src/modules/m_nationalchars.cpp

   1 /*       +------------------------------------+
   2  *       | Inspire Internet Relay Chat Daemon |
   3  *       +------------------------------------+
   4  *
   5  *  InspIRCd: (C) 2002-2009 InspIRCd Development Team
   6  * See: http://wiki.inspircd.org/Credits
   7  *
   8  * This program is free but copyrighted software; see
   9  *          the file COPYING for details.
  10  *
  11  * ---------------------------------------------------
  12  */
  13
  14 /* Contains a code of Unreal IRCd + Bynets patch ( http://www.unrealircd.com/ and http://www.bynets.org/ )
  15    Original patch is made by Dmitry "Killer{R}" Kononko. ( http://killprog.com/ )
  16    Changed at 2008-06-15 - 2009-02-11
  17    by Chernov-Phoenix Alexey (Phoenix@RusNet) mailto:phoenix /email address separator/ pravmail.ru */
  18
  19 #include "inspircd.h"
  20 #include "caller.h"
  21 #include <fstream>
  22
  23 /* $ModDesc: Provides an ability to have non-RFC1459 nicks & support for national CASEMAPPING */
  24
  25 class lwbNickHandler : public HandlerBase2<bool, const char*, size_t>
  26 {
  27  public:
  28         lwbNickHandler() { }
  29         virtual ~lwbNickHandler() { }
  30         virtual bool Call(const char*, size_t);
  31 };
  32
  33                                                                  /*,m_reverse_additionalUp[256];*/
  34 static unsigned char m_reverse_additional[256],m_additionalMB[256],m_additionalUtf8[256],m_additionalUtf8range[256],m_additionalUtf8interval[256];
  35
  36 char utf8checkrest(unsigned char * mb, unsigned char cnt)
  37 {
  38         for (unsigned char * tmp=mb; tmp<mb+cnt; tmp++)
  39         {
  40                 /* & is faster! -- Phoenix (char & b11000000 == b10000000) */
  41                 if ((*tmp & 192) != 128)
  42                         return -1;
  43         }
  44         return cnt + 1;
  45 }
  46
  47
  48 char utf8size(unsigned char * mb)
  49 {
  50         if (!*mb)
  51                 return -1;
  52         if (!(*mb & 128))
  53                 return 1;
  54         if ((*mb & 224) == 192)
  55                 return utf8checkrest(mb + 1,1);
  56         if ((*mb & 240) == 224)
  57                 return utf8checkrest(mb + 1,2);
  58         if ((*mb & 248) == 240)
  59                 return utf8checkrest(mb + 1,3);
  60         return -1;
  61 }
  62
  63
  64 /* Conditions added */
  65 bool lwbNickHandler::Call(const char* n, size_t max)
  66 {
  67         if (!n || !*n)
  68                 return false;
  69
  70         unsigned int p = 0;
  71         for (const char* i = n; *i; i++, p++)
  72         {
  73                 /* 1. Multibyte encodings support:  */
  74                 /* 1.1. 16bit char. areas, e.g. chinese:*/
  75
  76                 /* if current character is the last, we DO NOT check it against multibyte table */
  77                 /* if there are mbtable ranges, use ONLY them. No 8bit at all */
  78                 if (i[1] && m_additionalMB[0])
  79                 {
  80                         /* otherwise let's take a look at the current character and the following one */
  81                         bool found = false;
  82                         for(unsigned char * mb = m_additionalMB; (*mb) && (mb < m_additionalMB + sizeof(m_additionalMB)); mb += 4)
  83                         {
  84                                 if ( (i[0] >= mb[0]) && (i[0] <= mb[1]) && (i[1] >= mb[2]) && (i[1] <= mb[3]) )
  85                                 {
  86                                         /* multibyte range character found */
  87                                         i++;
  88                                         p++;
  89                                         found = true;
  90                                         break;
  91                                 }
  92                         }
  93                         if (found)
  94                                 /* next char! */
  95                                 continue;
  96                         else
  97                                 /* there are ranges, but incorrect char (8bit?) given, sorry */
  98                                 return false;
  99                 }
 100
 101                 /* 2. 8bit character support */
 102                 if (((*i >= 'A') && (*i <= '}')) || m_reverse_additional[(unsigned char)*i])
 103                         /* "A"-"}" can occur anywhere in a nickname */
 104                         continue;
 105
 106                 if ((((*i >= '0') && (*i <= '9')) || (*i == '-')) && (i > n))
 107                         /* "0"-"9", "-" can occur anywhere BUT the first char of a nickname */
 108                         continue;
 109
 110                 /* 3.1. Check against a simple UTF-8 characters enumeration */
 111                 int cursize, cursize2, ncursize = utf8size((unsigned char *)i);
 112                 /* do check only if current multibyte character is valid UTF-8 only */
