standards/rfctohtml.c

   1 /*
   2  * "$Id: rfctohtml.c 6649 2007-07-11 21:46:42Z mike $"
   3  *
   4  *   RFC file to HTML conversion program.
   5  *
   6  *   Copyright 2007 by Apple Inc.
   7  *   Copyright 2006 by Easy Software Products.
   8  *
   9  *   These coded instructions, statements, and computer programs are the
  10  *   property of Apple Inc. and are protected by Federal copyright
  11  *   law.  Distribution and use rights are outlined in the file "LICENSE.txt"
  12  *   which should have been included with this file.  If this file is
  13  *   file is missing or damaged, see the license at "http://www.cups.org/".
  14  *
  15  * Contents:
  16  *
  17  *   main()       - Convert a man page to HTML.
  18  *   put_entity() - Put a single character, using entities as needed.
  19  *   put_line()   - Put a whole string for a line.
  20  */
  21
  22 /*
  23  * Include necessary headers.
  24  */
  25
  26 #include <cups/string.h>
  27 #include <stdlib.h>
  28 #include <cups/file.h>
  29
  30
  31 /*
  32  * Local functions...
  33  */
  34
  35 void    put_entity(cups_file_t *fp, int ch);
  36 void    put_line(cups_file_t *fp, const char *line);
  37
  38
  39 /*
  40  * 'main()' - Convert a man page to HTML.
  41  */
  42
  43 int                                     /* O - Exit status */
  44 main(int  argc,                         /* I - Number of command-line args */
  45      char *argv[])                      /* I - Command-line arguments */
  46 {
  47   cups_file_t   *infile,                /* Input file */
  48                 *outfile;               /* Output file */
  49   char          line[1024],             /* Line from file */
  50                 *lineptr,               /* Pointer into line */
  51                 name[1024],             /* Heading anchor name */
  52                 *nameptr;               /* Pointer into anchor name */
  53   int           rfc,                    /* RFC # */
  54                 inheading,              /* Inside a heading? */
  55                 inpre,                  /* Inside preformatted text? */
  56                 intoc,                  /* Inside table-of-contents? */
  57                 toclevel,               /* Current table-of-contents level */
  58                 linenum;                /* Current line on page */
  59
  60
  61  /*
  62   * Check arguments...
  63   */
  64
  65   if (argc > 3)
  66   {
  67     fputs("Usage: rfctohtml [rfcNNNN.txt [rfcNNNN.html]]\n", stderr);
  68     return (1);
  69   }
  70
  71  /*
  72   * Open files as needed...
  73   */
  74
  75   if (argc > 1)
  76   {
  77     if ((infile = cupsFileOpen(argv[1], "r")) == NULL)
  78     {
  79       perror(argv[1]);
  80       return (1);
  81     }
  82   }
  83   else
  84     infile = cupsFileOpenFd(0, "r");
  85
  86   if (argc > 2)
  87   {
  88     if ((outfile = cupsFileOpen(argv[2], "w")) == NULL)
  89     {
  90       perror(argv[2]);
  91       cupsFileClose(infile);
  92       return (1);
  93     }
  94   }
  95   else
  96     outfile = cupsFileOpenFd(1, "w");
  97
  98  /*
  99   * Read from input and write the output...
 100   */
 101
 102   cupsFilePuts(outfile,
 103                "<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.0 Transitional//EN\" "
 104                "\"http://www.w3.org/TR/REC-html40/loose.dtd\">\n"
 105                "<html>\n"
 106                "<!-- SECTION: Specifications -->\n"
 107                "<head>\n"
 108                "\t<style type='text/css'><!--\n"
 109                "\th1, h2, h3 { font-family: sans-serif; }\n"
 110                "\tp, pre { font-family: monospace; }\n"
 111                "\th2.title, h3.title, h3.title { border-bottom: solid "
 112                "2px #000000; }\n"
 113                "\t--></style>\n");
 114
 115  /*
 116   * Skip the initial header stuff (working group ID, RFC #, authors, and
 117   * copyright...
 118   */
 119
 120   linenum = 0;
 121   rfc     = 0;
 122
 123   while (cupsFileGets(infile, line, sizeof(line)))
 124   {
 125     linenum ++;
 126
 127     if (line[0])
 128       break;
 129   }
 130
 131   while (cupsFileGets(infile, line, sizeof(line)))
 132   {
 133     linenum ++;
 134
 135     if (!line[0])
 136       break;
 137     else if (!strncasecmp(line, "Request for Comments:", 21))
 138       rfc = atoi(line + 21);
 139   }
 140
 141  /*
 142   * Read the document title...
