tools/wmc/mcl.c

   1 /*
   2  * Wine Message Compiler lexical scanner
   3  *
   4  * Copyright 2000 Bertho A. Stultiens (BS)
   5  *
   6  * This library is free software; you can redistribute it and/or
   7  * modify it under the terms of the GNU Lesser General Public
   8  * License as published by the Free Software Foundation; either
   9  * version 2.1 of the License, or (at your option) any later version.
  10  *
  11  * This library is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14  * Lesser General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU Lesser General Public
  17  * License along with this library; if not, write to the Free Software
  18  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
  19  */
  20
  21 #include "config.h"
  22
  23 #include <stdio.h>
  24 #include <stdlib.h>
  25 #include <ctype.h>
  26 #include <assert.h>
  27 #include <string.h>
  28
  29 #include "utils.h"
  30 #include "wmc.h"
  31 #include "lang.h"
  32
  33 #include "mcy.tab.h"
  34
  35 /*
  36  * Keywords are case insensitive. All normal input is treated as
  37  * being in codepage iso-8859-1 for ascii input files (unicode
  38  * page 0) and as equivalent unicode if unicode input is selected.
  39  * All normal input, which is not part of a message text, is
  40  * enforced to be unicode page 0. Otherwise an error will be
  41  * generated. The normal file data should only be ASCII because
  42  * that is the basic definition of the grammar.
  43  *
  44  * Byteorder or unicode input is determined automatically by
  45  * reading the first 8 bytes and checking them against unicode
  46  * page 0 byteorder (hibyte must be 0).
  47  * -- FIXME --
  48  * Alternatively, the input is checked against a special byte
  49  * sequence to identify the file.
  50  * -- FIXME --
  51  *
  52  *
  53  * Keywords:
  54  *      Codepages
  55  *      Facility
  56  *      FacilityNames
  57  *      LanguageNames
  58  *      MessageId
  59  *      MessageIdTypedef
  60  *      Severity
  61  *      SeverityNames
  62  *      SymbolicName
  63  *
  64  * Default added identifiers for classes:
  65  * SeverityNames:
  66  *      Success         = 0x0
  67  *      Informational   = 0x1
  68  *      Warning         = 0x2
  69  *      Error           = 0x3
  70  * FacilityNames:
  71  *      System          = 0x0FF
  72  *      Application     = 0xFFF
  73  *
  74  * The 'Codepages' keyword is a wmc extension.
  75  */
  76
  77 static const WCHAR ustr_application[]   = { 'A', 'p', 'p', 'l', 'i', 'c', 'a', 't', 'i', 'o', 'n', 0 };
  78 static const WCHAR ustr_codepages[]     = { 'C', 'o', 'd', 'e', 'p', 'a', 'g', 'e', 's', 0 };
  79 static const WCHAR ustr_english[]       = { 'E', 'n', 'g', 'l', 'i', 's', 'h', 0 };
  80 static const WCHAR ustr_error[]         = { 'E', 'r', 'r', 'o', 'r', 0 };
  81 static const WCHAR ustr_facility[]      = { 'F', 'a', 'c', 'i', 'l', 'i', 't', 'y', 0 };
  82 static const WCHAR ustr_facilitynames[] = { 'F', 'a', 'c', 'i', 'l', 'i', 't', 'y', 'N', 'a', 'm', 'e', 's', 0 };
  83 static const WCHAR ustr_informational[] = { 'I', 'n', 'f', 'o', 'r', 'm', 'a', 't', 'i', 'o', 'n', 'a', 'l', 0 };
  84 static const WCHAR ustr_language[]      = { 'L', 'a', 'n', 'g', 'u', 'a', 'g', 'e', 0};
  85 static const WCHAR ustr_languagenames[] = { 'L', 'a', 'n', 'g', 'u', 'a', 'g', 'e', 'N', 'a', 'm', 'e', 's', 0};
  86 static const WCHAR ustr_messageid[]     = { 'M', 'e', 's', 's', 'a', 'g', 'e', 'I', 'd', 0 };
  87 static const WCHAR ustr_messageidtypedef[] = { 'M', 'e', 's', 's', 'a', 'g', 'e', 'I', 'd', 'T', 'y', 'p', 'e', 'd', 'e', 'f', 0 };
  88 static const WCHAR ustr_outputbase[]    = { 'O', 'u', 't', 'p', 'u', 't', 'B', 'a', 's', 'e', 0 };
  89 static const WCHAR ustr_severity[]      = { 'S', 'e', 'v', 'e', 'r', 'i', 't', 'y', 0 };
  90 static const WCHAR ustr_severitynames[] = { 'S', 'e', 'v', 'e', 'r', 'i', 't', 'y', 'N', 'a', 'm', 'e', 's', 0 };
  91 static const WCHAR ustr_success[]       = { 'S', 'u', 'c', 'c', 'e', 's', 's', 0 };
  92 static const WCHAR ustr_symbolicname[]  = { 'S', 'y', 'm', 'b', 'o', 'l', 'i', 'c', 'N', 'a', 'm', 'e', 0 };
  93 static const WCHAR ustr_system[]        = { 'S', 'y', 's', 't', 'e', 'm', 0 };
  94 static const WCHAR ustr_warning[]       = { 'W', 'a', 'r', 'n', 'i', 'n', 'g', 0 };
