]> git.cameronkatri.com Git - mandoc.git/blob - html.c
Fix missing support for \N'n' when calculating string widths in -Tascii
[mandoc.git] / html.c
1 /* $Id: html.c,v 1.139 2011/05/15 15:47:46 kristaps Exp $ */
2 /*
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2011 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 #ifdef HAVE_CONFIG_H
19 #include "config.h"
20 #endif
21
22 #include <sys/types.h>
23
24 #include <assert.h>
25 #include <ctype.h>
26 #include <stdarg.h>
27 #include <stdio.h>
28 #include <stdint.h>
29 #include <stdlib.h>
30 #include <string.h>
31 #include <unistd.h>
32
33 #include "mandoc.h"
34 #include "libmandoc.h"
35 #include "out.h"
36 #include "html.h"
37 #include "main.h"
38
39 struct htmldata {
40 const char *name;
41 int flags;
42 #define HTML_CLRLINE (1 << 0)
43 #define HTML_NOSTACK (1 << 1)
44 #define HTML_AUTOCLOSE (1 << 2) /* Tag has auto-closure. */
45 };
46
47 static const struct htmldata htmltags[TAG_MAX] = {
48 {"html", HTML_CLRLINE}, /* TAG_HTML */
49 {"head", HTML_CLRLINE}, /* TAG_HEAD */
50 {"body", HTML_CLRLINE}, /* TAG_BODY */
51 {"meta", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_META */
52 {"title", HTML_CLRLINE}, /* TAG_TITLE */
53 {"div", HTML_CLRLINE}, /* TAG_DIV */
54 {"h1", 0}, /* TAG_H1 */
55 {"h2", 0}, /* TAG_H2 */
56 {"span", 0}, /* TAG_SPAN */
57 {"link", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_LINK */
58 {"br", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_BR */
59 {"a", 0}, /* TAG_A */
60 {"table", HTML_CLRLINE}, /* TAG_TABLE */
61 {"tbody", HTML_CLRLINE}, /* TAG_TBODY */
62 {"col", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_COL */
63 {"tr", HTML_CLRLINE}, /* TAG_TR */
64 {"td", HTML_CLRLINE}, /* TAG_TD */
65 {"li", HTML_CLRLINE}, /* TAG_LI */
66 {"ul", HTML_CLRLINE}, /* TAG_UL */
67 {"ol", HTML_CLRLINE}, /* TAG_OL */
68 {"dl", HTML_CLRLINE}, /* TAG_DL */
69 {"dt", HTML_CLRLINE}, /* TAG_DT */
70 {"dd", HTML_CLRLINE}, /* TAG_DD */
71 {"blockquote", HTML_CLRLINE}, /* TAG_BLOCKQUOTE */
72 {"p", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_P */
73 {"pre", HTML_CLRLINE }, /* TAG_PRE */
74 {"b", 0 }, /* TAG_B */
75 {"i", 0 }, /* TAG_I */
76 {"code", 0 }, /* TAG_CODE */
77 {"small", 0 }, /* TAG_SMALL */
78 };
79
80 static const char *const htmlattrs[ATTR_MAX] = {
81 "http-equiv", /* ATTR_HTTPEQUIV */
82 "content", /* ATTR_CONTENT */
83 "name", /* ATTR_NAME */
84 "rel", /* ATTR_REL */
85 "href", /* ATTR_HREF */
86 "type", /* ATTR_TYPE */
87 "media", /* ATTR_MEDIA */
88 "class", /* ATTR_CLASS */
89 "style", /* ATTR_STYLE */
90 "width", /* ATTR_WIDTH */
91 "id", /* ATTR_ID */
92 "summary", /* ATTR_SUMMARY */
93 "align", /* ATTR_ALIGN */
94 "colspan", /* ATTR_COLSPAN */
95 };
96
97 static void print_num(struct html *, const char *, size_t);
98 static void print_spec(struct html *, const char *, size_t);
99 static void print_res(struct html *, const char *, size_t);
100 static void print_ctag(struct html *, enum htmltag);
101 static void print_doctype(struct html *);
102 static void print_xmltype(struct html *);
103 static int print_encode(struct html *, const char *, int);
104 static void print_metaf(struct html *, enum mandoc_esc);
105 static void print_attr(struct html *,
106 const char *, const char *);
107 static void *ml_alloc(char *, enum htmltype);
108
109
110 static void *
111 ml_alloc(char *outopts, enum htmltype type)
112 {
113 struct html *h;
114 const char *toks[4];
115 char *v;