 143   */
 144
 145   while (cupsFileGets(infile, line, sizeof(line)))
 146   {
 147     linenum ++;
 148
 149     if (line[0])
 150       break;
 151   }
 152
 153   for (lineptr = line; isspace(*lineptr & 255); lineptr ++);
 154
 155   cupsFilePrintf(outfile, "<title>RFC %d: %s", rfc, lineptr);
 156
 157   while (cupsFileGets(infile, line, sizeof(line)))
 158   {
 159     linenum ++;
 160
 161     if (!line[0])
 162       break;
 163     else
 164     {
 165       for (lineptr = line; isspace(*lineptr & 255); lineptr ++);
 166       cupsFilePrintf(outfile, " %s", lineptr);
 167     }
 168   }
 169
 170   cupsFilePuts(outfile, "</title>\n"
 171                         "</head>\n"
 172                         "<body>\n");
 173
 174  /*
 175   * Read the rest of the file...
 176   */
 177
 178   inheading = 0;
 179   inpre     = 0;
 180   intoc     = 0;
 181   toclevel  = 0;
 182
 183   while (cupsFileGets(infile, line, sizeof(line)))
 184   {
 185     linenum ++;
 186
 187     if (!line[0])
 188     {
 189       if (linenum > 50)
 190         continue;
 191
 192       if (inpre)
 193       {
 194         cupsFilePuts(outfile, "</pre>\n");
 195         inpre = 0;
 196       }
 197
 198       if (inheading)
 199       {
 200         if (inheading < 0)
 201           cupsFilePuts(outfile, "</h1>\n");
 202         else
 203           cupsFilePrintf(outfile, "</a></h%d>\n", inheading);
 204
 205         inheading = 0;
 206       }
 207     }
 208     else if ((line[0] == ' ' ||
 209               (!isupper(line[0] & 255) && !isdigit(line[0] & 255) &&
 210                !strstr(line, "[Page "))) && !inheading)
 211     {
 212       if (inheading)
 213       {
 214         if (inheading < 0)
 215           cupsFilePuts(outfile, "</h1>\n");
 216         else
 217           cupsFilePrintf(outfile, "</a></h%d>\n", inheading);
 218
 219         inheading = 0;
 220       }
 221
 222       for (lineptr = line; *lineptr == ' '; lineptr ++);
 223
 224       if (intoc)
 225       {
 226         char    *temp;                  /* Temporary pointer into line */
 227         int     level;                  /* Heading level */
 228
 229
 230         if (isdigit(*lineptr & 255))
 231         {
 232           strlcpy(name, lineptr, sizeof(name));
 233
 234           for (nameptr = name, level = -1; *nameptr;)
 235             if (isdigit(*nameptr & 255))
 236             {
 237               while (isdigit(*nameptr & 255))
 238                 nameptr ++;
 239
 240               level ++;
 241             }
 242             else if (*nameptr == ' ')
 243             {
 244               *nameptr = '\0';
 245               break;
 246             }
 247             else
 248               nameptr ++;
 249
 250           while (toclevel > level)
 251           {
 252             cupsFilePuts(outfile, "\n</ul>");
 253             toclevel --;
 254           }
 255
 256           while (toclevel < level)
 257           {
 258             cupsFilePuts(outfile, "\n<ul style=\"list-style-type: none;\">\n");
 259             toclevel ++;
 260           }
 261
 262           cupsFilePrintf(outfile, "\n<%s><a href=\"#s%s\">", toclevel ? "li" : "p",
 263                          name);
 264         }
 265
 266         temp = lineptr + strlen(lineptr) - 1;
 267
 268         while (temp > lineptr)
 269           if (*temp == ' ' || !isdigit(*temp & 255))
 270             break;
 271           else
 272             temp --;
 273
 274         if (*temp == ' ')
 275         {
 276           while (temp > lineptr)
 277             if (*temp != ' ' && *temp != '.')
 278               break;
 279             else
 280               *temp-- = '\0';
 281         }
 282         else
 283           temp = NULL;
 284
 285         if (isdigit(*lineptr & 255))
 286           put_line(outfile, lineptr);
 287         else
 288           put_line(outfile, lineptr - 1);
 289
 290         if (temp)
 291           cupsFilePuts(outfile, "</a>");
 292       }
 293       else if (!inpre)
 294       {
 295         cupsFilePuts(outfile, "\n<pre>");
 296         put_line(outfile, line);
 297         inpre = 1;
 298       }
 299       else
 300       {
 301         cupsFilePutChar(outfile, '\n');
 302         put_line(outfile, line);
 303       }
 304     }
 305     else if (strstr(line, "[Page "))
 306     {
 307      /*
 308       * Skip page footer and header...
