html.c

   1 /*      $Id: html.c,v 1.62 2009/10/09 06:54:11 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 <stdlib.h>
  25 #include <string.h>
  26 #include <unistd.h>
  27
  28 #include "out.h"
  29 #include "chars.h"
  30 #include "html.h"
  31
  32 #define DOCTYPE         "-//W3C//DTD HTML 4.01//EN"
  33 #define DTD             "http://www.w3.org/TR/html4/strict.dtd"
  34
  35 struct  htmldata {
  36         char             *name;
  37         int               flags;
  38 #define HTML_CLRLINE     (1 << 0)
  39 #define HTML_NOSTACK     (1 << 1)
  40 };
  41
  42 static  const struct htmldata htmltags[TAG_MAX] = {
  43         {"html",        HTML_CLRLINE}, /* TAG_HTML */
  44         {"head",        HTML_CLRLINE}, /* TAG_HEAD */
  45         {"body",        HTML_CLRLINE}, /* TAG_BODY */
  46         {"meta",        HTML_CLRLINE | HTML_NOSTACK}, /* TAG_META */
  47         {"title",       HTML_CLRLINE}, /* TAG_TITLE */
  48         {"div",         HTML_CLRLINE}, /* TAG_DIV */
  49         {"h1",          0}, /* TAG_H1 */
  50         {"h2",          0}, /* TAG_H2 */
  51         {"p",           HTML_CLRLINE}, /* TAG_P */
  52         {"span",        0}, /* TAG_SPAN */
  53         {"link",        HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
  54         {"br",          HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
  55         {"a",           0}, /* TAG_A */
  56         {"table",       HTML_CLRLINE}, /* TAG_TABLE */
  57         {"col",         HTML_CLRLINE | HTML_NOSTACK}, /* TAG_COL */
  58         {"tr",          HTML_CLRLINE}, /* TAG_TR */
  59         {"td",          HTML_CLRLINE}, /* TAG_TD */
  60         {"li",          HTML_CLRLINE}, /* TAG_LI */
  61         {"ul",          HTML_CLRLINE}, /* TAG_UL */
  62         {"ol",          HTML_CLRLINE}, /* TAG_OL */
  63         {"base",        HTML_CLRLINE | HTML_NOSTACK}, /* TAG_BASE */
  64 };
  65
  66 static  const char       *const htmlattrs[ATTR_MAX] = {
  67         "http-equiv",
  68         "content",
  69         "name",
  70         "rel",
  71         "href",
  72         "type",
  73         "media",
  74         "class",
  75         "style",
  76         "width",
  77         "valign",
  78         "target",
  79         "id",
  80 };
  81
  82 #ifdef __linux__
  83 extern  int               getsubopt(char **, char * const *, char **);
  84 #endif
  85
  86 void *
  87 html_alloc(char *outopts)
  88 {
  89         struct html     *h;
  90         char            *toks[4], *v;
  91
  92         toks[0] = "style";
  93         toks[1] = "man";
  94         toks[2] = "includes";
  95         toks[3] = NULL;
  96
  97         if (NULL == (h = calloc(1, sizeof(struct html))))
  98                 return(NULL);
  99
 100         SLIST_INIT(&h->tags);
 101         SLIST_INIT(&h->ords);
 102
 103         if (NULL == (h->symtab = chars_init(CHARS_HTML))) {
 104                 free(h);
 105                 return(NULL);
 106         }
 107
 108         while (outopts && *outopts)
 109                 switch (getsubopt(&outopts, toks, &v)) {
 110                 case (0):
 111                         h->style = v;
 112                         break;
 113                 case (1):
 114                         h->base_man = v;
 115                         break;
 116                 case (2):
 117                         h->base_includes = v;
 118                         break;
 119                 default:
 120                         break;
 121                 }
 122
 123         return(h);
 124 }
 125
 126
 127 void
 128 html_free(void *p)
 129 {
 130         struct tag      *tag;
 131         struct ord      *ord;
 132         struct html     *h;
 133
 134         h = (struct html *)p;
 135
 136         while ( ! SLIST_EMPTY(&h->ords)) {
 137                 ord = SLIST_FIRST(&h->ords);
 138                 SLIST_REMOVE_HEAD(&h->ords, entry);
 139                 free(ord);
 140         }
 141
 142         while ( ! SLIST_EMPTY(&h->tags)) {
 143                 tag = SLIST_FIRST(&h->tags);
 144                 SLIST_REMOVE_HEAD(&h->tags, entry);
 145                 free(tag);
 146         }
 147
 148         if (h->symtab)
 149                 chars_free(h->symtab);
 150
 151         free(h);
 152 }
 153
 154
 155 void
 156 print_gen_head(struct html *h)
 157 {
 158         struct htmlpair  tag[4];
 159
 160         tag[0].key = ATTR_HTTPEQUIV;
 161         tag[0].val = "Content-Type";
 162         tag[1].key = ATTR_CONTENT;
 163         tag[1].val = "text/html; charset=utf-8";
 164         print_otag(h, TAG_META, 2, tag);
 165
 166         tag[0].key = ATTR_NAME;
 167         tag[0].val = "resource-type";
 168         tag[1].key = ATTR_CONTENT;
 169         tag[1].val = "document";
 170         print_otag(h, TAG_META, 2, tag);
 171
 172         if (h->style) {
 173                 tag[0].key = ATTR_REL;
 174                 tag[0].val = "stylesheet";
 175                 tag[1].key = ATTR_HREF;