116
117 toks[0] = "style";
118 toks[1] = "man";
119 toks[2] = "includes";
120 toks[3] = NULL;
121
122 h = mandoc_calloc(1, sizeof(struct html));
123
124 h->type = type;
125 h->tags.head = NULL;
126 h->symtab = mchars_alloc();
127
128 while (outopts && *outopts)
129 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
130 case (0):
131 h->style = v;
132 break;
133 case (1):
134 h->base_man = v;
135 break;
136 case (2):
137 h->base_includes = v;
138 break;
139 default:
140 break;
141 }
142
143 return(h);
144 }
145
146 void *
147 html_alloc(char *outopts)
148 {
149
150 return(ml_alloc(outopts, HTML_HTML_4_01_STRICT));
151 }
152
153
154 void *
155 xhtml_alloc(char *outopts)
156 {
157
158 return(ml_alloc(outopts, HTML_XHTML_1_0_STRICT));
159 }
160
161
162 void
163 html_free(void *p)
164 {
165 struct tag *tag;
166 struct html *h;
167
168 h = (struct html *)p;
169
170 while ((tag = h->tags.head) != NULL) {
171 h->tags.head = tag->next;
172 free(tag);
173 }
174
175 if (h->symtab)
176 mchars_free(h->symtab);
177
178 free(h);
179 }
180
181
182 void
183 print_gen_head(struct html *h)
184 {
185 struct htmlpair tag[4];
186
187 tag[0].key = ATTR_HTTPEQUIV;
188 tag[0].val = "Content-Type";
189 tag[1].key = ATTR_CONTENT;
190 tag[1].val = "text/html; charset=utf-8";
191 print_otag(h, TAG_META, 2, tag);
192
193 tag[0].key = ATTR_NAME;
194 tag[0].val = "resource-type";
195 tag[1].key = ATTR_CONTENT;
196 tag[1].val = "document";
197 print_otag(h, TAG_META, 2, tag);
198
199 if (h->style) {
200 tag[0].key = ATTR_REL;
201 tag[0].val = "stylesheet";
202 tag[1].key = ATTR_HREF;
203 tag[1].val = h->style;
204 tag[2].key = ATTR_TYPE;
205 tag[2].val = "text/css";
206 tag[3].key = ATTR_MEDIA;
207 tag[3].val = "all";
208 print_otag(h, TAG_LINK, 4, tag);
209 }
210 }
211
212 /* ARGSUSED */
213 static void
214 print_num(struct html *h, const char *p, size_t len)
215 {
216 char c;
217
218 if ('\0' != (c = mchars_num2char(p, len)))
219 putchar((int)c);
220 }
221
222 static void
223 print_spec(struct html *h, const char *p, size_t len)
224 {
225 int cp;
226 const char *rhs;
227 size_t sz;
228
229 if ((cp = mchars_spec2cp(h->symtab, p, len)) > 0) {
230 printf("&#%d;", cp);
231 return;
232 } else if (-1 == cp && 1 == len) {
233 fwrite(p, 1, len, stdout);
234 return;
235 } else if (-1 == cp)
236 return;
237
238 if (NULL != (rhs = mchars_spec2str(h->symtab, p, len, &sz)))
239 fwrite(rhs, 1, sz, stdout);
240 }
241
242
243 static void
244 print_res(struct html *h, const char *p, size_t len)
245 {
246 int cp;
247 const char *rhs;
248 size_t sz;
249
250 if ((cp = mchars_res2cp(h->symtab, p, len)) > 0) {
251 printf("&#%d;", cp);
252 return;
253 } else if (-1 == cp)
254 return;
255
256 if (NULL != (rhs = mchars_res2str(h->symtab, p, len, &sz)))
257 fwrite(rhs, 1, sz, stdout);
258 }
259
260
261 static void
262 print_metaf(struct html *h, enum mandoc_esc deco)
263 {
264 enum htmlfont font;
265
266 switch (deco) {
267 case (ESCAPE_FONTPREV):
268 font = h->metal;
269 break;
270 case (ESCAPE_FONTITALIC):
271 font = HTMLFONT_ITALIC;
272 break;
273 case (ESCAPE_FONTBOLD):
274 font = HTMLFONT_BOLD;
275 break;
276 case (ESCAPE_FONTROMAN):
277 font = HTMLFONT_NONE;
278 break;
279 default:
280 abort();
281 /* NOTREACHED */
282 }
283
284 if (h->metaf) {
285 print_tagq(h, h->metaf);
286 h->metaf = NULL;
287 }
288
289 h->metal = h->metac;
290 h->metac = font;
291
292 if (HTMLFONT_NONE != font)
293 h->metaf = HTMLFONT_BOLD == font ?
