]> git.cameronkatri.com Git - mandoc.git/blob - html.c
Remove <p> in favour of <div class="spacer">.
[mandoc.git] / html.c
1 /* $Id: html.c,v 1.171 2014/09/27 10:56:18 kristaps Exp $ */
2 /*
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2011, 2012, 2013, 2014 Ingo Schwarze <schwarze@openbsd.org>
5 *
6 * Permission to use, copy, modify, and distribute this software for any
7 * purpose with or without fee is hereby granted, provided that the above
8 * copyright notice and this permission notice appear in all copies.
9 *
10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
13 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17 */
18 #include "config.h"
19
20 #include <sys/types.h>
21
22 #include <assert.h>
23 #include <ctype.h>
24 #include <stdarg.h>
25 #include <stdio.h>
26 #include <stdint.h>
27 #include <stdlib.h>
28 #include <string.h>
29 #include <unistd.h>
30
31 #include "mandoc.h"
32 #include "mandoc_aux.h"
33 #include "libmandoc.h"
34 #include "out.h"
35 #include "html.h"
36 #include "main.h"
37
38 struct htmldata {
39 const char *name;
40 int flags;
41 #define HTML_CLRLINE (1 << 0)
42 #define HTML_NOSTACK (1 << 1)
43 #define HTML_AUTOCLOSE (1 << 2) /* Tag has auto-closure. */
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 | HTML_AUTOCLOSE}, /* 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 {"span", 0}, /* TAG_SPAN */
56 {"link", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_LINK */
57 {"br", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_BR */
58 {"a", 0}, /* TAG_A */
59 {"table", HTML_CLRLINE}, /* TAG_TABLE */
60 {"tbody", HTML_CLRLINE}, /* TAG_TBODY */
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 {"dl", HTML_CLRLINE}, /* TAG_DL */
68 {"dt", HTML_CLRLINE}, /* TAG_DT */
69 {"dd", HTML_CLRLINE}, /* TAG_DD */
70 {"blockquote", HTML_CLRLINE}, /* TAG_BLOCKQUOTE */
71 {"pre", HTML_CLRLINE }, /* TAG_PRE */
72 {"b", 0 }, /* TAG_B */
73 {"i", 0 }, /* TAG_I */
74 {"code", 0 }, /* TAG_CODE */
75 {"small", 0 }, /* TAG_SMALL */
76 {"style", HTML_CLRLINE}, /* TAG_STYLE */
77 };
78
79 static const char *const htmlattrs[ATTR_MAX] = {
80 "name", /* ATTR_NAME */
81 "rel", /* ATTR_REL */
82 "href", /* ATTR_HREF */
83 "type", /* ATTR_TYPE */
84 "media", /* ATTR_MEDIA */
85 "class", /* ATTR_CLASS */
86 "style", /* ATTR_STYLE */
87 "id", /* ATTR_ID */
88 "colspan", /* ATTR_COLSPAN */
89 "charset", /* ATTR_CHARSET */
90 };
91
92 static const char *const roffscales[SCALE_MAX] = {
93 "cm", /* SCALE_CM */
94 "in", /* SCALE_IN */
95 "pc", /* SCALE_PC */
96 "pt", /* SCALE_PT */
97 "em", /* SCALE_EM */
98 "em", /* SCALE_MM */
99 "ex", /* SCALE_EN */
100 "ex", /* SCALE_BU */
101 "em", /* SCALE_VS */
102 "ex", /* SCALE_FS */
103 };
104
105 static void bufncat(struct html *, const char *, size_t);
106 static void print_ctag(struct html *, enum htmltag);
107 static int print_escape(char);
108 static int print_encode(struct html *, const char *, int);
109 static void print_metaf(struct html *, enum mandoc_esc);
110 static void print_attr(struct html *, const char *, const char *);
111 static void *ml_alloc(char *, enum htmltype);
112
113
114 static void *
115 ml_alloc(char *outopts, enum htmltype type)
116 {
117 struct html *h;
118 const char *toks[5];
119 char *v;
120
121 toks[0] = "style";
122 toks[1] = "man";
123 toks[2] = "includes";
