html.c

   1 /*      $Id: html.c,v 1.88 2009/11/14 19:23:58 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the above
   7  * copyright notice and this permission notice appear in all copies.
   8  *
   9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  16  */
  17 #include <sys/types.h>
  18
  19 #include <assert.h>
  20 #include <ctype.h>
  21 #include <stdarg.h>
  22 #include <stdio.h>
  23 #include <stdint.h>
  24 #include <stdlib.h>
  25 #include <string.h>
  26 #include <unistd.h>
  27
  28 #include "out.h"
  29 #include "chars.h"
  30 #include "html.h"
  31 #include "main.h"
  32
  33 #define UNCONST(a)      ((void *)(uintptr_t)(const void *)(a))
  34
  35 #define DOCTYPE         "-//W3C//DTD HTML 4.01//EN"
  36 #define DTD             "http://www.w3.org/TR/html4/strict.dtd"
  37
  38 struct  htmldata {
  39         const char       *name;
  40         int               flags;
  41 #define HTML_CLRLINE     (1 << 0)
  42 #define HTML_NOSTACK     (1 << 1)
  43 };
  44
  45 static  const struct htmldata htmltags[TAG_MAX] = {
  46         {"html",        HTML_CLRLINE}, /* TAG_HTML */
  47         {"head",        HTML_CLRLINE}, /* TAG_HEAD */
  48         {"body",        HTML_CLRLINE}, /* TAG_BODY */
  49         {"meta",        HTML_CLRLINE | HTML_NOSTACK}, /* TAG_META */
  50         {"title",       HTML_CLRLINE}, /* TAG_TITLE */
  51         {"div",         HTML_CLRLINE}, /* TAG_DIV */
  52         {"h1",          0}, /* TAG_H1 */
  53         {"h2",          0}, /* TAG_H2 */
  54         {"p",           HTML_CLRLINE}, /* TAG_P */
  55         {"span",        0}, /* TAG_SPAN */
  56         {"link",        HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
  57         {"br",          HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
  58         {"a",           0}, /* TAG_A */
  59         {"table",       HTML_CLRLINE}, /* TAG_TABLE */
  60         {"col",         HTML_CLRLINE | HTML_NOSTACK}, /* TAG_COL */
  61         {"tr",          HTML_CLRLINE}, /* TAG_TR */
  62         {"td",          HTML_CLRLINE}, /* TAG_TD */
  63         {"li",          HTML_CLRLINE}, /* TAG_LI */
  64         {"ul",          HTML_CLRLINE}, /* TAG_UL */
  65         {"ol",          HTML_CLRLINE}, /* TAG_OL */
  66         {"base",        HTML_CLRLINE | HTML_NOSTACK}, /* TAG_BASE */
  67 };
  68
  69 static  const char      *const htmlattrs[ATTR_MAX] = {
  70         "http-equiv",
  71         "content",
  72         "name",
  73         "rel",
  74         "href",
  75         "type",
  76         "media",
  77         "class",
  78         "style",
  79         "width",
  80         "valign",
  81         "target",
  82         "id",
  83         "summary",
  84 };
  85
  86 #ifdef __linux__
  87 extern  int               getsubopt(char **, char * const *, char **);
  88 #endif
  89
  90
  91 static  void              print_spec(struct html *, const char *, size_t);
  92 static  void              print_res(struct html *, const char *, size_t);
  93 static  void              print_ctag(struct html *, enum htmltag);
  94 static  int               print_encode(struct html *, const char *, int);
  95 static  void              print_metaf(struct html *, enum roffdeco);
  96
  97
  98 void *
  99 html_alloc(char *outopts)
 100 {
 101         struct html     *h;
 102         const char      *toks[4];
 103         char            *v;
 104
 105         toks[0] = "style";
 106         toks[1] = "man";
 107         toks[2] = "includes";
 108         toks[3] = NULL;
 109
 110         h = calloc(1, sizeof(struct html));
 111         if (NULL == h) {
 112                 perror(NULL);
 113                 exit(EXIT_FAILURE);
 114         }
 115
 116         h->tags.head = NULL;
 117         h->ords.head = NULL;
 118         h->symtab = chars_init(CHARS_HTML);
 119
 120         while (outopts && *outopts)
 121                 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
 122                 case (0):
 123                         h->style = v;
 124                         break;
 125                 case (1):
 126                         h->base_man = v;
 127                         break;
 128                 case (2):
 129                         h->base_includes = v;
 130                         break;
 131                 default:
 132                         break;
 133                 }
 134
 135         return(h);
 136 }
 137
 138
 139 void
 140 html_free(void *p)
 141 {
 142         struct tag      *tag;
 143         struct ord      *ord;
 144         struct html     *h;
 145
 146         h = (struct html *)p;
 147
 148         while ((ord = h->ords.head) != NULL) {
 149                 h->ords.head = ord->next;
 150                 free(ord);
 151         }
 152
 153         while ((tag = h->tags.head) != NULL) {