  95 static const WCHAR ustr_msg00001[]      = { 'm', 's', 'g', '0', '0', '0', '0', '1', 0 };
  96 /*
  97  * This table is to beat any form of "expression building" to check for
  98  * correct filename characters. It is also used for ident checks.
  99  * FIXME: use it more consistently.
 100  */
 101
 102 #define CH_SHORTNAME    0x01
 103 #define CH_LONGNAME     0x02
 104 #define CH_IDENT        0x04
 105 #define CH_NUMBER       0x08
 106 /*#define CH_WILDCARD   0x10*/
 107 /*#define CH_DOT        0x20*/
 108 #define CH_PUNCT        0x40
 109 #define CH_INVALID      0x80
 110
 111 static const char char_table[256] = {
 112         0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, /* 0x00 - 0x07 */
 113         0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, /* 0x08 - 0x0F */
 114         0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, /* 0x10 - 0x17 */
 115         0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, /* 0x18 - 0x1F */
 116         0x80, 0x03, 0x80, 0x03, 0x03, 0x03, 0x03, 0x03, /* 0x20 - 0x27 " !"#$%&'" */
 117         0x43, 0x43, 0x10, 0x80, 0x03, 0x03, 0x22, 0x80, /* 0x28 - 0x2F "()*+,-./" */
 118         0x0b, 0x0b, 0x0b, 0x0b, 0x0b, 0x0b, 0x0b, 0x0b, /* 0x30 - 0x37 "01234567" */
 119         0x0b, 0x0b, 0xc0, 0x80, 0x80, 0x80, 0x80, 0x10, /* 0x38 - 0x3F "89:;<=>?" */
 120         0x03, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, /* 0x40 - 0x47 "@ABCDEFG" */
 121         0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, /* 0x48 - 0x4F "HIJKLMNO" */
 122         0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, /* 0x50 - 0x57 "PQRSTUVW" */
 123         0x07, 0x07, 0x07, 0x80, 0x80, 0x80, 0x80, 0x07, /* 0x58 - 0x5F "XYZ[\]^_" */
 124         0x03, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, /* 0x60 - 0x67 "`abcdefg" */
 125         0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, /* 0x68 - 0x6F "hijklmno" */
 126         0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, /* 0x70 - 0x77 "pqrstuvw" */
 127         0x07, 0x07, 0x07, 0x03, 0x80, 0x03, 0x03, 0x80, /* 0x78 - 0x7F "xyz{|}~ " */
 128         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0x80 - 0x87 */
 129         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0x88 - 0x8F */
 130         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0x90 - 0x97 */
 131         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0x98 - 0x9F */
 132         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0xA0 - 0xA7 */
 133         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0xA8 - 0xAF */
 134         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0xB0 - 0xB7 */
 135         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0xB8 - 0xBF */
 136         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0xC0 - 0xC7 */
 137         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0xC8 - 0xCF */
 138         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0xD0 - 0xD7 */
 139         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0xD8 - 0xDF */
 140         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0xE0 - 0xE7 */
 141         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0xE8 - 0xEF */
 142         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, /* 0xF0 - 0xF7 */
 143         0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x80, /* 0xF8 - 0xFF */
 144 };
 145
 146 static int isisochar(int ch)
 147 {
 148         return !(ch & (~0xff));
 149 }
 150
 151 static int codepage;
 152 static const union cptable *codepage_def;
 153
 154 void set_codepage(int cp)
 155 {
 156         codepage = cp;
 157         codepage_def = find_codepage(codepage);
 158         if(!codepage_def)
 159                 xyyerror("Codepage %d not found; cannot process\n", codepage);
 160 }
 161
 162 /*
 163  * Input functions
 164  */
 165 static int nungetstack = 0;
 166 static int allocungetstack = 0;
 167 static char *ungetstack = NULL;
 168 static int ninputbuffer = 0;
 169 static WCHAR *inputbuffer = NULL;
 170 static char *xlatebuffer = NULL;
 171
 172 #define INPUTBUFFER_SIZE        2048    /* Must be larger than 4 and approx. large enough to hold a line */
 173
 174 /*
 175  * Fill the input buffer with *one* line of input.
