmp & 192) != 128)
  51                         return -1;
  52         }
  53         return cnt + 1;
  54 }
  55
  56
  57 char utf8size(unsigned char * mb)
  58 {
  59         if (!*mb)
  60                 return -1;
  61         if (!(*mb & 128))
  62                 return 1;
  63         if ((*mb & 224) == 192)
  64                 return utf8checkrest(mb + 1,1);
  65         if ((*mb & 240) == 224)
  66                 return utf8checkrest(mb + 1,2);
  67         if ((*mb & 248) == 240)
  68                 return utf8checkrest(mb + 1,3);
  69         return -1;
  70 }
  71
  72
  73 /* Conditions added */
  74 bool lwbNickHandler::Call(const char* n, size_t max)
  75 {
  76         if (!n || !*n)
  77                 return false;
  78
  79         unsigned int p = 0;
  80         for (const char* i = n; *i; i++, p++)
  81         {
  82                 /* 1. Multibyte encodings support:  */
  83                 /* 1.1. 16bit char. areas, e.g. chinese:*/
  84
  85                 /* if current character is the last, we DO NOT check it against multibyte table */
  86                 /* if there are mbtable ranges, use ONLY them. No 8bit at all */
  87                 if (i[1] && m_additionalMB[0])
  88                 {
  89                         /* otherwise let's take a look at the current character and the following one */
  90                         bool found = false;
  91                         for(unsigned char * mb = m_additionalMB; (*mb) && (mb < m_additionalMB + sizeof(m_additionalMB)); mb += 4)
  92                         {
  93                                 if ( (i[0] >= mb[0]) && (i[0] <= mb[1]) && (i[1] >= mb[2]) && (i[1] <= mb[3]) )
  94                                 {
  95                                         /* multibyte range character found */
  96                                         i++;
  97                                         p++;
  98                                         found = true;
  99                                         break;
 100                                 }
 101                         }
 102                         if (found)
 103                                 /* next char! */
 104                                 continue;
 105                         else
 106                                 /* there are ranges, but incorrect char (8bit?) given, sorry */
 107                                 return false;
 108                 }
 109
 110                 /* 2. 8bit character support */
 111                 if (((*i >= 'A') && (*i <= '}')) || m_reverse_additional[(unsigned char)*i])
 112                         /* "A"-"}" can occur anywhere in a nickname */
 113                         continue;
 114
 115                 if ((((*i >= '0') && (*i <= '9')) || (*i == '-')) && (i > n))
 116                         /* "0"-"9", "-" can occur anywhere BUT the first char of a nickname */
 117                         continue;
 118
 119                 /* 3.1. Check against a simple UTF-8 characters enumeration */
 120                 int cursize, cursize2, ncursize = utf8size((unsigned char *)i);
 121                 /* do check only if current multibyte character is valid UTF-8 only */
 122                 if (ncursize != -1)
 123                 {
 124                         bool found = false;
 125                         for (unsigned char * mb = m_additionalUtf8; (utf8size(mb) != -1) && (mb < m_additionalUtf8 + sizeof(m_additionalUtf8)); mb += cursize)
 126                         {
 127                                 cursize = utf8size(mb);
 128                                 /* Size differs? Pick the next! */
 129                                 if (cursize != ncursize)
 130                                         continue;
 131
 132                                 if (!strncmp(i, (char *)mb, cursize))
 133                                 {
 134                                         i += cursize - 1;
 135                                         p += cursize - 1;
 136                                         found = true;
 137                                         break;
 138                                 }
 139                         }
 140                         if (found)
 141                                 continue;
 142
 143                         /* 3.2. Check against an UTF-8 ranges: <start character> and <length of the range>. */
 144                         found = false;
 145                         for (unsigned char * mb = m_additionalUtf8range; (utf8size(mb) != -1) && (mb < m_additionalUtf8range + sizeof(m_additionalUtf8range)); mb += cursize + 1)
 146                         {
 147                                 cursize = utf8size(mb);
 148                                 /* Size differs (or lengthbyte is zero)? Pick the next! */
 149                                 if ((cursize != ncursize) || (!mb[cursize]))
 150                                         continue;
 151
 152                                 unsigned char uright[5] = {0,0,0,0,0}, range = mb[cursize] - 1;
