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 std::string&, size_t>
  35 {
  36  public:
  37         lwbNickHandler() { }
  38         virtual ~lwbNickHandler() { }
  39         virtual bool Call(const std::string&, 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 std::string& nick, size_t max)
  75 {
  76         if (nick.empty())
  77                 return false;
  78
  79         const char* n = nick.c_str();
  80         unsigned int p = 0;
  81         for (const char* i = n; *i; i++, p++)
  82         {
  83                 /* 1. Multibyte encodings support:  */
  84                 /* 1.1. 16bit char. areas, e.g. chinese:*/
  85
  86                 /* if current character is the last, we DO NOT check it against multibyte table */
  87                 /* if there are mbtable ranges, use ONLY them. No 8bit at all */
  88                 if (i[1] && m_additionalMB[0])
  89                 {
  90                         /* otherwise let's take a look at the current character and the following one */
  91                         bool found = false;
  92                         for(unsigned char * mb = m_additionalMB; (*mb) && (mb < m_additionalMB + sizeof(m_additionalMB)); mb += 4)
  93                         {
  94                                 if ( (i[0] >= mb[0]) && (i[0] <= mb[1]) && (i[1] >= mb[2]) && (i[1] <= mb[3]) )
  95                                 {
  96                                         /* multibyte range character found */
  97                                         i++;
  98                                         p++;
  99                                         found = true;
 100                                         break;
 101                                 }
 102                         }
 103                         if (found)
 104                                 /* next char! */
 105                                 continue;
 106                         else
 107                                 /* there are ranges, but incorrect char (8bit?) given, sorry */
 108                                 return false;
 109                 }
 110
 111                 /* 2. 8bit character support */
 112                 if (((*i >= 'A') && (*i <= '}')) || m_reverse_additional[(unsigned char)*i])
 113                         /* "A"-"}" can occur anywhere in a nickname */
 114                         continue;
 115
 116                 if ((((*i >= '0') && (*i <= '9')) || (*i == '-')) && (i > n))
 117                         /* "0"-"9", "-" can occur anywhere BUT the first char of a nickname */
 118                         continue;
 119
 120                 /* 3.1. Check against a simple UTF-8 characters enumeration */
 121                 int cursize, cursize2, ncursize = utf8size((unsigned char *)i);
 122                 /* do check only if current multibyte character is valid UTF-8 only */
 123                 if (ncursize != -1)
 124                 {
 125                         bool found = false;
 126                         for (unsigned char * mb = m_additionalUtf8; (utf8size(mb) != -1) && (mb < m_additionalUtf8 + sizeof(m_additionalUtf8)); mb += cursize)
 127                         {
 128                                 cursize = utf8size(mb);
 129                                 /* Size differs? Pick the next! */
 130                                 if (cursize != ncursize)
 131                                         continue;
 132
 133                                 if (!strncmp(i, (char *)mb, cursize))
 134                                 {
 135                                         i += cursize - 1;
 136                                         p += cursize - 1;
 137                                         found = true;
 138                                         break;
 139                                 }
 140                         }
 141                         if (found)
 142                                 continue;
 143
 144                         /* 3.2. Check against an UTF-8 ranges: <start character> and <length of the range>. */
 145                         found = false;
 146                         for (unsigned char * mb = m_additionalUtf8range; (utf8size(mb) != -1) && (mb < m_additionalUtf8range + sizeof(m_additionalUtf8range)); mb += cursize + 1)
 147                         {
 148                                 cursize = utf8size(mb);
 149                                 /* Size differs (or lengthbyte is zero)? Pick the next! */
 150                                 if ((cursize != ncursize) || (!mb[cursize]))
 151                                         continue;
 152
 153                                 unsigned char uright[5] = {0,0,0,0,0}, range = mb[cursize] - 1;
 154                                 strncpy((char* ) uright, (char *) mb, cursize);
 155
 156                                 for (int temp = cursize - 1; (temp >= 0) && range; --temp)
 157                                 {
 158                                         /* all but the first char are 64-based */
 159                                         if (temp)
 160                                         {
 161                                                 char part64 = range & 63; /* i.e. % 64 */
 162                                                 /* handle carrying over */
 163                                                 if (uright[temp] + part64 - 1 > 191)
 164                                                 {
 165                                                         uright[temp] -= 64;
 166                                                         range += 64;
 167                                                 }
 168                                                 uright[temp] += part64;
 169                                                 range >>= 6; /* divide it on a 64 */
 170                                         }
 171                                         /* the first char of UTF-8 doesn't follow the rule */
