src/scan-gram.l

   1 /* Bison Grammar Scanner                             -*- C -*-
   2
   3    Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007 Free Software Foundation,
   4    Inc.
   5
   6    This file is part of Bison, the GNU Compiler Compiler.
   7
   8    This program is free software; you can redistribute it and/or modify
   9    it under the terms of the GNU General Public License as published by
  10    the Free Software Foundation; either version 2 of the License, or
  11    (at your option) any later version.
  12
  13    This program is distributed in the hope that it will be useful,
  14    but WITHOUT ANY WARRANTY; without even the implied warranty of
  15    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16    GNU General Public License for more details.
  17
  18    You should have received a copy of the GNU General Public License
  19    along with this program; if not, write to the Free Software
  20    Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
  21    02110-1301  USA
  22 */
  23
  24 %option debug nodefault nounput noyywrap never-interactive
  25 %option prefix="gram_" outfile="lex.yy.c"
  26
  27 %{
  28 /* Work around a bug in flex 2.5.31.  See Debian bug 333231
  29    <http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=333231>.  */
  30 #undef gram_wrap
  31 #define gram_wrap() 1
  32
  33 #define FLEX_PREFIX(Id) gram_ ## Id
  34 #include "flex-scanner.h"
  35
  36 #include "complain.h"
  37 #include "files.h"
  38 #include "gram.h"
  39 #include "quotearg.h"
  40 #include "reader.h"
  41 #include "uniqstr.h"
  42
  43 #include <mbswidth.h>
  44 #include <quote.h>
  45
  46 #include "scan-gram.h"
  47
  48 #define YY_DECL GRAM_LEX_DECL
  49
  50 #define YY_USER_INIT                                    \
  51    code_start = scanner_cursor = loc->start;            \
  52
  53 /* Location of scanner cursor.  */
  54 static boundary scanner_cursor;
  55
  56 #define YY_USER_ACTION  location_compute (loc, &scanner_cursor, yytext, yyleng);
  57
  58 static size_t no_cr_read (FILE *, char *, size_t);
  59 #define YY_INPUT(buf, result, size) ((result) = no_cr_read (yyin, buf, size))
  60
  61 /* A string representing the most recently saved token.  */
  62 static char *last_string;
  63
  64 void
  65 gram_scanner_last_string_free (void)
  66 {
  67   STRING_FREE;
  68 }
  69
  70 static void handle_syncline (char *, location);
  71 static unsigned long int scan_integer (char const *p, int base, location loc);
  72 static int convert_ucn_to_byte (char const *hex_text);
  73 static void unexpected_eof (boundary, char const *);
  74 static void unexpected_newline (boundary, char const *);
  75
  76 %}
  77  /* A C-like comment in directives/rules. */
  78 %x SC_YACC_COMMENT
  79  /* Strings and characters in directives/rules. */
  80 %x SC_ESCAPED_STRING SC_ESCAPED_CHARACTER
  81  /* A identifier was just read in directives/rules.  Special state
  82     to capture the sequence `identifier :'. */
  83 %x SC_AFTER_IDENTIFIER
  84
  85  /* Three types of user code:
  86     - prologue (code between `%{' `%}' in the first section, before %%);
  87     - actions, printers, union, etc, (between braced in the middle section);
  88     - epilogue (everything after the second %%). */
  89 %x SC_PROLOGUE SC_BRACED_CODE SC_EPILOGUE
  90  /* C and C++ comments in code. */
  91 %x SC_COMMENT SC_LINE_COMMENT
  92  /* Strings and characters in code. */
  93 %x SC_STRING SC_CHARACTER
  94
  95 letter    [.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_]
  96 id        {letter}({letter}|[0-9])*
  97 directive %{letter}({letter}|[0-9]|-)*
  98 int       [0-9]+
  99
 100 /* POSIX says that a tag must be both an id and a C union member, but
 101    historically almost any character is allowed in a tag.  We disallow
 102    NUL and newline, as this simplifies our implementation.  */
 103 tag      [^\0\n>]+
 104
 105 /* Zero or more instances of backslash-newline.  Following GCC, allow
 106    white space between the backslash and the newline.  */
 107 splice   (\\[ \f\t\v]*\n)*
 108
 109 %%
 110 %{
 111   /* Nesting level of the current code in braces.  */
 112   int braces_level IF_LINT (= 0);
 113
 114   /* Parent context state, when applicable.  */
 115   int context_state IF_LINT (= 0);
 116
 117   /* Location of most recent identifier, when applicable.  */
 118   location id_loc IF_LINT (= empty_location);
 119
 120   /* Where containing code started, when applicable.  Its initial
 121      value is relevant only when yylex is invoked in the SC_EPILOGUE
 122      start condition.  */
 123   boundary code_start = scanner_cursor;
 124
 125   /* Where containing comment or string or character literal started,
 126      when applicable.  */
 127   boundary token_start IF_LINT (= scanner_cursor);
 128 %}
 129
 130
 131   /*-----------------------.
