granicus.if.org Git - imagemagick/blob - MagickCore/token.c

   1 /*
   2 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
   3 %                                                                             %
   4 %                                                                             %
   5 %                                                                             %
   6 %                    TTTTT   OOO   K   K  EEEEE  N   N                        %
   7 %                      T    O   O  K  K   E      NN  N                        %
   8 %                      T    O   O  KKK    EEE    N N N                        %
   9 %                      T    O   O  K  K   E      N  NN                        %
  10 %                      T     OOO   K   K  EEEEE  N   N                        %
  11 %                                                                             %
  12 %                                                                             %
  13 %                         MagickCore Token Methods                            %
  14 %                                                                             %
  15 %                             Software Design                                 %
  16 %                               John Cristy                                   %
  17 %                              January 1993                                   %
  18 %                                                                             %
  19 %                                                                             %
  20 %  Copyright 1999-2013 ImageMagick Studio LLC, a non-profit organization      %
  21 %  dedicated to making software imaging solutions freely available.           %
  22 %                                                                             %
  23 %  You may not use this file except in compliance with the License.  You may  %
  24 %  obtain a copy of the License at                                            %
  25 %                                                                             %
  26 %    http://www.imagemagick.org/script/license.php                            %
  27 %                                                                             %
  28 %  Unless required by applicable law or agreed to in writing, software        %
  29 %  distributed under the License is distributed on an "AS IS" BASIS,          %
  30 %  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.   %
  31 %  See the License for the specific language governing permissions and        %
  32 %  limitations under the License.                                             %
  33 %                                                                             %
  34 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  35 %
  36 %
  37 %
  38 */
  39 \f
  40 /*
  41   Include declarations.
  42 */
  43 #include "MagickCore/studio.h"
  44 #include "MagickCore/exception.h"
  45 #include "MagickCore/exception-private.h"
  46 #include "MagickCore/image.h"
  47 #include "MagickCore/memory_.h"
  48 #include "MagickCore/string_.h"
  49 #include "MagickCore/string-private.h"
  50 #include "MagickCore/token.h"
  51 #include "MagickCore/token-private.h"
  52 #include "MagickCore/utility.h"
  53 #include "MagickCore/utility-private.h"
  54 \f
  55 /*
  56   Typedef declaractions.
  57 */
  58 struct _TokenInfo
  59 {
  60   int
  61     state;
  62
  63   MagickStatusType
  64     flag;
  65
  66   ssize_t
  67     offset;
  68
  69   char
  70     quote;
  71
  72   size_t
  73     signature;
  74 };
  75 \f
  76 /*
  77 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  78 %                                                                             %
  79 %                                                                             %
  80 %                                                                             %
  81 %   A c q u i r e T o k e n I n f o                                           %
  82 %                                                                             %
  83 %                                                                             %
  84 %                                                                             %
  85 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  86 %
  87 %  AcquireTokenInfo() allocates the TokenInfo structure.
  88 %
  89 %  The format of the AcquireTokenInfo method is:
  90 %
  91 %      TokenInfo *AcquireTokenInfo()
  92 %
  93 */
  94 MagickExport TokenInfo *AcquireTokenInfo(void)
  95 {
  96   TokenInfo
  97     *token_info;
  98
  99   token_info=(TokenInfo *) AcquireMagickMemory(sizeof(*token_info));
 100   if (token_info == (TokenInfo *) NULL)
 101     ThrowFatalException(ResourceLimitFatalError,"MemoryAllocationFailed");
 102   token_info->signature=MagickSignature;
 103   return(token_info);
 104 }
 105 \f
 106 /*
 107 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 108 %                                                                             %
 109 %                                                                             %
 110 %                                                                             %
 111 %   D e s t r o y T o k e n I n f o                                           %
 112 %                                                                             %
 113 %                                                                             %
 114 %                                                                             %
 115 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 116 %
 117 %  DestroyTokenInfo() deallocates memory associated with an TokenInfo
 118 %  structure.
