html.c

   1 /*      $Id: html.c,v 1.64 2009/10/13 10:57:25 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 #include <sys/queue.h>
  19
  20 #include <assert.h>
  21 #include <err.h>
  22 #include <stdio.h>
  23 #include <stdarg.h>
  24 #include <stdint.h>
  25 #include <stdlib.h>
  26 #include <string.h>
  27 #include <unistd.h>
  28
  29 #include "out.h"
  30 #include "chars.h"
  31 #include "html.h"
  32 #include "main.h"
  33
  34 #define UNCONST(a)      ((void *)(uintptr_t)(const void *)(a))
  35
  36 #define DOCTYPE         "-//W3C//DTD HTML 4.01//EN"
  37 #define DTD             "http://www.w3.org/TR/html4/strict.dtd"
  38
  39 struct  htmldata {
  40         const char       *name;
  41         int               flags;
  42 #define HTML_CLRLINE     (1 << 0)
  43 #define HTML_NOSTACK     (1 << 1)
  44 };
  45
  46 static  const struct htmldata htmltags[TAG_MAX] = {
  47         {"html",        HTML_CLRLINE}, /* TAG_HTML */
  48         {"head",        HTML_CLRLINE}, /* TAG_HEAD */
  49         {"body",        HTML_CLRLINE}, /* TAG_BODY */
  50         {"meta",        HTML_CLRLINE | HTML_NOSTACK}, /* TAG_META */
  51         {"title",       HTML_CLRLINE}, /* TAG_TITLE */
  52         {"div",         HTML_CLRLINE}, /* TAG_DIV */
  53         {"h1",          0}, /* TAG_H1 */
  54         {"h2",          0}, /* TAG_H2 */
  55         {"p",           HTML_CLRLINE}, /* TAG_P */
  56         {"span",        0}, /* TAG_SPAN */
  57         {"link",        HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
  58         {"br",          HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
  59         {"a",           0}, /* TAG_A */
  60         {"table",       HTML_CLRLINE}, /* TAG_TABLE */
  61         {"col",         HTML_CLRLINE | HTML_NOSTACK}, /* 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         {"base",        HTML_CLRLINE | HTML_NOSTACK}, /* TAG_BASE */
  68 };
  69
  70 static  const char       *const htmlattrs[ATTR_MAX] = {
  71         "http-equiv",
  72         "content",
  73         "name",
  74         "rel",
  75         "href",
  76         "type",
  77         "media",
  78         "class",
  79         "style",
  80         "width",
  81         "valign",
  82         "target",
  83         "id",
  84 };
  85
  86 #ifdef __linux__
  87 extern  int               getsubopt(char **, char * const *, char **);
  88 #endif
  89
  90 void *
  91 html_alloc(char *outopts)
  92 {
  93         struct html     *h;
  94         const char      *toks[4];
  95         char            *v;
  96
  97         toks[0] = "style";
  98         toks[1] = "man";
  99         toks[2] = "includes";
 100         toks[3] = NULL;
 101
 102         if (NULL == (h = calloc(1, sizeof(struct html))))
 103                 return(NULL);
 104
 105         SLIST_INIT(&h->tags);
 106         SLIST_INIT(&h->ords);
 107
 108         if (NULL == (h->symtab = chars_init(CHARS_HTML))) {
 109                 free(h);
 110                 return(NULL);
 111         }
 112
 113         while (outopts && *outopts)
 114                 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
 115                 case (0):
 116                         h->style = v;
 117                         break;
 118                 case (1):
 119                         h->base_man = v;
 120                         break;
 121                 case (2):
 122                         h->base_includes = v;
 123                         break;
 124                 default:
 125                         break;
 126                 }
 127
 128         return(h);
 129 }
 130
 131
 132 void
 133 html_free(void *p)
 134 {
 135         struct tag      *tag;
 136         struct ord      *ord;
 137         struct html     *h;
 138
 139         h = (struct html *)p;
 140
 141         while ( ! SLIST_EMPTY(&h->ords)) {
 142                 ord = SLIST_FIRST(&h->ords);
 143                 SLIST_REMOVE_HEAD(&h->ords, entry);
 144                 free(ord);
 145         }
 146
 147         while ( ! SLIST_EMPTY(&h->tags)) {
 148                 tag = SLIST_FIRST(&h->tags);
 149                 SLIST_REMOVE_HEAD(&h->tags, entry);
 150                 free(tag);
 151         }
 152
 153         if (h->symtab)
 154                 chars_free(h->symtab);
 155
 156         free(h);
 157 }
 158
 159
 160 void
 161 print_gen_head(struct html *h)
 162 {
 163         struct htmlpair  tag[4];
 164
 165         tag[0].key = ATTR_HTTPEQUIV;
 166         tag[0].val = "Content-Type";
 167         tag[1].key = ATTR_CONTENT;
 168         tag[1].val = "text/html; charset=utf-8";
 169         print_otag(h, TAG_META, 2, tag);
 170
 171         tag[0].key = ATTR_NAME;
 172         tag[0].val = "resource-type";
 173         tag[1].key = ATTR_CONTENT;
 174         tag[1].val = "document";
 175         print_otag(h, TAG_META, 2, tag);
 176
 177         if (h->style) {
 178                 tag[0].key = ATTR_REL;
 179                 tag[0].val = "stylesheet";
