html.c

   1 /*      $Id: html.c,v 1.93 2010/01/29 14:39:37 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 #ifdef HAVE_CONFIG_H
  18 #include "config.h"
  19 #endif
  20
  21 #include <sys/types.h>
  22
  23 #include <assert.h>
  24 #include <ctype.h>
  25 #include <stdarg.h>
  26 #include <stdio.h>
  27 #include <stdint.h>
  28 #include <stdlib.h>
  29 #include <string.h>
  30 #include <unistd.h>
  31
  32 #include "out.h"
  33 #include "chars.h"
  34 #include "html.h"
  35 #include "main.h"
  36
  37 #define UNCONST(a)      ((void *)(uintptr_t)(const void *)(a))
  38
  39 struct  htmldata {
  40         const char       *name;
  41         int               flags;
  42 #define HTML_CLRLINE     (1 << 0)
  43 #define HTML_NOSTACK     (1 << 1)
  44 #define HTML_AUTOCLOSE   (1 << 2) /* Tag has auto-closure. */
  45 };
  46
  47 static  const struct htmldata htmltags[TAG_MAX] = {
  48         {"html",        HTML_CLRLINE}, /* TAG_HTML */
  49         {"head",        HTML_CLRLINE}, /* TAG_HEAD */
  50         {"body",        HTML_CLRLINE}, /* TAG_BODY */
  51         {"meta",        HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_META */
  52         {"title",       HTML_CLRLINE}, /* TAG_TITLE */
  53         {"div",         HTML_CLRLINE}, /* TAG_DIV */
  54         {"h1",          0}, /* TAG_H1 */
  55         {"h2",          0}, /* TAG_H2 */
  56         {"span",        0}, /* TAG_SPAN */
  57         {"link",        HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
  58         {"br",          HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_BR */
  59         {"a",           0}, /* TAG_A */
  60         {"table",       HTML_CLRLINE}, /* TAG_TABLE */
  61         {"col",         HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_COL */
  62         {"tr",          HTML_CLRLINE}, /* TAG_TR */
  63         {"td",          HTML_CLRLINE}, /* TAG_TD */
  64         {"li",          HTML_CLRLINE}, /* TAG_LI */
  65         {"ul",          HTML_CLRLINE}, /* TAG_UL */
  66         {"ol",          HTML_CLRLINE}, /* TAG_OL */
  67 };
  68
  69 static  const char      *const htmlfonts[HTMLFONT_MAX] = {
  70         "roman",
  71         "bold",
  72         "italic"
  73 };
  74
  75 static  const char      *const htmlattrs[ATTR_MAX] = {
  76         "http-equiv",
  77         "content",
  78         "name",
  79         "rel",
  80         "href",
  81         "type",
  82         "media",
  83         "class",
  84         "style",
  85         "width",
  86         "valign",
  87         "target",
  88         "id",
  89         "summary",
  90 };
  91
  92 static  void              print_spec(struct html *, const char *, size_t);
  93 static  void              print_res(struct html *, const char *, size_t);
  94 static  void              print_ctag(struct html *, enum htmltag);
  95 static  void              print_doctype(struct html *);
  96 static  void              print_xmltype(struct html *);
  97 static  int               print_encode(struct html *, const char *, int);
  98 static  void              print_metaf(struct html *, enum roffdeco);
  99 static  void             *ml_alloc(char *, enum htmltype);
 100
 101
 102 static void *
 103 ml_alloc(char *outopts, enum htmltype type)
 104 {
 105         struct html     *h;
 106         const char      *toks[4];
 107         char            *v;
 108
 109         toks[0] = "style";
 110         toks[1] = "man";
 111         toks[2] = "includes";
 112         toks[3] = NULL;
 113
 114         h = calloc(1, sizeof(struct html));
 115         if (NULL == h) {
 116                 perror(NULL);
 117                 exit(EXIT_FAILURE);
 118         }
 119
 120         h->type = type;
 121         h->tags.head = NULL;
 122         h->ords.head = NULL;
 123         h->symtab = chars_init(CHARS_HTML);
 124
 125         while (outopts && *outopts)
 126                 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
 127                 case (0):
 128                         h->style = v;
 129                         break;
 130                 case (1):
 131                         h->base_man = v;
 132                         break;
 133                 case (2):
