]> git.cameronkatri.com Git - mandoc.git/blob - html.c
5a63ea5db78bd21369b64258255f44034252eedb
[mandoc.git] / html.c
1 /* $Id: html.c,v 1.115 2010/12/15 15:59:23 kristaps Exp $ */
2 /*
3 * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
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 "mandoc.h"
33 #include "out.h"
34 #include "chars.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 };
72
73 static const char *const htmlfonts[HTMLFONT_MAX] = {
74 "roman",
75 "bold",
76 "italic"
77 };
78
79 static const char *const htmlattrs[ATTR_MAX] = {
80 "http-equiv",
81 "content",
82 "name",
83 "rel",
84 "href",
85 "type",
86 "media",
87 "class",
88 "style",
89 "width",
90 "valign",
91 "target",
92 "id",
93 "summary",
94 };
95
96 static void print_spec(struct html *, enum roffdeco,
97 const char *, size_t);
98 static void print_res(struct html *, const char *, size_t);
99 static void print_ctag(struct html *, enum htmltag);
100 static void print_doctype(struct html *);
101 static void print_xmltype(struct html *);
102 static int print_encode(struct html *, const char *, int);
103 static void print_metaf(struct html *, enum roffdeco);
104 static void print_attr(struct html *,
105 const char *, const char *);
106 static void *ml_alloc(char *, enum htmltype);
107
108
109 static void *
110 ml_alloc(char *outopts, enum htmltype type)
111 {
112 struct html *h;
113 const char *toks[4];
114 char *v;
115
116 toks[0] = "style";
117 toks[1] = "man";
118 toks[2] = "includes";
119 toks[3] = NULL;
120
121 h = calloc(1, sizeof(struct html));
122 if (NULL == h) {
123 perror(NULL);
124 exit((int)MANDOCLEVEL_SYSERR);
125 }
126
127 h->type = type;
128 h->tags.head = NULL;
129 h->symtab = chars_init(CHARS_HTML);
130
131 while (outopts && *outopts)
132 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
133 case (0):
134 h->style = v;
135 break;
136 case (1):
137 h->base_man = v;
138 break;
139 case (2):
140 h->base_includes = v;
141 break;
142 default:
143 break;
144 }
145
146 return(h);
147 }
148
149 void *
150 html_alloc(char *outopts)
151 {
152
153 return(ml_alloc(outopts, HTML_HTML_4_01_STRICT));
154 }
155
156
157 void *
158 xhtml_alloc(char *outopts)
159 {
160
161 return(ml_alloc(outopts, HTML_XHTML_1_0_STRICT));
162 }
163
164
165 void
166 html_free(void *p)
167 {
168 struct tag *tag;
169 struct html *h;
170
171 h = (struct html *)p;
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, enum roffdeco d, const char *p, size_t len)
218 {
219 int cp;
220 const char *rhs;
221 size_t sz;
222
223 if ((cp = chars_spec2cp(h->symtab, p, len)) > 0) {
224 printf("&#%d;", cp);
225 return;
226 } else if (-1 == cp && DECO_SSPECIAL == d) {
227 fwrite(p, 1, len, stdout);
228 return;
229 } else if (-1 == cp)
230 return;
231
232 if (NULL != (rhs = chars_spec2str(h->symtab, p, len, &sz)))
233 fwrite(rhs, 1, sz, stdout);
234 }
235
236
237 static void
238 print_res(struct html *h, const char *p, size_t len)
239 {
240 int cp;
241 const char *rhs;
242 size_t sz;
243
244 if ((cp = chars_res2cp(h->symtab, p, len)) > 0) {
245 printf("&#%d;", cp);
246 return;
247 } else if (-1 == cp)
248 return;
249
250 if (NULL != (rhs = chars_res2str(h->symtab, p, len, &sz)))
251 fwrite(rhs, 1, sz, stdout);
252 }
253
254
255 struct tag *
256 print_ofont(struct html *h, enum htmlfont font)
257 {
258 struct htmlpair tag;
259
260 h->metal = h->metac;
261 h->metac = font;
262
263 /* FIXME: DECO_ROMAN should just close out preexisting. */
264
265 if (h->metaf && h->tags.head == h->metaf)
266 print_tagq(h, h->metaf);
267
268 PAIR_CLASS_INIT(&tag, htmlfonts[font]);
269 h->metaf = print_otag(h, TAG_SPAN, 1, &tag);