 119 %
 120 %  The format of the DestroyTokenInfo method is:
 121 %
 122 %      TokenInfo *DestroyTokenInfo(TokenInfo *token_info)
 123 %
 124 %  A description of each parameter follows:
 125 %
 126 %    o token_info: Specifies a pointer to an TokenInfo structure.
 127 %
 128 */
 129 MagickExport TokenInfo *DestroyTokenInfo(TokenInfo *token_info)
 130 {
 131   (void) LogMagickEvent(TraceEvent,GetMagickModule(),"...");
 132   assert(token_info != (TokenInfo *) NULL);
 133   assert(token_info->signature == MagickSignature);
 134   token_info->signature=(~MagickSignature);
 135   token_info=(TokenInfo *) RelinquishMagickMemory(token_info);
 136   return(token_info);
 137 }
 138 \f
 139 /*
 140 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 141 %                                                                             %
 142 %                                                                             %
 143 %                                                                             %
 144 +   G e t M a g i c k T o k e n                                               %
 145 %                                                                             %
 146 %                                                                             %
 147 %                                                                             %
 148 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 149 %
 150 %  GetMagickToken() gets a token from the token stream.  A token is defined as
 151 %  a sequence of characters delimited by whitespace (e.g. clip-path), a
 152 %  sequence delimited with quotes (.e.g "Quote me"), or a sequence enclosed in
 153 %  parenthesis (e.g. rgb(0,0,0)).  GetMagickToken() also recognizes these
 154 %  separator characters: ':', '=', ',', and ';'.
 155 %
 156 %  The format of the GetMagickToken method is:
 157 %
 158 %      void GetMagickToken(const char *start,const char **end,char *token)
 159 %
 160 %  A description of each parameter follows:
 161 %
 162 %    o start: the start of the token sequence.
 163 %
 164 %    o end: point to the end of the token sequence.
 165 %
 166 %    o token: copy the token to this buffer.
 167 %
 168 */
 169 MagickExport void GetMagickToken(const char *start,const char **end,char *token)
 170 {
 171   double
 172     value;
 173
 174   register const char
 175     *p;
 176
 177   register ssize_t
 178     i;
 179
 180   assert(start != (const char *) NULL);
 181   assert(token != (char *) NULL);
 182   i=0;
 183   for (p=start; *p != '\0'; )
 184   {
 185     while ((isspace((int) ((unsigned char) *p)) != 0) && (*p != '\0'))
 186       p++;
 187     if (*p == '\0')
 188       break;
 189     switch (*p)
 190     {
 191       case '"':
 192       case '\'':
 193       case '`':
 194       case '{':
 195       {
 196         register char
 197           escape;
 198
 199         switch (*p)
 200         {
 201           case '"': escape='"'; break;
 202           case '\'': escape='\''; break;
 203           case '`': escape='\''; break;
 204           case '{': escape='}'; break;
 205           default: escape=(*p); break;
 206         }
 207         for (p++; *p != '\0'; p++)
 208         {
 209           if ((*p == '\\') && ((*(p+1) == escape) || (*(p+1) == '\\')))
 210             p++;
 211           else
 212             if (*p == escape)
 213               {
 214                 p++;
 215                 break;
 216               }
 217           token[i++]=(*p);
 218         }
 219         break;
 220       }
 221       case '/':
 222       {
 223         token[i++]=(*p++);
 224         if ((*p == '>') || (*p == '/'))
 225           token[i++]=(*p++);
 226         break;
 227       }
 228       default:
 229       {
 230         char
 231           *q;
 232
 233         value=StringToDouble(p,&q);
 234         (void) value;
 235         if ((p != q) && (*p != ','))
 236           {
 237             for ( ; (p < q) && (*p != ','); p++)
 238               token[i++]=(*p);
 239             if (*p == '%')
 240               token[i++]=(*p++);
 241             break;
 242           }
 243         if ((*p != '\0') && (isalpha((int) ((unsigned char) *p)) == 0) &&