 113                 if (ncursize != -1)
 114                 {
 115                         bool found = false;
 116                         for (unsigned char * mb = m_additionalUtf8; (utf8size(mb) != -1) && (mb < m_additionalUtf8 + sizeof(m_additionalUtf8)); mb += cursize)
 117                         {
 118                                 cursize = utf8size(mb);
 119                                 /* Size differs? Pick the next! */
 120                                 if (cursize != ncursize)
 121                                         continue;
 122
 123                                 if (!strncmp(i, (char *)mb, cursize))
 124                                 {
 125                                         i += cursize - 1;
 126                                         p += cursize - 1;
 127                                         found = true;
 128                                         break;
 129                                 }
 130                         }
 131                         if (found)
 132                                 continue;
 133
 134                         /* 3.2. Check against an UTF-8 ranges: <start character> and <length of the range>. */
 135                         found = false;
 136                         for (unsigned char * mb = m_additionalUtf8range; (utf8size(mb) != -1) && (mb < m_additionalUtf8range + sizeof(m_additionalUtf8range)); mb += cursize + 1)
 137                         {
 138                                 cursize = utf8size(mb);
 139                                 /* Size differs (or lengthbyte is zero)? Pick the next! */
 140                                 if ((cursize != ncursize) || (!mb[cursize]))
 141                                         continue;
 142
 143                                 unsigned char uright[5] = {0,0,0,0,0}, range = mb[cursize] - 1;
 144                                 strncpy((char* ) uright, (char *) mb, cursize);
 145
 146                                 for (int temp = cursize - 1; (temp >= 0) && range; --temp)
 147                                 {
 148                                         /* all but the first char are 64-based */
 149                                         if (temp)
 150                                         {
 151                                                 char part64 = range & 63; /* i.e. % 64 */
 152                                                 /* handle carrying over */
 153                                                 if (uright[temp] + part64 - 1 > 191)
 154                                                 {
 155                                                         uright[temp] -= 64;
 156                                                         range += 64;
 157                                                 }
 158                                                 uright[temp] += part64;
 159                                                 range >>= 6; /* divide it on a 64 */
 160                                         }
 161                                         /* the first char of UTF-8 doesn't follow the rule */
 162                                         else
 163                                         {
 164                                                 uright[temp] += range;
 165                                         }
 166                                 }
 167
 168                                 if ((strncmp(i, (char *) mb, cursize) >= 0) && (strncmp(i, (char *) uright, cursize) <= 0))
 169                                 {
 170                                         i += cursize - 1;
 171                                         p += cursize - 1;
 172                                         found = true;
 173                                         break;
 174                                 }
 175                         }
 176                         if (found)
 177                                 continue;
 178
 179                         /* 3.3. Check against an UTF-8 intervals: <start character> and <end character>. */
 180                         found = false;
 181                         for (unsigned char * mb = m_additionalUtf8interval; (utf8size(mb) != -1) && (utf8size(mb+utf8size(mb)) != -1)
 182                                 && (mb < m_additionalUtf8interval + sizeof(m_additionalUtf8interval)); mb += (cursize+cursize2) )
 183                         {
 184                                 cursize = utf8size(mb);
 185                                 cursize2= utf8size(mb+cursize);
 186
 187                                 int minlen  = cursize  > ncursize ? ncursize : cursize;
 188                                 int minlen2 = cursize2 > ncursize ? ncursize : cursize2;
 189
 190                                 unsigned char* uright = mb + cursize;
 191
 192                                 if ((strncmp(i, (char *) mb, minlen) >= 0) && (strncmp(i, (char *) uright, minlen2) <= 0))
 193                                 {
 194                                         i += cursize - 1;