 154                 h->tags.head = tag->next;
 155                 free(tag);
 156         }
 157
 158         if (h->symtab)
 159                 chars_free(h->symtab);
 160
 161         free(h);
 162 }
 163
 164
 165 void
 166 print_gen_head(struct html *h)
 167 {
 168         struct htmlpair  tag[4];
 169
 170         tag[0].key = ATTR_HTTPEQUIV;
 171         tag[0].val = "Content-Type";
 172         tag[1].key = ATTR_CONTENT;
 173         tag[1].val = "text/html; charset=utf-8";
 174         print_otag(h, TAG_META, 2, tag);
 175
 176         tag[0].key = ATTR_NAME;
 177         tag[0].val = "resource-type";
 178         tag[1].key = ATTR_CONTENT;
 179         tag[1].val = "document";
 180         print_otag(h, TAG_META, 2, tag);
 181
 182         if (h->style) {
 183                 tag[0].key = ATTR_REL;
 184                 tag[0].val = "stylesheet";
 185                 tag[1].key = ATTR_HREF;
 186                 tag[1].val = h->style;
 187                 tag[2].key = ATTR_TYPE;
 188                 tag[2].val = "text/css";
 189                 tag[3].key = ATTR_MEDIA;
 190                 tag[3].val = "all";
 191                 print_otag(h, TAG_LINK, 4, tag);
 192         }
 193 }
 194
 195
 196 static void
 197 print_spec(struct html *h, const char *p, size_t len)
 198 {
 199         const char      *rhs;
 200         size_t           sz;
 201
 202         rhs = chars_a2ascii(h->symtab, p, len, &sz);
 203
 204         if (NULL == rhs)
 205                 return;
 206         fwrite(rhs, 1, sz, stdout);
 207 }
 208
 209
 210 static void
 211 print_res(struct html *h, const char *p, size_t len)
 212 {
 213         const char      *rhs;
 214         size_t           sz;
 215
 216         rhs = chars_a2res(h->symtab, p, len, &sz);
 217
 218         if (NULL == rhs)
 219                 return;
 220         fwrite(rhs, 1, sz, stdout);
 221 }
 222
 223
 224 static void
 225 print_metaf(struct html *h, enum roffdeco deco)
 226 {
 227         const char      *class;
 228         struct htmlpair  tag;
 229
 230         switch (deco) {
 231         case (DECO_BOLD):
 232                 class = "bold";
 233                 break;
 234         case (DECO_ITALIC):
 235                 class = "italic";
 236                 break;
 237         case (DECO_ROMAN):
 238                 class = "roman";
 239                 break;
 240         default:
 241                 abort();
 242                 /* NOTREACHED */
 243         }
 244
 245         if (h->metaf) {
 246                 assert(h->tags.head);
 247                 assert(h->metaf == h->tags.head);
 248                 print_tagq(h, h->metaf);
 249         }
 250
 251         PAIR_CLASS_INIT(&tag, class);
 252         h->metaf = print_otag(h, TAG_SPAN, 1, &tag);
 253 }
 254
 255
 256 static int
 257 print_encode(struct html *h, const char *p, int norecurse)
 258 {
 259         size_t           sz;
 260         int              len, nospace;
 261         const char      *seq;
 262         enum roffdeco    deco;
 263
 264         nospace = 0;
 265
 266         for (; *p; p++) {
 267                 sz = strcspn(p, "\\<>&");
 268
 269                 fwrite(p, 1, sz, stdout);
 270                 p += /* LINTED */
 271                         sz;
 272
 273                 if ('<' == *p) {
 274                         printf("&lt;");
 275                         continue;
 276                 } else if ('>' == *p) {
 277                         printf("&gt;");
 278                         continue;
 279                 } else if ('&' == *p) {
 280                         printf("&amp;");
 281                         continue;
 282                 } else if ('\0' == *p)
 283                         break;
 284
 285                 seq = ++p;
 286                 len = a2roffdeco(&deco, &seq, &sz);
 287
 288                 switch (deco) {
 289                 case (DECO_RESERVED):
 290                         print_res(h, seq, sz);
 291                         break;
 292                 case (DECO_SPECIAL):
 293                         print_spec(h, seq, sz);
 294                         break;
 295                 case (DECO_BOLD):
 296                         /* FALLTHROUGH */
 297                 case (DECO_ITALIC):
 298                         /* FALLTHROUGH */
 299                 case (DECO_ROMAN):
 300                         if (norecurse)
 301                                 break;
 302                         print_metaf(h, deco);
 303                         break;
 304                 default:
 305                         break;
 306                 }
 307
 308                 p += len - 1;
 309
 310                 if (DECO_NOSPACE == deco && '\0' == *(p + 1))
 311                         nospace = 1;
 312         }
 313
 314         return(nospace);
