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 class lwbNickHandler : public HandlerBase1<bool, const std::string&>
  33 {
  34  public:
  35         lwbNickHandler() { }
  36         ~lwbNickHandler() { }
  37         bool Call(const std::string&);
  38 };
  39
  40                                                                  /*,m_reverse_additionalUp[256];*/
  41 static unsigned char m_reverse_additional[256],m_additionalMB[256],m_additionalUtf8[256],m_additionalUtf8range[256],m_additionalUtf8interval[256];
  42
  43 char utf8checkrest(unsigned char * mb, unsigned char cnt)
  44 {
  45         for (unsigned char * tmp=mb; tmp<mb+cnt; tmp++)
  46         {
  47                 /* & is faster! -- Phoenix (char & b11000000 == b10000000) */
  48                 if ((*tmp & 192) != 128)
  49                         return -1;
  50         }
  51         return cnt + 1;
  52 }
  53
  54
  55 char utf8size(unsigned char * mb)
  56 {
  57         if (!*mb)
  58                 return -1;
  59         if (!(*mb & 128))
  60                 return 1;
  61         if ((*mb & 224) == 192)
  62                 return utf8checkrest(mb + 1,1);
  63         if ((*mb & 240) == 224)
  64                 return utf8checkrest(mb + 1,2);
  65         if ((*mb & 248) == 240)
  66                 return utf8checkrest(mb + 1,3);
  67         return -1;
  68 }
  69
  70
  71 /* Conditions added */
  72 bool lwbNickHandler::Call(const std::string& nick)
  73 {
  74         if (nick.empty())
  75                 return false;
  76
  77         const char* n = nick.c_str();
  78         unsigned int p = 0;
  79         for (const char* i = n; *i; i++, p++)
  80         {
  81                 /* 1. Multibyte encodings support:  */
  82                 /* 1.1. 16bit char. areas, e.g. chinese:*/
  83
  84                 /* if current character is the last, we DO NOT check it against multibyte table */
  85                 /* if there are mbtable ranges, use ONLY them. No 8bit at all */
  86                 if (i[1] && m_additionalMB[0])
  87                 {
  88                         /* otherwise let's take a look at the current character and the following one */
  89                         bool found = false;
  90                         for(unsigned char * mb = m_additionalMB; (*mb) && (mb < m_additionalMB + sizeof(m_additionalMB)); mb += 4)
  91                         {
  92                                 if ( (i[0] >= mb[0]) && (i[0] <= mb[1]) && (i[1] >= mb[2]) && (i[1] <= mb[3]) )
  93                                 {
  94                                         /* multibyte range character found */
  95                                         i++;
  96                                         p++;
  97                                         found = true;
  98                                         break;
  99                                 }
 100                         }
 101                         if (found)
 102                                 /* next char! */
 103                                 continue;
 104                         else
 105                                 /* there are ranges, but incorrect char (8bit?) given, sorry */
 106                                 return false;
 107                 }
 108
 109                 /* 2. 8bit character support */
 110                 if (((*i >= 'A') && (*i <= '}')) || m_reverse_additional[(unsigned char)*i])
 111                         /* "A"-"}" can occur anywhere in a nickname */
 112                         continue;
 113
 114                 if ((((*i >= '0') && (*i <= '9')) || (*i == '-')) && (i > n))
 115                         /* "0"-"9", "-" can occur anywhere BUT the first char of a nickname */
 116                         continue;
 117
 118                 /* 3.1. Check against a simple UTF-8 characters enumeration */
 119                 int cursize, cursize2, ncursize = utf8size((unsigned char *)i);
 120                 /* do check only if current multibyte character is valid UTF-8 only */
 121                 if (ncursize != -1)
 122                 {
 123                         bool found = false;
 124                         for (unsigned char * mb = m_additionalUtf8; (utf8size(mb) != -1) && (mb < m_additionalUtf8 + sizeof(m_additionalUtf8)); mb += cursize)
 125                         {
 126                                 cursize = utf8size(mb);
 127                                 /* Size differs? Pick the next! */
 128                                 if (cursize != ncursize)
 129                                         continue;
 130
 131                                 if (!strncmp(i, (char *)mb, cursize))
 132                                 {
 133                                         i += cursize - 1;
 134                                         p += cursize - 1;
 135                                         found = true;
 136                                         break;
 137                                 }
 138                         }
 139                         if (found)
 140                                 continue;
 141
 142                         /* 3.2. Check against an UTF-8 ranges: <start character> and <length of the range>. */
 143                         found = false;
 144                         for (unsigned char * mb = m_additionalUtf8range; (utf8size(mb) != -1) && (mb < m_additionalUtf8range + sizeof(m_additionalUtf8range)); mb += cursize + 1)
 145                         {
 146                                 cursize = utf8size(mb);
 147                                 /* Size differs (or lengthbyte is zero)? Pick the next! */
 148                                 if ((cursize != ncursize) || (!mb[cursize]))
 149                                         continue;
 150
 151                                 unsigned char uright[5] = {0,0,0,0,0}, range = mb[cursize] - 1;
 152                                 strncpy((char* ) uright, (char *) mb, cursize);
 153
 154                                 for (int temp = cursize - 1; (temp >= 0) && range; --temp)
 155                                 {
 156                                         /* all but the first char are 64-based */
 157                                         if (temp)
 158                                         {