 195                                         p += cursize - 1;
 196                                         found = true;
 197                                         break;
 198                                 }
 199                         }
 200                         if (found)
 201                                 continue;
 202                 }
 203
 204                 /* invalid character! abort */
 205                 return false;
 206         }
 207
 208         /* too long? or not -- pointer arithmetic rocks */
 209         return (p < max);
 210 }
 211
 212
 213 class ModuleNationalChars : public Module
 214 {
 215  private:
 216         lwbNickHandler myhandler;
 217         std::string charset, casemapping;
 218         unsigned char m_additional[256], m_additionalUp[256], m_lower[256], m_upper[256];
 219         caller2<bool, const char*, size_t> rememberer;
 220         bool forcequit;
 221         const unsigned char * lowermap_rememberer;
 222
 223  public:
 224         ModuleNationalChars() : rememberer(ServerInstance->IsNick)
 225         {
 226                 lowermap_rememberer = national_case_insensitive_map;
 227                 memcpy(m_lower, rfc_case_insensitive_map, 256);
 228                 national_case_insensitive_map = m_lower;
 229
 230                 ServerInstance->IsNick = &myhandler;
 231
 232                 Implementation eventlist[] = { I_OnRehash, I_On005Numeric };
 233                 ServerInstance->Modules->Attach(eventlist, this, 2);
 234                 OnRehash(NULL);
 235         }
 236
 237         virtual void On005Numeric(std::string &output)
 238         {
 239                 std::string tmp(casemapping);
 240                 tmp.insert(0, "CASEMAPPING=");
 241                 SearchAndReplace(output, std::string("CASEMAPPING=rfc1459"), tmp);
 242         }
 243
 244         virtual void OnRehash(User* user)
 245         {
 246                 ConfigReader* conf = new ConfigReader;
 247                 charset = conf->ReadValue("nationalchars", "file", 0);
 248                 casemapping = conf->ReadValue("nationalchars", "casemapping", charset, 0, false);
 249                 charset.insert(0, "../locales/");
 250                 unsigned char * tables[8] = { m_additional, m_additionalMB, m_additionalUp, m_lower, m_upper, m_additionalUtf8, m_additionalUtf8range, m_additionalUtf8interval };
 251                 loadtables(charset, tables, 8, 5);
 252                 forcequit = conf->ReadFlag("nationalchars", "forcequit", 0);
 253                 CheckForceQuit("National character set changed");
 254                 delete conf;
 255         }
 256
 257         void CheckForceQuit(const char * message)
 258         {
 259                 if (!forcequit)
 260                         return;
 261
 262                 for (std::vector<User*>::iterator iter = ServerInstance->Users->local_users.begin(); iter != ServerInstance->Users->local_users.end(); ++iter)
 263                 {
 264                         /* Fix by Brain: Dont quit UID users */
 265                         User* n = *iter;
 266                         if (!isdigit(n->nick[0]) && !ServerInstance->IsNick(n->nick.c_str(), ServerInstance->Config->Limits.NickMax))
 267                                 ServerInstance->Users->QuitUser(n, message);
 268                 }
 269         }
 270
 271         virtual ~ModuleNationalChars()
 272         {
 273                 ServerInstance->IsNick = rememberer;
 274                 national_case_insensitive_map = lowermap_rememberer;
 275                 CheckForceQuit("National characters module unloaded");
 276         }
 277
 278         virtual Version GetVersion()
 279         {
 280                 return Version("Provides an ability to have non-RFC1459 nicks & support for national CASEMAPPING",VF_COMMON,API_VERSION);
 281         }
 282
 283         /*make an array to check against it 8bit characters a bit faster. Whether allowed or uppercase (for your needs).*/
 284         void makereverse(unsigned char * from, unsigned  char * to, unsigned int cnt)
 285         {
 286                 memset(to, 0, cnt);
 287                 for(unsigned char * n=from; (*n) && ((*n)<cnt) && (n<from+cnt); n++)
 288                         to[*n] = 1;
 289         }
 290
 291         /*so Bynets Unreal distribution stuff*/
 292         void loadtables(std::string filename, unsigned char ** tables, unsigned char cnt, char faillimit)
 293         {
 294                 std::ifstream ifs(filename.c_str());
 295                 if (ifs.fail())
 296                 {
 297                         ServerInstance->Logs->Log("m_nationalchars",DEFAULT,"loadtables() called for missing file: %s", filename.c_str());
 298                         return;
 299                 }
 300
 301                 for (unsigned char n=0; n< cnt; n++)
 302                 {
 303                         memset(tables[n], 0, 256);
 304                 }
 305
 306                 memcpy(m_lower, rfc_case_insensitive_map, 256);