 315 }
 316
 317
 318 struct tag *
 319 print_otag(struct html *h, enum htmltag tag,
 320                 int sz, const struct htmlpair *p)
 321 {
 322         int              i;
 323         struct tag      *t;
 324
 325         if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
 326                 t = malloc(sizeof(struct tag));
 327                 if (NULL == t) {
 328                         perror(NULL);
 329                         exit(EXIT_FAILURE);
 330                 }
 331                 t->tag = tag;
 332                 t->next = h->tags.head;
 333                 h->tags.head = t;
 334         } else
 335                 t = NULL;
 336
 337         if ( ! (HTML_NOSPACE & h->flags))
 338                 if ( ! (HTML_CLRLINE & htmltags[tag].flags))
 339                         putchar(' ');
 340
 341         printf("<%s", htmltags[tag].name);
 342         for (i = 0; i < sz; i++) {
 343                 printf(" %s=\"", htmlattrs[p[i].key]);
 344                 assert(p->val);
 345                 (void)print_encode(h, p[i].val, 1);
 346                 putchar('\"');
 347         }
 348         putchar('>');
 349
 350         h->flags |= HTML_NOSPACE;
 351         return(t);
 352 }
 353
 354
 355 /* ARGSUSED */
 356 static void
 357 print_ctag(struct html *h, enum htmltag tag)
 358 {
 359
 360         printf("</%s>", htmltags[tag].name);
 361         if (HTML_CLRLINE & htmltags[tag].flags) {
 362                 h->flags |= HTML_NOSPACE;
 363                 putchar('\n');
 364         }
 365 }
 366
 367
 368 /* ARGSUSED */
 369 void
 370 print_gen_doctype(struct html *h)
 371 {
 372
 373         printf("<!DOCTYPE HTML PUBLIC \"%s\" \"%s\">", DOCTYPE, DTD);
 374 }
 375
 376
 377 void
 378 print_text(struct html *h, const char *p)
 379 {
 380
 381         if (*p && 0 == *(p + 1))
 382                 switch (*p) {
 383                 case('.'):
 384                         /* FALLTHROUGH */
 385                 case(','):
 386                         /* FALLTHROUGH */
 387                 case(';'):
 388                         /* FALLTHROUGH */
 389                 case(':'):
 390                         /* FALLTHROUGH */
 391                 case('?'):
 392                         /* FALLTHROUGH */
 393                 case('!'):
 394                         /* FALLTHROUGH */
 395                 case(')'):
 396                         /* FALLTHROUGH */
 397                 case(']'):
 398                         /* FALLTHROUGH */
 399                 case('}'):
 400                         if ( ! (HTML_IGNDELIM & h->flags))
 401                                 h->flags |= HTML_NOSPACE;
 402                         break;
 403                 default:
 404                         break;
 405                 }
 406
 407         if ( ! (h->flags & HTML_NOSPACE))
 408                 putchar(' ');
 409
 410         assert(p);
 411         if ( ! print_encode(h, p, 0))
 412                 h->flags &= ~HTML_NOSPACE;
 413
 414         if (*p && 0 == *(p + 1))
 415                 switch (*p) {
 416                 case('('):
 417                         /* FALLTHROUGH */
 418                 case('['):
 419                         /* FALLTHROUGH */
 420                 case('{'):
 421                         h->flags |= HTML_NOSPACE;
 422                         break;
 423                 default:
 424                         break;
 425                 }
 426 }
 427
 428
 429 void
 430 print_tagq(struct html *h, const struct tag *until)
 431 {
 432         struct tag      *tag;
 433
 434         while ((tag = h->tags.head) != NULL) {
 435                 print_ctag(h, tag->tag);
 436                 h->tags.head = tag->next;
 437                 free(tag);
 438                 if (until && tag == until)
 439                         return;
 440         }
 441 }
 442
 443
 444 void
 445 print_stagq(struct html *h, const struct tag *suntil)
 446 {
 447         struct tag      *tag;
 448
 449         while ((tag = h->tags.head) != NULL) {
 450                 if (suntil && tag == suntil)
 451                         return;
 452                 print_ctag(h, tag->tag);
 453                 h->tags.head = tag->next;
 454                 free(tag);
 455         }
 456 }
 457
 458
 459 void
 460 bufinit(struct html *h)
 461 {
 462
 463         h->buf[0] = '\0';
 464         h->buflen = 0;
 465 }
 466
 467
 468 void
 469 bufcat_style(struct html *h, const char *key, const char *val)
 470 {
 471
 472         bufcat(h, key);
 473         bufncat(h, ":", 1);
 474         bufcat(h, val);
 475         bufncat(h, ";", 1);
 476 }
 477
 478
 479 void
 480 bufcat(struct html *h, const char *p)
 481 {
 482
 483         bufncat(h, p, strlen(p));
 484 }
 485
 486
 487 void
 488 buffmt(struct html *h, const char *fmt, ...)