 134                         h->base_includes = v;
 135                         break;
 136                 default:
 137                         break;
 138                 }
 139
 140         return(h);
 141 }
 142
 143 void *
 144 html_alloc(char *outopts)
 145 {
 146
 147         return(ml_alloc(outopts, HTML_HTML_4_01_STRICT));
 148 }
 149
 150
 151 void *
 152 xhtml_alloc(char *outopts)
 153 {
 154
 155         return(ml_alloc(outopts, HTML_XHTML_1_0_STRICT));
 156 }
 157
 158
 159 void
 160 html_free(void *p)
 161 {
 162         struct tag      *tag;
 163         struct ord      *ord;
 164         struct html     *h;
 165
 166         h = (struct html *)p;
 167
 168         while ((ord = h->ords.head) != NULL) {
 169                 h->ords.head = ord->next;
 170                 free(ord);
 171         }
 172
 173         while ((tag = h->tags.head) != NULL) {
 174                 h->tags.head = tag->next;
 175                 free(tag);
 176         }
 177
 178         if (h->symtab)
 179                 chars_free(h->symtab);
 180
 181         free(h);
 182 }
 183
 184
 185 void
 186 print_gen_head(struct html *h)
 187 {
 188         struct htmlpair  tag[4];
 189
 190         tag[0].key = ATTR_HTTPEQUIV;
 191         tag[0].val = "Content-Type";
 192         tag[1].key = ATTR_CONTENT;
 193         tag[1].val = "text/html; charset=utf-8";
 194         print_otag(h, TAG_META, 2, tag);
 195
 196         tag[0].key = ATTR_NAME;
 197         tag[0].val = "resource-type";
 198         tag[1].key = ATTR_CONTENT;
 199         tag[1].val = "document";
 200         print_otag(h, TAG_META, 2, tag);
 201
 202         if (h->style) {
 203                 tag[0].key = ATTR_REL;
 204                 tag[0].val = "stylesheet";
 205                 tag[1].key = ATTR_HREF;
 206                 tag[1].val = h->style;
 207                 tag[2].key = ATTR_TYPE;
 208                 tag[2].val = "text/css";
 209                 tag[3].key = ATTR_MEDIA;
 210                 tag[3].val = "all";
 211                 print_otag(h, TAG_LINK, 4, tag);
 212         }
 213 }
 214
 215
 216 static void
 217 print_spec(struct html *h, const char *p, size_t len)
 218 {
 219         const char      *rhs;
 220         size_t           sz;
 221
 222         rhs = chars_a2ascii(h->symtab, p, len, &sz);
 223
 224         if (NULL == rhs)
 225                 return;
 226         fwrite(rhs, 1, sz, stdout);
 227 }
 228
 229
 230 static void
 231 print_res(struct html *h, const char *p, size_t len)
 232 {
 233         const char      *rhs;
 234         size_t           sz;
 235
 236         rhs = chars_a2res(h->symtab, p, len, &sz);
 237
 238         if (NULL == rhs)
 239                 return;
 240         fwrite(rhs, 1, sz, stdout);
 241 }
 242
 243
 244 struct tag *
 245 print_ofont(struct html *h, enum htmlfont font)
 246 {
 247         struct htmlpair  tag;
 248
 249         h->metal = h->metac;
 250         h->metac = font;
 251
 252         /* FIXME: DECO_ROMAN should just close out preexisting. */
 253
 254         if (h->metaf && h->tags.head == h->metaf)
 255                 print_tagq(h, h->metaf);
 256
 257         PAIR_CLASS_INIT(&tag, htmlfonts[font]);
 258         h->metaf = print_otag(h, TAG_SPAN, 1, &tag);
 259         return(h->metaf);
 260 }
 261
 262
 263 static void
 264 print_metaf(struct html *h, enum roffdeco deco)
 265 {
 266         enum htmlfont    font;
 267
 268         switch (deco) {
 269         case (DECO_PREVIOUS):
 270                 font = h->metal;
 271                 break;
 272         case (DECO_ITALIC):
 273                 font = HTMLFONT_ITALIC;
 274                 break;
 275         case (DECO_BOLD):
 276                 font = HTMLFONT_BOLD;
 277                 break;
 278         case (DECO_ROMAN):
 279                 font = HTMLFONT_NONE;
 280                 break;
 281         default:
 282                 abort();
 283                 /* NOTREACHED */
 284         }
 285
 286         (void)print_ofont(h, font);
 287 }
 288
 289
 290 static int
 291 print_encode(struct html *h, const char *p, int norecurse)
 292 {
 293         size_t           sz;
 294         int              len, nospace;