 307
 308                 for (unsigned char n = 0; n < cnt; n++)
 309                 {
 310                         if (loadtable(ifs, tables[n], 255) && (n < faillimit))
 311                         {
 312                                 ServerInstance->Logs->Log("m_nationalchars",DEFAULT,"loadtables() called for illegal file: %s (line %d)", filename.c_str(), n+1);
 313                                 return;
 314                         }
 315                 }
 316
 317                 makereverse(m_additional, m_reverse_additional, sizeof(m_additional));
 318         }
 319
 320         unsigned char symtoi(const char *t,unsigned char base)
 321         /* base = 16 for hexadecimal, 10 for decimal, 8 for octal ;) */
 322         {
 323                 unsigned char tmp = 0, current;
 324                 while ((*t) && (*t !=' ') && (*t != 13) && (*t != 10) && (*t != ','))
 325                 {
 326                         tmp *= base;
 327                         current = ascii_case_insensitive_map[(unsigned char)*t];
 328                         if (current >= 'a')
 329                                 current = current - 'a' + 10;
 330                         else
 331                                 current = current - '0';
 332                         tmp+=current;
 333                         t++;
 334                 }
 335                 return tmp;
 336         }
 337
 338         int loadtable(std::ifstream &ifs , unsigned char *chartable, unsigned int maxindex)
 339         {
 340                 std::string buf;
 341                 getline(ifs, buf);
 342
 343                 unsigned int i = 0;
 344                 int fail = 0;
 345
 346                 buf.erase(buf.find_last_not_of("\n") + 1);
 347
 348                 if (buf[0] == '.')      /* simple plain-text string after dot */
 349                 {
 350                         i = buf.size() - 1;
 351
 352                         if (i > (maxindex + 1))
 353                                 i = maxindex + 1;
 354
 355                         memcpy(chartable, buf.c_str() + 1, i);
 356                 }
 357                 else
 358                 {
 359                         const char * p = buf.c_str();
 360                         while (*p)
 361                         {
 362                                 if (i > maxindex)
 363                                 {
 364                                         fail = 1;
 365                                         break;
 366                                 }
 367
 368                                 if (*p != '\'')         /* decimal or hexadecimal char code */
 369                                 {
 370                                         if (*p == '0')
 371                                         {
 372                                                 if (p[1] == 'x')
 373                                                          /* hex with the leading "0x" */
 374                                                         chartable[i] = symtoi(p + 2, 16);
 375                                                 else
 376                                                         chartable[i] = symtoi(p + 1, 8);
 377                                         }
 378                                         /* hex form */
 379                                         else if (*p == 'x')
 380                                         {
 381                                                 chartable[i] = symtoi(p + 1, 16);
 382                                         }else    /* decimal form */
 383                                         {
 384                                                 chartable[i] = symtoi(p, 10);
 385                                         }
 386                                 }
 387                                 else             /* plain-text char between '' */
 388                                 {
 389                                         if (*(p + 1) == '\\')
 390                                         {
 391                                                 chartable[i] = *(p + 2);
 392                                                 p += 3;
 393                                         }else
 394                                         {
 395                                                 chartable[i] = *(p + 1);
 396                                                 p += 2;
 397                                         }
 398                                 }
 399                                 while (*p && (*p != ',') && (*p != ' ') && (*p != 13) && (*p != 10))
 400                                         p++;
 401                                 while (*p && ((*p == ',') || (*p == ' ') || (*p == 13) || (*p == 10)))
 402                                         p++;
 403                                 i++;
 404                         }
 405                 }
 406                 return fail;
 407         }
 408 };
 409
 410 MODULE_INIT(ModuleNationalChars)