 180                 tag[1].key = ATTR_HREF;
 181                 tag[1].val = h->style;
 182                 tag[2].key = ATTR_TYPE;
 183                 tag[2].val = "text/css";
 184                 tag[3].key = ATTR_MEDIA;
 185                 tag[3].val = "all";
 186                 print_otag(h, TAG_LINK, 4, tag);
 187         }
 188 }
 189
 190
 191 static void
 192 print_spec(struct html *h, const char *p, int len)
 193 {
 194         const char      *rhs;
 195         int              i;
 196         size_t           sz;
 197
 198         rhs = chars_a2ascii(h->symtab, p, (size_t)len, &sz);
 199
 200         if (NULL == rhs)
 201                 return;
 202         for (i = 0; i < (int)sz; i++)
 203                 putchar(rhs[i]);
 204 }
 205
 206
 207 static void
 208 print_res(struct html *h, const char *p, int len)
 209 {
 210         const char      *rhs;
 211         int              i;
 212         size_t           sz;
 213
 214         rhs = chars_a2res(h->symtab, p, (size_t)len, &sz);
 215
 216         if (NULL == rhs)
 217                 return;
 218         for (i = 0; i < (int)sz; i++)
 219                 putchar(rhs[i]);
 220 }
 221
 222
 223 static void
 224 print_escape(struct html *h, const char **p)
 225 {
 226         int              j, type;
 227         const char      *wp;
 228
 229         wp = *p;
 230         type = 1;
 231
 232         if (0 == *(++wp)) {
 233                 *p = wp;
 234                 return;
 235         }
 236
 237         if ('(' == *wp) {
 238                 wp++;
 239                 if (0 == *wp || 0 == *(wp + 1)) {
 240                         *p = 0 == *wp ? wp : wp + 1;
 241                         return;
 242                 }
 243
 244                 print_spec(h, wp, 2);
 245                 *p = ++wp;
 246                 return;
 247
 248         } else if ('*' == *wp) {
 249                 if (0 == *(++wp)) {
 250                         *p = wp;
 251                         return;
 252                 }
 253
 254                 switch (*wp) {
 255                 case ('('):
 256                         wp++;
 257                         if (0 == *wp || 0 == *(wp + 1)) {
 258                                 *p = 0 == *wp ? wp : wp + 1;
 259                                 return;
 260                         }
 261
 262                         print_res(h, wp, 2);
 263                         *p = ++wp;
 264                         return;
 265                 case ('['):
 266                         type = 0;
 267                         break;
 268                 default:
 269                         print_res(h, wp, 1);
 270                         *p = wp;
 271                         return;
 272                 }
 273
 274         } else if ('f' == *wp) {
 275                 if (0 == *(++wp)) {
 276                         *p = wp;
 277                         return;
 278                 }
 279
 280                 switch (*wp) {
 281                 case ('B'):
 282                         /* TODO */
 283                         break;
 284                 case ('I'):
 285                         /* TODO */
 286                         break;
 287                 case ('P'):
 288                         /* FALLTHROUGH */
 289                 case ('R'):
 290                         /* TODO */
 291                         break;
 292                 default:
 293                         break;
 294                 }
 295
 296                 *p = wp;
 297                 return;
 298
 299         } else if ('[' != *wp) {
 300                 print_spec(h, wp, 1);
 301                 *p = wp;
 302                 return;
 303         }
 304
 305         wp++;
 306         for (j = 0; *wp && ']' != *wp; wp++, j++)
 307                 /* Loop... */ ;
 308
 309         if (0 == *wp) {
 310                 *p = wp;
 311                 return;
 312         }
 313
 314         if (type)
 315                 print_spec(h, wp - j, j);
 316         else
 317                 print_res(h, wp - j, j);
 318
 319         *p = wp;
 320 }
 321
 322
 323 static void
 324 print_encode(struct html *h, const char *p)
 325 {
 326
 327         for (; *p; p++) {
 328                 if ('\\' == *p) {
 329                         print_escape(h, &p);
 330                         continue;
 331                 }
 332                 switch (*p) {
 333                 case ('<'):
 334                         printf("&lt;");
 335                         break;
 336                 case ('>'):
 337                         printf("&gt;");
 338                         break;
 339                 case ('&'):
 340                         printf("&amp;");
 341                         break;
 342                 default:
 343                         putchar(*p);
 344                         break;
 345                 }
 346         }
 347 }
 348
 349
 350 struct tag *
 351 print_otag(struct html *h, enum htmltag tag,
 352                 int sz, const struct htmlpair *p)