294 print_otag(h, TAG_B, 0, NULL) :
295 print_otag(h, TAG_I, 0, NULL);
296 }
297
298 int
299 html_strlen(const char *cp)
300 {
301 int ssz, sz;
302 const char *seq, *p;
303
304 /*
305 * Account for escaped sequences within string length
306 * calculations. This follows the logic in term_strlen() as we
307 * must calculate the width of produced strings.
308 * Assume that characters are always width of "1". This is
309 * hacky, but it gets the job done for approximation of widths.
310 */
311
312 sz = 0;
313 while (NULL != (p = strchr(cp, '\\'))) {
314 sz += (int)(p - cp);
315 ++cp;
316 switch (mandoc_escape(&cp, &seq, &ssz)) {
317 case (ESCAPE_ERROR):
318 return(sz);
319 case (ESCAPE_NUMBERED):
320 /* FALLTHROUGH */
321 case (ESCAPE_PREDEF):
322 /* FALLTHROUGH */
323 case (ESCAPE_SPECIAL):
324 sz++;
325 break;
326 default:
327 break;
328 }
329 }
330
331 assert(sz >= 0);
332 return(sz + strlen(cp));
333 }
334
335 static int
336 print_encode(struct html *h, const char *p, int norecurse)
337 {
338 size_t sz;
339 int len, nospace;
340 const char *seq;
341 enum mandoc_esc esc;
342 static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' };
343
344 nospace = 0;
345
346 while ('\0' != *p) {
347 sz = strcspn(p, rejs);
348
349 fwrite(p, 1, sz, stdout);
350 p += (int)sz;
351
352 if ('\0' == *p)
353 break;
354
355 switch (*p++) {
356 case ('<'):
357 printf("&lt;");
358 continue;
359 case ('>'):
360 printf("&gt;");
361 continue;
362 case ('&'):
363 printf("&amp;");
364 continue;
365 case (ASCII_HYPH):
366 putchar('-');
367 continue;
368 default:
369 break;
370 }
371
372 esc = mandoc_escape(&p, &seq, &len);
373 if (ESCAPE_ERROR == esc)
374 break;
375
376 switch (esc) {
377 case (ESCAPE_NUMBERED):
378 print_num(h, seq, len);
379 break;
380 case (ESCAPE_PREDEF):
381 print_res(h, seq, len);
382 break;
383 case (ESCAPE_SPECIAL):
384 print_spec(h, seq, len);
385 break;
386 case (ESCAPE_FONTPREV):
387 /* FALLTHROUGH */
388 case (ESCAPE_FONTBOLD):
389 /* FALLTHROUGH */
390 case (ESCAPE_FONTITALIC):
391 /* FALLTHROUGH */
392 case (ESCAPE_FONTROMAN):
393 if (norecurse)
394 break;
395 print_metaf(h, esc);
396 break;
397 case (ESCAPE_NOSPACE):
398 if ('\0' == *p)
399 nospace = 1;
400 break;
401 default:
402 break;
403 }
404 }
405
406 return(nospace);
407 }
408
409
410 static void
411 print_attr(struct html *h, const char *key, const char *val)
412 {
413 printf(" %s=\"", key);
414 (void)print_encode(h, val, 1);
415 putchar('\"');
416 }
417
418
419 struct tag *
420 print_otag(struct html *h, enum htmltag tag,
421 int sz, const struct htmlpair *p)
422 {
423 int i;
424 struct tag *t;
425
426 /* Push this tags onto the stack of open scopes. */
427
428 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
429 t = mandoc_malloc(sizeof(struct tag));
430 t->tag = tag;