 132   | Scanning white space.  |
 133   `-----------------------*/
 134
 135 <INITIAL,SC_AFTER_IDENTIFIER>
 136 {
 137   /* Comments and white space.  */
 138   ","          warn_at (*loc, _("stray `,' treated as white space"));
 139   [ \f\n\t\v]  |
 140   "//".*       ;
 141   "/*" {
 142     token_start = loc->start;
 143     context_state = YY_START;
 144     BEGIN SC_YACC_COMMENT;
 145   }
 146
 147   /* #line directives are not documented, and may be withdrawn or
 148      modified in future versions of Bison.  */
 149   ^"#line "{int}" \"".*"\"\n" {
 150     handle_syncline (yytext + sizeof "#line " - 1, *loc);
 151   }
 152 }
 153
 154
 155   /*----------------------------.
 156   | Scanning Bison directives.  |
 157   `----------------------------*/
 158 <INITIAL>
 159 {
 160   "%binary"                         return PERCENT_NONASSOC;
 161   "%code"                           return PERCENT_CODE;
 162   "%debug"                          return PERCENT_DEBUG;
 163   "%default"[-_]"prec"              return PERCENT_DEFAULT_PREC;
 164   "%define"                         return PERCENT_DEFINE;
 165   "%defines"                        return PERCENT_DEFINES;
 166   "%destructor"                     return PERCENT_DESTRUCTOR;
 167   "%dprec"                          return PERCENT_DPREC;
 168   "%error"[-_]"verbose"             return PERCENT_ERROR_VERBOSE;
 169   "%expect"                         return PERCENT_EXPECT;
 170   "%expect"[-_]"rr"                 return PERCENT_EXPECT_RR;
 171   "%file-prefix"                    return PERCENT_FILE_PREFIX;
 172   "%fixed"[-_]"output"[-_]"files"   return PERCENT_YACC;
 173   "%initial-action"                 return PERCENT_INITIAL_ACTION;
 174   "%glr-parser"                     return PERCENT_GLR_PARSER;
 175   "%language"                       return PERCENT_LANGUAGE;
 176   "%left"                           return PERCENT_LEFT;
 177   "%lex-param"                      return PERCENT_LEX_PARAM;
 178   "%locations"                      return PERCENT_LOCATIONS;
 179   "%merge"                          return PERCENT_MERGE;
 180   "%name"[-_]"prefix"               return PERCENT_NAME_PREFIX;
 181   "%no"[-_]"default"[-_]"prec"      return PERCENT_NO_DEFAULT_PREC;
 182   "%no"[-_]"lines"                  return PERCENT_NO_LINES;
 183   "%nonassoc"                       return PERCENT_NONASSOC;
 184   "%nondeterministic-parser"        return PERCENT_NONDETERMINISTIC_PARSER;
 185   "%nterm"                          return PERCENT_NTERM;
 186   "%output"                         return PERCENT_OUTPUT;
 187   "%parse-param"                    return PERCENT_PARSE_PARAM;
 188   "%prec"                           return PERCENT_PREC;
 189   "%printer"                        return PERCENT_PRINTER;
 190   "%pure"[-_]"parser"               return PERCENT_PURE_PARSER;
 191   "%push-parser"                    return PERCENT_PUSH_PARSER;
 192   "%push-pull-parser"               return PERCENT_PUSH_PULL_PARSER;
 193   "%require"                        return PERCENT_REQUIRE;
 194   "%right"                          return PERCENT_RIGHT;
 195   "%skeleton"                       return PERCENT_SKELETON;
 196   "%start"                          return PERCENT_START;
 197   "%term"                           return PERCENT_TOKEN;
 198   "%token"                          return PERCENT_TOKEN;
 199   "%token"[-_]"table"               return PERCENT_TOKEN_TABLE;
 200   "%type"                           return PERCENT_TYPE;
 201   "%union"                          return PERCENT_UNION;
