]> git.cameronkatri.com Git - mandoc.git/blob - html.c
Don't pretend we have a separate XHTML and HTML mode any more.
[mandoc.git] / html.c
1 /* $Id: html.c,v 1.172 2014/09/27 11:17:19 kristaps Exp $ */
2 /*
3 * Copyright (c) 2008, 2009, 2010, 2011, 2014 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 *);
112
113
114 static void *
115 ml_alloc(char *outopts)
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->tags.head = NULL;
130 h->symtab = mchars_alloc();
131
132 while (outopts && *outopts)
133 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
134 case 0:
135 h->style = v;
136 break;
137 case 1:
138 h->base_man = v;
139 break;
140 case 2:
141 h->base_includes = v;
142 break;
143 case 3:
144 h->oflags |= HTML_FRAGMENT;
145 break;
146 default:
147 break;
148 }
149
150 return(h);
151 }
152
153 void *
154 html_alloc(char *outopts)
155 {
156
157 return(ml_alloc(outopts));
158 }
159
160 void *
161 xhtml_alloc(char *outopts)
162 {
163
164 return(ml_alloc(outopts));
165 }
166
167 void
168 html_free(void *p)
169 {
170 struct tag *tag;
171 struct html *h;
172
173 h = (struct html *)p;
174
175 while ((tag = h->tags.head) != NULL) {
176 h->tags.head = tag->next;
177 free(tag);
178 }
179
180 if (h->symtab)
181 mchars_free(h->symtab);
182
183 free(h);
184 }
185
186 void
187 print_gen_head(struct html *h)
188 {
189 struct htmlpair tag[4];
190 struct tag *t;
191
192 tag[0].key = ATTR_CHARSET;
193 tag[0].val = "utf-8";
194 print_otag(h, TAG_META, 1, tag);
195
196 /*
197 * Print a default style-sheet.
198 */
199 t = print_otag(h, TAG_STYLE, 0, NULL);
200 print_text(h, "table.head, table.foot { width: 100%; }\n"
201 "td.head-rtitle, td.foot-os { text-align: right; }\n"
202 "td.head-vol { text-align: center; }\n"
203 "table.foot td { width: 50%; }\n"
204 "table.head td { width: 33%; }\n"
205 "div.spacer { margin: 1em 0; }\n");
206 print_tagq(h, t);
207
208 if (h->style) {
209 tag[0].key = ATTR_REL;
210 tag[0].val = "stylesheet";
211 tag[1].key = ATTR_HREF;
212 tag[1].val = h->style;
213 tag[2].key = ATTR_TYPE;
214 tag[2].val = "text/css";
215 tag[3].key = ATTR_MEDIA;
216 tag[3].val = "all";
217 print_otag(h, TAG_LINK, 4, tag);
218 }
219 }
220
221 static void
222 print_metaf(struct html *h, enum mandoc_esc deco)
223 {
224 enum htmlfont font;
225
226 switch (deco) {
227 case ESCAPE_FONTPREV:
228 font = h->metal;
229 break;
230 case ESCAPE_FONTITALIC:
231 font = HTMLFONT_ITALIC;
232 break;
233 case ESCAPE_FONTBOLD:
234 font = HTMLFONT_BOLD;
235 break;
236 case ESCAPE_FONTBI:
237 font = HTMLFONT_BI;
238 break;
239 case ESCAPE_FONT:
240 /* FALLTHROUGH */
241 case ESCAPE_FONTROMAN:
242 font = HTMLFONT_NONE;
243 break;
244 default:
245 abort();
246 /* NOTREACHED */
247 }
248
249 if (h->metaf) {
250 print_tagq(h, h->metaf);
251 h->metaf = NULL;
252 }
253
254 h->metal = h->metac;
255 h->metac = font;
256
257 switch (font) {
258 case HTMLFONT_ITALIC:
259 h->metaf = print_otag(h, TAG_I, 0, NULL);
260 break;
261 case HTMLFONT_BOLD:
262 h->metaf = print_otag(h, TAG_B, 0, NULL);
263 break;
264 case HTMLFONT_BI:
265 h->metaf = print_otag(h, TAG_B, 0, NULL);
266 print_otag(h, TAG_I, 0, NULL);
267 break;
268 default:
269 break;
270 }
271 }
272
273 int
274 html_strlen(const char *cp)
275 {
276 size_t rsz;
277 int skip, sz;
278
279 /*
280 * Account for escaped sequences within string length
281 * calculations. This follows the logic in term_strlen() as we
282 * must calculate the width of produced strings.