 176  * The line is '\n' terminated so that scanning
 177  * messages with translation works as expected
 178  * (otherwise we cannot pre-translate because the
 179  * language is first known one line before the
 180  * actual message).
 181  */
 182 static int fill_inputbuffer(void)
 183 {
 184         int n;
 185         static const char err_fatalread[] = "Fatal: reading input failed";
 186         static int endian = -1;
 187
 188         if(!inputbuffer)
 189         {
 190                 inputbuffer = xmalloc(INPUTBUFFER_SIZE*sizeof(WCHAR));
 191                 xlatebuffer = xmalloc(INPUTBUFFER_SIZE);
 192         }
 193
 194 try_again:
 195         if(!unicodein)
 196         {
 197                 char *cptr;
 198                 cptr = fgets(xlatebuffer, INPUTBUFFER_SIZE, yyin);
 199                 if(!cptr && ferror(yyin))
 200                         xyyerror(err_fatalread);
 201                 else if(!cptr)
 202                         return 0;
 203                 assert(codepage_def != NULL);
 204                 n = wine_cp_mbstowcs(codepage_def, 0, xlatebuffer, strlen(xlatebuffer)+1, inputbuffer, INPUTBUFFER_SIZE);
 205                 if(n < 0)
 206                         internal_error(__FILE__, __LINE__, "Could not translate to unicode (%d)\n", n);
 207                 if(n <= 1)
 208                         goto try_again; /* Should not happen */
 209                 n--;    /* Strip added conversion '\0' from input length */
 210                 /*
 211                  * FIXME:
 212                  * Detect UTF-8 in the first time we read some bytes by
 213                  * checking the special sequence "FE..." or something like
 214                  * that. I need to check www.unicode.org for details.
 215                  */
 216         }
 217         else
 218         {
 219                 if(endian == -1)
 220                 {
 221                         n = fread(inputbuffer, 1, 8, yyin);
 222                         if(n != 8)
 223                         {
 224                                 if(!n && ferror(yyin))
 225                                         xyyerror(err_fatalread);
 226                                 else
 227                                         xyyerror("Fatal: file to short to determine byteorder (should never happen)\n");
 228                         }
 229                         if(isisochar(inputbuffer[0]) &&
 230                                 isisochar(inputbuffer[1]) &&
 231                                 isisochar(inputbuffer[2]) &&
 232                                 isisochar(inputbuffer[3]))
 233                         {
 234 #ifdef WORDS_BIGENDIAN
 235                                 endian = WMC_BO_BIG;
 236 #else
 237                                 endian = WMC_BO_LITTLE;
 238 #endif
 239                         }
 240                         else if(isisochar(BYTESWAP_WORD(inputbuffer[0])) &&
 241                                 isisochar(BYTESWAP_WORD(inputbuffer[1])) &&
 242                                 isisochar(BYTESWAP_WORD(inputbuffer[2])) &&
 243                                 isisochar(BYTESWAP_WORD(inputbuffer[3])))
 244                         {
 245 #ifdef WORDS_BIGENDIAN
 246                                 endian = WMC_BO_LITTLE;
 247 #else
 248                                 endian = WMC_BO_BIG;
 249 #endif
 250                         }
 251                         else
 252                                 xyyerror("Fatal: cannot determine file's byteorder\n");
 253                         /* FIXME:
 254                          * Determine the file-endian with the leader-bytes
 255                          * "FF FE..."; can't remember the exact sequence.