 202   "%verbose"                        return PERCENT_VERBOSE;
 203   "%yacc"                           return PERCENT_YACC;
 204
 205   {directive} {
 206     complain_at (*loc, _("invalid directive: %s"), quote (yytext));
 207   }
 208
 209   "="                     return EQUAL;
 210   "|"                     return PIPE;
 211   ";"                     return SEMICOLON;
 212   "<*>"                   return TYPE_TAG_ANY;
 213   "<>"                    return TYPE_TAG_NONE;
 214
 215   {id} {
 216     val->uniqstr = uniqstr_new (yytext);
 217     id_loc = *loc;
 218     BEGIN SC_AFTER_IDENTIFIER;
 219   }
 220
 221   {int} {
 222     val->integer = scan_integer (yytext, 10, *loc);
 223     return INT;
 224   }
 225   0[xX][0-9abcdefABCDEF]+ {
 226     val->integer = scan_integer (yytext, 16, *loc);
 227     return INT;
 228   }
 229
 230   /* Characters.  We don't check there is only one.  */
 231   "'"         STRING_GROW; token_start = loc->start; BEGIN SC_ESCAPED_CHARACTER;
 232
 233   /* Strings. */
 234   "\""        token_start = loc->start; BEGIN SC_ESCAPED_STRING;
 235
 236   /* Prologue. */
 237   "%{"        code_start = loc->start; BEGIN SC_PROLOGUE;
 238
 239   /* Code in between braces.  */
 240   "{" {
 241     STRING_GROW;
 242     braces_level = 0;
 243     code_start = loc->start;
 244     BEGIN SC_BRACED_CODE;
 245   }
 246
 247   /* A type. */
 248   "<"{tag}">" {
 249     obstack_grow (&obstack_for_string, yytext + 1, yyleng - 2);
 250     STRING_FINISH;
 251     val->uniqstr = uniqstr_new (last_string);
 252     STRING_FREE;
 253     return TYPE;
 254   }
 255
 256   "%%" {
 257     static int percent_percent_count;
 258     if (++percent_percent_count == 2)
 259       BEGIN SC_EPILOGUE;
 260     return PERCENT_PERCENT;
 261   }
 262
 263   . {
 264     complain_at (*loc, _("invalid character: %s"), quote (yytext));
 265   }
 266
 267   <<EOF>> {
 268     loc->start = loc->end = scanner_cursor;
 269     yyterminate ();
 270   }
 271 }
 272
 273
 274   /*-----------------------------------------------------------------.
 275   | Scanning after an identifier, checking whether a colon is next.  |
 276   `-----------------------------------------------------------------*/
 277
 278 <SC_AFTER_IDENTIFIER>
 279 {
 280   ":" {
 281     *loc = id_loc;
 282     BEGIN INITIAL;
 283     return ID_COLON;
 284   }
 285   . {
 286     scanner_cursor.column -= mbsnwidth (yytext, yyleng, 0);
 287     yyless (0);
 288     *loc = id_loc;
 289     BEGIN INITIAL;
 290     return ID;
 291   }
 292   <<EOF>> {
 293     *loc = id_loc;
 294     BEGIN INITIAL;
 295     return ID;
 296   }
 297 }
 298
 299
 300   /*---------------------------------------------------------------.
 301   | Scanning a Yacc comment.  The initial `/ *' is already eaten.  |
 302   `---------------------------------------------------------------*/
 303
 304 <SC_YACC_COMMENT>
 305 {
 306   "*/"     BEGIN context_state;
 307   .|\n     ;
 308   <<EOF>>  unexpected_eof (token_start, "*/"); BEGIN context_state;
 309 }
 310
 311
 312   /*------------------------------------------------------------.
 313   | Scanning a C comment.  The initial `/ *' is already eaten.  |
 314   `------------------------------------------------------------*/
 315
 316 <SC_COMMENT>
 317 {
 318   "*"{splice}"/"  STRING_GROW; BEGIN context_state;
 319   <<EOF>>         unexpected_eof (token_start, "*/"); BEGIN context_state;
 320 }
 321
 322
 323   /*--------------------------------------------------------------.