 309       */
 310
 311       cupsFileGets(infile, line, sizeof(line));
 312       cupsFileGets(infile, line, sizeof(line));
 313       cupsFileGets(infile, line, sizeof(line));
 314       cupsFileGets(infile, line, sizeof(line));
 315       linenum = 3;
 316     }
 317     else if (isdigit(line[0] & 255) && !inheading)
 318     {
 319       int level;                        /* Heading level */
 320
 321
 322       if (intoc)
 323       {
 324         while (toclevel > 0)
 325         {
 326           cupsFilePuts(outfile, "\n</ul>");
 327           toclevel --;
 328         }
 329
 330         cupsFilePutChar(outfile, '\n');
 331         intoc = 0;
 332       }
 333
 334       if (inpre)
 335       {
 336         cupsFilePuts(outfile, "</pre>\n");
 337         inpre = 0;
 338       }
 339
 340       strlcpy(name, line, sizeof(name));
 341       for (nameptr = name, level = 1; *nameptr;)
 342         if (isdigit(*nameptr & 255))
 343         {
 344           while (isdigit(*nameptr & 255))
 345             nameptr ++;
 346
 347           level ++;
 348         }
 349         else if (*nameptr == ' ')
 350         {
 351           *nameptr = '\0';
 352           break;
 353         }
 354         else
 355           nameptr ++;
 356
 357       cupsFilePrintf(outfile, "\n<h%d class='title'><a name='s%s'>", level,
 358                      name);
 359       put_line(outfile, line);
 360
 361       intoc     = 0;
 362       inheading = level;
 363     }
 364     else
 365     {
 366       if (intoc)
 367       {
 368         while (toclevel > 0)
 369         {
 370           cupsFilePuts(outfile, "\n</ul>");
 371           toclevel --;
 372         }
 373
 374         cupsFilePutChar(outfile, '\n');
 375         intoc = 0;
 376       }
 377
 378       if (!inheading)
 379       {
 380         cupsFilePuts(outfile, "\n<h2 class='title'>");
 381         inheading = -1;
 382       }
 383
 384       put_line(outfile, line);
 385
 386       intoc    = !strcasecmp(line, "Table of Contents");
 387       toclevel = 0;
 388     }
 389   }
 390
 391   if (inpre)
 392     cupsFilePuts(outfile, "</pre>\n");
 393
 394   if (inheading)
 395   {
 396     if (inheading < 0)
 397       cupsFilePuts(outfile, "</h2>\n");
 398     else
 399       cupsFilePrintf(outfile, "</a></h%d>\n", inheading);
 400   }
 401
 402   cupsFilePuts(outfile, "</body>\n"
 403                         "</html>\n");
 404
 405  /*
 406   * Close files...
 407   */
 408
 409   cupsFileClose(infile);
 410   cupsFileClose(outfile);
 411
 412  /*
 413   * Return with no errors...
 414   */
 415
 416   return (0);
 417 }
 418
 419
 420 /*
 421  * 'put_entity()' - Put a single character, using entities as needed.
 422  */
 423
 424 void
 425 put_entity(cups_file_t *fp,             /* I - File */
 426            int         ch)              /* I - Character */
 427 {
 428   if (ch == '&')
 429     cupsFilePuts(fp, "&amp;");
 430   else if (ch == '<')
 431     cupsFilePuts(fp, "&lt;");
 432   else
 433     cupsFilePutChar(fp, ch);
 434 }
 435
 436
 437 /*
 438  * 'put_line()' - Put a whole string for a line.
 439  */
 440
 441 void
 442 put_line(cups_file_t *fp,               /* I - File */
 443          const char  *s)                /* I - String */
 444 {
 445   int   whitespace,                     /* Saw whitespace */
 446         i,                              /* Looping var */
 447         len;                            /* Length of keyword */
 448   static const char * const keywords[] =/* Special keywords to boldface */
 449   {
 450     "MAY",
 451     "MUST",
 452     "NOT",
 453     "SHALL",
 454     "SHOULD"
 455   };
 456
 457
 458   whitespace = 1;
 459
 460   while (*s)
 461   {
 462     if (*s == ' ')
 463       whitespace = 1;
 464
 465     if (whitespace && isupper(*s & 255))
 466     {
 467       whitespace = 0;
 468
 469       for (i = 0; i < (int)(sizeof(keywords) / sizeof(sizeof(keywords[0]))); i ++)
 470       {
 471         len = strlen(keywords[i]);
 472         if (!strncmp(s, keywords[i], len) && (isspace(s[len] & 255) || !*s))
 473         {
 474           cupsFilePrintf(fp, "<b>%s</b>", keywords[i]);
 475           s += len;
 476           break;
 477         }
 478       }
 479
 480       if (i >= (int)(sizeof(keywords) / sizeof(sizeof(keywords[0]))))
 481         put_entity(fp, *s++);
 482     }
 483     else
 484     {
 485       if (*s != ' ')
 486         whitespace = 0;
 487
 488       put_entity(fp, *s++);
 489     }
 490   }
 491 }
 492
 493
 494 /*
 495  * End of "$Id: rfctohtml.c 6649 2007-07-11 21:46:42Z mike $".
 496  */