 172                                         else
 173                                         {
 174                                                 uright[temp] += range;
 175                                         }
 176                                 }
 177
 178                                 if ((strncmp(i, (char *) mb, cursize) >= 0) && (strncmp(i, (char *) uright, cursize) <= 0))
 179                                 {
 180                                         i += cursize - 1;
 181                                         p += cursize - 1;
 182                                         found = true;
 183                                         break;
 184                                 }
 185                         }
 186                         if (found)
 187                                 continue;
 188
 189                         /* 3.3. Check against an UTF-8 intervals: <start character> and <end character>. */
 190                         found = false;
 191                         for (unsigned char * mb = m_additionalUtf8interval; (utf8size(mb) != -1) && (utf8size(mb+utf8size(mb)) != -1)
 192                                 && (mb < m_additionalUtf8interval + sizeof(m_additionalUtf8interval)); mb += (cursize+cursize2) )
 193                         {
 194                                 cursize = utf8size(mb);
 195                                 cursize2= utf8size(mb+cursize);
 196
 197                                 int minlen  = cursize  > ncursize ? ncursize : cursize;
 198                                 int minlen2 = cursize2 > ncursize ? ncursize : cursize2;
 199
 200                                 unsigned char* uright = mb + cursize;
 201
 202                                 if ((strncmp(i, (char *) mb, minlen) >= 0) && (strncmp(i, (char *) uright, minlen2) <= 0))
 203                                 {
 204                                         i += cursize - 1;
 205                                         p += cursize - 1;
 206                                         found = true;
 207                                         break;
 208                                 }
 209                         }
 210                         if (found)
 211                                 continue;
 212                 }
 213
 214                 /* invalid character! abort */
 215                 return false;
 216         }
 217
 218         /* too long? or not -- pointer arithmetic rocks */
 219         return (p < max);
 220 }
 221
 222
 223 class ModuleNationalChars : public Module
 224 {
 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 std::string&, size_t> rememberer;
 229         bool forcequit;
 230         const unsigned char * lowermap_rememberer;
 231
 232  public:
 233         ModuleNationalChars()
 234                 : rememberer(ServerInstance->IsNick), lowermap_rememberer(national_case_insensitive_map)
 235         {
 236         }
 237
 238         void init()
 239         {
 240                 memcpy(m_lower, rfc_case_insensitive_map, 256);
 241                 national_case_insensitive_map = m_lower;
 242
 243                 ServerInstance->IsNick = &myhandler;
 244
 245                 Implementation eventlist[] = { I_OnRehash, I_On005Numeric };
 246                 ServerInstance->Modules->Attach(eventlist, this, sizeof(eventlist)/sizeof(Implementation));
 247                 OnRehash(NULL);
 248         }
 249
 250         virtual void On005Numeric(std::string &output)
 251         {
 252                 std::string tmp(casemapping);
 253                 tmp.insert(0, "CASEMAPPING=");
 254                 SearchAndReplace(output, std::string("CASEMAPPING=rfc1459"), tmp);
 255         }
 256
 257         virtual void OnRehash(User* user)
 258         {
 259                 ConfigTag* tag = ServerInstance->Config->ConfValue("nationalchars");
 260                 charset = tag->getString("file");
 261                 casemapping = tag->getString("casemapping", charset);
 262                 if(charset[0] != '/')
 263                         charset.insert(0, "../locales/");
 264                 unsigned char * tables[8] = { m_additional, m_additionalMB, m_additionalUp, m_lower, m_upper, m_additionalUtf8, m_additionalUtf8range, m_additionalUtf8interval };
 265                 loadtables(charset, tables, 8, 5);
 266                 forcequit = tag->getBool("forcequit");
 267                 CheckForceQuit("National character set changed");
 268         }
 269
 270         void CheckForceQuit(const char * message)
 271         {
 272                 if (!forcequit)
 273                         return;
 274
 275                 for (LocalUserList::const_iterator iter = ServerInstance->Users->local_users.begin(); iter != ServerInstance->Users->local_users.end(); ++iter)
 276                 {
 277                         /* Fix by Brain: Dont quit UID users */
 278                         User* n = *iter;
 279                         if (!isdigit(n->nick[0]) && !ServerInstance->IsNick(n->nick, ServerInstance->Config->Limits.NickMax))
 280                                 ServerInstance->Users->QuitUser(n, message);
 281                 }
 282         }
 283
 284         virtual ~ModuleNationalChars()
 285         {
 286                 ServerInstance->IsNick = rememberer;
 287                 national_case_insensitive_map = lowermap_rememberer;
 288                 CheckForceQuit("National characters module unloaded");
 289         }
 290
 291         virtual Version GetVersion()
 292         {
 293                 return Version("Provides an ability to have non-RFC1459 nicks & support for national CASEMAPPING", VF_VENDOR | VF_COMMON, charset);
 294         }
 295
 296         /*make an array to check against it 8bit characters a bit faster. Whether allowed or uppercase (for your needs).*/
 297         void makereverse(unsigned char * from, unsigned  char * to, unsigned int cnt)
