]> git.cameronkatri.com Git - mandoc.git/blob - html.c
When in a <PRE>, don't print out the <BR> before lines that have leading
[mandoc.git] / html.c
1 /* $Id: html.c,v 1.125 2011/01/13 14:30:13 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 {"p", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_P */
72 {"pre", HTML_CLRLINE }, /* TAG_PRE */
73 {"b", 0 }, /* TAG_B */
74 {"i", 0 }, /* TAG_I */
75 {"code", 0 }, /* TAG_CODE */
76 {"small", 0 }, /* TAG_SMALL */
77 };
78
79 static const char *const htmlattrs[ATTR_MAX] = {
80 "http-equiv", /* ATTR_HTTPEQUIV */
81 "content", /* ATTR_CONTENT */
82 "name", /* ATTR_NAME */
83 "rel", /* ATTR_REL */
84 "href", /* ATTR_HREF */
85 "type", /* ATTR_TYPE */
86 "media", /* ATTR_MEDIA */
87 "class", /* ATTR_CLASS */
88 "style", /* ATTR_STYLE */
89 "width", /* ATTR_WIDTH */
90 "id", /* ATTR_ID */
91 "summary", /* ATTR_SUMMARY */
92 "align", /* ATTR_ALIGN */
93 "colspan", /* ATTR_COLSPAN */
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 static void
256 print_metaf(struct html *h, enum roffdeco deco)
257 {
258 enum htmlfont font;
259
260 switch (deco) {
261 case (DECO_PREVIOUS):
262 font = h->metal;
263 break;
264 case (DECO_ITALIC):
265 font = HTMLFONT_ITALIC;
266 break;
267 case (DECO_BOLD):
268 font = HTMLFONT_BOLD;
269 break;
270 case (DECO_ROMAN):
271 font = HTMLFONT_NONE;
272 break;
273 default:
274 abort();
275 /* NOTREACHED */
276 }
277
278 if (h->metaf) {
279 print_tagq(h, h->metaf);
280 h->metaf = NULL;
281 }
282
283 h->metal = h->metac;
284 h->metac = font;
285
286 if (HTMLFONT_NONE != font)
287 h->metaf = HTMLFONT_BOLD == font ?
288 print_otag(h, TAG_B, 0, NULL) :
289 print_otag(h, TAG_I, 0, NULL);
290 }
291
292
293 static int
294 print_encode(struct html *h, const char *p, int norecurse)
295 {
296 size_t sz;
297 int len, nospace;
298 const char *seq;
299 enum roffdeco deco;
300 static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' };
301
302 nospace = 0;
303
304 for (; *p; p++) {
305 sz = strcspn(p, rejs);
306
307 fwrite(p, 1, sz, stdout);
308 p += /* LINTED */
309 sz;
310
311 if ('<' == *p) {
312 printf("&lt;");
313 continue;
314 } else if ('>' == *p) {
315 printf("&gt;");
316 continue;
317 } else if ('&' == *p) {
318 printf("&amp;");
319 continue;
320 } else if (ASCII_HYPH == *p) {
321 /*
322 * Note: "soft hyphens" aren't graphically
323 * displayed when not breaking the text; we want
324 * them to be displayed.