 244             (*p != *DirectorySeparator) && (*p != '#') && (*p != '<'))
 245           {
 246             token[i++]=(*p++);
 247             break;
 248           }
 249         for ( ; *p != '\0'; p++)
 250         {
 251           if (((isspace((int) ((unsigned char) *p)) != 0) || (*p == '=') ||
 252               (*p == ':') || (*p == ',') || (*p == '|') || (*p == ';')) &&
 253               (*(p-1) != '\\'))
 254             break;
 255           if ((i > 0) && (*p == '<'))
 256             break;
 257           token[i++]=(*p);
 258           if (*p == '>')
 259             break;
 260           if (*p == '(')
 261             for (p++; *p != '\0'; p++)
 262             {
 263               token[i++]=(*p);
 264               if ((*p == ')') && (*(p-1) != '\\'))
 265                 break;
 266             }
 267         }
 268         break;
 269       }
 270     }
 271     break;
 272   }
 273   token[i]='\0';
 274   if (LocaleNCompare(token,"url(",4) == 0)
 275     {
 276       ssize_t
 277         offset;
 278
 279       offset=4;
 280       if (token[offset] == '#')
 281         offset++;
 282       i=(ssize_t) strlen(token);
 283       (void) CopyMagickString(token,token+offset,MaxTextExtent);
 284       token[i-offset-1]='\0';
 285     }
 286   while (isspace((int) ((unsigned char) *p)) != 0)
 287     p++;
 288   if (end != (const char **) NULL)
 289     *end=(const char *) p;
 290 }
 291 \f
 292 /*
 293 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 294 %                                                                             %
 295 %                                                                             %
 296 %                                                                             %
 297 %   G l o b E x p r e s s i o n                                               %
 298 %                                                                             %
 299 %                                                                             %
 300 %                                                                             %
 301 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 302 %
 303 %  GlobExpression() returns MagickTrue if the expression matches the pattern.
 304 %
 305 %  The format of the GlobExpression function is:
 306 %
 307 %      MagickBooleanType GlobExpression(const char *expression,
 308 %        const char *pattern,const MagickBooleanType case_insensitive)
 309 %
 310 %  A description of each parameter follows:
 311 %
 312 %    o expression: Specifies a pointer to a text string containing a file name.
 313 %
 314 %    o pattern: Specifies a pointer to a text string containing a pattern.
 315 %
 316 %    o case_insensitive: set to MagickTrue to ignore the case when matching
 317 %      an expression.
 318 %
 319 */
 320 MagickExport MagickBooleanType GlobExpression(const char *expression,
 321   const char *pattern,const MagickBooleanType case_insensitive)
 322 {
 323   MagickBooleanType
 324     done,
 325     match;
 326
 327   register const char
 328     *p;
 329
 330   /*
 331     Return on empty pattern or '*'.
 332   */
 333   if (pattern == (char *) NULL)
 334     return(MagickTrue);
 335   if (GetUTFCode(pattern) == 0)
 336     return(MagickTrue);
 337   if (LocaleCompare(pattern,"*") == 0)
 338     return(MagickTrue);
 339   p=pattern+strlen(pattern)-1;
 340   if ((GetUTFCode(p) == ']') && (strchr(pattern,'[') != (char *) NULL))
 341     {
 342       ExceptionInfo
 343         *exception;
 344
 345       ImageInfo
 346         *image_info;
 347
 348       /*
 349         Determine if pattern is a scene, i.e. img0001.pcd[2].
 350       */
 351       image_info=AcquireImageInfo();
 352       (void) CopyMagickString(image_info->filename,pattern,MaxTextExtent);
 353       exception=AcquireExceptionInfo();
 354       (void) SetImageInfo(image_info,0,exception);
 355       exception=DestroyExceptionInfo(exception);
 356       if (LocaleCompare(image_info->filename,pattern) != 0)
 357         {
 358           image_info=DestroyImageInfo(image_info);
 359           return(MagickFalse);
 360         }
 361       image_info=DestroyImageInfo(image_info);
