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 <fstream>
  30
  31 class lwbNickHandler : public HandlerBase1<bool, const std::string&>
  32 {
  33  public:
  34         bool Call(const std::string&);
  35 };
  36
  37                                                                  /*,m_reverse_additionalUp[256];*/
  38 static unsigned char m_reverse_additional[256],m_additionalMB[256],m_additionalUtf8[256],m_additionalUtf8range[256],m_additionalUtf8interval[256];
  39
  40 char utf8checkrest(unsigned char * mb, unsigned char cnt)
  41 {
  42         for (unsigned char * tmp=mb; tmp<mb+cnt; tmp++)
  43         {
  44                 /* & is faster! -- Phoenix (char & b11000000 == b10000000) */
  45                 if ((*tmp & 192) != 128)
  46                         return -1;
  47         }
  48         return cnt + 1;
  49 }
  50
  51
  52 char utf8size(unsigned char * mb)
  53 {
  54         if (!*mb)
  55                 return -1;
  56         if (!(*mb & 128))
  57                 return 1;
  58         if ((*mb & 224) == 192)
  59                 return utf8checkrest(mb + 1,1);
  60         if ((*mb & 240) == 224)
  61                 return utf8checkrest(mb + 1,2);
  62         if ((*mb & 248) == 240)
  63                 return utf8checkrest(mb + 1,3);
  64         return -1;
  65 }
  66
  67
  68 /* Conditions added */
  69 bool lwbNickHandler::Call(const std::string& nick)
  70 {
  71         if (nick.empty())
  72                 return false;
  73
  74         const char* n = nick.c_str();
  75         unsigned int p = 0;
  76         for (const char* i = n; *i; i++, p++)
  77         {
  78                 /* 1. Multibyte encodings support:  */
  79                 /* 1.1. 16bit char. areas, e.g. chinese:*/
  80
  81                 /* if current character is the last, we DO NOT check it against multibyte table */
  82                 /* if there are mbtable ranges, use ONLY them. No 8bit at all */
  83                 if (i[1] && m_additionalMB[0])
  84                 {
  85                         /* otherwise let's take a look at the current character and the following one */
  86                         bool found = false;
  87                         for(unsigned char * mb = m_additionalMB; (*mb) && (mb < m_additionalMB + sizeof(m_additionalMB)); mb += 4)
  88                         {
  89                                 if ( (i[0] >= mb[0]) && (i[0] <= mb[1]) && (i[1] >= mb[2]) && (i[1] <= mb[3]) )
  90                                 {
  91                                         /* multibyte range character found */
  92                                         i++;
  93                                         p++;
  94                                         found = true;
  95                                         break;
  96                                 }
  97                         }
  98                         if (found)
  99                                 /* next char! */
 100                                 continue;
 101                         else
 102                                 /* there are ranges, but incorrect char (8bit?) given, sorry */
 103                                 return false;
 104                 }
 105
 106                 /* 2. 8bit character support */
 107                 if (((*i >= 'A') && (*i <= '}')) || m_reverse_additional[(unsigned char)*i])
 108                         /* "A"-"}" can occur anywhere in a nickname */
 109                         continue;
 110
 111                 if ((((*i >= '0') && (*i <= '9')) || (*i == '-')) && (i > n))
 112                         /* "0"-"9", "-" can occur anywhere BUT the first char of a nickname */
 113                         continue;
 114
 115                 /* 3.1. Check against a simple UTF-8 characters enumeration */
 116                 int cursize, cursize2, ncursize = utf8size((unsigned char *)i);
 117                 /* do check only if current multibyte character is valid UTF-8 only */
 118                 if (ncursize != -1)
 119                 {
 120                         bool found = false;
 121                         for (unsigned char * mb = m_additionalUtf8; (utf8size(mb) != -1) && (mb < m_additionalUtf8 + sizeof(m_additionalUtf8)); mb += cursize)
 122                         {
 123                                 cursize = utf8size(mb);
 124                                 /* Size differs? Pick the next! */
 125                                 if (cursize != ncursize)
 126                                         continue;
 127
 128                                 if (!strncmp(i, (char *)mb, cursize))
 129                                 {
 130                                         i += cursize - 1;
 131                                         p += cursize - 1;
 132                                         found = true;
 133                                         break;
 134                                 }
 135                         }
 136                         if (found)
 137                                 continue;
 138
 139                         /* 3.2. Check against an UTF-8 ranges: <start character> and <length of the range>. */
 140                         found = false;
 141                         for (unsigned char * mb = m_additionalUtf8range; (utf8size(mb) != -1) && (mb < m_additionalUtf8range + sizeof(m_additionalUtf8range)); mb += cursize + 1)
 142                         {
 143                                 cursize = utf8size(mb);
 144                                 /* Size differs (or lengthbyte is zero)? Pick the next! */
 145                                 if ((cursize != ncursize) || (!mb[cursize]))
 146                                         continue;
 147
 148                                 unsigned char uright[5] = {0,0,0,0,0}, range = mb[cursize] - 1;
 149                                 strncpy((char* ) uright, (char *) mb, cursize);
 150
 151                                 for (int temp = cursize - 1; (temp >= 0) && range; --temp)
 152                                 {
 153                                         /* all but the first char are 64-based */
 154                                         if (temp)
 155                                         {
 156                                                 char part64 = range & 63; /* i.e. % 64 */