124 toks[3] = "fragment";
125 toks[4] = NULL;
126
127 h = mandoc_calloc(1, sizeof(struct html));
128
129 h->type = type;
130 h->tags.head = NULL;
131 h->symtab = mchars_alloc();
132
133 while (outopts && *outopts)
134 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
135 case 0:
136 h->style = v;
137 break;
138 case 1:
139 h->base_man = v;
140 break;
141 case 2:
142 h->base_includes = v;
143 break;
144 case 3:
145 h->oflags |= HTML_FRAGMENT;
146 break;
147 default:
148 break;
149 }
150
151 return(h);
152 }
153
154 void *
155 html_alloc(char *outopts)
156 {
157
158 return(ml_alloc(outopts, HTML_HTML_4_01_STRICT));
159 }
160
161 void *
162 xhtml_alloc(char *outopts)
163 {
164
165 return(ml_alloc(outopts, HTML_XHTML_1_0_STRICT));
166 }
167
168 void
169 html_free(void *p)
170 {
171 struct tag *tag;
172 struct html *h;
173
174 h = (struct html *)p;
175
176 while ((tag = h->tags.head) != NULL) {
177 h->tags.head = tag->next;
178 free(tag);
179 }
180
181 if (h->symtab)
182 mchars_free(h->symtab);
183
184 free(h);
185 }
186
187 void
188 print_gen_head(struct html *h)
189 {
190 struct htmlpair tag[4];
191 struct tag *t;
192
193 tag[0].key = ATTR_CHARSET;
194 tag[0].val = "utf-8";
195 print_otag(h, TAG_META, 1, tag);
196
197 /*
198 * Print a default style-sheet.
199 */
200 t = print_otag(h, TAG_STYLE, 0, NULL);
201 print_text(h, "table.head, table.foot { width: 100%; }\n"
202 "td.head-rtitle, td.foot-os { text-align: right; }\n"
203 "td.head-vol { text-align: center; }\n"
204 "table.foot td { width: 50%; }\n"
205 "table.head td { width: 33%; }\n"
206 "div.spacer { margin: 1em 0; }\n");
207 print_tagq(h, t);
208
209 if (h->style) {
210 tag[0].key = ATTR_REL;
211 tag[0].val = "stylesheet";
212 tag[1].key = ATTR_HREF;
213 tag[1].val = h->style;
214 tag[2].key = ATTR_TYPE;
215 tag[2].val = "text/css";
216 tag[3].key = ATTR_MEDIA;
217 tag[3].val = "all";
218 print_otag(h, TAG_LINK, 4, tag);
219 }
220 }
221
222 static void
223 print_metaf(struct html *h, enum mandoc_esc deco)
224 {
225 enum htmlfont font;
226
227 switch (deco) {
228 case ESCAPE_FONTPREV:
229 font = h->metal;
230 break;
231 case ESCAPE_FONTITALIC:
232 font = HTMLFONT_ITALIC;
233 break;
234 case ESCAPE_FONTBOLD:
235 font = HTMLFONT_BOLD;
236 break;
237 case ESCAPE_FONTBI:
238 font = HTMLFONT_BI;
239 break;
240 case ESCAPE_FONT:
241 /* FALLTHROUGH */
242 case ESCAPE_FONTROMAN:
243 font = HTMLFONT_NONE;
244 break;
245 default:
246 abort();
247 /* NOTREACHED */
248 }
249
250 if (h->metaf) {
251 print_tagq(h, h->metaf);
252 h->metaf = NULL;
253 }
254
255 h->metal = h->metac;
256 h->metac = font;
257
258 switch (font) {
259 case HTMLFONT_ITALIC:
260 h->metaf = print_otag(h, TAG_I, 0, NULL);
261 break;
262 case HTMLFONT_BOLD:
263 h->metaf = print_otag(h, TAG_B, 0, NULL);
264 break;
265 case HTMLFONT_BI:
266 h->metaf = print_otag(h, TAG_B, 0, NULL);
267 print_otag(h, TAG_I, 0, NULL);
268 break;
269 default:
270 break;
271 }
272 }
273
274 int
275 html_strlen(const char *cp)
276 {
277 size_t rsz;
278 int skip, sz;
279
280 /*
281 * Account for escaped sequences within string length
282 * calculations. This follows the logic in term_strlen() as we
283 * must calculate the width of produced strings.
284 * Assume that characters are always width of "1". This is
285 * hacky, but it gets the job done for approximation of widths.