 353 {
 354         int              i;
 355         struct tag      *t;
 356
 357         if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
 358                 if (NULL == (t = malloc(sizeof(struct tag))))
 359                         err(EXIT_FAILURE, "malloc");
 360                 t->tag = tag;
 361                 SLIST_INSERT_HEAD(&h->tags, t, entry);
 362         } else
 363                 t = NULL;
 364
 365         if ( ! (HTML_NOSPACE & h->flags))
 366                 if ( ! (HTML_CLRLINE & htmltags[tag].flags))
 367                         printf(" ");
 368
 369         printf("<%s", htmltags[tag].name);
 370         for (i = 0; i < sz; i++) {
 371                 printf(" %s=\"", htmlattrs[p[i].key]);
 372                 assert(p->val);
 373                 print_encode(h, p[i].val);
 374                 printf("\"");
 375         }
 376         printf(">");
 377
 378         h->flags |= HTML_NOSPACE;
 379         if (HTML_CLRLINE & htmltags[tag].flags)
 380                 h->flags |= HTML_NEWLINE;
 381         else
 382                 h->flags &= ~HTML_NEWLINE;
 383
 384         return(t);
 385 }
 386
 387
 388 /* ARGSUSED */
 389 static void
 390 print_ctag(struct html *h, enum htmltag tag)
 391 {
 392
 393         printf("</%s>", htmltags[tag].name);
 394         if (HTML_CLRLINE & htmltags[tag].flags)
 395                 h->flags |= HTML_NOSPACE;
 396         if (HTML_CLRLINE & htmltags[tag].flags)
 397                 h->flags |= HTML_NEWLINE;
 398         else
 399                 h->flags &= ~HTML_NEWLINE;
 400 }
 401
 402
 403 /* ARGSUSED */
 404 void
 405 print_gen_doctype(struct html *h)
 406 {
 407
 408         printf("<!DOCTYPE HTML PUBLIC \"%s\" \"%s\">", DOCTYPE, DTD);
 409 }
 410
 411
 412 void
 413 print_text(struct html *h, const char *p)
 414 {
 415
 416         if (*p && 0 == *(p + 1))
 417                 switch (*p) {
 418                 case('.'):
 419                         /* FALLTHROUGH */
 420                 case(','):
 421                         /* FALLTHROUGH */
 422                 case(';'):
 423                         /* FALLTHROUGH */
 424                 case(':'):
 425                         /* FALLTHROUGH */
 426                 case('?'):
 427                         /* FALLTHROUGH */
 428                 case('!'):
 429                         /* FALLTHROUGH */
 430                 case(')'):
 431                         /* FALLTHROUGH */
 432                 case(']'):
 433                         /* FALLTHROUGH */
 434                 case('}'):
 435                         if ( ! (HTML_IGNDELIM & h->flags))
 436                                 h->flags |= HTML_NOSPACE;
 437                         break;
 438                 default:
 439                         break;
 440                 }
 441
 442         if ( ! (h->flags & HTML_NOSPACE))
 443                 printf(" ");
 444
 445         h->flags &= ~HTML_NOSPACE;
 446         h->flags &= ~HTML_NEWLINE;
 447
 448         if (p)
 449                 print_encode(h, p);
 450
 451         if (*p && 0 == *(p + 1))
 452                 switch (*p) {
 453                 case('('):
 454                         /* FALLTHROUGH */
 455                 case('['):
 456                         /* FALLTHROUGH */
 457                 case('{'):
 458                         h->flags |= HTML_NOSPACE;
 459                         break;
 460                 default:
 461                         break;
 462                 }
 463 }
 464
 465
 466 void
 467 print_tagq(struct html *h, const struct tag *until)
 468 {
 469         struct tag      *tag;
 470
 471         while ( ! SLIST_EMPTY(&h->tags)) {
 472                 tag = SLIST_FIRST(&h->tags);
 473                 print_ctag(h, tag->tag);
 474                 SLIST_REMOVE_HEAD(&h->tags, entry);
 475                 free(tag);
 476                 if (until && tag == until)
 477                         return;
 478         }
 479 }
 480
 481
 482 void
 483 print_stagq(struct html *h, const struct tag *suntil)
 484 {
 485         struct tag      *tag;
 486
 487         while ( ! SLIST_EMPTY(&h->tags)) {
 488                 tag = SLIST_FIRST(&h->tags);
 489                 if (suntil && tag == suntil)
 490                         return;
 491                 print_ctag(h, tag->tag);
 492                 SLIST_REMOVE_HEAD(&h->tags, entry);
 493                 free(tag);
 494         }
 495 }
 496
 497
 498 void
 499 bufinit(struct html *h)
 500 {
 501
 502         h->buf[0] = '\0';
 503         h->buflen = 0;
 504 }
 505
 506
 507 void
 508 bufcat_style(struct html *h, const char *key, const char *val)
 509 {
 510
 511         bufcat(h, key);
 512         bufncat(h, ":", 1);
 513         bufcat(h, val);
 514         bufncat(h, ";", 1);
 515 }
 516
 517
 518 void
 519 bufcat(struct html *h, const char *p)
 520 {
 521
 522         bufncat(h, p, strlen(p));
 523 }
 524
 525
 526 void
 527 buffmt(struct html *h, const char *fmt, ...)