 298         {
 299                 memset(to, 0, cnt);
 300                 for(unsigned char * n=from; (*n) && ((*n)<cnt) && (n<from+cnt); n++)
 301                         to[*n] = 1;
 302         }
 303
 304         /*so Bynets Unreal distribution stuff*/
 305         void loadtables(std::string filename, unsigned char ** tables, unsigned char cnt, char faillimit)
 306         {
 307                 std::ifstream ifs(filename.c_str());
 308                 if (ifs.fail())
 309                 {
 310                         ServerInstance->Logs->Log("m_nationalchars",DEFAULT,"loadtables() called for missing file: %s", filename.c_str());
 311                         return;
 312                 }
 313
 314                 for (unsigned char n=0; n< cnt; n++)
 315                 {
 316                         memset(tables[n], 0, 256);
 317                 }
 318
 319                 memcpy(m_lower, rfc_case_insensitive_map, 256);
 320
 321                 for (unsigned char n = 0; n < cnt; n++)
 322                 {
 323                         if (loadtable(ifs, tables[n], 255) && (n < faillimit))
 324                         {
 325                                 ServerInstance->Logs->Log("m_nationalchars",DEFAULT,"loadtables() called for illegal file: %s (line %d)", filename.c_str(), n+1);
 326                                 return;
 327                         }
 328                 }
 329
 330                 makereverse(m_additional, m_reverse_additional, sizeof(m_additional));
 331         }
 332
 333         unsigned char symtoi(const char *t,unsigned char base)
 334         /* base = 16 for hexadecimal, 10 for decimal, 8 for octal ;) */
 335         {
 336                 unsigned char tmp = 0, current;
 337                 while ((*t) && (*t !=' ') && (*t != 13) && (*t != 10) && (*t != ','))
 338                 {
 339                         tmp *= base;
 340                         current = ascii_case_insensitive_map[(unsigned char)*t];
 341                         if (current >= 'a')
 342                                 current = current - 'a' + 10;
 343                         else
 344                                 current = current - '0';
 345                         tmp+=current;
 346                         t++;
 347                 }
 348                 return tmp;
 349         }
 350
 351         int loadtable(std::ifstream &ifs , unsigned char *chartable, unsigned int maxindex)
 352         {
 353                 std::string buf;
 354                 getline(ifs, buf);
 355
 356                 unsigned int i = 0;
 357                 int fail = 0;
 358
 359                 buf.erase(buf.find_last_not_of("\n") + 1);
 360
 361                 if (buf[0] == '.')      /* simple plain-text string after dot */
 362                 {
 363                         i = buf.size() - 1;
 364
 365                         if (i > (maxindex + 1))
 366                                 i = maxindex + 1;
 367
 368                         memcpy(chartable, buf.c_str() + 1, i);
 369                 }
 370                 else
 371                 {
 372                         const char * p = buf.c_str();
 373                         while (*p)
 374                         {
 375                                 if (i > maxindex)
 376                                 {
 377                                         fail = 1;
 378                                         break;
 379                                 }
 380
 381                                 if (*p != '\'')         /* decimal or hexadecimal char code */
 382                                 {
 383                                         if (*p == '0')
 384                                         {
 385                                                 if (p[1] == 'x')
 386                                                          /* hex with the leading "0x" */
 387                                                         chartable[i] = symtoi(p + 2, 16);
 388                                                 else
 389                                                         chartable[i] = symtoi(p + 1, 8);
 390                                         }
 391                                         /* hex form */
 392                                         else if (*p == 'x')
 393                                         {
 394                                                 chartable[i] = symtoi(p + 1, 16);
 395                                         }else    /* decimal form */
 396                                         {
 397                                                 chartable[i] = symtoi(p, 10);
 398                                         }
 399                                 }
 400                                 else             /* plain-text char between '' */
 401                                 {
 402                                         if (*(p + 1) == '\\')
 403                                         {
 404                                                 chartable[i] = *(p + 2);
 405                                                 p += 3;
 406                                         }else
 407                                         {
 408                                                 chartable[i] = *(p + 1);
 409                                                 p += 2;
 410                                         }
 411                                 }
 412                                 while (*p && (*p != ',') && (*p != ' ') && (*p != 13) && (*p != 10))
 413                                         p++;
 414                                 while (*p && ((*p == ',') || (*p == ' ') || (*p == 13) || (*p == 10)))
 415                                         p++;
 416                                 i++;
 417                         }
 418                 }
 419                 return fail;
 420         }
 421 };
 422
 423 MODULE_INIT(ModuleNationalChars)