 295         const char      *seq;
 296         enum roffdeco    deco;
 297
 298         nospace = 0;
 299
 300         for (; *p; p++) {
 301                 sz = strcspn(p, "\\<>&");
 302
 303                 fwrite(p, 1, sz, stdout);
 304                 p += /* LINTED */
 305                         sz;
 306
 307                 if ('<' == *p) {
 308                         printf("&lt;");
 309                         continue;
 310                 } else if ('>' == *p) {
 311                         printf("&gt;");
 312                         continue;
 313                 } else if ('&' == *p) {
 314                         printf("&amp;");
 315                         continue;
 316                 } else if ('\0' == *p)
 317                         break;
 318
 319                 seq = ++p;
 320                 len = a2roffdeco(&deco, &seq, &sz);
 321
 322                 switch (deco) {
 323                 case (DECO_RESERVED):
 324                         print_res(h, seq, sz);
 325                         break;
 326                 case (DECO_SPECIAL):
 327                         print_spec(h, seq, sz);
 328                         break;
 329                 case (DECO_PREVIOUS):
 330                         /* FALLTHROUGH */
 331                 case (DECO_BOLD):
 332                         /* FALLTHROUGH */
 333                 case (DECO_ITALIC):
 334                         /* FALLTHROUGH */
 335                 case (DECO_ROMAN):
 336                         if (norecurse)
 337                                 break;
 338                         print_metaf(h, deco);
 339                         break;
 340                 default:
 341                         break;
 342                 }
 343
 344                 p += len - 1;
 345
 346                 if (DECO_NOSPACE == deco && '\0' == *(p + 1))
 347                         nospace = 1;
 348         }
 349
 350         return(nospace);
 351 }
 352
 353
 354 struct tag *
 355 print_otag(struct html *h, enum htmltag tag,
 356                 int sz, const struct htmlpair *p)
 357 {
 358         int              i;
 359         struct tag      *t;
 360
 361         if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
 362                 t = malloc(sizeof(struct tag));
 363                 if (NULL == t) {
 364                         perror(NULL);
 365                         exit(EXIT_FAILURE);
 366                 }
 367                 t->tag = tag;
 368                 t->next = h->tags.head;
 369                 h->tags.head = t;
 370         } else
 371                 t = NULL;
 372
 373         if ( ! (HTML_NOSPACE & h->flags))
 374                 if ( ! (HTML_CLRLINE & htmltags[tag].flags))
 375                         putchar(' ');
 376
 377         printf("<%s", htmltags[tag].name);
 378         for (i = 0; i < sz; i++) {
 379                 printf(" %s=\"", htmlattrs[p[i].key]);
 380                 assert(p->val);
 381                 (void)print_encode(h, p[i].val, 1);
 382                 putchar('\"');
 383         }
 384
 385         if (HTML_AUTOCLOSE & htmltags[tag].flags)
 386                 switch (h->type) {
 387                 case (HTML_XHTML_1_0_STRICT):
 388                         putchar('/');
 389                         break;
 390                 default:
 391                         break;
 392                 }
 393
 394         putchar('>');
 395
 396         h->flags |= HTML_NOSPACE;
 397         return(t);
 398 }
 399
 400
 401 static void
 402 print_ctag(struct html *h, enum htmltag tag)
 403 {
 404
 405         printf("</%s>", htmltags[tag].name);
 406         if (HTML_CLRLINE & htmltags[tag].flags) {
 407                 h->flags |= HTML_NOSPACE;
 408                 putchar('\n');
 409         }
 410 }
 411
 412
 413 void
 414 print_gen_decls(struct html *h)
 415 {
 416
 417         print_xmltype(h);
 418         print_doctype(h);
 419 }
 420
 421
 422 static void
 423 print_xmltype(struct html *h)
 424 {
 425         const char      *decl;
 426
 427         switch (h->type) {
 428         case (HTML_XHTML_1_0_STRICT):
 429                 decl = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>";