 528 {
 529         va_list          ap;
 530
 531         va_start(ap, fmt);
 532         (void)vsnprintf(h->buf + (int)h->buflen,
 533                         BUFSIZ - h->buflen - 1, fmt, ap);
 534         va_end(ap);
 535         h->buflen = strlen(h->buf);
 536 }
 537
 538
 539 void
 540 bufncat(struct html *h, const char *p, size_t sz)
 541 {
 542
 543         if (h->buflen + sz > BUFSIZ - 1)
 544                 sz = BUFSIZ - 1 - h->buflen;
 545
 546         (void)strncat(h->buf, p, sz);
 547         h->buflen += sz;
 548 }
 549
 550
 551 void
 552 buffmt_includes(struct html *h, const char *name)
 553 {
 554         const char      *p, *pp;
 555
 556         pp = h->base_includes;
 557
 558         while (NULL != (p = strchr(pp, '%'))) {
 559                 bufncat(h, pp, (size_t)(p - pp));
 560                 switch (*(p + 1)) {
 561                 case('I'):
 562                         bufcat(h, name);
 563                         break;
 564                 default:
 565                         bufncat(h, p, 2);
 566                         break;
 567                 }
 568                 pp = p + 2;
 569         }
 570         if (pp)
 571                 bufcat(h, pp);
 572 }
 573
 574
 575 void
 576 buffmt_man(struct html *h,
 577                 const char *name, const char *sec)
 578 {
 579         const char      *p, *pp;
 580
 581         pp = h->base_man;
 582
 583         /* LINTED */
 584         while (NULL != (p = strchr(pp, '%'))) {
 585                 bufncat(h, pp, (size_t)(p - pp));
 586                 switch (*(p + 1)) {
 587                 case('S'):
 588                         bufcat(h, sec ? sec : "1");
 589                         break;
 590                 case('N'):
 591                         buffmt(h, name);
 592                         break;
 593                 default:
 594                         bufncat(h, p, 2);
 595                         break;
 596                 }
 597                 pp = p + 2;
 598         }
 599         if (pp)
 600                 bufcat(h, pp);
 601 }
 602
 603
 604 void
 605 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
 606 {
 607         double           v;
 608         const char      *u;
 609
 610         v = su->scale;
 611
 612         switch (su->unit) {
 613         case (SCALE_CM):
 614                 u = "cm";
 615                 break;
 616         case (SCALE_IN):
 617                 u = "in";
 618                 break;
 619         case (SCALE_PC):
 620                 u = "pc";
 621                 break;
 622         case (SCALE_PT):
 623                 u = "pt";
 624                 break;
 625         case (SCALE_EM):
 626                 u = "em";
 627                 break;
 628         case (SCALE_MM):
 629                 if (0 == (v /= 100))
 630                         v = 1;
 631                 u = "em";
 632                 break;
 633         case (SCALE_EN):
 634                 u = "ex";
 635                 break;
 636         case (SCALE_BU):
 637                 u = "ex";
 638                 break;
 639         case (SCALE_VS):
 640                 u = "em";
 641                 break;
 642         default:
 643                 u = "ex";
 644                 break;
 645         }
 646
 647         if (su->pt)
 648                 buffmt(h, "%s: %f%s;", p, v, u);
 649         else
 650                 /* LINTED */
 651                 buffmt(h, "%s: %d%s;", p, (int)v, u);
 652 }