 256                          */
 257                         n /= 2;
 258 #ifdef WORDS_BIGENDIAN
 259                         if(endian == WMC_BO_LITTLE)
 260 #else
 261                         if(endian == WMC_BO_BIG)
 262 #endif
 263                         {
 264                                 inputbuffer[0] = BYTESWAP_WORD(inputbuffer[0]);
 265                                 inputbuffer[1] = BYTESWAP_WORD(inputbuffer[1]);
 266                                 inputbuffer[2] = BYTESWAP_WORD(inputbuffer[2]);
 267                                 inputbuffer[3] = BYTESWAP_WORD(inputbuffer[3]);
 268                         }
 269
 270                 }
 271                 else
 272                 {
 273                         int i;
 274                         n = 0;
 275                         for(i = 0; i < INPUTBUFFER_SIZE; i++)
 276                         {
 277                                 int t;
 278                                 t = fread(&inputbuffer[i], 2, 1, yyin);
 279                                 if(!t && ferror(yyin))
 280                                         xyyerror(err_fatalread);
 281                                 else if(!t && n)
 282                                         break;
 283                                 n++;
 284 #ifdef WORDS_BIGENDIAN
 285                                 if(endian == WMC_BO_LITTLE)
 286 #else
 287                                 if(endian == WMC_BO_BIG)
 288 #endif
 289                                 {
 290                                         if((inputbuffer[i] = BYTESWAP_WORD(inputbuffer[i])) == '\n')
 291                                                 break;
 292                                 }
 293                                 else
 294                                 {
 295                                         if(inputbuffer[i] == '\n')
 296                                                 break;
 297                                 }
 298                         }
 299                 }
 300
 301         }
 302
 303         if(!n)
 304         {
 305                 mcy_warning("Re-read line (input was or converted to zilch)\n");
 306                 goto try_again; /* Should not happen, but could be due to stdin reading and a signal */
 307         }
 308
 309         ninputbuffer += n;
 310         return 1;
 311 }
 312
 313 static int get_unichar(void)
 314 {
 315         static WCHAR *b = NULL;
 316         char_number++;
 317
 318         if(nungetstack)
 319                 return ungetstack[--nungetstack];
 320
 321         if(!ninputbuffer)
 322         {
 323                 if(!fill_inputbuffer())
 324                         return EOF;
 325                 b = inputbuffer;
 326         }
 327
 328         ninputbuffer--;
 329         return (int)(*b++ & 0xffff);
 330 }
 331
 332 static void unget_unichar(int ch)
 333 {
 334         if(ch == EOF)
 335                 return;
 336
 337         char_number--;
 338
 339         if(nungetstack == allocungetstack)
 340         {
 341                 allocungetstack += 32;
 342                 ungetstack = xrealloc(ungetstack, allocungetstack * sizeof(*ungetstack));
 343         }
 344
 345         ungetstack[nungetstack++] = (WCHAR)ch;
 346 }
 347
 348
 349 /*
 350  * Normal character stack.
 351  * Used for number scanning.
 352  */
 353 static int ncharstack = 0;
 354 static int alloccharstack = 0;
 355 static char *charstack = NULL;
 356
 357 static void empty_char_stack(void)
 358 {
 359         ncharstack = 0;
 360 }
 361
 362 static void push_char(int ch)
 363 {
 364         if(ncharstack == alloccharstack)
 365         {
 366                 alloccharstack += 32;
 367                 charstack = xrealloc(charstack, alloccharstack * sizeof(*charstack));
 368         }
 369         charstack[ncharstack++] = (char)ch;
 370 }
 371
 372 static int tos_char_stack(void)
 373 {
 374         if(!ncharstack)
 375                 return 0;
 376         else
 377                 return (int)(charstack[ncharstack-1] & 0xff);
 378 }
 379
 380 static char *get_char_stack(void)
 381 {
 382         return charstack;
 383 }
 384
 385 /*
 386  * Unicode character stack.
 387  * Used for general scanner.