 153                                 strncpy((char* ) uright, (char *) mb, cursize);
 154
 155                                 for (int temp = cursize - 1; (temp >= 0) && range; --temp)
 156                                 {
 157                                         /* all but the first char are 64-based */
 158                                         if (temp)
 159                                         {
 160                                                 char part64 = range & 63; /* i.e. % 64 */
 161                                                 /* handle carrying over */
 162                                                 if (uright[temp] + part64 - 1 > 191)
 163                                                 {
 164                                                         uright[temp] -= 64;
 165                                                         range += 64;
 166                                                 }
 167                                                 uright[temp] += part64;
 168                                                 range >>= 6; /* divide it on a 64 */
 169                                         }
 170                                         /* the first char of UTF-8 doesn't follow the rule */
 171                                         else
 172                                         {
 173                                                 uright[temp] += range;
 174                                         }
 175                                 }
 176
 177                                 if ((strncmp(i, (char *) mb, cursize) >= 0) && (strncmp(i, (char *) uright, cursize) <= 0))
 178                                 {
 179                                         i += cursize - 1;
 180                                         p += cursize - 1;
 181                                         found = true;
 182                                         break;
 183                                 }
 184                         }
 185                         if (found)
 186                                 continue;
 187
 188                         /* 3.3. Check against an UTF-8 intervals: <start character> and <end character>. */
 189                         found = false;
 190                         for (unsigned char * mb = m_additionalUtf8interval; (utf8size(mb) != -1) && (utf8size(mb+utf8size(mb)) != -1)
 191                                 && (mb < m_additionalUtf8interval + sizeof(m_additionalUtf8interval)); mb += (cursize+cursize2) )
 192                         {
 193                                 cursize = utf8size(mb);
 194                                 cursize2= utf8size(mb+cursize);
 195
 196                                 int minlen  = cursize  > ncursize ? ncursize : cursize;
 197                                 int minlen2 = cursize2 > ncursize ? ncursize : cursize2;
 198
 199                                 unsigned char* uright = mb + cursize;
 200
 201                                 if ((strncmp(i, (char *) mb, minlen) >= 0) && (strncmp(i, (char *) uright, minlen2) <= 0))
 202                                 {
 203                                         i += cursize - 1;
 204                                         p += cursize - 1;
 205                                         found = true;
 206                                         break;
 207                                 }
 208                         }
 209                         if (found)
 210                                 continue;
 211                 }
 212
 213                 /* invalid character! abort */
 214                 return false;
 215         }
 216
 217         /* too long? or not -- pointer arithmetic rocks */
 218         return (p < max);
 219 }
 220
 221
 222 class ModuleNationalChars : public Module
 223 {
 224  private:
 225         lwbNickHandler myhandler;
 226         std::string charset, casemapping;
 227         unsigned char m_additional[256], m_additionalUp[256], m_lower[256], m_upper[256];
 228         caller2<bool, const char*, size_t> rememberer;
 229         bool forcequit;
 230         const unsigned char * lowermap_rememberer;
 231         unsigned char prev_map[256];
 232
 233         void CheckRehash()
 234         {
 235                 // See if anything changed
 236                 if (!memcmp(prev_map, national_case_insensitive_map, sizeof(prev_map)))
 237                         return;
 238
 239                 memcpy(prev_map, national_case_insensitive_map, sizeof(prev_map));
 240
 241                 ServerInstance->RehashUsersAndChans();
 242
 243                 // The OnGarbageCollect() method in m_watch rebuilds the hashmap used by it
 244                 Module* mod = ServerInstance->Modules->Find("m_watch.so");
 245                 if (mod)
 246                         mod->OnGarbageCollect();
 247
 248                 // Send a Request to m_spanningtree asking it to rebuild its hashmaps
 249                 mod = ServerInstance->Modules->Find("m_spanningtree.so");