270 return(h->metaf);
271 }
272
273
274 static void
275 print_metaf(struct html *h, enum roffdeco deco)
276 {
277 enum htmlfont font;
278
279 switch (deco) {
280 case (DECO_PREVIOUS):
281 font = h->metal;
282 break;
283 case (DECO_ITALIC):
284 font = HTMLFONT_ITALIC;
285 break;
286 case (DECO_BOLD):
287 font = HTMLFONT_BOLD;
288 break;
289 case (DECO_ROMAN):
290 font = HTMLFONT_NONE;
291 break;
292 default:
293 abort();
294 /* NOTREACHED */
295 }
296
297 (void)print_ofont(h, font);
298 }
299
300
301 static int
302 print_encode(struct html *h, const char *p, int norecurse)
303 {
304 size_t sz;
305 int len, nospace;
306 const char *seq;
307 enum roffdeco deco;
308 static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' };
309
310 nospace = 0;
311
312 for (; *p; p++) {
313 sz = strcspn(p, rejs);
314
315 fwrite(p, 1, sz, stdout);
316 p += /* LINTED */
317 sz;
318
319 if ('<' == *p) {
320 printf("&lt;");
321 continue;
322 } else if ('>' == *p) {
323 printf("&gt;");
324 continue;
325 } else if ('&' == *p) {
326 printf("&amp;");
327 continue;
328 } else if (ASCII_HYPH == *p) {
329 /*
330 * Note: "soft hyphens" aren't graphically
331 * displayed when not breaking the text; we want
332 * them to be displayed.
333 */
334 /*printf("&#173;");*/
335 putchar('-');
336 continue;
337 } else if ('\0' == *p)
338 break;
339
340 seq = ++p;
341 len = a2roffdeco(&deco, &seq, &sz);
342
343 switch (deco) {
344 case (DECO_RESERVED):
345 print_res(h, seq, sz);
346 break;
347 case (DECO_SSPECIAL):
348 /* FALLTHROUGH */
349 case (DECO_SPECIAL):
350 print_spec(h, deco, seq, sz);
351 break;
352 case (DECO_PREVIOUS):
353 /* FALLTHROUGH */
354 case (DECO_BOLD):
355 /* FALLTHROUGH */
356 case (DECO_ITALIC):
357 /* FALLTHROUGH */
358 case (DECO_ROMAN):
359 if (norecurse)
360 break;
361 print_metaf(h, deco);
362 break;
363 default:
364 break;
365 }
366
367 p += len - 1;
368
369 if (DECO_NOSPACE == deco && '\0' == *(p + 1))
370 nospace = 1;
371 }
372
373 return(nospace);
374 }
375
376
377 static void
378 print_attr(struct html *h, const char *key, const char *val)
379 {
380 printf(" %s=\"", key);
381 (void)print_encode(h, val, 1);
382 putchar('\"');
383 }
384
385
386 struct tag *
387 print_otag(struct html *h, enum htmltag tag,
388 int sz, const struct htmlpair *p)
389 {
390 int i;
391 struct tag *t;
392
393 /* Push this tags onto the stack of open scopes. */
394
395 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
396 t = malloc(sizeof(struct tag));
397 if (NULL == t) {
398 perror(NULL);
399 exit((int)MANDOCLEVEL_SYSERR);
400 }
401 t->tag = tag;
402 t->next = h->tags.head;
403 h->tags.head = t;
404 } else
405 t = NULL;
406
407 if ( ! (HTML_NOSPACE & h->flags))
408 if ( ! (HTML_CLRLINE & htmltags[tag].flags)) {
409 /* Manage keeps! */
410 if ( ! (HTML_KEEP & h->flags)) {
411 if (HTML_PREKEEP & h->flags)
412 h->flags |= HTML_KEEP;
413 putchar(' ');
414 } else
415 printf("&#160;");
416 }
417
418 if ( ! (h->flags & HTML_NONOSPACE))
419 h->flags &= ~HTML_NOSPACE;
420 else
421 h->flags |= HTML_NOSPACE;
422
423 /* Print out the tag name and attributes. */
424
425 printf("<%s", htmltags[tag].name);
426 for (i = 0; i < sz; i++)
427 print_attr(h, htmlattrs[p[i].key], p[i].val);
428
429 /* Add non-overridable attributes. */
430
431 if (TAG_HTML == tag && HTML_XHTML_1_0_STRICT == h->type) {
432 print_attr(h, "xmlns", "http://www.w3.org/1999/xhtml");
433 print_attr(h, "xml:lang", "en");
434 print_attr(h, "lang", "en");
435 }
436
437 /* Accomodate for XML "well-formed" singleton escaping. */
438
439 if (HTML_AUTOCLOSE & htmltags[tag].flags)
440 switch (h->type) {