 176                 tag[1].val = h->style;
 177                 tag[2].key = ATTR_TYPE;
 178                 tag[2].val = "text/css";
 179                 tag[3].key = ATTR_MEDIA;
 180                 tag[3].val = "all";
 181                 print_otag(h, TAG_LINK, 4, tag);
 182         }
 183 }
 184
 185
 186 static void
 187 print_spec(struct html *h, const char *p, int len)
 188 {
 189         const char      *rhs;
 190         int              i;
 191         size_t           sz;
 192
 193         rhs = chars_a2ascii(h->symtab, p, (size_t)len, &sz);
 194
 195         if (NULL == rhs)
 196                 return;
 197         for (i = 0; i < (int)sz; i++)
 198                 putchar(rhs[i]);
 199 }
 200
 201
 202 static void
 203 print_res(struct html *h, const char *p, int len)
 204 {
 205         const char      *rhs;
 206         int              i;
 207         size_t           sz;
 208
 209         rhs = chars_a2res(h->symtab, p, (size_t)len, &sz);
 210
 211         if (NULL == rhs)
 212                 return;
 213         for (i = 0; i < (int)sz; i++)
 214                 putchar(rhs[i]);
 215 }
 216
 217
 218 static void
 219 print_escape(struct html *h, const char **p)
 220 {
 221         int              j, type;
 222         const char      *wp;
 223
 224         wp = *p;
 225         type = 1;
 226
 227         if (0 == *(++wp)) {
 228                 *p = wp;
 229                 return;
 230         }
 231
 232         if ('(' == *wp) {
 233                 wp++;
 234                 if (0 == *wp || 0 == *(wp + 1)) {
 235                         *p = 0 == *wp ? wp : wp + 1;
 236                         return;
 237                 }
 238
 239                 print_spec(h, wp, 2);
 240                 *p = ++wp;
 241                 return;
 242
 243         } else if ('*' == *wp) {
 244                 if (0 == *(++wp)) {
 245                         *p = wp;
 246                         return;
 247                 }
 248
 249                 switch (*wp) {
 250                 case ('('):
 251                         wp++;
 252                         if (0 == *wp || 0 == *(wp + 1)) {
 253                                 *p = 0 == *wp ? wp : wp + 1;
 254                                 return;
 255                         }
 256
 257                         print_res(h, wp, 2);
 258                         *p = ++wp;
 259                         return;
 260                 case ('['):
 261                         type = 0;
 262                         break;
 263                 default:
 264                         print_res(h, wp, 1);
 265                         *p = wp;
 266                         return;
 267                 }
 268
 269         } else if ('f' == *wp) {
 270                 if (0 == *(++wp)) {
 271                         *p = wp;
 272                         return;
 273                 }
 274
 275                 switch (*wp) {
 276                 case ('B'):
 277                         /* TODO */
 278                         break;
 279                 case ('I'):
 280                         /* TODO */
 281                         break;
 282                 case ('P'):
 283                         /* FALLTHROUGH */
 284                 case ('R'):
 285                         /* TODO */
 286                         break;
 287                 default:
 288                         break;
 289                 }
 290
 291                 *p = wp;
 292                 return;
 293
 294         } else if ('[' != *wp) {
 295                 print_spec(h, wp, 1);
 296                 *p = wp;
 297                 return;
 298         }
 299
 300         wp++;
 301         for (j = 0; *wp && ']' != *wp; wp++, j++)
 302                 /* Loop... */ ;
 303
 304         if (0 == *wp) {
 305                 *p = wp;
 306                 return;
 307         }
 308
 309         if (type)
 310                 print_spec(h, wp - j, j);
 311         else
 312                 print_res(h, wp - j, j);
 313
 314         *p = wp;
 315 }
 316
 317
 318 static void
 319 print_encode(struct html *h, const char *p)
 320 {
 321
 322         for (; *p; p++) {
 323                 if ('\\' == *p) {
 324                         print_escape(h, &p);
 325                         continue;
 326                 }
 327                 switch (*p) {
 328                 case ('<'):
 329                         printf("&lt;");
 330                         break;
 331                 case ('>'):
 332                         printf("&gt;");
 333                         break;
 334                 case ('&'):
 335                         printf("&amp;");
 336                         break;
 337                 default:
 338                         putchar(*p);
 339                         break;
 340                 }
 341         }
 342 }
 343
 344
 345 struct tag *
 346 print_otag(struct html *h, enum htmltag tag,
 347                 int sz, const struct htmlpair *p)
 348 {