 324   | Scanning a line comment.  The initial `//' is already eaten.  |
 325   `--------------------------------------------------------------*/
 326
 327 <SC_LINE_COMMENT>
 328 {
 329   "\n"           STRING_GROW; BEGIN context_state;
 330   {splice}       STRING_GROW;
 331   <<EOF>>        BEGIN context_state;
 332 }
 333
 334
 335   /*------------------------------------------------.
 336   | Scanning a Bison string, including its escapes. |
 337   | The initial quote is already eaten.             |
 338   `------------------------------------------------*/
 339
 340 <SC_ESCAPED_STRING>
 341 {
 342   "\""|"\n" {
 343     if (yytext[0] == '\n')
 344       unexpected_newline (token_start, "\"");
 345     STRING_FINISH;
 346     loc->start = token_start;
 347     val->chars = last_string;
 348     BEGIN INITIAL;
 349     return STRING;
 350   }
 351   <<EOF>> {
 352     unexpected_eof (token_start, "\"");
 353     STRING_FINISH;
 354     loc->start = token_start;
 355     val->chars = last_string;
 356     BEGIN INITIAL;
 357     return STRING;
 358   }
 359 }
 360
 361   /*----------------------------------------------------------.
 362   | Scanning a Bison character literal, decoding its escapes. |
 363   | The initial quote is already eaten.                       |
 364   `----------------------------------------------------------*/
 365
 366 <SC_ESCAPED_CHARACTER>
 367 {
 368   "'"|"\n" {
 369     if (yytext[0] == '\n')
 370       unexpected_newline (token_start, "'");
 371     STRING_GROW;
 372     STRING_FINISH;
 373     loc->start = token_start;
 374     val->character = last_string[1];
 375     STRING_FREE;
 376     BEGIN INITIAL;
 377     return CHAR;
 378   }
 379   <<EOF>> {
 380     unexpected_eof (token_start, "'");
 381     STRING_FINISH;
 382     loc->start = token_start;
 383     if (strlen(last_string) > 1)
 384       val->character = last_string[1];
 385     else
 386       val->character = last_string[0];
 387     STRING_FREE;
 388     BEGIN INITIAL;
 389     return CHAR;
 390   }
 391 }
 392
 393 <SC_ESCAPED_CHARACTER,SC_ESCAPED_STRING>
 394 {
 395   \0        complain_at (*loc, _("invalid null character"));
 396 }
 397
 398
 399   /*----------------------------.
 400   | Decode escaped characters.  |
 401   `----------------------------*/
 402
 403 <SC_ESCAPED_STRING,SC_ESCAPED_CHARACTER>
 404 {
 405   \\[0-7]{1,3} {
 406     unsigned long int c = strtoul (yytext + 1, NULL, 8);
 407     if (UCHAR_MAX < c)
 408       complain_at (*loc, _("invalid escape sequence: %s"), quote (yytext));
 409     else if (! c)
 410       complain_at (*loc, _("invalid null character: %s"), quote (yytext));
 411     else
 412       obstack_1grow (&obstack_for_string, c);
 413   }
 414
 415   \\x[0-9abcdefABCDEF]+ {
 416     verify (UCHAR_MAX < ULONG_MAX);
 417     unsigned long int c = strtoul (yytext + 2, NULL, 16);
 418     if (UCHAR_MAX < c)
 419       complain_at (*loc, _("invalid escape sequence: %s"), quote (yytext));
 420     else if (! c)
 421       complain_at (*loc, _("invalid null character: %s"), quote (yytext));
 422     else
 423       obstack_1grow (&obstack_for_string, c);
 424   }
 425
 426   \\a   obstack_1grow (&obstack_for_string, '\a');
 427   \\b   obstack_1grow (&obstack_for_string, '\b');
 428   \\f   obstack_1grow (&obstack_for_string, '\f');
 429   \\n   obstack_1grow (&obstack_for_string, '\n');
 430   \\r   obstack_1grow (&obstack_for_string, '\r');
 431   \\t   obstack_1grow (&obstack_for_string, '\t');
 432   \\v   obstack_1grow (&obstack_for_string, '\v');
 433
 434   /* \\[\"\'?\\] would be shorter, but it confuses xgettext.  */
 435   \\("\""|"'"|"?"|"\\")  obstack_1grow (&obstack_for_string, yytext[1]);
 436
 437   \\(u|U[0-9abcdefABCDEF]{4})[0-9abcdefABCDEF]{4} {
 438     int c = convert_ucn_to_byte (yytext);
 439     if (c < 0)
 440       complain_at (*loc, _("invalid escape sequence: %s"), quote (yytext));
 441     else if (! c)
 442       complain_at (*loc, _("invalid null character: %s"), quote (yytext));
 443     else
 444       obstack_1grow (&obstack_for_string, c);
 445   }
 446   \\(.|\n)      {
 447     complain_at (*loc, _("unrecognized escape sequence: %s"), quote (yytext));
 448     STRING_GROW;
 449   }
 450 }
 451
 452   /*--------------------------------------------.