 430                 break;
 431         default:
 432                 decl = NULL;
 433                 break;
 434         }
 435
 436         if (NULL == decl)
 437                 return;
 438
 439         printf("%s\n", decl);
 440 }
 441
 442
 443 static void
 444 print_doctype(struct html *h)
 445 {
 446         const char      *doctype;
 447         const char      *dtd;
 448
 449         switch (h->type) {
 450         case (HTML_HTML_4_01_STRICT):
 451                 doctype = "-//W3C//DTD HTML 4.01//EN";
 452                 dtd = "http://www.w3.org/TR/html4/strict.dtd";
 453                 break;
 454         default:
 455                 doctype = "-//W3C//DTD XHTML 1.0 Strict//EN";
 456                 dtd = "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
 457                 break;
 458         }
 459
 460         printf("<!DOCTYPE HTML PUBLIC \"%s\" \"%s\">\n", doctype, dtd);
 461 }
 462
 463
 464 void
 465 print_text(struct html *h, const char *p)
 466 {
 467
 468         if (*p && 0 == *(p + 1))
 469                 switch (*p) {
 470                 case('.'):
 471                         /* FALLTHROUGH */
 472                 case(','):
 473                         /* FALLTHROUGH */
 474                 case(';'):
 475                         /* FALLTHROUGH */
 476                 case(':'):
 477                         /* FALLTHROUGH */
 478                 case('?'):
 479                         /* FALLTHROUGH */
 480                 case('!'):
 481                         /* FALLTHROUGH */
 482                 case(')'):
 483                         /* FALLTHROUGH */
 484                 case(']'):
 485                         /* FALLTHROUGH */
 486                 case('}'):
 487                         if ( ! (HTML_IGNDELIM & h->flags))
 488                                 h->flags |= HTML_NOSPACE;
 489                         break;
 490                 default:
 491                         break;
 492                 }
 493
 494         if ( ! (h->flags & HTML_NOSPACE))
 495                 putchar(' ');
 496
 497         assert(p);
 498         if ( ! print_encode(h, p, 0))
 499                 h->flags &= ~HTML_NOSPACE;
 500
 501         if (*p && 0 == *(p + 1))
 502                 switch (*p) {
 503                 case('('):
 504                         /* FALLTHROUGH */
 505                 case('['):
 506                         /* FALLTHROUGH */
 507                 case('{'):
 508                         h->flags |= HTML_NOSPACE;
 509                         break;
 510                 default:
 511                         break;
 512                 }
 513 }
 514
 515
 516 void
 517 print_tagq(struct html *h, const struct tag *until)
 518 {
 519         struct tag      *tag;
 520
 521         while ((tag = h->tags.head) != NULL) {
 522                 if (tag == h->metaf)
 523                         h->metaf = NULL;
 524                 print_ctag(h, tag->tag);
 525                 h->tags.head = tag->next;
 526                 free(tag);
 527                 if (until && tag == until)
 528                         return;
 529         }
 530 }
 531
 532
 533 void
 534 print_stagq(struct html *h, const struct tag *suntil)
 535 {
 536         struct tag      *tag;
 537
 538         while ((tag = h->tags.head) != NULL) {
 539                 if (suntil && tag == suntil)
 540                         return;
 541                 if (tag == h->metaf)
 542                         h->metaf = NULL;
 543                 print_ctag(h, tag->tag);
 544                 h->tags.head = tag->next;
 545                 free(tag);
 546         }
 547 }
 548
 549
 550 void
 551 bufinit(struct html *h)
 552 {
 553
 554         h->buf[0] = '\0';
 555         h->buflen = 0;
 556 }
 557
 558
 559 void
 560 bufcat_style(struct html *h, const char *key, const char *val)
 561 {
 562
 563         bufcat(h, key);
 564         bufncat(h, ":", 1);
 565         bufcat(h, val);
 566         bufncat(h, ";", 1);
 567 }
 568
 569
 570 void
 571 bufcat(struct html *h, const char *p)
 572 {
 573
 574         bufncat(h, p, strlen(p));
 575 }
 576
 577
 578 void
 579 buffmt(struct html *h, const char *fmt, ...)