325 */
326 /*printf("&#173;");*/
327 putchar('-');
328 continue;
329 } else if ('\0' == *p)
330 break;
331
332 seq = ++p;
333 len = a2roffdeco(&deco, &seq, &sz);
334
335 switch (deco) {
336 case (DECO_RESERVED):
337 print_res(h, seq, sz);
338 break;
339 case (DECO_SSPECIAL):
340 /* FALLTHROUGH */
341 case (DECO_SPECIAL):
342 print_spec(h, deco, seq, sz);
343 break;
344 case (DECO_PREVIOUS):
345 /* FALLTHROUGH */
346 case (DECO_BOLD):
347 /* FALLTHROUGH */
348 case (DECO_ITALIC):
349 /* FALLTHROUGH */
350 case (DECO_ROMAN):
351 if (norecurse)
352 break;
353 print_metaf(h, deco);
354 break;
355 default:
356 break;
357 }
358
359 p += len - 1;
360
361 if (DECO_NOSPACE == deco && '\0' == *(p + 1))
362 nospace = 1;
363 }
364
365 return(nospace);
366 }
367
368
369 static void
370 print_attr(struct html *h, const char *key, const char *val)
371 {
372 printf(" %s=\"", key);
373 (void)print_encode(h, val, 1);
374 putchar('\"');
375 }
376
377
378 struct tag *
379 print_otag(struct html *h, enum htmltag tag,
380 int sz, const struct htmlpair *p)
381 {
382 int i;
383 struct tag *t;
384
385 /* Push this tags onto the stack of open scopes. */
386
387 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
388 t = malloc(sizeof(struct tag));
389 if (NULL == t) {
390 perror(NULL);
391 exit((int)MANDOCLEVEL_SYSERR);
392 }
393 t->tag = tag;
394 t->next = h->tags.head;
395 h->tags.head = t;
396 } else
397 t = NULL;
398
399 if ( ! (HTML_NOSPACE & h->flags))
400 if ( ! (HTML_CLRLINE & htmltags[tag].flags)) {
401 /* Manage keeps! */
402 if ( ! (HTML_KEEP & h->flags)) {
403 if (HTML_PREKEEP & h->flags)
404 h->flags |= HTML_KEEP;
405 putchar(' ');
406 } else
407 printf("&#160;");
408 }
409
410 if ( ! (h->flags & HTML_NONOSPACE))
411 h->flags &= ~HTML_NOSPACE;
412 else
413 h->flags |= HTML_NOSPACE;
414
415 /* Print out the tag name and attributes. */
416
417 printf("<%s", htmltags[tag].name);
418 for (i = 0; i < sz; i++)
419 print_attr(h, htmlattrs[p[i].key], p[i].val);
420
421 /* Add non-overridable attributes. */
422
423 if (TAG_HTML == tag && HTML_XHTML_1_0_STRICT == h->type) {
424 print_attr(h, "xmlns", "http://www.w3.org/1999/xhtml");
425 print_attr(h, "xml:lang", "en");
426 print_attr(h, "lang", "en");
427 }
428
429 /* Accomodate for XML "well-formed" singleton escaping. */
430
431 if (HTML_AUTOCLOSE & htmltags[tag].flags)
432 switch (h->type) {
433 case (HTML_XHTML_1_0_STRICT):
434 putchar('/');
435 break;
436 default:
437 break;
438 }
439
440 putchar('>');
441
442 h->flags |= HTML_NOSPACE;
443
444 if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags)
445 putchar('\n');
446
447 return(t);
448 }
449
450
451 static void
452 print_ctag(struct html *h, enum htmltag tag)
453 {
454
455 printf("</%s>", htmltags[tag].name);
456 if (HTML_CLRLINE & htmltags[tag].flags) {
457 h->flags |= HTML_NOSPACE;
458 putchar('\n');
459 }
460 }
461
462
463 void
464 print_gen_decls(struct html *h)
465 {
466
467 print_xmltype(h);
468 print_doctype(h);
469 }
470
471
472 static void
473 print_xmltype(struct html *h)
474 {
475
476 if (HTML_XHTML_1_0_STRICT == h->type)
477 puts("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
478 }
479
480
481 static void
482 print_doctype(struct html *h)
483 {
484 const char *doctype;
485 const char *dtd;
486 const char *name;
487
488 switch (h->type) {