 362     }
 363   /*
 364     Evaluate glob expression.
 365   */
 366   done=MagickFalse;
 367   while ((GetUTFCode(pattern) != 0) && (done == MagickFalse))
 368   {
 369     if (GetUTFCode(expression) == 0)
 370       if ((GetUTFCode(pattern) != '{') && (GetUTFCode(pattern) != '*'))
 371         break;
 372     switch (GetUTFCode(pattern))
 373     {
 374       case '*':
 375       {
 376         MagickBooleanType
 377           status;
 378
 379         status=MagickFalse;
 380         pattern+=GetUTFOctets(pattern);
 381         while ((GetUTFCode(expression) != 0) && (status == MagickFalse))
 382         {
 383           status=GlobExpression(expression,pattern,case_insensitive);
 384           expression+=GetUTFOctets(expression);
 385         }
 386         if (status != MagickFalse)
 387           {
 388             while (GetUTFCode(expression) != 0)
 389               expression+=GetUTFOctets(expression);
 390             while (GetUTFCode(pattern) != 0)
 391               pattern+=GetUTFOctets(pattern);
 392           }
 393         break;
 394       }
 395       case '[':
 396       {
 397         int
 398           c;
 399
 400         pattern+=GetUTFOctets(pattern);
 401         for ( ; ; )
 402         {
 403           if ((GetUTFCode(pattern) == 0) || (GetUTFCode(pattern) == ']'))
 404             {
 405               done=MagickTrue;
 406               break;
 407             }
 408           if (GetUTFCode(pattern) == '\\')
 409             {
 410               pattern+=GetUTFOctets(pattern);
 411               if (GetUTFCode(pattern) == 0)
 412                 {
 413                   done=MagickTrue;
 414                   break;
 415                 }
 416              }
 417           if (GetUTFCode(pattern+GetUTFOctets(pattern)) == '-')
 418             {
 419               c=GetUTFCode(pattern);
 420               pattern+=GetUTFOctets(pattern);
 421               pattern+=GetUTFOctets(pattern);
 422               if (GetUTFCode(pattern) == ']')
 423                 {
 424                   done=MagickTrue;
 425                   break;
 426                 }
 427               if (GetUTFCode(pattern) == '\\')
 428                 {
 429                   pattern+=GetUTFOctets(pattern);
 430                   if (GetUTFCode(pattern) == 0)
 431                     {
 432                       done=MagickTrue;
 433                       break;
 434                     }
 435                 }
 436               if ((GetUTFCode(expression) < c) ||
 437                   (GetUTFCode(expression) > GetUTFCode(pattern)))
 438                 {
 439                   pattern+=GetUTFOctets(pattern);
 440                   continue;
 441                 }
 442             }
 443           else
 444             if (GetUTFCode(pattern) != GetUTFCode(expression))
 445               {
 446                 pattern+=GetUTFOctets(pattern);
 447                 continue;
 448               }
 449           pattern+=GetUTFOctets(pattern);
 450           while ((GetUTFCode(pattern) != ']') && (GetUTFCode(pattern) != 0))
 451           {
 452             if ((GetUTFCode(pattern) == '\\') &&
 453                 (GetUTFCode(pattern+GetUTFOctets(pattern)) > 0))
 454               pattern+=GetUTFOctets(pattern);
 455             pattern+=GetUTFOctets(pattern);
 456           }
 457           if (GetUTFCode(pattern) != 0)
 458             {
 459               pattern+=GetUTFOctets(pattern);
 460               expression+=GetUTFOctets(expression);
 461             }
 462           break;
 463         }
 464         break;
 465       }
 466       case '?':
 467       {
 468         pattern+=GetUTFOctets(pattern);
 469         expression+=GetUTFOctets(expression);
 470         break;
 471       }
 472       case '{':
 473       {
 474         register const char
 475           *p;
 476
 477         pattern+=GetUTFOctets(pattern);
 478         while ((GetUTFCode(pattern) != '}') && (GetUTFCode(pattern) != 0))
 479         {
 480           p=expression;
 481           match=MagickTrue;
 482           while ((GetUTFCode(p) != 0) && (GetUTFCode(pattern) != 0) &&