 157                                                 /* handle carrying over */
 158                                                 if (uright[temp] + part64 - 1 > 191)
 159                                                 {
 160                                                         uright[temp] -= 64;
 161                                                         range += 64;
 162                                                 }
 163                                                 uright[temp] += part64;
 164                                                 range >>= 6; /* divide it on a 64 */
 165                                         }
 166                                         /* the first char of UTF-8 doesn't follow the rule */
 167                                         else
 168                                         {
 169                                                 uright[temp] += range;
 170                                         }
 171                                 }
 172
 173                                 if ((strncmp(i, (char *) mb, cursize) >= 0) && (strncmp(i, (char *) uright, cursize) <= 0))
 174                                 {
 175                                         i += cursize - 1;
 176                                         p += cursize - 1;
 177                                         found = true;
 178                                         break;
 179                                 }
 180                         }
 181                         if (found)
 182                                 continue;
 183
 184                         /* 3.3. Check against an UTF-8 intervals: <start character> and <end character>. */
 185                         found = false;
 186                         for (unsigned char * mb = m_additionalUtf8interval; (utf8size(mb) != -1) && (utf8size(mb+utf8size(mb)) != -1)
 187                                 && (mb < m_additionalUtf8interval + sizeof(m_additionalUtf8interval)); mb += (cursize+cursize2) )
 188                         {
 189                                 cursize = utf8size(mb);
 190                                 cursize2= utf8size(mb+cursize);
 191
 192                                 int minlen  = cursize  > ncursize ? ncursize : cursize;
 193                                 int minlen2 = cursize2 > ncursize ? ncursize : cursize2;
 194
 195                                 unsigned char* uright = mb + cursize;
 196
 197                                 if ((strncmp(i, (char *) mb, minlen) >= 0) && (strncmp(i, (char *) uright, minlen2) <= 0))
 198                                 {
 199                                         i += cursize - 1;
 200                                         p += cursize - 1;
 201                                         found = true;
 202                                         break;
 203                                 }
 204                         }
 205                         if (found)
 206                                 continue;
 207                 }
 208
 209                 /* invalid character! abort */
 210                 return false;
 211         }
 212
 213         /* too long? or not -- pointer arithmetic rocks */
 214         return (p < ServerInstance->Config->Limits.NickMax);
 215 }
 216
 217
 218 class ModuleNationalChars : public Module
 219 {
 220         lwbNickHandler myhandler;
 221         std::string charset, casemapping;
 222         unsigned char m_additional[256], m_additionalUp[256], m_lower[256], m_upper[256];
 223         caller1<bool, const std::string&> rememberer;
 224         bool forcequit;
 225         const unsigned char * lowermap_rememberer;
 226
 227  public:
 228         ModuleNationalChars()
 229                 : rememberer(ServerInstance->IsNick), lowermap_rememberer(national_case_insensitive_map)
 230         {
 231         }
 232
 233         void init() CXX11_OVERRIDE
 234         {
 235                 memcpy(m_lower, rfc_case_insensitive_map, 256);
 236                 national_case_insensitive_map = m_lower;
 237
 238                 ServerInstance->IsNick = &myhandler;
 239         }
 240
 241         void On005Numeric(std::map<std::string, std::string>& tokens) CXX11_OVERRIDE
 242         {
 243                 tokens["CASEMAPPING"] = casemapping;
 244         }
 245
 246         void ReadConfig(ConfigStatus& status) CXX11_OVERRIDE
 247         {
 248                 ConfigTag* tag = ServerInstance->Config->ConfValue("nationalchars");
 249                 charset = tag->getString("file");
 250                 casemapping = tag->getString("casemapping", charset);
 251                 if(charset[0] != '/')
 252                         charset.insert(0, "../locales/");
 253                 unsigned char * tables[8] = { m_additional, m_additionalMB, m_additionalUp, m_lower, m_upper, m_additionalUtf8, m_additionalUtf8range, m_additionalUtf8interval };
 254                 loadtables(charset, tables, 8, 5);
 255                 forcequit = tag->getBool("forcequit");
 256                 CheckForceQuit("National character set changed");
 257         }
 258
 259         void CheckForceQuit(const char * message)
 260         {
 261                 if (!forcequit)
 262                         return;
 263
 264                 const UserManager::LocalList& list = ServerInstance->Users.GetLocalUsers();
 265                 for (UserManager::LocalList::const_iterator iter = list.begin(); iter != list.end(); ++iter)
 266                 {
 267                         /* Fix by Brain: Dont quit UID users */
 268                         User* n = *iter;
 269                         if (!isdigit(n->nick[0]) && !ServerInstance->IsNick(n->nick))
 270                                 ServerInstance->Users->QuitUser(n, message);
 271                 }
 272         }
 273
 274         ~ModuleNationalChars()
 275         {
 276                 ServerInstance->IsNick = rememberer;
 277                 national_case_insensitive_map = lowermap_rememberer;
 278                 CheckForceQuit("National characters module unloaded");
 279         }
 280
 281         Version GetVersion() CXX11_OVERRIDE
 282         {
 283                 return Version("Provides an ability to have non-RFC1459 nicks & support for national CASEMAPPING", VF_VENDOR | VF_COMMON, charset);