 388  */
 389 static int nunicharstack = 0;
 390 static int allocunicharstack = 0;
 391 static WCHAR *unicharstack = NULL;
 392
 393 static void empty_unichar_stack(void)
 394 {
 395         nunicharstack = 0;
 396 }
 397
 398 static void push_unichar(int ch)
 399 {
 400         if(nunicharstack == allocunicharstack)
 401         {
 402                 allocunicharstack += 128;
 403                 unicharstack = xrealloc(unicharstack, allocunicharstack * sizeof(*unicharstack));
 404         }
 405         unicharstack[nunicharstack++] = (WCHAR)ch;
 406 }
 407
 408 #if 0
 409 static int tos_unichar_stack(void)
 410 {
 411         if(!nunicharstack)
 412                 return 0;
 413         else
 414                 return (int)(unicharstack[nunicharstack-1] & 0xffff);
 415 }
 416 #endif
 417
 418 static WCHAR *get_unichar_stack(void)
 419 {
 420         return unicharstack;
 421 }
 422
 423 /*
 424  * Number scanner
 425  *
 426  * state |      ch         | next state
 427  * ------+-----------------+--------------------------
 428  *   0   | [0]             | 1
 429  *   0   | [1-9]           | 4
 430  *   0   | .               | error (should never occur)
 431  *   1   | [xX]            | 2
 432  *   1   | [0-7]           | 3
 433  *   1   | [89a-wyzA-WYZ_] | error invalid digit
 434  *   1   | .               | return 0
 435  *   2   | [0-9a-fA-F]     | 2
 436  *   2   | [g-zG-Z_]       | error invalid hex digit
 437  *   2   | .               | return (hex-number) if TOS != [xX] else error
 438  *   3   | [0-7]           | 3
 439  *   3   | [89a-zA-Z_]     | error invalid octal digit
 440  *   3   | .               | return (octal-number)
 441  *   4   | [0-9]           | 4
 442  *   4   | [a-zA-Z_]       | error invalid decimal digit
 443  *   4   | .               | return (decimal-number)
 444  *
 445  * All non-identifier characters [^a-zA-Z_0-9] terminate the scan
 446  * and return the value. This is not entirely correct, but close
 447  * enough (should check punctuators as trailing context, but the
 448  * char_table is not adapted to that and it is questionable whether
 449  * it is worth the trouble).
 450  * All non-iso-8859-1 characters are an error.
 451  */
 452 static int scan_number(int ch)
 453 {
 454         int state = 0;
 455         int base = 10;
 456         empty_char_stack();
 457
 458         while(1)
 459         {
 460                 if(!isisochar(ch))
 461                         xyyerror("Invalid digit\n");
 462
 463                 switch(state)
 464                 {
 465                 case 0:
 466                         if(isdigit(ch))
 467                         {
 468                                 push_char(ch);
 469                                 if(ch == '0')
 470                                         state = 1;
 471                                 else
 472                                         state = 4;
 473                         }
 474                         else
 475                                 internal_error(__FILE__, __LINE__, "Non-digit in first number-scanner state\n");
 476                         break;
 477                 case 1:
 478                         if(ch == 'x' || ch == 'X')
 479                         {
 480                                 push_char(ch);
 481                                 state = 2;
 482                         }
 483                         else if(ch >= '0' && ch <= '7')
 484                         {
 485                                 push_char(ch);
 486                                 state = 3;
 487                         }
 488                         else if(isalpha(ch) || ch == '_')
 489                                 xyyerror("Invalid number digit\n");
 490                         else
 491                         {
 492                                 unget_unichar(ch);
 493                                 mcy_lval.num = 0;
 494                                 return tNUMBER;
 495                         }
 496                         break;
 497                 case 2:
 498                         if(isxdigit(ch))
 499                                 push_char(ch);
 500                         else if(isalpha(ch) || ch == '_' || !isxdigit(tos_char_stack()))
 501                                 xyyerror("Invalid hex digit\n");
 502                         else
 503                         {
 504                                 base = 16;
 505                                 goto finish;
 506                         }
 507                         break;
 508                 case 3:
 509                         if(ch >= '0' && ch <= '7')
 510                                 push_char(ch);
 511                         else if(isalnum(ch) || ch == '_')
 512                                 xyyerror("Invalid octal digit\n");
 513                         else
 514                         {
 515                                 base = 8;
 516                                 goto finish;
 517                         }
 518                         break;
 519                 case 4:
 520                         if(isdigit(ch))
 521                                 push_char(ch);
 522                         else if(isalnum(ch) || ch == '_')
 523                                 xyyerror("Invalid decimal digit\n");
 524                         else
 525                         {
 526                                 base = 10;
 527                                 goto finish;
 528                         }
 529                         break;
 530                 default:
 531                         internal_error(__FILE__, __LINE__, "Invalid state in number-scanner\n");
 532                 }
 533                 ch = get_unichar();
 534         }
 535 finish:
 536         unget_unichar(ch);
 537         push_char(0);
 538         mcy_lval.num = strtoul(get_char_stack(), NULL, base);
 539         return tNUMBER;
 540 }
 541
 542 static void newline(void)
 543 {
 544         line_number++;
 545         char_number = 1;
 546 }
 547
 548 static int unisort(const void *p1, const void *p2)
 549 {
 550         return unistricmp(((const token_t *)p1)->name, ((const token_t *)p2)->name);
 551 }
 552
 553 static token_t *tokentable = NULL;
 554 static int ntokentable = 0;
 555
 556 token_t *lookup_token(const WCHAR *s)
 557 {
 558         token_t tok;
 559
 560         tok.name = s;
 561         return (token_t *)bsearch(&tok, tokentable, ntokentable, sizeof(*tokentable), unisort);
 562 }
 563
 564 void add_token(tok_e type, const WCHAR *name, int tok, int cp, const WCHAR *alias, int fix)
 565 {
 566         ntokentable++;
 567         tokentable = xrealloc(tokentable, ntokentable * sizeof(*tokentable));
 568         tokentable[ntokentable-1].type = type;
 569         tokentable[ntokentable-1].name = name;
 570         tokentable[ntokentable-1].token = tok;
 571         tokentable[ntokentable-1].codepage = cp;
 572         tokentable[ntokentable-1].alias = alias;
 573         tokentable[ntokentable-1].fixed = fix;
 574         qsort(tokentable, ntokentable, sizeof(*tokentable), unisort);
 575 }
 576
 577 void get_tokentable(token_t **tab, int *len)
 578 {
 579         assert(tab != NULL);
 580         assert(len != NULL);
 581         *tab = tokentable;
 582         *len = ntokentable;
 583 }
 584
 585 /*
 586  * The scanner
 587  *
 588  */
 589 int mcy_lex(void)
 590 {
 591         static const WCHAR ustr_dot1[] = { '.', '\n', 0 };
 592         static const WCHAR ustr_dot2[] = { '.', '\r', '\n', 0 };
 593         static int isinit = 0;
 594         int ch;
 595
 596         if(!isinit)
 597         {
 598                 isinit++;
 599                 set_codepage(WMC_DEFAULT_CODEPAGE);
 600                 add_token(tok_keyword,  ustr_codepages,         tCODEPAGE,      0, NULL, 0);
 601                 add_token(tok_keyword,  ustr_facility,          tFACILITY,      0, NULL, 1);
 602                 add_token(tok_keyword,  ustr_facilitynames,     tFACNAMES,      0, NULL, 1);
 603                 add_token(tok_keyword,  ustr_language,          tLANGUAGE,      0, NULL, 1);
 604                 add_token(tok_keyword,  ustr_languagenames,     tLANNAMES,      0, NULL, 1);
 605                 add_token(tok_keyword,  ustr_messageid,         tMSGID,         0, NULL, 1);
 606                 add_token(tok_keyword,  ustr_messageidtypedef,  tTYPEDEF,       0, NULL, 1);
 607                 add_token(tok_keyword,  ustr_outputbase,        tBASE,          0, NULL, 1);
 608                 add_token(tok_keyword,  ustr_severity,          tSEVERITY,      0, NULL, 1);
 609                 add_token(tok_keyword,  ustr_severitynames,     tSEVNAMES,      0, NULL, 1);
 610                 add_token(tok_keyword,  ustr_symbolicname,      tSYMNAME,       0, NULL, 1);
 611                 add_token(tok_severity, ustr_error,             0x03,           0, NULL, 0);
 612                 add_token(tok_severity, ustr_warning,           0x02,           0, NULL, 0);
 613                 add_token(tok_severity, ustr_informational,     0x01,           0, NULL, 0);
 614                 add_token(tok_severity, ustr_success,           0x00,           0, NULL, 0);
 615                 add_token(tok_facility, ustr_application,       0xFFF,          0, NULL, 0);
 616                 add_token(tok_facility, ustr_system,            0x0FF,          0, NULL, 0);
 617                 add_token(tok_language, ustr_english,           0x409,          437, ustr_msg00001, 0);
 618         }
 619
 620         empty_unichar_stack();
 621
 622         while(1)
 623         {
 624                 if(want_line)
 625                 {
 626                         while((ch = get_unichar()) != '\n')
 627                         {
 628                                 if(ch == EOF)
 629                                         xyyerror("Unexpected EOF\n");
 630                                 push_unichar(ch);
 631                         }
 632                         newline();
 633                         push_unichar(ch);
 634                         push_unichar(0);
 635                         if(!unistrcmp(ustr_dot1, get_unichar_stack()) || !unistrcmp(ustr_dot2, get_unichar_stack()))
