html.c

   1 /*      $Id: html.c,v 1.84 2009/11/12 08:21:05 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  void              print_encode(struct html *, const char *);
  95
  96
  97 void *
  98 html_alloc(char *outopts)
  99 {
 100         struct html     *h;
 101         const char      *toks[4];
 102         char            *v;
 103
 104         toks[0] = "style";
 105         toks[1] = "man";
 106         toks[2] = "includes";
 107         toks[3] = NULL;
 108
 109         h = calloc(1, sizeof(struct html));
 110         if (NULL == h) {
 111                 perror(NULL);
 112                 exit(EXIT_FAILURE);
 113         }
 114
 115         h->tags.head = NULL;
 116         h->ords.head = NULL;
 117         h->symtab = chars_init(CHARS_HTML);
 118
 119         while (outopts && *outopts)
 120                 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
 121                 case (0):
 122                         h->style = v;
 123                         break;
 124                 case (1):
 125                         h->base_man = v;
 126                         break;
 127                 case (2):
 128                         h->base_includes = v;
 129                         break;
 130                 default:
 131                         break;
 132                 }
 133
 134         return(h);
 135 }
 136
 137
 138 void
 139 html_free(void *p)
 140 {
 141         struct tag      *tag;
 142         struct ord      *ord;
 143         struct html     *h;
 144
 145         h = (struct html *)p;
 146
 147         while ((ord = h->ords.head) != NULL) {
 148                 h->ords.head = ord->next;
 149                 free(ord);
 150         }
 151
 152         while ((tag = h->tags.head) != NULL) {
 153                 h->tags.head = tag->next;
 154                 free(tag);
 155         }
 156
 157         if (h->symtab)
 158                 chars_free(h->symtab);
 159
 160         free(h);
 161 }
 162
 163
 164 void
 165 print_gen_head(struct html *h)
 166 {
 167         struct htmlpair  tag[4];
 168
 169         tag[0].key = ATTR_HTTPEQUIV;
 170         tag[0].val = "Content-Type";
 171         tag[1].key = ATTR_CONTENT;
 172         tag[1].val = "text/html; charset=utf-8";
 173         print_otag(h, TAG_META, 2, tag);
 174
 175         tag[0].key = ATTR_NAME;
 176         tag[0].val = "resource-type";
 177         tag[1].key = ATTR_CONTENT;
 178         tag[1].val = "document";
 179         print_otag(h, TAG_META, 2, tag);
 180
 181         if (h->style) {
 182                 tag[0].key = ATTR_REL;
 183                 tag[0].val = "stylesheet";
 184                 tag[1].key = ATTR_HREF;
 185                 tag[1].val = h->style;
 186                 tag[2].key = ATTR_TYPE;
 187                 tag[2].val = "text/css";
 188                 tag[3].key = ATTR_MEDIA;
 189                 tag[3].val = "all";
 190                 print_otag(h, TAG_LINK, 4, tag);
 191         }
 192 }
 193
 194
 195 static void
 196 print_spec(struct html *h, const char *p, size_t len)
 197 {
 198         const char      *rhs;
 199         size_t           sz;
 200
 201         rhs = chars_a2ascii(h->symtab, p, len, &sz);
 202
 203         if (NULL == rhs)
 204                 return;
 205         fwrite(rhs, 1, sz, stdout);
 206 }
 207
 208
 209 static void
 210 print_res(struct html *h, const char *p, size_t len)
 211 {
 212         const char      *rhs;
 213         size_t           sz;
 214
 215         rhs = chars_a2res(h->symtab, p, len, &sz);
 216
 217         if (NULL == rhs)
 218                 return;
 219         fwrite(rhs, 1, sz, stdout);
 220 }
 221
 222
 223 static void
 224 print_encode(struct html *h, const char *p)
 225 {
 226         size_t           sz;
 227         int              len;
 228         const char      *seq;
 229         enum roffdeco    deco;
 230
 231         for (; *p; p++) {
 232                 sz = strcspn(p, "\\<>&");
 233
 234                 fwrite(p, 1, sz, stdout);
 235                 p += /* LINTED */
 236                         sz;
 237
 238                 if ('<' == *p) {
 239                         printf("&lt;");
 240                         continue;
 241                 } else if ('>' == *p) {
 242                         printf("&gt;");
 243                         continue;
 244                 } else if ('&' == *p) {
 245                         printf("&amp;");
 246                         continue;
 247                 } else if ('\0' == *p)
 248                         break;
 249
 250                 seq = ++p;
 251                 len = a2roffdeco(&deco, &seq, &sz);
 252
 253                 switch (deco) {
 254                 case (DECO_RESERVED):
 255                         print_res(h, seq, sz);
 256                         break;
 257                 case (DECO_SPECIAL):
 258                         print_spec(h, seq, sz);
 259                         break;
 260                 default:
 261                         break;
 262                 }
 263
 264                 p += len - 1;
 265
 266                 if (DECO_NOSPACE == deco && '\0' == *(p + 1))