283 * Assume that characters are always width of "1". This is
284 * hacky, but it gets the job done for approximation of widths.
285 */
286
287 sz = 0;
288 skip = 0;
289 while (1) {
290 rsz = strcspn(cp, "\\");
291 if (rsz) {
292 cp += rsz;
293 if (skip) {
294 skip = 0;
295 rsz--;
296 }
297 sz += rsz;
298 }
299 if ('\0' == *cp)
300 break;
301 cp++;
302 switch (mandoc_escape(&cp, NULL, NULL)) {
303 case ESCAPE_ERROR:
304 return(sz);
305 case ESCAPE_UNICODE:
306 /* FALLTHROUGH */
307 case ESCAPE_NUMBERED:
308 /* FALLTHROUGH */
309 case ESCAPE_SPECIAL:
310 if (skip)
311 skip = 0;
312 else
313 sz++;
314 break;
315 case ESCAPE_SKIPCHAR:
316 skip = 1;
317 break;
318 default:
319 break;
320 }
321 }
322 return(sz);
323 }
324
325 static int
326 print_escape(char c)
327 {
328
329 switch (c) {
330 case '<':
331 printf("&lt;");
332 break;
333 case '>':
334 printf("&gt;");
335 break;
336 case '&':
337 printf("&amp;");
338 break;
339 case '"':
340 printf("&quot;");
341 break;
342 case ASCII_NBRSP:
343 putchar('-');
344 break;
345 case ASCII_HYPH:
346 putchar('-');
347 /* FALLTHROUGH */
348 case ASCII_BREAK:
349 break;
350 default:
351 return(0);
352 }
353 return(1);
354 }
355
356 static int
357 print_encode(struct html *h, const char *p, int norecurse)
358 {
359 size_t sz;
360 int c, len, nospace;
361 const char *seq;
362 enum mandoc_esc esc;
363 static const char rejs[9] = { '\\', '<', '>', '&', '"',
364 ASCII_NBRSP, ASCII_HYPH, ASCII_BREAK, '\0' };
365
366 nospace = 0;
367
368 while ('\0' != *p) {
369 if (HTML_SKIPCHAR & h->flags && '\\' != *p) {
370 h->flags &= ~HTML_SKIPCHAR;
371 p++;
372 continue;
373 }
374
375 sz = strcspn(p, rejs);
376
377 fwrite(p, 1, sz, stdout);
378 p += (int)sz;
379
380 if ('\0' == *p)
381 break;
382
383 if (print_escape(*p++))
384 continue;
385
386 esc = mandoc_escape(&p, &seq, &len);
387 if (ESCAPE_ERROR == esc)
388 break;
389
390 switch (esc) {
391 case ESCAPE_FONT:
392 /* FALLTHROUGH */
393 case ESCAPE_FONTPREV:
394 /* FALLTHROUGH */
395 case ESCAPE_FONTBOLD:
396 /* FALLTHROUGH */
397 case ESCAPE_FONTITALIC:
398 /* FALLTHROUGH */
399 case ESCAPE_FONTBI:
400 /* FALLTHROUGH */
401 case ESCAPE_FONTROMAN:
402 if (0 == norecurse)
403 print_metaf(h, esc);
404 continue;
405 case ESCAPE_SKIPCHAR:
406 h->flags |= HTML_SKIPCHAR;
407 continue;
408 default:
409 break;
410 }
411
412 if (h->flags & HTML_SKIPCHAR) {
413 h->flags &= ~HTML_SKIPCHAR;
414 continue;
415 }
416
417 switch (esc) {
418 case ESCAPE_UNICODE:
419 /* Skip past "u" header. */
420 c = mchars_num2uc(seq + 1, len - 1);
421 if ('\0' != c)
422 printf("&#x%x;", c);
423 break;
424 case ESCAPE_NUMBERED:
425 c = mchars_num2char(seq, len);
426 if ( ! ('\0' == c || print_escape(c)))
427 putchar(c);
428 break;
429 case ESCAPE_SPECIAL:
430 c = mchars_spec2cp(h->symtab, seq, len);
431 if (c > 0)
432 printf("&#%d;", c);
433 else if (-1 == c && 1 == len &&
434 !print_escape(*seq))
435 putchar((int)*seq);
436 break;
437 case ESCAPE_NOSPACE:
438 if ('\0' == *p)
439 nospace = 1;
440 break;
441 default:
442 break;
443 }
444 }
445
446 return(nospace);
447 }
448
449 static void
450 print_attr(struct html *h, const char *key, const char *val)
451 {
452 printf(" %s=\"", key);
453 (void)print_encode(h, val, 1);
454 putchar('\"');
455 }
456
457 struct tag *
458 print_otag(struct html *h, enum htmltag tag,
459 int sz, const struct htmlpair *p)
460 {
461 int i;
462 struct tag *t;
463
464 /* Push this tags onto the stack of open scopes. */
465
466 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
467 t = mandoc_malloc(sizeof(struct tag));
468 t->tag = tag;
469 t->next = h->tags.head;
470 h->tags.head = t;
471 } else
472 t = NULL;
473
474 if ( ! (HTML_NOSPACE & h->flags))
475 if ( ! (HTML_CLRLINE & htmltags[tag].flags)) {
476 /* Manage keeps! */
477 if ( ! (HTML_KEEP & h->flags)) {
478 if (HTML_PREKEEP & h->flags)
479 h->flags |= HTML_KEEP;
480 putchar(' ');
481 } else
482 printf("&#160;");
483 }
484
485 if ( ! (h->flags & HTML_NONOSPACE))
486 h->flags &= ~HTML_NOSPACE;
487 else
488 h->flags |= HTML_NOSPACE;
489
490 /* Print out the tag name and attributes. */
491
492 printf("<%s", htmltags[tag].name);
493 for (i = 0; i < sz; i++)
494 print_attr(h, htmlattrs[p[i].key], p[i].val);
495
496 /* Accommodate for "well-formed" singleton escaping. */
497
498 if (HTML_AUTOCLOSE & htmltags[tag].flags)
499 putchar('/');
500
501 putchar('>');
502
503 h->flags |= HTML_NOSPACE;
504
505 if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags)
506 putchar('\n');
507
508 return(t);
509 }
510
511 static void
512 print_ctag(struct html *h, enum htmltag tag)
513 {
514
515 printf("</%s>", htmltags[tag].name);
516 if (HTML_CLRLINE & htmltags[tag].flags) {
517 h->flags |= HTML_NOSPACE;
518 putchar('\n');
519 }
520 }
521
522 void
523 print_gen_decls(struct html *h)
524 {
525
526 puts("<!DOCTYPE html>");
527 }
528
529 void
530 print_text(struct html *h, const char *word)
531 {
532
533 if ( ! (HTML_NOSPACE & h->flags)) {
534 /* Manage keeps! */
535 if ( ! (HTML_KEEP & h->flags)) {
536 if (HTML_PREKEEP & h->flags)
537 h->flags |= HTML_KEEP;
538 putchar(' ');
539 } else
540 printf("&#160;");
541 }
542
543 assert(NULL == h->metaf);
544 switch (h->metac) {
545 case HTMLFONT_ITALIC:
546 h->metaf = print_otag(h, TAG_I, 0, NULL);
547 break;
548 case HTMLFONT_BOLD:
549 h->metaf = print_otag(h, TAG_B, 0, NULL);
550 break;
551 case HTMLFONT_BI:
552 h->metaf = print_otag(h, TAG_B, 0, NULL);
553 print_otag(h, TAG_I, 0, NULL);
554 break;
555 default:
556 break;
557 }
558
559 assert(word);
560 if ( ! print_encode(h, word, 0)) {
561 if ( ! (h->flags & HTML_NONOSPACE))
562 h->flags &= ~HTML_NOSPACE;
563 } else
564 h->flags |= HTML_NOSPACE;
565
566 if (h->metaf) {
567 print_tagq(h, h->metaf);
568 h->metaf = NULL;
569 }
570
571 h->flags &= ~HTML_IGNDELIM;
572 }
573
574 void
575 print_tagq(struct html *h, const struct tag *until)
576 {
577 struct tag *tag;
578
579 while ((tag = h->tags.head) != NULL) {
580 /*
581 * Remember to close out and nullify the current
582 * meta-font and table, if applicable.