 453   | Scanning user-code characters and strings.  |
 454   `--------------------------------------------*/
 455
 456 <SC_CHARACTER,SC_STRING>
 457 {
 458   {splice}|\\{splice}[^\n\[\]]  STRING_GROW;
 459 }
 460
 461 <SC_CHARACTER>
 462 {
 463   "'"           STRING_GROW; BEGIN context_state;
 464   \n            unexpected_newline (token_start, "'"); BEGIN context_state;
 465   <<EOF>>       unexpected_eof (token_start, "'"); BEGIN context_state;
 466 }
 467
 468 <SC_STRING>
 469 {
 470   "\""          STRING_GROW; BEGIN context_state;
 471   \n            unexpected_newline (token_start, "\""); BEGIN context_state;
 472   <<EOF>>       unexpected_eof (token_start, "\""); BEGIN context_state;
 473 }
 474
 475
 476   /*---------------------------------------------------.
 477   | Strings, comments etc. can be found in user code.  |
 478   `---------------------------------------------------*/
 479
 480 <SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
 481 {
 482   "'" {
 483     STRING_GROW;
 484     context_state = YY_START;
 485     token_start = loc->start;
 486     BEGIN SC_CHARACTER;
 487   }
 488   "\"" {
 489     STRING_GROW;
 490     context_state = YY_START;
 491     token_start = loc->start;
 492     BEGIN SC_STRING;
 493   }
 494   "/"{splice}"*" {
 495     STRING_GROW;
 496     context_state = YY_START;
 497     token_start = loc->start;
 498     BEGIN SC_COMMENT;
 499   }
 500   "/"{splice}"/" {
 501     STRING_GROW;
 502     context_state = YY_START;
 503     BEGIN SC_LINE_COMMENT;
 504   }
 505 }
 506
 507
 508
 509   /*-----------------------------------------------------------.
 510   | Scanning some code in braces (actions). The initial "{" is |
 511   | already eaten.                                             |
 512   `-----------------------------------------------------------*/
 513
 514 <SC_BRACED_CODE>
 515 {
 516   "{"|"<"{splice}"%"  STRING_GROW; braces_level++;
 517   "%"{splice}">"      STRING_GROW; braces_level--;
 518   "}" {
 519     obstack_1grow (&obstack_for_string, '}');
 520
 521     --braces_level;
 522     if (braces_level < 0)
 523       {
 524         STRING_FINISH;
 525         loc->start = code_start;
 526         val->code = last_string;
 527         BEGIN INITIAL;
 528         return BRACED_CODE;
 529       }
 530   }
 531
 532   /* Tokenize `<<%' correctly (as `<<' `%') rather than incorrrectly
 533      (as `<' `<%').  */
 534   "<"{splice}"<"  STRING_GROW;
 535
 536   <<EOF>> {
 537     unexpected_eof (code_start, "}");
 538     STRING_FINISH;
 539     loc->start = code_start;
 540     val->code = last_string;
 541     BEGIN INITIAL;
 542     return BRACED_CODE;
 543   }
 544 }
 545
 546
 547   /*--------------------------------------------------------------.
 548   | Scanning some prologue: from "%{" (already scanned) to "%}".  |
 549   `--------------------------------------------------------------*/
 550
 551 <SC_PROLOGUE>
 552 {
 553   "%}" {
 554     STRING_FINISH;
 555     loc->start = code_start;
 556     val->chars = last_string;
 557     BEGIN INITIAL;
 558     return PROLOGUE;
 559   }
 560
 561   <<EOF>> {
 562     unexpected_eof (code_start, "%}");
 563     STRING_FINISH;
 564     loc->start = code_start;
 565     val->chars = last_string;
 566     BEGIN INITIAL;
 567     return PROLOGUE;
 568   }
 569 }
 570
 571
 572   /*---------------------------------------------------------------.