 267                         h->flags |= HTML_NOSPACE;
 268         }
 269 }
 270
 271
 272 struct tag *
 273 print_otag(struct html *h, enum htmltag tag,
 274                 int sz, const struct htmlpair *p)
 275 {
 276         int              i;
 277         struct tag      *t;
 278
 279         if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
 280                 t = malloc(sizeof(struct tag));
 281                 if (NULL == t) {
 282                         perror(NULL);
 283                         exit(EXIT_FAILURE);
 284                 }
 285                 t->tag = tag;
 286                 t->next = h->tags.head;
 287                 h->tags.head = t;
 288         } else
 289                 t = NULL;
 290
 291         if ( ! (HTML_NOSPACE & h->flags))
 292                 if ( ! (HTML_CLRLINE & htmltags[tag].flags))
 293                         putchar(' ');
 294
 295         printf("<%s", htmltags[tag].name);
 296         for (i = 0; i < sz; i++) {
 297                 printf(" %s=\"", htmlattrs[p[i].key]);
 298                 assert(p->val);
 299                 print_encode(h, p[i].val);
 300                 putchar('\"');
 301         }
 302         putchar('>');
 303
 304         h->flags |= HTML_NOSPACE;
 305         if (HTML_CLRLINE & htmltags[tag].flags)
 306                 h->flags |= HTML_NEWLINE;
 307         else
 308                 h->flags &= ~HTML_NEWLINE;
 309
 310         return(t);
 311 }
 312
 313
 314 /* ARGSUSED */
 315 static void
 316 print_ctag(struct html *h, enum htmltag tag)
 317 {
 318
 319         printf("</%s>", htmltags[tag].name);
 320         if (HTML_CLRLINE & htmltags[tag].flags) {
 321                 h->flags |= HTML_NOSPACE;
 322                 h->flags |= HTML_NEWLINE;
 323                 putchar('\n');
 324         } else
 325                 h->flags &= ~HTML_NEWLINE;
 326 }
 327
 328
 329 /* ARGSUSED */
 330 void
 331 print_gen_doctype(struct html *h)
 332 {
 333
 334         printf("<!DOCTYPE HTML PUBLIC \"%s\" \"%s\">", DOCTYPE, DTD);
 335 }
 336
 337
 338 void
 339 print_text(struct html *h, const char *p)
 340 {
 341
 342         if (*p && 0 == *(p + 1))
 343                 switch (*p) {
 344                 case('.'):
 345                         /* FALLTHROUGH */
 346                 case(','):
 347                         /* FALLTHROUGH */
 348                 case(';'):
 349                         /* FALLTHROUGH */
 350                 case(':'):
 351                         /* FALLTHROUGH */
 352                 case('?'):
 353                         /* FALLTHROUGH */
 354                 case('!'):
 355                         /* FALLTHROUGH */
 356                 case(')'):
 357                         /* FALLTHROUGH */
 358                 case(']'):
 359                         /* FALLTHROUGH */
 360                 case('}'):
 361                         if ( ! (HTML_IGNDELIM & h->flags))
 362                                 h->flags |= HTML_NOSPACE;
 363                         break;
 364                 default:
 365                         break;
 366                 }
 367
 368         if ( ! (h->flags & HTML_NOSPACE))
 369                 putchar(' ');
 370
 371         h->flags &= ~HTML_NOSPACE;
 372         h->flags &= ~HTML_NEWLINE;
 373
 374         if (p)
 375                 print_encode(h, p);
 376
 377         if (*p && 0 == *(p + 1))
 378                 switch (*p) {
 379                 case('('):
 380                         /* FALLTHROUGH */
 381                 case('['):
 382                         /* FALLTHROUGH */
 383                 case('{'):
 384                         h->flags |= HTML_NOSPACE;
 385                         break;
 386                 default:
 387                         break;
 388                 }
 389 }
 390
 391
 392 void
 393 print_tagq(struct html *h, const struct tag *until)
 394 {
 395         struct tag      *tag;
 396
 397         while ((tag = h->tags.head) != NULL) {
 398                 print_ctag(h, tag->tag);
 399                 h->tags.head = tag->next;
 400                 free(tag);
 401                 if (until && tag == until)
 402                         return;
 403         }
 404 }
 405
 406
 407 void
 408 print_stagq(struct html *h, const struct tag *suntil)
 409 {
 410         struct tag      *tag;
 411
 412         while ((tag = h->tags.head) != NULL) {
 413                 if (suntil && tag == suntil)
 414                         return;
 415                 print_ctag(h, tag->tag);
 416                 h->tags.head = tag->next;
 417                 free(tag);
 418         }
 419 }
 420
 421
 422 void
 423 bufinit(struct html *h)
 424 {
 425
 426         h->buf[0] = '\0';
 427         h->buflen = 0;
 428 }
 429
 430
 431 void
 432 bufcat_style(struct html *h, const char *key, const char *val)
 433 {
 434
 435         bufcat(h, key);
 436         bufncat(h, ":", 1);
 437         bufcat(h, val);
 438         bufncat(h, ";", 1);
 439 }
 440
 441
 442 void
 443 bufcat(struct html *h, const char *p)
 444 {
 445
 446         bufncat(h, p, strlen(p));
 447 }
 448
 449
 450 void
 451 buffmt(struct html *h, const char *fmt, ...)