583 */
584 if (tag == h->metaf)
585 h->metaf = NULL;
586 if (tag == h->tblt)
587 h->tblt = NULL;
588 print_ctag(h, tag->tag);
589 h->tags.head = tag->next;
590 free(tag);
591 if (until && tag == until)
592 return;
593 }
594 }
595
596 void
597 print_stagq(struct html *h, const struct tag *suntil)
598 {
599 struct tag *tag;
600
601 while ((tag = h->tags.head) != NULL) {
602 if (suntil && tag == suntil)
603 return;
604 /*
605 * Remember to close out and nullify the current
606 * meta-font and table, if applicable.
607 */
608 if (tag == h->metaf)
609 h->metaf = NULL;
610 if (tag == h->tblt)
611 h->tblt = NULL;
612 print_ctag(h, tag->tag);
613 h->tags.head = tag->next;
614 free(tag);
615 }
616 }
617
618 void
619 print_paragraph(struct html *h)
620 {
621 struct tag *t;
622 struct htmlpair tag;
623
624 PAIR_CLASS_INIT(&tag, "spacer");
625 t = print_otag(h, TAG_DIV, 1, &tag);
626 print_tagq(h, t);
627 }
628
629
630 void
631 bufinit(struct html *h)
632 {
633
634 h->buf[0] = '\0';
635 h->buflen = 0;
636 }
637
638 void
639 bufcat_style(struct html *h, const char *key, const char *val)
640 {
641
642 bufcat(h, key);
643 bufcat(h, ":");
644 bufcat(h, val);
645 bufcat(h, ";");
646 }
647
648 void
649 bufcat(struct html *h, const char *p)
650 {
651
652 /*
653 * XXX This is broken and not easy to fix.
654 * When using the -Oincludes option, buffmt_includes()
655 * may pass in strings overrunning BUFSIZ, causing a crash.
656 */
657
658 h->buflen = strlcat(h->buf, p, BUFSIZ);
659 assert(h->buflen < BUFSIZ);
660 }
661
662 void
663 bufcat_fmt(struct html *h, const char *fmt, ...)
664 {
665 va_list ap;
666
667 va_start(ap, fmt);
668 (void)vsnprintf(h->buf + (int)h->buflen,
669 BUFSIZ - h->buflen - 1, fmt, ap);
670 va_end(ap);
671 h->buflen = strlen(h->buf);
672 }
673
674 static void
675 bufncat(struct html *h, const char *p, size_t sz)
676 {
677
678 assert(h->buflen + sz + 1 < BUFSIZ);
679 strncat(h->buf, p, sz);
680 h->buflen += sz;
681 }
682
683 void
684 buffmt_includes(struct html *h, const char *name)
685 {
686 const char *p, *pp;
687
688 pp = h->base_includes;
689
690 bufinit(h);
691 while (NULL != (p = strchr(pp, '%'))) {
692 bufncat(h, pp, (size_t)(p - pp));
693 switch (*(p + 1)) {
694 case'I':
695 bufcat(h, name);
696 break;
697 default:
698 bufncat(h, p, 2);
699 break;
700 }
701 pp = p + 2;
702 }
703 if (pp)
704 bufcat(h, pp);
705 }
706
707 void
708 buffmt_man(struct html *h, const char *name, const char *sec)
709 {
710 const char *p, *pp;
711
712 pp = h->base_man;
713
714 bufinit(h);
715 while (NULL != (p = strchr(pp, '%'))) {
716 bufncat(h, pp, (size_t)(p - pp));
717 switch (*(p + 1)) {
718 case 'S':
719 bufcat(h, sec ? sec : "1");
720 break;
721 case 'N':
722 bufcat_fmt(h, "%s", name);
723 break;
724 default:
725 bufncat(h, p, 2);
726 break;
727 }
728 pp = p + 2;
729 }
730 if (pp)
731 bufcat(h, pp);
732 }
733
734 void
735 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
736 {
737 double v;
738
739 v = su->scale;
740 if (SCALE_MM == su->unit && 0.0 == (v /= 100.0))
741 v = 1.0;
742 else if (SCALE_BU == su->unit)
743 v /= 24.0;
744
745 bufcat_fmt(h, "%s: %.2f%s;", p, v, roffscales[su->unit]);
746 }
747
748 void
749 bufcat_id(struct html *h, const char *src)
750 {
751
752 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
753
754 while ('\0' != *src)
755 bufcat_fmt(h, "%.2x", *src++);
756 }