286 */
287
288 sz = 0;
289 skip = 0;
290 while (1) {
291 rsz = strcspn(cp, "\\");
292 if (rsz) {
293 cp += rsz;
294 if (skip) {
295 skip = 0;
296 rsz--;
297 }
298 sz += rsz;
299 }
300 if ('\0' == *cp)
301 break;
302 cp++;
303 switch (mandoc_escape(&cp, NULL, NULL)) {
304 case ESCAPE_ERROR:
305 return(sz);
306 case ESCAPE_UNICODE:
307 /* FALLTHROUGH */
308 case ESCAPE_NUMBERED:
309 /* FALLTHROUGH */
310 case ESCAPE_SPECIAL:
311 if (skip)
312 skip = 0;
313 else
314 sz++;
315 break;
316 case ESCAPE_SKIPCHAR:
317 skip = 1;
318 break;
319 default:
320 break;
321 }
322 }
323 return(sz);
324 }
325
326 static int
327 print_escape(char c)
328 {
329
330 switch (c) {
331 case '<':
332 printf("&lt;");
333 break;
334 case '>':
335 printf("&gt;");
336 break;
337 case '&':
338 printf("&amp;");
339 break;
340 case '"':
341 printf("&quot;");
342 break;
343 case ASCII_NBRSP:
344 putchar('-');
345 break;
346 case ASCII_HYPH:
347 putchar('-');
348 /* FALLTHROUGH */
349 case ASCII_BREAK:
350 break;
351 default:
352 return(0);
353 }
354 return(1);
355 }
356
357 static int
358 print_encode(struct html *h, const char *p, int norecurse)
359 {
360 size_t sz;
361 int c, len, nospace;
362 const char *seq;
363 enum mandoc_esc esc;
364 static const char rejs[9] = { '\\', '<', '>', '&', '"',
365 ASCII_NBRSP, ASCII_HYPH, ASCII_BREAK, '\0' };
366
367 nospace = 0;
368
369 while ('\0' != *p) {
370 if (HTML_SKIPCHAR & h->flags && '\\' != *p) {
371 h->flags &= ~HTML_SKIPCHAR;
372 p++;
373 continue;
374 }
375
376 sz = strcspn(p, rejs);
377
378 fwrite(p, 1, sz, stdout);
379 p += (int)sz;
380
381 if ('\0' == *p)
382 break;
383
384 if (print_escape(*p++))
385 continue;
386
387 esc = mandoc_escape(&p, &seq, &len);
388 if (ESCAPE_ERROR == esc)
389 break;
390
391 switch (esc) {
392 case ESCAPE_FONT:
393 /* FALLTHROUGH */
394 case ESCAPE_FONTPREV:
395 /* FALLTHROUGH */
396 case ESCAPE_FONTBOLD:
397 /* FALLTHROUGH */
398 case ESCAPE_FONTITALIC:
399 /* FALLTHROUGH */
400 case ESCAPE_FONTBI:
401 /* FALLTHROUGH */
402 case ESCAPE_FONTROMAN:
403 if (0 == norecurse)
404 print_metaf(h, esc);
405 continue;
406 case ESCAPE_SKIPCHAR:
407 h->flags |= HTML_SKIPCHAR;
408 continue;
409 default:
410 break;
411 }
412
413 if (h->flags & HTML_SKIPCHAR) {
414 h->flags &= ~HTML_SKIPCHAR;
415 continue;
416 }
417
418 switch (esc) {
419 case ESCAPE_UNICODE:
420 /* Skip past "u" header. */
421 c = mchars_num2uc(seq + 1, len - 1);
422 if ('\0' != c)
423 printf("&#x%x;", c);
424 break;
425 case ESCAPE_NUMBERED:
426 c = mchars_num2char(seq, len);
427 if ( ! ('\0' == c || print_escape(c)))
428 putchar(c);
429 break;
430 case ESCAPE_SPECIAL:
431 c = mchars_spec2cp(h->symtab, seq, len);
432 if (c > 0)
433 printf("&#%d;", c);
434 else if (-1 == c && 1 == len &&
435 !print_escape(*seq))
436 putchar((int)*seq);
437 break;
438 case ESCAPE_NOSPACE:
439 if ('\0' == *p)
440 nospace = 1;
441 break;
442 default:
443 break;
444 }
445 }
446
447 return(nospace);
448 }
449
450 static void
451 print_attr(struct html *h, const char *key, const char *val)