 349         int              i;
 350         struct tag      *t;
 351
 352         if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
 353                 if (NULL == (t = malloc(sizeof(struct tag))))
 354                         err(EXIT_FAILURE, "malloc");
 355                 t->tag = tag;
 356                 SLIST_INSERT_HEAD(&h->tags, t, entry);
 357         } else
 358                 t = NULL;
 359
 360         if ( ! (HTML_NOSPACE & h->flags))
 361                 if ( ! (HTML_CLRLINE & htmltags[tag].flags))
 362                         printf(" ");
 363
 364         printf("<%s", htmltags[tag].name);
 365         for (i = 0; i < sz; i++) {
 366                 printf(" %s=\"", htmlattrs[p[i].key]);
 367                 assert(p->val);
 368                 print_encode(h, p[i].val);
 369                 printf("\"");
 370         }
 371         printf(">");
 372
 373         h->flags |= HTML_NOSPACE;
 374         if (HTML_CLRLINE & htmltags[tag].flags)
 375                 h->flags |= HTML_NEWLINE;
 376         else
 377                 h->flags &= ~HTML_NEWLINE;
 378
 379         return(t);
 380 }
 381
 382
 383 /* ARGSUSED */
 384 static void
 385 print_ctag(struct html *h, enum htmltag tag)
 386 {
 387
 388         printf("</%s>", htmltags[tag].name);
 389         if (HTML_CLRLINE & htmltags[tag].flags)
 390                 h->flags |= HTML_NOSPACE;
 391         if (HTML_CLRLINE & htmltags[tag].flags)
 392                 h->flags |= HTML_NEWLINE;
 393         else
 394                 h->flags &= ~HTML_NEWLINE;
 395 }
 396
 397
 398 /* ARGSUSED */
 399 void
 400 print_gen_doctype(struct html *h)
 401 {
 402
 403         printf("<!DOCTYPE HTML PUBLIC \"%s\" \"%s\">", DOCTYPE, DTD);
 404 }
 405
 406
 407 void
 408 print_text(struct html *h, const char *p)
 409 {
 410
 411         if (*p && 0 == *(p + 1))
 412                 switch (*p) {
 413                 case('.'):
 414                         /* FALLTHROUGH */
 415                 case(','):
 416                         /* FALLTHROUGH */
 417                 case(';'):
 418                         /* FALLTHROUGH */
 419                 case(':'):
 420                         /* FALLTHROUGH */
 421                 case('?'):
 422                         /* FALLTHROUGH */
 423                 case('!'):
 424                         /* FALLTHROUGH */
 425                 case(')'):
 426                         /* FALLTHROUGH */
 427                 case(']'):
 428                         /* FALLTHROUGH */
 429                 case('}'):
 430                         if ( ! (HTML_IGNDELIM & h->flags))
 431                                 h->flags |= HTML_NOSPACE;
 432                         break;
 433                 default:
 434                         break;
 435                 }
 436
 437         if ( ! (h->flags & HTML_NOSPACE))
 438                 printf(" ");
 439
 440         h->flags &= ~HTML_NOSPACE;
 441         h->flags &= ~HTML_NEWLINE;
 442
 443         if (p)
 444                 print_encode(h, p);
 445
 446         if (*p && 0 == *(p + 1))
 447                 switch (*p) {
 448                 case('('):
 449                         /* FALLTHROUGH */
 450                 case('['):
 451                         /* FALLTHROUGH */
 452                 case('{'):
 453                         h->flags |= HTML_NOSPACE;
 454                         break;
 455                 default:
 456                         break;
 457                 }
 458 }
 459
 460
 461 void
 462 print_tagq(struct html *h, const struct tag *until)
 463 {
 464         struct tag      *tag;
 465
 466         while ( ! SLIST_EMPTY(&h->tags)) {
 467                 tag = SLIST_FIRST(&h->tags);
 468                 print_ctag(h, tag->tag);
 469                 SLIST_REMOVE_HEAD(&h->tags, entry);
 470                 free(tag);
 471                 if (until && tag == until)
 472                         return;
 473         }
 474 }
 475
 476
 477 void
 478 print_stagq(struct html *h, const struct tag *suntil)
 479 {
 480         struct tag      *tag;
 481
 482         while ( ! SLIST_EMPTY(&h->tags)) {
 483                 tag = SLIST_FIRST(&h->tags);
 484                 if (suntil && tag == suntil)
 485                         return;
 486                 print_ctag(h, tag->tag);
 487                 SLIST_REMOVE_HEAD(&h->tags, entry);
 488                 free(tag);
 489         }
 490 }
 491
 492
 493 void
 494 bufinit(struct html *h)
 495 {
 496
 497         h->buf[0] = '\0';
 498         h->buflen = 0;
 499 }
 500
 501
 502 void
 503 bufcat_style(struct html *h, const char *key, const char *val)
 504 {
 505
 506         bufcat(h, key);
 507         bufncat(h, ":", 1);
 508         bufcat(h, val);
 509         bufncat(h, ";", 1);
 510 }
 511
 512
 513 void
 514 bufcat(struct html *h, const char *p)
 515 {
 516
 517         bufncat(h, p, strlen(p));
 518 }
 519
 520
 521 void
 522 buffmt(struct html *h, const char *fmt, ...)