 483                  (GetUTFCode(pattern) != ',') && (GetUTFCode(pattern) != '}') &&
 484                  (match != MagickFalse))
 485           {
 486             if (GetUTFCode(pattern) == '\\')
 487               pattern+=GetUTFOctets(pattern);
 488             match=(GetUTFCode(pattern) == GetUTFCode(p)) ? MagickTrue :
 489               MagickFalse;
 490             p+=GetUTFOctets(p);
 491             pattern+=GetUTFOctets(pattern);
 492           }
 493           if (GetUTFCode(pattern) == 0)
 494             {
 495               match=MagickFalse;
 496               done=MagickTrue;
 497               break;
 498             }
 499           else
 500             if (match != MagickFalse)
 501               {
 502                 expression=p;
 503                 while ((GetUTFCode(pattern) != '}') &&
 504                        (GetUTFCode(pattern) != 0))
 505                 {
 506                   pattern+=GetUTFOctets(pattern);
 507                   if (GetUTFCode(pattern) == '\\')
 508                     {
 509                       pattern+=GetUTFOctets(pattern);
 510                       if (GetUTFCode(pattern) == '}')
 511                         pattern+=GetUTFOctets(pattern);
 512                     }
 513                 }
 514               }
 515             else
 516               {
 517                 while ((GetUTFCode(pattern) != '}') &&
 518                        (GetUTFCode(pattern) != ',') &&
 519                        (GetUTFCode(pattern) != 0))
 520                 {
 521                   pattern+=GetUTFOctets(pattern);
 522                   if (GetUTFCode(pattern) == '\\')
 523                     {
 524                       pattern+=GetUTFOctets(pattern);
 525                       if ((GetUTFCode(pattern) == '}') ||
 526                           (GetUTFCode(pattern) == ','))
 527                         pattern+=GetUTFOctets(pattern);
 528                     }
 529                 }
 530               }
 531             if (GetUTFCode(pattern) != 0)
 532               pattern+=GetUTFOctets(pattern);
 533           }
 534         break;
 535       }
 536       case '\\':
 537       {
 538         pattern+=GetUTFOctets(pattern);
 539         if (GetUTFCode(pattern) == 0)
 540           break;
 541       }
 542       default:
 543       {
 544         if (case_insensitive != MagickFalse)
 545           {
 546             if (tolower((int) GetUTFCode(expression)) !=
 547                 tolower((int) GetUTFCode(pattern)))
 548               {
 549                 done=MagickTrue;
 550                 break;
 551               }
 552           }
 553         else
 554           if (GetUTFCode(expression) != GetUTFCode(pattern))
 555             {
 556               done=MagickTrue;
 557               break;
 558             }
 559         expression+=GetUTFOctets(expression);
 560         pattern+=GetUTFOctets(pattern);
 561       }
 562     }
 563   }
 564   while (GetUTFCode(pattern) == '*')
 565     pattern+=GetUTFOctets(pattern);
 566   match=(GetUTFCode(expression) == 0) && (GetUTFCode(pattern) == 0) ?
 567     MagickTrue : MagickFalse;
 568   return(match);
 569 }
 570 \f
 571 /*
 572 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 573 %                                                                             %
 574 %                                                                             %
 575 %                                                                             %
 576 +     I s G l o b                                                             %
 577 %                                                                             %
 578 %                                                                             %
 579 %                                                                             %
 580 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 581 %
 582 %  IsGlob() returns MagickTrue if the path specification contains a globbing
 583 %  pattern.
 584 %
 585 %  The format of the IsGlob method is:
 586 %
 587 %      MagickBooleanType IsGlob(const char *geometry)
 588 %
 589 %  A description of each parameter follows:
 590 %
 591 %    o path: the path.