452 {
453 printf(" %s=\"", key);
454 (void)print_encode(h, val, 1);
455 putchar('\"');
456 }
457
458 struct tag *
459 print_otag(struct html *h, enum htmltag tag,
460 int sz, const struct htmlpair *p)
461 {
462 int i;
463 struct tag *t;
464
465 /* Push this tags onto the stack of open scopes. */
466
467 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
468 t = mandoc_malloc(sizeof(struct tag));
469 t->tag = tag;
470 t->next = h->tags.head;
471 h->tags.head = t;
472 } else
473 t = NULL;
474
475 if ( ! (HTML_NOSPACE & h->flags))
476 if ( ! (HTML_CLRLINE & htmltags[tag].flags)) {
477 /* Manage keeps! */
478 if ( ! (HTML_KEEP & h->flags)) {
479 if (HTML_PREKEEP & h->flags)
480 h->flags |= HTML_KEEP;
481 putchar(' ');
482 } else
483 printf("&#160;");
484 }
485
486 if ( ! (h->flags & HTML_NONOSPACE))
487 h->flags &= ~HTML_NOSPACE;
488 else
489 h->flags |= HTML_NOSPACE;
490
491 /* Print out the tag name and attributes. */
492
493 printf("<%s", htmltags[tag].name);
494 for (i = 0; i < sz; i++)
495 print_attr(h, htmlattrs[p[i].key], p[i].val);
496
497 /* Accommodate for XML "well-formed" singleton escaping. */
498
499 if (HTML_AUTOCLOSE & htmltags[tag].flags)
500 switch (h->type) {
501 case HTML_XHTML_1_0_STRICT:
502 putchar('/');
503 break;
504 default:
505 break;
506 }
507
508 putchar('>');
509
510 h->flags |= HTML_NOSPACE;
511
512 if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags)
513 putchar('\n');
514
515 return(t);
516 }
517
518 static void
519 print_ctag(struct html *h, enum htmltag tag)
520 {
521
522 printf("</%s>", htmltags[tag].name);
523 if (HTML_CLRLINE & htmltags[tag].flags) {
524 h->flags |= HTML_NOSPACE;
525 putchar('\n');
526 }
527 }
528
529 void
530 print_gen_decls(struct html *h)
531 {
532
533 puts("<!DOCTYPE html>");
534 }
535
536 void
537 print_text(struct html *h, const char *word)
538 {
539
540 if ( ! (HTML_NOSPACE & h->flags)) {
541 /* Manage keeps! */
542 if ( ! (HTML_KEEP & h->flags)) {
543 if (HTML_PREKEEP & h->flags)
544 h->flags |= HTML_KEEP;
545 putchar(' ');
546 } else
547 printf("&#160;");
548 }
549
550 assert(NULL == h->metaf);
551 switch (h->metac) {
552 case HTMLFONT_ITALIC:
553 h->metaf = print_otag(h, TAG_I, 0, NULL);
554 break;
555 case HTMLFONT_BOLD:
556 h->metaf = print_otag(h, TAG_B, 0, NULL);
557 break;
558 case HTMLFONT_BI:
559 h->metaf = print_otag(h, TAG_B, 0, NULL);
560 print_otag(h, TAG_I, 0, NULL);
561 break;
562 default:
563 break;
564 }
565
566 assert(word);
567 if ( ! print_encode(h, word, 0)) {
568 if ( ! (h->flags & HTML_NONOSPACE))
569 h->flags &= ~HTML_NOSPACE;
570 } else
571 h->flags |= HTML_NOSPACE;
572
573 if (h->metaf) {
574 print_tagq(h, h->metaf);
575 h->metaf = NULL;
576 }
577
578 h->flags &= ~HTML_IGNDELIM;
579 }
580
581 void
582 print_tagq(struct html *h, const struct tag *until)
583 {
584 struct tag *tag;
585
586 while ((tag = h->tags.head) != NULL) {
587 /*
588 * Remember to close out and nullify the current
589 * meta-font and table, if applicable.