489 case (HTML_HTML_4_01_STRICT):
490 name = "HTML";
491 doctype = "-//W3C//DTD HTML 4.01//EN";
492 dtd = "http://www.w3.org/TR/html4/strict.dtd";
493 break;
494 default:
495 name = "html";
496 doctype = "-//W3C//DTD XHTML 1.0 Strict//EN";
497 dtd = "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
498 break;
499 }
500
501 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
502 name, doctype, dtd);
503 }
504
505
506 void
507 print_text(struct html *h, const char *word)
508 {
509
510 if (word[0] && '\0' == word[1])
511 switch (word[0]) {
512 case('.'):
513 /* FALLTHROUGH */
514 case(','):
515 /* FALLTHROUGH */
516 case(';'):
517 /* FALLTHROUGH */
518 case(':'):
519 /* FALLTHROUGH */
520 case('?'):
521 /* FALLTHROUGH */
522 case('!'):
523 /* FALLTHROUGH */
524 case(')'):
525 /* FALLTHROUGH */
526 case(']'):
527 if ( ! (HTML_IGNDELIM & h->flags))
528 h->flags |= HTML_NOSPACE;
529 break;
530 default:
531 break;
532 }
533
534 if ( ! (HTML_NOSPACE & h->flags)) {
535 /* Manage keeps! */
536 if ( ! (HTML_KEEP & h->flags)) {
537 if (HTML_PREKEEP & h->flags)
538 h->flags |= HTML_KEEP;
539 putchar(' ');
540 } else
541 printf("&#160;");
542 }
543
544 assert(NULL == h->metaf);
545 if (HTMLFONT_NONE != h->metac)
546 h->metaf = HTMLFONT_BOLD == h->metac ?
547 print_otag(h, TAG_B, 0, NULL) :
548 print_otag(h, TAG_I, 0, NULL);
549
550 assert(word);
551 if ( ! print_encode(h, word, 0))
552 if ( ! (h->flags & HTML_NONOSPACE))
553 h->flags &= ~HTML_NOSPACE;
554
555 if (h->metaf) {
556 print_tagq(h, h->metaf);
557 h->metaf = NULL;
558 }
559
560 h->flags &= ~HTML_IGNDELIM;
561
562 /*
563 * Note that we don't process the pipe: the parser sees it as
564 * punctuation, but we don't in terms of typography.
565 */
566 if (word[0] && '\0' == word[1])
567 switch (word[0]) {
568 case('('):
569 /* FALLTHROUGH */
570 case('['):
571 h->flags |= HTML_NOSPACE;
572 break;
573 default:
574 break;
575 }
576 }
577
578
579 void
580 print_tagq(struct html *h, const struct tag *until)
581 {
582 struct tag *tag;
583
584 while ((tag = h->tags.head) != NULL) {
585 /*
586 * Remember to close out and nullify the current
587 * meta-font and table, if applicable.
588 */
589 if (tag == h->metaf)
590 h->metaf = NULL;
591 if (tag == h->tblt)
592 h->tblt = NULL;
593 print_ctag(h, tag->tag);
594 h->tags.head = tag->next;
595 free(tag);
596 if (until && tag == until)
597 return;
598 }
599 }
600
601
602 void
603 print_stagq(struct html *h, const struct tag *suntil)
604 {
605 struct tag *tag;
606
607 while ((tag = h->tags.head) != NULL) {
608 if (suntil && tag == suntil)
609 return;
610 /*
611 * Remember to close out and nullify the current
612 * meta-font and table, if applicable.
613 */
614 if (tag == h->metaf)
615 h->metaf = NULL;
616 if (tag == h->tblt)
617 h->tblt = NULL;
618 print_ctag(h, tag->tag);
619 h->tags.head = tag->next;
620 free(tag);
621 }
622 }
623
624
625 void
626 bufinit(struct html *h)
627 {
628
629 h->buf[0] = '\0';
630 h->buflen = 0;
631 }
632
633
634 void
635 bufcat_style(struct html *h, const char *key, const char *val)
636 {
637
638 bufcat(h, key);
639 bufncat(h, ":", 1);
640 bufcat(h, val);
641 bufncat(h, ";", 1);
642 }
643
644
645 void
646 bufcat(struct html *h, const char *p)
647 {
648
649 bufncat(h, p, strlen(p));
650 }
651
652
653 void
654 buffmt(struct html *h, const char *fmt, ...)