 592 %
 593 */
 594 MagickPrivate MagickBooleanType IsGlob(const char *path)
 595 {
 596   MagickBooleanType
 597     status = MagickFalse;
 598
 599   register const char
 600     *p;
 601
 602   if (IsPathAccessible(path) != MagickFalse)
 603     return(MagickFalse);
 604   for (p=path; *p != '\0'; p++)
 605   {
 606     switch (*p)
 607     {
 608       case '*':
 609       case '?':
 610       case '{':
 611       case '}':
 612       case '[':
 613       case ']':
 614       {
 615         status=MagickTrue;
 616         break;
 617       }
 618       default:
 619         break;
 620     }
 621   }
 622   return(status);
 623 }
 624 \f
 625 /*
 626 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 627 %                                                                             %
 628 %                                                                             %
 629 %                                                                             %
 630 %   T o k e n i z e r                                                         %
 631 %                                                                             %
 632 %                                                                             %
 633 %                                                                             %
 634 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 635 %
 636 %  Tokenizer() is a generalized, finite state token parser.  It extracts tokens
 637 %  one at a time from a string of characters.  The characters used for white
 638 %  space, for break characters, and for quotes can be specified.  Also,
 639 %  characters in the string can be preceded by a specifiable escape character
 640 %  which removes any special meaning the character may have.
 641 %
 642 %  Here is some terminology:
 643 %
 644 %    o token: A single unit of information in the form of a group of
 645 %      characters.
 646 %
 647 %    o white space: Apace that gets ignored (except within quotes or when
 648 %      escaped), like blanks and tabs. in addition, white space terminates a
 649 %      non-quoted token.
 650 %
 651 %    o break set: One or more characters that separates non-quoted tokens.
 652 %      Commas are a common break character. The usage of break characters to
 653 %      signal the end of a token is the same as that of white space, except
 654 %      multiple break characters with nothing or only white space between
 655 %      generate a null token for each two break characters together.
 656 %
 657 %      For example, if blank is set to be the white space and comma is set to
 658 %      be the break character, the line
 659 %
 660 %        A, B, C ,  , DEF
 661 %
 662 %        ... consists of 5 tokens:
 663 %
 664 %        1)  "A"
 665 %        2)  "B"
 666 %        3)  "C"
 667 %        4)  "" (the null string)
 668 %        5)  "DEF"
 669 %
 670 %    o Quote character: A character that, when surrounding a group of other
 671 %      characters, causes the group of characters to be treated as a single
 672 %      token, no matter how many white spaces or break characters exist in
 673 %      the group. Also, a token always terminates after the closing quote.
 674 %      For example, if ' is the quote character, blank is white space, and
 675 %      comma is the break character, the following string
 676 %
 677 %        A, ' B, CD'EF GHI
 678 %
 679 %        ... consists of 4 tokens:
 680 %
 681 %        1)  "A"
 682 %        2)  " B, CD" (note the blanks & comma)
 683 %        3)  "EF"
 684 %        4)  "GHI"
 685 %
 686 %      The quote characters themselves do not appear in the resultant
 687 %      tokens.  The double quotes are delimiters i use here for
 688 %      documentation purposes only.
 689 %
 690 %    o Escape character: A character which itself is ignored but which
 691 %      causes the next character to be used as is.  ^ and \ are often used
 692 %      as escape characters. An escape in the last position of the string
 693 %      gets treated as a "normal" (i.e., non-quote, non-white, non-break,
 694 %      and non-escape) character. For example, assume white space, break
 695 %      character, and quote are the same as in the above examples, and
 696 %      further, assume that ^ is the escape character. Then, in the string
 697 %
 698 %        ABC, ' DEF ^' GH' I ^ J K^ L ^
 699 %
 700 %        ... there are 7 tokens:
 701 %
 702 %        1)  "ABC"
 703 %        2)  " DEF ' GH"
 704 %        3)  "I"
 705 %        4)  " "     (a lone blank)
 706 %        5)  "J"
 707 %        6)  "K L"
 708 %        7)  "^"     (passed as is at end of line)
 709 %
 710 %  The format of the Tokenizer method is:
 711 %
 712 %      int Tokenizer(TokenInfo *token_info,const unsigned flag,char *token,
 713 %        const size_t max_token_length,const char *line,const char *white,
 714 %        const char *break_set,const char *quote,const char escape,
 715 %        char *breaker,int *next,char *quoted)
 716 %
 717 %  A description of each parameter follows:
 718 %
 719 %    o flag: right now, only the low order 3 bits are used.