 636                         {
 637                                 want_line = 0;
 638                                 /* Reset the codepage to our default after each message */
 639                                 set_codepage(WMC_DEFAULT_CODEPAGE);
 640                                 return tMSGEND;
 641                         }
 642                         mcy_lval.str = xunistrdup(get_unichar_stack());
 643                         return tLINE;
 644                 }
 645
 646                 ch = get_unichar();
 647
 648                 if(ch == EOF)
 649                         return EOF;
 650
 651                 if(ch == '\n')
 652                 {
 653                         newline();
 654                         if(want_nl)
 655                         {
 656                                 want_nl = 0;
 657                                 return tNL;
 658                         }
 659                         continue;
 660                 }
 661
 662                 if(isisochar(ch))
 663                 {
 664                         if(want_file)
 665                         {
 666                                 int n = 0;
 667                                 while(n < 8 && isisochar(ch))
 668                                 {
 669                                         int t = char_table[ch];
 670                                         if((t & CH_PUNCT) || !(t & CH_SHORTNAME))
 671                                                 break;
 672
 673                                         push_unichar(ch);
 674                                         n++;
 675                                         ch = get_unichar();
 676                                 }
 677                                 unget_unichar(ch);
 678                                 push_unichar(0);
 679                                 want_file = 0;
 680                                 mcy_lval.str = xunistrdup(get_unichar_stack());
 681                                 return tFILE;
 682                         }
 683
 684                         if(char_table[ch] & CH_IDENT)
 685                         {
 686                                 token_t *tok;
 687                                 while(isisochar(ch) && (char_table[ch] & (CH_IDENT|CH_NUMBER)))
 688                                 {
 689                                         push_unichar(ch);
 690                                         ch = get_unichar();
 691                                 }
 692                                 unget_unichar(ch);
 693                                 push_unichar(0);
 694                                 if(!(tok = lookup_token(get_unichar_stack())))
 695                                 {
 696                                         mcy_lval.str = xunistrdup(get_unichar_stack());
 697                                         return tIDENT;
 698                                 }
 699                                 switch(tok->type)
 700                                 {
 701                                 case tok_keyword:
 702                                         return tok->token;
 703
 704                                 case tok_language:
 705                                         codepage = tok->codepage;
 706                                         /* Fall through */
 707                                 case tok_severity:
 708                                 case tok_facility:
 709                                         mcy_lval.tok = tok;
 710                                         return tTOKEN;
 711
 712                                 default:
 713                                         internal_error(__FILE__, __LINE__, "Invalid token type encountered\n");
 714                                 }
 715                         }
 716
 717                         if(isspace(ch)) /* Ignore space */
 718                                 continue;
 719
 720                         if(isdigit(ch))
 721                                 return scan_number(ch);
 722                 }
 723
 724                 switch(ch)
 725                 {
 726                 case ':':
 727                 case '=':
 728                 case '+':
 729                 case '(':
 730                 case ')':
 731                         return ch;
 732                 case ';':
 733                         while(ch != '\n' && ch != EOF)
 734                         {
 735                                 push_unichar(ch);
 736                                 ch = get_unichar();
 737                         }
 738                         newline();
 739                         push_unichar(ch);       /* Include the newline */
 740                         push_unichar(0);
 741                         mcy_lval.str = xunistrdup(get_unichar_stack());
 742                         return tCOMMENT;
 743                 default:
 744                         xyyerror("Invalid character '%c' (0x%04x)\n", isisochar(ch) && isprint(ch) ? ch : '.', ch);
 745                 }
 746         }
 747 }