tools/wmc/mcl.c

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