 452 {
 453         va_list          ap;
 454
 455         va_start(ap, fmt);
 456         (void)vsnprintf(h->buf + (int)h->buflen,
 457                         BUFSIZ - h->buflen - 1, fmt, ap);
 458         va_end(ap);
 459         h->buflen = strlen(h->buf);
 460 }
 461
 462
 463 void
 464 bufncat(struct html *h, const char *p, size_t sz)
 465 {
 466
 467         if (h->buflen + sz > BUFSIZ - 1)
 468                 sz = BUFSIZ - 1 - h->buflen;
 469
 470         (void)strncat(h->buf, p, sz);
 471         h->buflen += sz;
 472 }
 473
 474
 475 void
 476 buffmt_includes(struct html *h, const char *name)
 477 {
 478         const char      *p, *pp;
 479
 480         pp = h->base_includes;
 481
 482         while (NULL != (p = strchr(pp, '%'))) {
 483                 bufncat(h, pp, (size_t)(p - pp));
 484                 switch (*(p + 1)) {
 485                 case('I'):
 486                         bufcat(h, name);
 487                         break;
 488                 default:
 489                         bufncat(h, p, 2);
 490                         break;
 491                 }
 492                 pp = p + 2;
 493         }
 494         if (pp)
 495                 bufcat(h, pp);
 496 }
 497
 498
 499 void
 500 buffmt_man(struct html *h,
 501                 const char *name, const char *sec)
 502 {
 503         const char      *p, *pp;
 504
 505         pp = h->base_man;
 506
 507         /* LINTED */
 508         while (NULL != (p = strchr(pp, '%'))) {
 509                 bufncat(h, pp, (size_t)(p - pp));
 510                 switch (*(p + 1)) {
 511                 case('S'):
 512                         bufcat(h, sec ? sec : "1");
 513                         break;
 514                 case('N'):
 515                         buffmt(h, name);
 516                         break;
 517                 default:
 518                         bufncat(h, p, 2);
 519                         break;
 520                 }
 521                 pp = p + 2;
 522         }
 523         if (pp)
 524                 bufcat(h, pp);
 525 }
 526
 527
 528 void
 529 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
 530 {
 531         double           v;
 532         const char      *u;
 533
 534         v = su->scale;
 535
 536         switch (su->unit) {
 537         case (SCALE_CM):
 538                 u = "cm";
 539                 break;
 540         case (SCALE_IN):
 541                 u = "in";
 542                 break;
 543         case (SCALE_PC):
 544                 u = "pc";
 545                 break;
 546         case (SCALE_PT):
 547                 u = "pt";
 548                 break;
 549         case (SCALE_EM):
 550                 u = "em";
 551                 break;
 552         case (SCALE_MM):
 553                 if (0 == (v /= 100))
 554                         v = 1;
 555                 u = "em";
 556                 break;
 557         case (SCALE_EN):
 558                 u = "ex";
 559                 break;
 560         case (SCALE_BU):
 561                 u = "ex";
 562                 break;
 563         case (SCALE_VS):
 564                 u = "em";
 565                 break;
 566         default:
 567                 u = "ex";
 568                 break;
 569         }
 570
 571         if (su->pt)
 572                 buffmt(h, "%s: %f%s;", p, v, u);
 573         else
 574                 /* LINTED */
 575                 buffmt(h, "%s: %d%s;", p, (int)v, u);
 576 }
 577
 578
 579 void
 580 html_idcat(char *dst, const char *src, int sz)
 581 {
 582         int              ssz;
 583
 584         assert(sz);
 585
 586         /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
 587
 588         for ( ; *dst != '\0' && sz; dst++, sz--)
 589                 /* Jump to end. */ ;
 590
 591         assert(sz > 2);
 592
 593         /* We can't start with a number (bah). */
 594
 595         *dst++ = 'x';
 596         *dst = '\0';
 597         sz--;
 598
 599         for ( ; *src != '\0' && sz > 1; src++) {
 600                 ssz = snprintf(dst, (size_t)sz, "%.2x", *src);
 601                 sz -= ssz;
 602                 dst += ssz;
 603         }
 604 }