 573   | Scanning the epilogue (everything after the second "%%", which |
 574   | has already been eaten).                                       |
 575   `---------------------------------------------------------------*/
 576
 577 <SC_EPILOGUE>
 578 {
 579   <<EOF>> {
 580     STRING_FINISH;
 581     loc->start = code_start;
 582     val->chars = last_string;
 583     BEGIN INITIAL;
 584     return EPILOGUE;
 585   }
 586 }
 587
 588
 589   /*-----------------------------------------------------.
 590   | By default, grow the string obstack with the input.  |
 591   `-----------------------------------------------------*/
 592
 593 <SC_COMMENT,SC_LINE_COMMENT,SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE,SC_STRING,SC_CHARACTER,SC_ESCAPED_STRING,SC_ESCAPED_CHARACTER>.      |
 594 <SC_COMMENT,SC_LINE_COMMENT,SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>\n   STRING_GROW;
 595
 596 %%
 597
 598 /* Read bytes from FP into buffer BUF of size SIZE.  Return the
 599    number of bytes read.  Remove '\r' from input, treating \r\n
 600    and isolated \r as \n.  */
 601
 602 static size_t
 603 no_cr_read (FILE *fp, char *buf, size_t size)
 604 {
 605   size_t bytes_read = fread (buf, 1, size, fp);
 606   if (bytes_read)
 607     {
 608       char *w = memchr (buf, '\r', bytes_read);
 609       if (w)
 610         {
 611           char const *r = ++w;
 612           char const *lim = buf + bytes_read;
 613
 614           for (;;)
 615             {
 616               /* Found an '\r'.  Treat it like '\n', but ignore any
 617                  '\n' that immediately follows.  */
 618               w[-1] = '\n';
 619               if (r == lim)
 620                 {
 621                   int ch = getc (fp);
 622                   if (ch != '\n' && ungetc (ch, fp) != ch)
 623                     break;
 624                 }
 625               else if (*r == '\n')
 626                 r++;
 627
 628               /* Copy until the next '\r'.  */
 629               do
 630                 {
 631                   if (r == lim)
 632                     return w - buf;
 633                 }
 634               while ((*w++ = *r++) != '\r');
 635             }
 636
 637           return w - buf;
 638         }
 639     }
 640
 641   return bytes_read;
 642 }
 643
 644
 645
 646 /*------------------------------------------------------.
 647 | Scan NUMBER for a base-BASE integer at location LOC.  |
 648 `------------------------------------------------------*/
 649
 650 static unsigned long int
 651 scan_integer (char const *number, int base, location loc)
 652 {
 653   verify (INT_MAX < ULONG_MAX);
 654   unsigned long int num = strtoul (number, NULL, base);
 655
 656   if (INT_MAX < num)
 657     {
 658       complain_at (loc, _("integer out of range: %s"), quote (number));
 659       num = INT_MAX;
 660     }
 661
 662   return num;
 663 }
 664
 665
 666 /*------------------------------------------------------------------.
 667 | Convert universal character name UCN to a single-byte character,  |
 668 | and return that character.  Return -1 if UCN does not correspond  |
 669 | to a single-byte character.                                       |
 670 `------------------------------------------------------------------*/
 671
 672 static int
 673 convert_ucn_to_byte (char const *ucn)
 674 {
 675   verify (UCHAR_MAX <= INT_MAX);
 676   unsigned long int code = strtoul (ucn + 2, NULL, 16);
 677
 678   /* FIXME: Currently we assume Unicode-compatible unibyte characters
 679      on ASCII hosts (i.e., Latin-1 on hosts with 8-bit bytes).  On
 680      non-ASCII hosts we support only the portable C character set.