655 {
656 va_list ap;
657
658 va_start(ap, fmt);
659 (void)vsnprintf(h->buf + (int)h->buflen,
660 BUFSIZ - h->buflen - 1, fmt, ap);
661 va_end(ap);
662 h->buflen = strlen(h->buf);
663 }
664
665
666 void
667 bufncat(struct html *h, const char *p, size_t sz)
668 {
669
670 if (h->buflen + sz > BUFSIZ - 1)
671 sz = BUFSIZ - 1 - h->buflen;
672
673 (void)strncat(h->buf, p, sz);
674 h->buflen += sz;
675 }
676
677
678 void
679 buffmt_includes(struct html *h, const char *name)
680 {
681 const char *p, *pp;
682
683 pp = h->base_includes;
684
685 while (NULL != (p = strchr(pp, '%'))) {
686 bufncat(h, pp, (size_t)(p - pp));
687 switch (*(p + 1)) {
688 case('I'):
689 bufcat(h, name);
690 break;
691 default:
692 bufncat(h, p, 2);
693 break;
694 }
695 pp = p + 2;
696 }
697 if (pp)
698 bufcat(h, pp);
699 }
700
701
702 void
703 buffmt_man(struct html *h,
704 const char *name, const char *sec)
705 {
706 const char *p, *pp;
707
708 pp = h->base_man;
709
710 /* LINTED */
711 while (NULL != (p = strchr(pp, '%'))) {
712 bufncat(h, pp, (size_t)(p - pp));
713 switch (*(p + 1)) {
714 case('S'):
715 bufcat(h, sec ? sec : "1");
716 break;
717 case('N'):
718 buffmt(h, name);
719 break;
720 default:
721 bufncat(h, p, 2);
722 break;
723 }
724 pp = p + 2;
725 }
726 if (pp)
727 bufcat(h, pp);
728 }
729
730
731 void
732 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
733 {
734 double v;
735 const char *u;
736
737 v = su->scale;
738
739 switch (su->unit) {
740 case (SCALE_CM):
741 u = "cm";
742 break;
743 case (SCALE_IN):
744 u = "in";
745 break;
746 case (SCALE_PC):
747 u = "pc";
748 break;
749 case (SCALE_PT):
750 u = "pt";
751 break;
752 case (SCALE_EM):
753 u = "em";
754 break;
755 case (SCALE_MM):
756 if (0 == (v /= 100))
757 v = 1;
758 u = "em";
759 break;
760 case (SCALE_EN):
761 u = "ex";
762 break;
763 case (SCALE_BU):
764 u = "ex";
765 break;
766 case (SCALE_VS):
767 u = "em";
768 break;
769 default:
770 u = "ex";
771 break;
772 }
773
774 /*
775 * XXX: the CSS spec isn't clear as to which types accept
776 * integer or real numbers, so we just make them all decimals.
777 */
778 buffmt(h, "%s: %.2f%s;", p, v, u);
779 }
780
781
782 void
783 html_idcat(char *dst, const char *src, int sz)
784 {
785 int ssz;
786
787 assert(sz > 2);
788
789 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
790
791 /* We can't start with a number (bah). */
792
793 if ('#' == *dst) {
794 dst++;
795 sz--;
796 }
797 if ('\0' == *dst) {
798 *dst++ = 'x';
799 *dst = '\0';
800 sz--;
801 }
802
803 for ( ; *dst != '\0' && sz; dst++, sz--)
804 /* Jump to end. */ ;
805
806 for ( ; *src != '\0' && sz > 1; src++) {
807 ssz = snprintf(dst, (size_t)sz, "%.2x", *src);
808 sz -= ssz;
809 dst += ssz;
810 }
811 }