 284         }
 285
 286         /*make an array to check against it 8bit characters a bit faster. Whether allowed or uppercase (for your needs).*/
 287         void makereverse(unsigned char * from, unsigned  char * to, unsigned int cnt)
 288         {
 289                 memset(to, 0, cnt);
 290                 for(unsigned char * n=from; (*n) && ((*n)<cnt) && (n<from+cnt); n++)
 291                         to[*n] = 1;
 292         }
 293
 294         /*so Bynets Unreal distribution stuff*/
 295         void loadtables(std::string filename, unsigned char ** tables, unsigned char cnt, char faillimit)
 296         {
 297                 std::ifstream ifs(ServerInstance->Config->Paths.PrependConfig(filename).c_str());
 298                 if (ifs.fail())
 299                 {
 300                         ServerInstance->Logs->Log(MODNAME, LOG_DEFAULT, "loadtables() called for missing file: %s", filename.c_str());
 301                         return;
 302                 }
 303
 304                 for (unsigned char n=0; n< cnt; n++)
 305                 {
 306                         memset(tables[n], 0, 256);
 307                 }
 308
 309                 memcpy(m_lower, rfc_case_insensitive_map, 256);
 310
 311                 for (unsigned char n = 0; n < cnt; n++)
 312                 {
 313                         if (loadtable(ifs, tables[n], 255) && (n < faillimit))
 314                         {
 315                                 ServerInstance->Logs->Log(MODNAME, LOG_DEFAULT, "loadtables() called for illegal file: %s (line %d)", filename.c_str(), n+1);
 316                                 return;
 317                         }
 318                 }
 319
 320                 makereverse(m_additional, m_reverse_additional, sizeof(m_additional));
 321         }
 322
 323         unsigned char symtoi(const char *t,unsigned char base)
 324         /* base = 16 for hexadecimal, 10 for decimal, 8 for octal ;) */
 325         {
 326                 unsigned char tmp = 0, current;
 327                 while ((*t) && (*t !=' ') && (*t != 13) && (*t != 10) && (*t != ','))
 328                 {
 329                         tmp *= base;
 330                         current = ascii_case_insensitive_map[(unsigned char)*t];
 331                         if (current >= 'a')
 332                                 current = current - 'a' + 10;
 333                         else
 334                                 current = current - '0';
 335                         tmp+=current;
 336                         t++;
 337                 }
 338                 return tmp;
 339         }
 340
 341         int loadtable(std::ifstream &ifs , unsigned char *chartable, unsigned int maxindex)
 342         {
 343                 std::string buf;
 344                 getline(ifs, buf);
 345
 346                 unsigned int i = 0;
 347                 int fail = 0;
 348
 349                 buf.erase(buf.find_last_not_of("\n") + 1);
 350
 351                 if (buf[0] == '.')      /* simple plain-text string after dot */
 352                 {
 353                         i = buf.size() - 1;
 354
 355                         if (i > (maxindex + 1))
 356                                 i = maxindex + 1;
 357
 358                         memcpy(chartable, buf.c_str() + 1, i);
 359                 }
 360                 else
 361                 {
 362                         const char * p = buf.c_str();
 363                         while (*p)
 364                         {
 365                                 if (i > maxindex)
 366                                 {
 367                                         fail = 1;
 368                                         break;
 369                                 }
 370
 371                                 if (*p != '\'')         /* decimal or hexadecimal char code */
 372                                 {
 373                                         if (*p == '0')
 374                                         {
 375                                                 if (p[1] == 'x')
 376                                                          /* hex with the leading "0x" */
 377                                                         chartable[i] = symtoi(p + 2, 16);
 378                                                 else
 379                                                         chartable[i] = symtoi(p + 1, 8);
 380                                         }
 381                                         /* hex form */
 382                                         else if (*p == 'x')
 383                                         {
 384                                                 chartable[i] = symtoi(p + 1, 16);
 385                                         }else    /* decimal form */
 386                                         {
 387                                                 chartable[i] = symtoi(p, 10);
 388                                         }
 389                                 }
 390                                 else             /* plain-text char between '' */
 391                                 {
 392                                         if (*(p + 1) == '\\')
 393                                         {
 394                                                 chartable[i] = *(p + 2);
 395                                                 p += 3;
 396                                         }else
 397                                         {
 398                                                 chartable[i] = *(p + 1);
 399                                                 p += 2;
 400                                         }
 401                                 }
 402                                 while (*p && (*p != ',') && (*p != ' ') && (*p != 13) && (*p != 10))
 403                                         p++;
 404                                 while (*p && ((*p == ',') || (*p == ' ') || (*p == 13) || (*p == 10)))
 405                                         p++;
 406                                 i++;
 407                         }
 408                 }
 409                 return fail;
 410         }
 411 };
 412
 413 MODULE_INIT(ModuleNationalChars)