 489 {
 490         va_list          ap;
 491
 492         va_start(ap, fmt);
 493         (void)vsnprintf(h->buf + (int)h->buflen,
 494                         BUFSIZ - h->buflen - 1, fmt, ap);
 495         va_end(ap);
 496         h->buflen = strlen(h->buf);
 497 }
 498
 499
 500 void
 501 bufncat(struct html *h, const char *p, size_t sz)
 502 {
 503
 504         if (h->buflen + sz > BUFSIZ - 1)
 505                 sz = BUFSIZ - 1 - h->buflen;
 506
 507         (void)strncat(h->buf, p, sz);
 508         h->buflen += sz;
 509 }
 510
 511
 512 void
 513 buffmt_includes(struct html *h, const char *name)
 514 {
 515         const char      *p, *pp;
 516
 517         pp = h->base_includes;
 518
 519         while (NULL != (p = strchr(pp, '%'))) {
 520                 bufncat(h, pp, (size_t)(p - pp));
 521                 switch (*(p + 1)) {
 522                 case('I'):
 523                         bufcat(h, name);
 524                         break;
 525                 default:
 526                         bufncat(h, p, 2);
 527                         break;
 528                 }
 529                 pp = p + 2;
 530         }
 531         if (pp)
 532                 bufcat(h, pp);
 533 }
 534
 535
 536 void
 537 buffmt_man(struct html *h,
 538                 const char *name, const char *sec)
 539 {
 540         const char      *p, *pp;
 541
 542         pp = h->base_man;
 543
 544         /* LINTED */
 545         while (NULL != (p = strchr(pp, '%'))) {
 546                 bufncat(h, pp, (size_t)(p - pp));
 547                 switch (*(p + 1)) {
 548                 case('S'):
 549                         bufcat(h, sec ? sec : "1");
 550                         break;
 551                 case('N'):
 552                         buffmt(h, name);
 553                         break;
 554                 default:
 555                         bufncat(h, p, 2);
 556                         break;
 557                 }
 558                 pp = p + 2;
 559         }
 560         if (pp)
 561                 bufcat(h, pp);
 562 }
 563
 564
 565 void
 566 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
 567 {
 568         double           v;
 569         const char      *u;
 570
 571         v = su->scale;
 572
 573         switch (su->unit) {
 574         case (SCALE_CM):
 575                 u = "cm";
 576                 break;
 577         case (SCALE_IN):
 578                 u = "in";
 579                 break;
 580         case (SCALE_PC):
 581                 u = "pc";
 582                 break;
 583         case (SCALE_PT):
 584                 u = "pt";
 585                 break;
 586         case (SCALE_EM):
 587                 u = "em";
 588                 break;
 589         case (SCALE_MM):
 590                 if (0 == (v /= 100))
 591                         v = 1;
 592                 u = "em";
 593                 break;
 594         case (SCALE_EN):
 595                 u = "ex";
 596                 break;
 597         case (SCALE_BU):
 598                 u = "ex";
 599                 break;
 600         case (SCALE_VS):
 601                 u = "em";
 602                 break;
 603         default:
 604                 u = "ex";
 605                 break;
 606         }
 607
 608         if (su->pt)
 609                 buffmt(h, "%s: %f%s;", p, v, u);
 610         else
 611                 /* LINTED */
 612                 buffmt(h, "%s: %d%s;", p, (int)v, u);
 613 }
 614
 615
 616 void
 617 html_idcat(char *dst, const char *src, int sz)
 618 {
 619         int              ssz;
 620
 621         assert(sz);
 622
 623         /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
 624
 625         for ( ; *dst != '\0' && sz; dst++, sz--)
 626                 /* Jump to end. */ ;
 627
 628         assert(sz > 2);
 629
 630         /* We can't start with a number (bah). */
 631
 632         *dst++ = 'x';
 633         *dst = '\0';
 634         sz--;
 635
 636         for ( ; *src != '\0' && sz > 1; src++) {
 637                 ssz = snprintf(dst, (size_t)sz, "%.2x", *src);
 638                 sz -= ssz;
 639                 dst += ssz;
 640         }
 641 }