 250                 if (mod)
 251                 {
 252                         Request req(this, mod, "rehash");
 253                         req.Send();
 254                 }
 255         }
 256
 257  public:
 258         ModuleNationalChars()
 259                 : rememberer(ServerInstance->IsNick), lowermap_rememberer(national_case_insensitive_map)
 260         {
 261                 memcpy(prev_map, national_case_insensitive_map, sizeof(prev_map));
 262         }
 263
 264         void init()
 265         {
 266                 memcpy(m_lower, rfc_case_insensitive_map, 256);
 267                 national_case_insensitive_map = m_lower;
 268
 269                 ServerInstance->IsNick = &myhandler;
 270
 271                 Implementation eventlist[] = { I_OnRehash, I_On005Numeric };
 272                 ServerInstance->Modules->Attach(eventlist, this, sizeof(eventlist)/sizeof(Implementation));
 273                 OnRehash(NULL);
 274         }
 275
 276         virtual void On005Numeric(std::string &output)
 277         {
 278                 std::string tmp(casemapping);
 279                 tmp.insert(0, "CASEMAPPING=");
 280                 SearchAndReplace(output, std::string("CASEMAPPING=rfc1459"), tmp);
 281         }
 282
 283         virtual void OnRehash(User* user)
 284         {
 285                 ConfigTag* tag = ServerInstance->Config->ConfValue("nationalchars");
 286                 charset = tag->getString("file");
 287                 casemapping = tag->getString("casemapping", ServerConfig::CleanFilename(charset.c_str()));
 288                 if (casemapping.find(' ') != std::string::npos)
 289                         throw ModuleException("<nationalchars:casemapping> must not contain any spaces!");
 290 #if defined _WIN32
 291                 if (!ServerInstance->Config->StartsWithWindowsDriveLetter(charset))
 292                         charset.insert(0, "./locales/");
 293 #else
 294                 if(charset[0] != '/')
 295                         charset.insert(0, "../locales/");
 296 #endif
 297                 unsigned char * tables[8] = { m_additional, m_additionalMB, m_additionalUp, m_lower, m_upper, m_additionalUtf8, m_additionalUtf8range, m_additionalUtf8interval };
 298                 if (!loadtables(charset, tables, 8, 5))
 299                         throw ModuleException("The locale file failed to load. Check your log file for more information.");
 300                 forcequit = tag->getBool("forcequit");
 301                 CheckForceQuit("National character set changed");
 302                 CheckRehash();
 303         }
 304
 305         void CheckForceQuit(const char * message)
 306         {
 307                 if (!forcequit)
 308                         return;
 309
 310                 for (LocalUserList::const_iterator iter = ServerInstance->Users->local_users.begin(); iter != ServerInstance->Users->local_users.end(); ++iter)
 311                 {
 312                         /* Fix by Brain: Dont quit UID users */
 313                         User* n = *iter;
 314                         if (!isdigit(n->nick[0]) && !ServerInstance->IsNick(n->nick.c_str(), ServerInstance->Config->Limits.NickMax))
 315                                 ServerInstance->Users->QuitUser(n, message);
 316                 }
 317         }
 318
 319         virtual ~ModuleNationalChars()
 320         {
 321                 ServerInstance->IsNick = rememberer;
 322                 national_case_insensitive_map = lowermap_rememberer;
 323                 CheckForceQuit("National characters module unloaded");
 324                 CheckRehash();
 325         }
 326
 327         virtual Version GetVersion()
 328         {
 329                 return Version("Provides an ability to have non-RFC1459 nicks & support for national CASEMAPPING", VF_VENDOR | VF_COMMON, charset);
 330         }
 331
 332         /*make an array to check against it 8bit characters a bit faster. Whether allowed or uppercase (for your needs).*/
 333         void makereverse(unsigned char * from, unsigned  char * to, unsigned int cnt)
 334         {
 335                 memset(to, 0, cnt);
 336                 for(unsigned char * n=from; (*n) && ((*n)<cnt) && (n<from+cnt); n++)
 337                         to[*n] = 1;
 338         }
 339
 340         /*so Bynets Unreal distribution stuff*/
 341         bool loadtables(std::string filename, unsigned char ** tables, unsigned char cnt, char faillimit)
 342         {
 343                 std::ifstream ifs(filename.c_str());
 344                 if (ifs.fail())
 345                 {
 346                         ServerInstance->Logs->Log("m_nationalchars",DEFAULT,"loadtables() called for missing file: %s", filename.c_str());
 347                         return false;
 348                 }
 349
 350                 for (unsigned char n=0; n< cnt; n++)
 351                 {
 352                         memset(tables[n], 0, 256);
 353                 }
 354
 355                 memcpy(m_lower, rfc_case_insensitive_map, 256);