431 t->next = h->tags.head;
432 h->tags.head = t;
433 } else
434 t = NULL;
435
436 if ( ! (HTML_NOSPACE & h->flags))
437 if ( ! (HTML_CLRLINE & htmltags[tag].flags)) {
438 /* Manage keeps! */
439 if ( ! (HTML_KEEP & h->flags)) {
440 if (HTML_PREKEEP & h->flags)
441 h->flags |= HTML_KEEP;
442 putchar(' ');
443 } else
444 printf("&#160;");
445 }
446
447 if ( ! (h->flags & HTML_NONOSPACE))
448 h->flags &= ~HTML_NOSPACE;
449 else
450 h->flags |= HTML_NOSPACE;
451
452 /* Print out the tag name and attributes. */
453
454 printf("<%s", htmltags[tag].name);
455 for (i = 0; i < sz; i++)
456 print_attr(h, htmlattrs[p[i].key], p[i].val);
457
458 /* Add non-overridable attributes. */
459
460 if (TAG_HTML == tag && HTML_XHTML_1_0_STRICT == h->type) {
461 print_attr(h, "xmlns", "http://www.w3.org/1999/xhtml");
462 print_attr(h, "xml:lang", "en");
463 print_attr(h, "lang", "en");
464 }
465
466 /* Accommodate for XML "well-formed" singleton escaping. */
467
468 if (HTML_AUTOCLOSE & htmltags[tag].flags)
469 switch (h->type) {
470 case (HTML_XHTML_1_0_STRICT):
471 putchar('/');
472 break;
473 default:
474 break;
475 }
476
477 putchar('>');
478
479 h->flags |= HTML_NOSPACE;
480
481 if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags)
482 putchar('\n');
483
484 return(t);
485 }
486
487
488 static void
489 print_ctag(struct html *h, enum htmltag tag)
490 {
491
492 printf("</%s>", htmltags[tag].name);
493 if (HTML_CLRLINE & htmltags[tag].flags) {
494 h->flags |= HTML_NOSPACE;
495 putchar('\n');
496 }
497 }
498
499
500 void
501 print_gen_decls(struct html *h)
502 {
503
504 print_xmltype(h);
505 print_doctype(h);
506 }
507
508
509 static void
510 print_xmltype(struct html *h)
511 {
512
513 if (HTML_XHTML_1_0_STRICT == h->type)
514 puts("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
515 }
516
517
518 static void
519 print_doctype(struct html *h)
520 {
521 const char *doctype;
522 const char *dtd;
523 const char *name;
524
525 switch (h->type) {
526 case (HTML_HTML_4_01_STRICT):
527 name = "HTML";
528 doctype = "-//W3C//DTD HTML 4.01//EN";
529 dtd = "http://www.w3.org/TR/html4/strict.dtd";
530 break;
531 default:
532 name = "html";
533 doctype = "-//W3C//DTD XHTML 1.0 Strict//EN";
534 dtd = "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
535 break;
536 }
537
538 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
539 name, doctype, dtd);
540 }
541
542 void
543 print_text(struct html *h, const char *word)
544 {
545
546 if ( ! (HTML_NOSPACE & h->flags)) {
547 /* Manage keeps! */
548 if ( ! (HTML_KEEP & h->flags)) {
549 if (HTML_PREKEEP & h->flags)
550 h->flags |= HTML_KEEP;
551 putchar(' ');
552 } else
553 printf("&#160;");
554 }
555
556 assert(NULL == h->metaf);
557 if (HTMLFONT_NONE != h->metac)
558 h->metaf = HTMLFONT_BOLD == h->metac ?