441 case (HTML_XHTML_1_0_STRICT):
442 putchar('/');
443 break;
444 default:
445 break;
446 }
447
448 putchar('>');
449
450 h->flags |= HTML_NOSPACE;
451 return(t);
452 }
453
454
455 static void
456 print_ctag(struct html *h, enum htmltag tag)
457 {
458
459 printf("</%s>", htmltags[tag].name);
460 if (HTML_CLRLINE & htmltags[tag].flags) {
461 h->flags |= HTML_NOSPACE;
462 putchar('\n');
463 }
464 }
465
466
467 void
468 print_gen_decls(struct html *h)
469 {
470
471 print_xmltype(h);
472 print_doctype(h);
473 }
474
475
476 static void
477 print_xmltype(struct html *h)
478 {
479
480 if (HTML_XHTML_1_0_STRICT == h->type)
481 printf("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
482 }
483
484
485 static void
486 print_doctype(struct html *h)
487 {
488 const char *doctype;
489 const char *dtd;
490 const char *name;
491
492 switch (h->type) {
493 case (HTML_HTML_4_01_STRICT):
494 name = "HTML";
495 doctype = "-//W3C//DTD HTML 4.01//EN";
496 dtd = "http://www.w3.org/TR/html4/strict.dtd";
497 break;
498 default:
499 name = "html";
500 doctype = "-//W3C//DTD XHTML 1.0 Strict//EN";
501 dtd = "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
502 break;
503 }
504
505 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
506 name, doctype, dtd);
507 }
508
509
510 void
511 print_text(struct html *h, const char *word)
512 {
513
514 if (word[0] && '\0' == word[1])
515 switch (word[0]) {
516 case('.'):
517 /* FALLTHROUGH */
518 case(','):
519 /* FALLTHROUGH */
520 case(';'):
521 /* FALLTHROUGH */
522 case(':'):
523 /* FALLTHROUGH */
524 case('?'):
525 /* FALLTHROUGH */
526 case('!'):
527 /* FALLTHROUGH */
528 case(')'):
529 /* FALLTHROUGH */
530 case(']'):
531 if ( ! (HTML_IGNDELIM & h->flags))
532 h->flags |= HTML_NOSPACE;
533 break;
534 default:
535 break;
536 }
537
538 if ( ! (HTML_NOSPACE & h->flags)) {
539 /* Manage keeps! */
540 if ( ! (HTML_KEEP & h->flags)) {
541 if (HTML_PREKEEP & h->flags)
542 h->flags |= HTML_KEEP;
543 putchar(' ');
544 } else
545 printf("&#160;");
546 }
547
548 assert(word);
549 if ( ! print_encode(h, word, 0))
550 if ( ! (h->flags & HTML_NONOSPACE))
551 h->flags &= ~HTML_NOSPACE;
552
553 h->flags &= ~HTML_IGNDELIM;
554
555 /*
556 * Note that we don't process the pipe: the parser sees it as
557 * punctuation, but we don't in terms of typography.
558 */
559 if (word[0] && '\0' == word[1])
560 switch (word[0]) {
561 case('('):
562 /* FALLTHROUGH */
563 case('['):
564 h->flags |= HTML_NOSPACE;
565 break;
566 default:
567 break;
568 }
569 }
570
571
572 void
573 print_tagq(struct html *h, const struct tag *until)
574 {
575 struct tag *tag;
576
577 while ((tag = h->tags.head) != NULL) {
578 if (tag == h->metaf)
579 h->metaf = NULL;
580 print_ctag(h, tag->tag);
581 h->tags.head = tag->next;
582 free(tag);
583 if (until && tag == until)
584 return;
585 }
586 }
587
588
589 void
590 print_stagq(struct html *h, const struct tag *suntil)
591 {
592 struct tag *tag;
593
594 while ((tag = h->tags.head) != NULL) {
595 if (suntil && tag == suntil)
596 return;
597 if (tag == h->metaf)
598 h->metaf = NULL;
599 print_ctag(h, tag->tag);
600 h->tags.head = tag->next;
601 free(tag);
602 }
603 }
604
605
606 void
607 bufinit(struct html *h)
608 {
609
610 h->buf[0] = '\0';
611 h->buflen = 0;
612 }
613
614
615 void
616 bufcat_style(struct html *h, const char *key, const char *val)
617 {
618
619 bufcat(h, key);
620 bufncat(h, ":", 1);
621 bufcat(h, val);
622 bufncat(h, ";", 1);
623 }
624
625
626 void
627 bufcat(struct html *h, const char *p)
628 {
629
630 bufncat(h, p, strlen(p));
631 }
632
633
634 void
635 buffmt(struct html *h, const char *fmt, ...)