 356
 357                 for (unsigned char n = 0; n < cnt; n++)
 358                 {
 359                         if (loadtable(ifs, tables[n], 255) && (n < faillimit))
 360                         {
 361                                 ServerInstance->Logs->Log("m_nationalchars",DEFAULT,"loadtables() called for illegal file: %s (line %d)", filename.c_str(), n+1);
 362                                 return false;
 363                         }
 364                 }
 365
 366                 makereverse(m_additional, m_reverse_additional, sizeof(m_additional));
 367                 return true;
 368         }
 369
 370         unsigned char symtoi(const char *t,unsigned char base)
 371         /* base = 16 for hexadecimal, 10 for decimal, 8 for octal ;) */
 372         {
 373                 unsigned char tmp = 0, current;
 374                 while ((*t) && (*t !=' ') && (*t != 13) && (*t != 10) && (*t != ','))
 375                 {
 376                         tmp *= base;
 377                         current = ascii_case_insensitive_map[(unsigned char)*t];
 378                         if (current >= 'a')
 379                                 current = current - 'a' + 10;
 380                         else
 381                                 current = current - '0';
 382                         tmp+=current;
 383                         t++;
 384                 }
 385                 return tmp;
 386         }
 387
 388         int loadtable(std::ifstream &ifs , unsigned char *chartable, unsigned int maxindex)
 389         {
 390                 std::string buf;
 391                 getline(ifs, buf);
 392
 393                 unsigned int i = 0;
 394                 int fail = 0;
 395
 396                 buf.erase(buf.find_last_not_of("\n") + 1);
 397
 398                 if (buf[0] == '.')      /* simple plain-text string after dot */
 399                 {
 400                         i = buf.size() - 1;
 401
 402                         if (i > (maxindex + 1))
 403                                 i = maxindex + 1;
 404
 405                         memcpy(chartable, buf.c_str() + 1, i);
 406                 }
 407                 else
 408                 {
 409                         const char * p = buf.c_str();
 410                         while (*p)
 411                         {
 412                                 if (i > maxindex)
 413                                 {
 414                                         fail = 1;
 415                                         break;
 416                                 }
 417
 418                                 if (*p != '\'')         /* decimal or hexadecimal char code */
 419                                 {
 420                                         if (*p == '0')
 421                                         {
 422                                                 if (p[1] == 'x')
 423                                                          /* hex with the leading "0x" */
 424                                                         chartable[i] = symtoi(p + 2, 16);
 425                                                 else
 426                                                         chartable[i] = symtoi(p + 1, 8);
 427                                         }
 428                                         /* hex form */
 429                                         else if (*p == 'x')
 430                                         {
 431                                                 chartable[i] = symtoi(p + 1, 16);
 432                                         }else    /* decimal form */
 433                                         {
 434                                                 chartable[i] = symtoi(p, 10);
 435                                         }
 436                                 }
 437                                 else             /* plain-text char between '' */
 438                                 {
 439                                         if (*(p + 1) == '\\')
 440                                         {
 441                                                 chartable[i] = *(p + 2);
 442                                                 p += 3;
 443                                         }else
 444                                         {
 445                                                 chartable[i] = *(p + 1);
 446                                                 p += 2;
 447                                         }
 448                                 }
 449                                 while (*p && (*p != ',') && (*p != ' ') && (*p != 13) && (*p != 10))
 450                                         p++;
 451                                 while (*p && ((*p == ',') || (*p == ' ') || (*p == 13) || (*p == 10)))
 452                                         p++;
 453                                 i++;
 454                         }
 455                 }
 456                 return fail;
 457         }
 458 };
 459
 460 MODULE_INIT(ModuleNationalChars)