 580 {
 581         va_list          ap;
 582
 583         va_start(ap, fmt);
 584         (void)vsnprintf(h->buf + (int)h->buflen,
 585                         BUFSIZ - h->buflen - 1, fmt, ap);
 586         va_end(ap);
 587         h->buflen = strlen(h->buf);
 588 }
 589
 590
 591 void
 592 bufncat(struct html *h, const char *p, size_t sz)
 593 {
 594
 595         if (h->buflen + sz > BUFSIZ - 1)
 596                 sz = BUFSIZ - 1 - h->buflen;
 597
 598         (void)strncat(h->buf, p, sz);
 599         h->buflen += sz;
 600 }
 601
 602
 603 void
 604 buffmt_includes(struct html *h, const char *name)
 605 {
 606         const char      *p, *pp;
 607
 608         pp = h->base_includes;
 609
 610         while (NULL != (p = strchr(pp, '%'))) {
 611                 bufncat(h, pp, (size_t)(p - pp));
 612                 switch (*(p + 1)) {
 613                 case('I'):
 614                         bufcat(h, name);
 615                         break;
 616                 default:
 617                         bufncat(h, p, 2);
 618                         break;
 619                 }
 620                 pp = p + 2;
 621         }
 622         if (pp)
 623                 bufcat(h, pp);
 624 }
 625
 626
 627 void
 628 buffmt_man(struct html *h,
 629                 const char *name, const char *sec)
 630 {
 631         const char      *p, *pp;
 632
 633         pp = h->base_man;
 634
 635         /* LINTED */
 636         while (NULL != (p = strchr(pp, '%'))) {
 637                 bufncat(h, pp, (size_t)(p - pp));
 638                 switch (*(p + 1)) {
 639                 case('S'):
 640                         bufcat(h, sec ? sec : "1");
 641                         break;
 642                 case('N'):
 643                         buffmt(h, name);
 644                         break;
 645                 default:
 646                         bufncat(h, p, 2);
 647                         break;
 648                 }
 649                 pp = p + 2;
 650         }
 651         if (pp)
 652                 bufcat(h, pp);
 653 }
 654
 655
 656 void
 657 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
 658 {
 659         double           v;
 660         const char      *u;
 661
 662         v = su->scale;
 663
 664         switch (su->unit) {
 665         case (SCALE_CM):
 666                 u = "cm";
 667                 break;
 668         case (SCALE_IN):
 669                 u = "in";
 670                 break;
 671         case (SCALE_PC):
 672                 u = "pc";
 673                 break;
 674         case (SCALE_PT):
 675                 u = "pt";
 676                 break;
 677         case (SCALE_EM):
 678                 u = "em";
 679                 break;
 680         case (SCALE_MM):
 681                 if (0 == (v /= 100))
 682                         v = 1;
 683                 u = "em";
 684                 break;
 685         case (SCALE_EN):
 686                 u = "ex";
 687                 break;
 688         case (SCALE_BU):
 689                 u = "ex";
 690                 break;
 691         case (SCALE_VS):
 692                 u = "em";
 693                 break;
 694         default:
 695                 u = "ex";
 696                 break;
 697         }
 698
 699         if (su->pt)
 700                 buffmt(h, "%s: %f%s;", p, v, u);
 701         else
 702                 /* LINTED */
 703                 buffmt(h, "%s: %d%s;", p, (int)v, u);
 704 }
 705
 706
 707 void
 708 html_idcat(char *dst, const char *src, int sz)
 709 {
 710         int              ssz;
 711
 712         assert(sz);
 713
 714         /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
 715
 716         for ( ; *dst != '\0' && sz; dst++, sz--)
 717                 /* Jump to end. */ ;
 718
 719         assert(sz > 2);
 720
 721         /* We can't start with a number (bah). */
 722
 723         *dst++ = 'x';
 724         *dst = '\0';
 725         sz--;
 726
 727         for ( ; *src != '\0' && sz > 1; src++) {
 728                 ssz = snprintf(dst, (size_t)sz, "%.2x", *src);
 729                 sz -= ssz;
 730                 dst += ssz;
 731         }
 732 }