 720 %
 721 %        1 => convert non-quoted tokens to upper case
 722 %        2 => convert non-quoted tokens to lower case
 723 %        0 => do not convert non-quoted tokens
 724 %
 725 %    o token: a character string containing the returned next token
 726 %
 727 %    o max_token_length: the maximum size of "token".  Characters beyond
 728 %      "max_token_length" are truncated.
 729 %
 730 %    o string: the string to be parsed.
 731 %
 732 %    o white: a string of the valid white spaces.  example:
 733 %
 734 %        char whitesp[]={" \t"};
 735 %
 736 %      blank and tab will be valid white space.
 737 %
 738 %    o break: a string of the valid break characters. example:
 739 %
 740 %        char breakch[]={";,"};
 741 %
 742 %      semicolon and comma will be valid break characters.
 743 %
 744 %    o quote: a string of the valid quote characters. An example would be
 745 %
 746 %        char whitesp[]={"'\"");
 747 %
 748 %      (this causes single and double quotes to be valid) Note that a
 749 %      token starting with one of these characters needs the same quote
 750 %      character to terminate it.
 751 %
 752 %      for example:
 753 %
 754 %        "ABC '
 755 %
 756 %      is unterminated, but
 757 %
 758 %        "DEF" and 'GHI'
 759 %
 760 %      are properly terminated.  Note that different quote characters
 761 %      can appear on the same line; only for a given token do the quote
 762 %      characters have to be the same.
 763 %
 764 %    o escape: the escape character (NOT a string ... only one
 765 %      allowed). Use zero if none is desired.
 766 %
 767 %    o breaker: the break character used to terminate the current
 768 %      token.  If the token was quoted, this will be the quote used.  If
 769 %      the token is the last one on the line, this will be zero.
 770 %
 771 %    o next: this variable points to the first character of the
 772 %      next token.  it gets reset by "tokenizer" as it steps through the
 773 %      string.  Set it to 0 upon initialization, and leave it alone
 774 %      after that.  You can change it if you want to jump around in the
 775 %      string or re-parse from the beginning, but be careful.
 776 %
 777 %    o quoted: set to True if the token was quoted and MagickFalse
 778 %      if not.  You may need this information (for example:  in C, a
 779 %      string with quotes around it is a character string, while one
 780 %      without is an identifier).
 781 %
 782 %    o result: 0 if we haven't reached EOS (end of string), and 1
 783 %      if we have.
 784 %
 785 */
 786
 787 #define IN_WHITE 0
 788 #define IN_TOKEN 1
 789 #define IN_QUOTE 2
 790 #define IN_OZONE 3
 791
 792 static ssize_t sindex(int c,const char *string)
 793 {
 794   register const char
 795     *p;
 796
 797   for (p=string; *p != '\0'; p++)
 798     if (c == (int) (*p))
 799       return((ssize_t) (p-string));
 800   return(-1);
 801 }
 802
 803 static void StoreToken(TokenInfo *token_info,char *string,
 804   size_t max_token_length,int c)
 805 {
 806   register ssize_t
 807     i;
 808
 809   if ((token_info->offset < 0) ||
 810       ((size_t) token_info->offset >= (max_token_length-1)))
 811     return;
 812   i=token_info->offset++;
 813   string[i]=(char) c;
 814   if (token_info->state == IN_QUOTE)
 815     return;
 816   switch (token_info->flag & 0x03)
 817   {
 818     case 1:
 819     {
 820       string[i]=(char) toupper(c);
 821       break;
 822     }
 823     case 2:
 824     {
 825       string[i]=(char) tolower(c);
 826       break;
 827     }
 828     default:
 829       break;
 830   }
 831 }
 832
 833 MagickExport int Tokenizer(TokenInfo *token_info,const unsigned flag,
 834   char *token,const size_t max_token_length,const char *line,const char *white,
 835   const char *break_set,const char *quote,const char escape,char *breaker,
 836   int *next,char *quoted)
 837 {
 838   int
 839     c;