636 {
637 va_list ap;
638
639 va_start(ap, fmt);
640 (void)vsnprintf(h->buf + (int)h->buflen,
641 BUFSIZ - h->buflen - 1, fmt, ap);
642 va_end(ap);
643 h->buflen = strlen(h->buf);
644 }
645
646
647 void
648 bufncat(struct html *h, const char *p, size_t sz)
649 {
650
651 if (h->buflen + sz > BUFSIZ - 1)
652 sz = BUFSIZ - 1 - h->buflen;
653
654 (void)strncat(h->buf, p, sz);
655 h->buflen += sz;
656 }
657
658
659 void
660 buffmt_includes(struct html *h, const char *name)
661 {
662 const char *p, *pp;
663
664 pp = h->base_includes;
665
666 while (NULL != (p = strchr(pp, '%'))) {
667 bufncat(h, pp, (size_t)(p - pp));
668 switch (*(p + 1)) {
669 case('I'):
670 bufcat(h, name);
671 break;
672 default:
673 bufncat(h, p, 2);
674 break;
675 }
676 pp = p + 2;
677 }
678 if (pp)
679 bufcat(h, pp);
680 }
681
682
683 void
684 buffmt_man(struct html *h,
685 const char *name, const char *sec)
686 {
687 const char *p, *pp;
688
689 pp = h->base_man;
690
691 /* LINTED */
692 while (NULL != (p = strchr(pp, '%'))) {
693 bufncat(h, pp, (size_t)(p - pp));
694 switch (*(p + 1)) {
695 case('S'):
696 bufcat(h, sec ? sec : "1");
697 break;
698 case('N'):
699 buffmt(h, name);
700 break;
701 default:
702 bufncat(h, p, 2);
703 break;
704 }
705 pp = p + 2;
706 }
707 if (pp)
708 bufcat(h, pp);
709 }
710
711
712 void
713 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
714 {
715 double v;
716 const char *u;
717
718 v = su->scale;
719
720 switch (su->unit) {
721 case (SCALE_CM):
722 u = "cm";
723 break;
724 case (SCALE_IN):
725 u = "in";
726 break;
727 case (SCALE_PC):
728 u = "pc";
729 break;
730 case (SCALE_PT):
731 u = "pt";
732 break;
733 case (SCALE_EM):
734 u = "em";
735 break;
736 case (SCALE_MM):
737 if (0 == (v /= 100))
738 v = 1;
739 u = "em";
740 break;
741 case (SCALE_EN):
742 u = "ex";
743 break;
744 case (SCALE_BU):
745 u = "ex";
746 break;
747 case (SCALE_VS):
748 u = "em";
749 break;
750 default:
751 u = "ex";
752 break;
753 }
754
755 /*
756 * XXX: the CSS spec isn't clear as to which types accept
757 * integer or real numbers, so we just make them all decimals.
758 */
759 buffmt(h, "%s: %.2f%s;", p, v, u);
760 }
761
762
763 void
764 html_idcat(char *dst, const char *src, int sz)
765 {
766 int ssz;
767
768 assert(sz);
769
770 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
771
772 for ( ; *dst != '\0' && sz; dst++, sz--)
773 /* Jump to end. */ ;
774
775 assert(sz > 2);
776
777 /* We can't start with a number (bah). */
778
779 *dst++ = 'x';
780 *dst = '\0';
781 sz--;
782
783 for ( ; *src != '\0' && sz > 1; src++) {
784 ssz = snprintf(dst, (size_t)sz, "%.2x", *src);
785 sz -= ssz;
786 dst += ssz;
787 }
788 }