 159                                                 char part64 = range & 63; /* i.e. % 64 */
 160                                                 /* handle carrying over */
 161                                                 if (uright[temp] + part64 - 1 > 191)
 162                                                 {
 163                                                         uright[temp] -= 64;
 164                                                         range += 64;
 165                                                 }
 166                                                 uright[temp] += part64;
 167                                                 range >>= 6; /* divide it on a 64 */
 168                                         }
 169                                         /* the first char of UTF-8 doesn't follow the rule */
 170                                         else
 171                                         {
 172                                                 uright[temp] += range;
 173                                         }
 174                                 }
 175
 176                                 if ((strncmp(i, (char *) mb, cursize) >= 0) && (strncmp(i, (char *) uright, cursize) <= 0))
 177                                 {
 178                                         i += cursize - 1;
 179                                         p += cursize - 1;
 180                                         found = true;
 181                                         break;
 182                                 }
 183                         }
 184                         if (found)
 185                                 continue;
 186
 187                         /* 3.3. Check against an UTF-8 intervals: <start character> and <end character>. */
 188                         found = false;
 189                         for (unsigned char * mb = m_additionalUtf8interval; (utf8size(mb) != -1) && (utf8size(mb+utf8size(mb)) != -1)
 190                                 && (mb < m_additionalUtf8interval + sizeof(m_additionalUtf8interval)); mb += (cursize+cursize2) )
 191                         {
 192                                 cursize = utf8size(mb);
 193                                 cursize2= utf8size(mb+cursize);
 194
 195                                 int minlen  = cursize  > ncursize ? ncursize : cursize;
 196                                 int minlen2 = cursize2 > ncursize ? ncursize : cursize2;
 197
 198                                 unsigned char* uright = mb + cursize;
 199
 200                                 if ((strncmp(i, (char *) mb, minlen) >= 0) && (strncmp(i, (char *) uright, minlen2) <= 0))
 201                                 {
 202                                         i += cursize - 1;
 203                                         p += cursize - 1;
 204                                         found = true;
 205                                         break;
 206                                 }
 207                         }
 208                         if (found)
 209                                 continue;
 210                 }
 211
 212                 /* invalid character! abort */
 213                 return false;
 214         }
 215
 216         /* too long? or not -- pointer arithmetic rocks */
 217         return (p < ServerInstance->Config->Limits.NickMax);
 218 }
 219
 220
 221 class ModuleNationalChars : public Module
 222 {
 223         lwbNickHandler myhandler;
 224         std::string charset, casemapping;
 225         unsigned char m_additional[256], m_additionalUp[256], m_lower[256], m_upper[256];
 226         caller1<bool, const std::string&> rememberer;
 227         bool forcequit;
 228         const unsigned char * lowermap_rememberer;
 229
 230  public:
 231         ModuleNationalChars()
 232                 : rememberer(ServerInstance->IsNick), lowermap_rememberer(national_case_insensitive_map)
 233         {
 234         }
 235
 236         void init() CXX11_OVERRIDE
 237         {
 238                 memcpy(m_lower, rfc_case_insensitive_map, 256);
 239                 national_case_insensitive_map = m_lower;
 240
 241                 ServerInstance->IsNick = &myhandler;
 242
 243                 OnRehash(NULL);
 244         }
 245
 246         void On005Numeric(std::map<std::string, std::string>& tokens) CXX11_OVERRIDE
 247         {
 248                 tokens["CASEMAPPING"] = casemapping;
 249         }
 250
 251         void OnRehash(User* user) CXX11_OVERRIDE
 252         {
 253                 ConfigTag* tag = ServerInstance->Config->ConfValue("nationalchars");
 254                 charset = tag->getString("file");
 255                 casemapping = tag->getString("casemapping", charset);
 256                 if(charset[0] != '/')
 257                         charset.insert(0, "../locales/");
 258                 unsigned char * tables[8] = { m_additional, m_additionalMB, m_additionalUp, m_lower, m_upper, m_additionalUtf8, m_additionalUtf8range, m_additionalUtf8interval };
 259                 loadtables(charset, tables, 8, 5);
 260                 forcequit = tag->getBool("forcequit");
 261                 CheckForceQuit("National character set changed");
 262         }
 263
 264         void CheckForceQuit(const char * message)
 265         {
 266                 if (!forcequit)
 267                         return;
 268
 269                 for (LocalUserList::const_iterator iter = ServerInstance->Users->local_users.begin(); iter != ServerInstance->Users->local_users.end(); ++iter)
 270                 {
 271                         /* Fix by Brain: Dont quit UID users */
 272                         User* n = *iter;
 273                         if (!isdigit(n->nick[0]) && !ServerInstance->IsNick(n->nick))
 274                                 ServerInstance->Users->QuitUser(n, message);
 275                 }
 276         }
 277
 278         ~ModuleNationalChars()
 279         {
 280                 ServerInstance->IsNick = rememberer;
 281                 national_case_insensitive_map = lowermap_rememberer;
 282                 CheckForceQuit("National characters module unloaded");
 283         }
 284
 285         Version GetVersion() CXX11_OVERRIDE
 286         {