 840
 841   register ssize_t
 842     i;
 843
 844   *breaker='\0';
 845   *quoted='\0';
 846   if (line[*next] == '\0')
 847     return(1);
 848   token_info->state=IN_WHITE;
 849   token_info->quote=(char) MagickFalse;
 850   token_info->flag=flag;
 851   for (token_info->offset=0; (int) line[*next] != 0; (*next)++)
 852   {
 853     c=(int) line[*next];
 854     i=sindex(c,break_set);
 855     if (i >= 0)
 856       {
 857         switch (token_info->state)
 858         {
 859           case IN_WHITE:
 860           case IN_TOKEN:
 861           case IN_OZONE:
 862           {
 863             (*next)++;
 864             *breaker=break_set[i];
 865             token[token_info->offset]='\0';
 866             return(0);
 867           }
 868           case IN_QUOTE:
 869           {
 870             StoreToken(token_info,token,max_token_length,c);
 871             break;
 872           }
 873         }
 874         continue;
 875       }
 876     i=sindex(c,quote);
 877     if (i >= 0)
 878       {
 879         switch (token_info->state)
 880         {
 881           case IN_WHITE:
 882           {
 883             token_info->state=IN_QUOTE;
 884             token_info->quote=quote[i];
 885             *quoted=(char) MagickTrue;
 886             break;
 887           }
 888           case IN_QUOTE:
 889           {
 890             if (quote[i] != token_info->quote)
 891               StoreToken(token_info,token,max_token_length,c);
 892             else
 893               {
 894                 token_info->state=IN_OZONE;
 895                 token_info->quote='\0';
 896               }
 897             break;
 898           }
 899           case IN_TOKEN:
 900           case IN_OZONE:
 901           {
 902             *breaker=(char) c;
 903             token[token_info->offset]='\0';
 904             return(0);
 905           }
 906         }
 907         continue;
 908       }
 909     i=sindex(c,white);
 910     if (i >= 0)
 911       {
 912         switch (token_info->state)
 913         {
 914           case IN_WHITE:
 915           case IN_OZONE:
 916             break;
 917           case IN_TOKEN:
 918           {
 919             token_info->state=IN_OZONE;
 920             break;
 921           }
 922           case IN_QUOTE:
 923           {
 924             StoreToken(token_info,token,max_token_length,c);
 925             break;
 926           }
 927         }
 928         continue;
 929       }
 930     if (c == (int) escape)
 931       {
 932         if (line[(*next)+1] == '\0')
 933           {
 934             *breaker='\0';
 935             StoreToken(token_info,token,max_token_length,c);
 936             (*next)++;
 937             token[token_info->offset]='\0';
 938             return(0);
 939           }
 940         switch (token_info->state)
 941         {
 942           case IN_WHITE:
 943           {
 944             (*next)--;
 945             token_info->state=IN_TOKEN;
 946             break;
 947           }
 948           case IN_TOKEN:
 949           case IN_QUOTE:
 950           {
 951             (*next)++;
 952             c=(int) line[*next];
 953             StoreToken(token_info,token,max_token_length,c);
 954             break;
 955           }
 956           case IN_OZONE:
 957           {
 958             token[token_info->offset]='\0';
 959             return(0);
 960           }
 961         }
 962         continue;
 963       }
 964     switch (token_info->state)
 965     {
 966       case IN_WHITE:
 967         token_info->state=IN_TOKEN;
 968       case IN_TOKEN:
 969       case IN_QUOTE:
 970       {
 971         StoreToken(token_info,token,max_token_length,c);
 972         break;
 973       }
 974       case IN_OZONE:
 975       {
 976         token[token_info->offset]='\0';
 977         return(0);
 978       }
 979     }
 980   }
 981   token[token_info->offset]='\0';
 982   return(0);
 983 }