590 */
591 if (tag == h->metaf)
592 h->metaf = NULL;
593 if (tag == h->tblt)
594 h->tblt = NULL;
595 print_ctag(h, tag->tag);
596 h->tags.head = tag->next;
597 free(tag);
598 if (until && tag == until)
599 return;
600 }
601 }
602
603 void
604 print_stagq(struct html *h, const struct tag *suntil)
605 {
606 struct tag *tag;
607
608 while ((tag = h->tags.head) != NULL) {
609 if (suntil && tag == suntil)
610 return;
611 /*
612 * Remember to close out and nullify the current
613 * meta-font and table, if applicable.
614 */
615 if (tag == h->metaf)
616 h->metaf = NULL;
617 if (tag == h->tblt)
618 h->tblt = NULL;
619 print_ctag(h, tag->tag);
620 h->tags.head = tag->next;
621 free(tag);
622 }
623 }
624
625 void
626 print_paragraph(struct html *h)
627 {
628 struct tag *t;
629 struct htmlpair tag;
630
631 PAIR_CLASS_INIT(&tag, "spacer");
632 t = print_otag(h, TAG_DIV, 1, &tag);
633 print_tagq(h, t);
634 }
635
636
637 void
638 bufinit(struct html *h)
639 {
640
641 h->buf[0] = '\0';
642 h->buflen = 0;
643 }
644
645 void
646 bufcat_style(struct html *h, const char *key, const char *val)
647 {
648
649 bufcat(h, key);
650 bufcat(h, ":");
651 bufcat(h, val);
652 bufcat(h, ";");
653 }
654
655 void
656 bufcat(struct html *h, const char *p)
657 {
658
659 /*
660 * XXX This is broken and not easy to fix.
661 * When using the -Oincludes option, buffmt_includes()
662 * may pass in strings overrunning BUFSIZ, causing a crash.
663 */
664
665 h->buflen = strlcat(h->buf, p, BUFSIZ);
666 assert(h->buflen < BUFSIZ);
667 }
668
669 void
670 bufcat_fmt(struct html *h, const char *fmt, ...)
671 {
672 va_list ap;
673
674 va_start(ap, fmt);
675 (void)vsnprintf(h->buf + (int)h->buflen,
676 BUFSIZ - h->buflen - 1, fmt, ap);
677 va_end(ap);
678 h->buflen = strlen(h->buf);
679 }
680
681 static void
682 bufncat(struct html *h, const char *p, size_t sz)
683 {
684
685 assert(h->buflen + sz + 1 < BUFSIZ);
686 strncat(h->buf, p, sz);
687 h->buflen += sz;
688 }
689
690 void
691 buffmt_includes(struct html *h, const char *name)
692 {
693 const char *p, *pp;
694
695 pp = h->base_includes;
696
697 bufinit(h);
698 while (NULL != (p = strchr(pp, '%'))) {
699 bufncat(h, pp, (size_t)(p - pp));
700 switch (*(p + 1)) {
701 case'I':
702 bufcat(h, name);
703 break;
704 default:
705 bufncat(h, p, 2);
706 break;
707 }
708 pp = p + 2;
709 }
710 if (pp)
711 bufcat(h, pp);
712 }
713
714 void
715 buffmt_man(struct html *h, const char *name, const char *sec)
716 {
717 const char *p, *pp;
718
719 pp = h->base_man;
720
721 bufinit(h);
722 while (NULL != (p = strchr(pp, '%'))) {
723 bufncat(h, pp, (size_t)(p - pp));
724 switch (*(p + 1)) {
725 case 'S':
726 bufcat(h, sec ? sec : "1");
727 break;
728 case 'N':
729 bufcat_fmt(h, "%s", name);
730 break;
731 default:
732 bufncat(h, p, 2);
733 break;
734 }
735 pp = p + 2;
736 }
737 if (pp)
738 bufcat(h, pp);
739 }
740
741 void
742 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
743 {
744 double v;
745
746 v = su->scale;
747 if (SCALE_MM == su->unit && 0.0 == (v /= 100.0))
748 v = 1.0;
749 else if (SCALE_BU == su->unit)
750 v /= 24.0;
751
752 bufcat_fmt(h, "%s: %.2f%s;", p, v, roffscales[su->unit]);
753 }
754
755 void
756 bufcat_id(struct html *h, const char *src)
757 {
758
759 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
760
761 while ('\0' != *src)
762 bufcat_fmt(h, "%.2x", *src++);
763 }