 523 {
 524         va_list          ap;
 525
 526         va_start(ap, fmt);
 527         (void)vsnprintf(h->buf + (int)h->buflen,
 528                         BUFSIZ - h->buflen - 1, fmt, ap);
 529         va_end(ap);
 530         h->buflen = strlen(h->buf);
 531 }
 532
 533
 534 void
 535 bufncat(struct html *h, const char *p, size_t sz)
 536 {
 537
 538         if (h->buflen + sz > BUFSIZ - 1)
 539                 sz = BUFSIZ - 1 - h->buflen;
 540
 541         (void)strncat(h->buf, p, sz);
 542         h->buflen += sz;
 543 }
 544
 545
 546 void
 547 buffmt_includes(struct html *h, const char *name)
 548 {
 549         const char      *p, *pp;
 550
 551         pp = h->base_includes;
 552
 553         while (NULL != (p = strchr(pp, '%'))) {
 554                 bufncat(h, pp, (size_t)(p - pp));
 555                 switch (*(p + 1)) {
 556                 case('I'):
 557                         bufcat(h, name);
 558                         break;
 559                 default:
 560                         bufncat(h, p, 2);
 561                         break;
 562                 }
 563                 pp = p + 2;
 564         }
 565         if (pp)
 566                 bufcat(h, pp);
 567 }
 568
 569
 570 void
 571 buffmt_man(struct html *h,
 572                 const char *name, const char *sec)
 573 {
 574         const char      *p, *pp;
 575
 576         pp = h->base_man;
 577
 578         /* LINTED */
 579         while (NULL != (p = strchr(pp, '%'))) {
 580                 bufncat(h, pp, (size_t)(p - pp));
 581                 switch (*(p + 1)) {
 582                 case('S'):
 583                         bufcat(h, sec ? sec : "1");
 584                         break;
 585                 case('N'):
 586                         buffmt(h, name);
 587                         break;
 588                 default:
 589                         bufncat(h, p, 2);
 590                         break;
 591                 }
 592                 pp = p + 2;
 593         }
 594         if (pp)
 595                 bufcat(h, pp);
 596 }
 597
 598
 599 void
 600 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
 601 {
 602         double           v;
 603         char            *u;
 604
 605         v = su->scale;
 606
 607         switch (su->unit) {
 608         case (SCALE_CM):
 609                 u = "cm";
 610                 break;
 611         case (SCALE_IN):
 612                 u = "in";
 613                 break;
 614         case (SCALE_PC):
 615                 u = "pc";
 616                 break;
 617         case (SCALE_PT):
 618                 u = "pt";
 619                 break;
 620         case (SCALE_EM):
 621                 u = "em";
 622                 break;
 623         case (SCALE_MM):
 624                 if (0 == (v /= 100))
 625                         v = 1;
 626                 u = "em";
 627                 break;
 628         case (SCALE_EN):
 629                 u = "ex";
 630                 break;
 631         case (SCALE_BU):
 632                 u = "ex";
 633                 break;
 634         case (SCALE_VS):
 635                 u = "em";
 636                 break;
 637         default:
 638                 u = "ex";
 639                 break;
 640         }
 641
 642         if (su->pt)
 643                 buffmt(h, "%s: %f%s;", p, v, u);
 644         else
 645                 /* LINTED */
 646                 buffmt(h, "%s: %d%s;", p, (int)v, u);
 647 }