 681      These limitations should be removed once we add support for
 682      multibyte characters.  */
 683
 684   if (UCHAR_MAX < code)
 685     return -1;
 686
 687 #if ! ('$' == 0x24 && '@' == 0x40 && '`' == 0x60 && '~' == 0x7e)
 688   {
 689     /* A non-ASCII host.  Use CODE to index into a table of the C
 690        basic execution character set, which is guaranteed to exist on
 691        all Standard C platforms.  This table also includes '$', '@',
 692        and '`', which are not in the basic execution character set but
 693        which are unibyte characters on all the platforms that we know
 694        about.  */
 695     static signed char const table[] =
 696       {
 697         '\0',   -1,   -1,   -1,   -1,   -1,   -1, '\a',
 698         '\b', '\t', '\n', '\v', '\f', '\r',   -1,   -1,
 699           -1,   -1,   -1,   -1,   -1,   -1,   -1,   -1,
 700           -1,   -1,   -1,   -1,   -1,   -1,   -1,   -1,
 701          ' ',  '!',  '"',  '#',  '$',  '%',  '&', '\'',
 702          '(',  ')',  '*',  '+',  ',',  '-',  '.',  '/',
 703          '0',  '1',  '2',  '3',  '4',  '5',  '6',  '7',
 704          '8',  '9',  ':',  ';',  '<',  '=',  '>',  '?',
 705          '@',  'A',  'B',  'C',  'D',  'E',  'F',  'G',
 706          'H',  'I',  'J',  'K',  'L',  'M',  'N',  'O',
 707          'P',  'Q',  'R',  'S',  'T',  'U',  'V',  'W',
 708          'X',  'Y',  'Z',  '[', '\\',  ']',  '^',  '_',
 709          '`',  'a',  'b',  'c',  'd',  'e',  'f',  'g',
 710          'h',  'i',  'j',  'k',  'l',  'm',  'n',  'o',
 711          'p',  'q',  'r',  's',  't',  'u',  'v',  'w',
 712          'x',  'y',  'z',  '{',  '|',  '}',  '~'
 713       };
 714
 715     code = code < sizeof table ? table[code] : -1;
 716   }
 717 #endif
 718
 719   return code;
 720 }
 721
 722
 723 /*----------------------------------------------------------------.
 724 | Handle `#line INT "FILE"'.  ARGS has already skipped `#line '.  |
 725 `----------------------------------------------------------------*/
 726
 727 static void
 728 handle_syncline (char *args, location loc)
 729 {
 730   char *after_num;
 731   unsigned long int lineno = strtoul (args, &after_num, 10);
 732   char *file = strchr (after_num, '"') + 1;
 733   *strchr (file, '"') = '\0';
 734   if (INT_MAX <= lineno)
 735     {
 736       warn_at (loc, _("line number overflow"));
 737       lineno = INT_MAX;
 738     }
 739   current_file = uniqstr_new (file);
 740   boundary_set (&scanner_cursor, current_file, lineno, 1);
 741 }
 742
 743
 744 /*----------------------------------------------------------------.
 745 | For a token or comment starting at START, report message MSGID, |
 746 | which should say that an end marker was found before            |
 747 | the expected TOKEN_END.                                         |
 748 `----------------------------------------------------------------*/
 749
 750 static void
 751 unexpected_end (boundary start, char const *msgid, char const *token_end)
 752 {
 753   location loc;
 754   loc.start = start;
 755   loc.end = scanner_cursor;
 756   complain_at (loc, _(msgid), token_end);
 757 }
 758
 759
 760 /*------------------------------------------------------------------------.
 761 | Report an unexpected EOF in a token or comment starting at START.       |
 762 | An end of file was encountered and the expected TOKEN_END was missing.  |
 763 `------------------------------------------------------------------------*/
 764
 765 static void
 766 unexpected_eof (boundary start, char const *token_end)
 767 {
 768   unexpected_end (start, N_("missing `%s' at end of file"), token_end);
 769 }
 770
 771
 772 /*----------------------------------------.
 773 | Likewise, but for unexpected newlines.  |
 774 `----------------------------------------*/
 775
 776 static void
 777 unexpected_newline (boundary start, char const *token_end)
 778 {
 779   unexpected_end (start, N_("missing `%s' at end of line"), token_end);
 780 }
 781
 782
 783 /*-------------------------.
 784 | Initialize the scanner.  |
 785 `-------------------------*/
 786
 787 void
 788 gram_scanner_initialize (void)
 789 {
 790   obstack_init (&obstack_for_string);
 791 }
 792
 793
 794 /*-----------------------------------------------.
 795 | Free all the memory allocated to the scanner.  |
 796 `-----------------------------------------------*/
 797
 798 void
 799 gram_scanner_free (void)
 800 {
 801   obstack_free (&obstack_for_string, 0);
 802   /* Reclaim Flex's buffers.  */
 803   yylex_destroy ();
 804 }