 287                 return Version("Provides an ability to have non-RFC1459 nicks & support for national CASEMAPPING", VF_VENDOR | VF_COMMON, charset);
 288         }
 289
 290         /*make an array to check against it 8bit characters a bit faster. Whether allowed or uppercase (for your needs).*/
 291         void makereverse(unsigned char * from, unsigned  char * to, unsigned int cnt)
 292         {
 293                 memset(to, 0, cnt);
 294                 for(unsigned char * n=from; (*n) && ((*n)<cnt) && (n<from+cnt); n++)
 295                         to[*n] = 1;
 296         }
 297
 298         /*so Bynets Unreal distribution stuff*/
 299         void loadtables(std::string filename, unsigned char ** tables, unsigned char cnt, char faillimit)
 300         {
 301                 std::ifstream ifs(filename.c_str());
 302                 if (ifs.fail())
 303                 {
 304                         ServerInstance->Logs->Log(MODNAME, LOG_DEFAULT, "loadtables() called for missing file: %s", filename.c_str());
 305                         return;
 306                 }
 307
 308                 for (unsigned char n=0; n< cnt; n++)
 309                 {
 310                         memset(tables[n], 0, 256);
 311                 }
 312
 313                 memcpy(m_lower, rfc_case_insensitive_map, 256);
 314
 315                 for (unsigned char n = 0; n < cnt; n++)
 316                 {
 317                         if (loadtable(ifs, tables[n], 255) && (n < faillimit))
 318                         {
 319                                 ServerInstance->Logs->Log(MODNAME, LOG_DEFAULT, "loadtables() called for illegal file: %s (line %d)", filename.c_str(), n+1);
 320                                 return;
 321                         }
 322                 }
 323
 324                 makereverse(m_additional, m_reverse_additional, sizeof(m_additional));
 325         }
 326
 327         unsigned char symtoi(const char *t,unsigned char base)
 328         /* base = 16 for hexadecimal, 10 for decimal, 8 for octal ;) */
 329         {
 330                 unsigned char tmp = 0, current;
 331                 while ((*t) && (*t !=' ') && (*t != 13) && (*t != 10) && (*t != ','))
 332                 {
 333                         tmp *= base;
 334                         current = ascii_case_insensitive_map[(unsigned char)*t];
 335                         if (current >= 'a')
 336                                 current = current - 'a' + 10;
 337                         else
 338                                 current = current - '0';
 339                         tmp+=current;
 340                         t++;
 341                 }
 342                 return tmp;
 343         }
 344
 345         int loadtable(std::ifstream &ifs , unsigned char *chartable, unsigned int maxindex)
 346         {
 347                 std::string buf;
 348                 getline(ifs, buf);
 349
 350                 unsigned int i = 0;
 351                 int fail = 0;
 352
 353                 buf.erase(buf.find_last_not_of("\n") + 1);
 354
 355                 if (buf[0] == '.')      /* simple plain-text string after dot */
 356                 {
 357                         i = buf.size() - 1;
 358
 359                         if (i > (maxindex + 1))
 360                                 i = maxindex + 1;
 361
 362                         memcpy(chartable, buf.c_str() + 1, i);
 363                 }
 364                 else
 365                 {
 366                         const char * p = buf.c_str();
 367                         while (*p)
 368                         {
 369                                 if (i > maxindex)
 370                                 {
 371                                         fail = 1;
 372                                         break;
 373                                 }
 374
 375                                 if (*p != '\'')         /* decimal or hexadecimal char code */
 376                                 {
 377                                         if (*p == '0')
 378                                         {
 379                                                 if (p[1] == 'x')
 380                                                          /* hex with the leading "0x" */
 381                                                         chartable[i] = symtoi(p + 2, 16);
 382                                                 else
 383                                                         chartable[i] = symtoi(p + 1, 8);
 384                                         }
 385                                         /* hex form */
 386                                         else if (*p == 'x')
 387                                         {
 388                                                 chartable[i] = symtoi(p + 1, 16);
 389                                         }else    /* decimal form */
 390                                         {
 391                                                 chartable[i] = symtoi(p, 10);
 392                                         }
 393                                 }
 394                                 else             /* plain-text char between '' */
 395                                 {
 396                                         if (*(p + 1) == '\\')
 397                                         {
 398                                                 chartable[i] = *(p + 2);
 399                                                 p += 3;
 400                                         }else
 401                                         {
 402                                                 chartable[i] = *(p + 1);
 403                                                 p += 2;
 404                                         }
 405                                 }
 406                                 while (*p && (*p != ',') && (*p != ' ') && (*p != 13) && (*p != 10))
 407                                         p++;
 408                                 while (*p && ((*p == ',') || (*p == ' ') || (*p == 13) || (*p == 10)))
 409                                         p++;
 410                                 i++;
 411                         }
 412                 }
 413                 return fail;
 414         }
 415 };
 416
 417 MODULE_INIT(ModuleNationalChars)