559 print_otag(h, TAG_B, 0, NULL) :
560 print_otag(h, TAG_I, 0, NULL);
561
562 assert(word);
563 if ( ! print_encode(h, word, 0))
564 if ( ! (h->flags & HTML_NONOSPACE))
565 h->flags &= ~HTML_NOSPACE;
566
567 if (h->metaf) {
568 print_tagq(h, h->metaf);
569 h->metaf = NULL;
570 }
571
572 h->flags &= ~HTML_IGNDELIM;
573 }
574
575
576 void
577 print_tagq(struct html *h, const struct tag *until)
578 {
579 struct tag *tag;
580
581 while ((tag = h->tags.head) != NULL) {
582 /*
583 * Remember to close out and nullify the current
584 * meta-font and table, if applicable.
585 */
586 if (tag == h->metaf)
587 h->metaf = NULL;
588 if (tag == h->tblt)
589 h->tblt = NULL;
590 print_ctag(h, tag->tag);
591 h->tags.head = tag->next;
592 free(tag);
593 if (until && tag == until)
594 return;
595 }
596 }
597
598
599 void
600 print_stagq(struct html *h, const struct tag *suntil)
601 {
602 struct tag *tag;
603
604 while ((tag = h->tags.head) != NULL) {
605 if (suntil && tag == suntil)
606 return;
607 /*
608 * Remember to close out and nullify the current
609 * meta-font and table, if applicable.
610 */
611 if (tag == h->metaf)
612 h->metaf = NULL;
613 if (tag == h->tblt)
614 h->tblt = NULL;
615 print_ctag(h, tag->tag);
616 h->tags.head = tag->next;
617 free(tag);
618 }
619 }
620
621
622 void
623 bufinit(struct html *h)
624 {
625
626 h->buf[0] = '\0';
627 h->buflen = 0;
628 }
629
630
631 void
632 bufcat_style(struct html *h, const char *key, const char *val)
633 {
634
635 bufcat(h, key);
636 bufncat(h, ":", 1);
637 bufcat(h, val);
638 bufncat(h, ";", 1);
639 }
640
641
642 void
643 bufcat(struct html *h, const char *p)
644 {
645
646 bufncat(h, p, strlen(p));
647 }
648
649
650 void
651 buffmt(struct html *h, const char *fmt, ...)
652 {
653 va_list ap;
654
655 va_start(ap, fmt);
656 (void)vsnprintf(h->buf + (int)h->buflen,
657 BUFSIZ - h->buflen - 1, fmt, ap);
658 va_end(ap);
659 h->buflen = strlen(h->buf);
660 }
661
662
663 void
664 bufncat(struct html *h, const char *p, size_t sz)
665 {
666
667 if (h->buflen + sz > BUFSIZ - 1)
668 sz = BUFSIZ - 1 - h->buflen;
669
670 (void)strncat(h->buf, p, sz);
671 h->buflen += sz;
672 }
673
674
675 void
676 buffmt_includes(struct html *h, const char *name)
677 {
678 const char *p, *pp;
679
680 pp = h->base_includes;
681
682 while (NULL != (p = strchr(pp, '%'))) {
683 bufncat(h, pp, (size_t)(p - pp));
684 switch (*(p + 1)) {
685 case('I'):
686 bufcat(h, name);
687 break;
688 default:
689 bufncat(h, p, 2);
690 break;
691 }
692 pp = p + 2;
693 }
694 if (pp)
695 bufcat(h, pp);
696 }
697
698
699 void
700 buffmt_man(struct html *h,
701 const char *name, const char *sec)
702 {
703 const char *p, *pp;
704
705 pp = h->base_man;
706
707 /* LINTED */
708 while (NULL != (p = strchr(pp, '%'))) {
709 bufncat(h, pp, (size_t)(p - pp));
710 switch (*(p + 1)) {
711 case('S'):
712 bufcat(h, sec ? sec : "1");
713 break;
714 case('N'):
715 buffmt(h, name);
716 break;
717 default:
718 bufncat(h, p, 2);
719 break;
720 }
721 pp = p + 2;
722 }
723 if (pp)
724 bufcat(h, pp);
725 }
726
727
728 void
729 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
730 {
731 double v;
732 const char *u;
733
734 v = su->scale;
735
736 switch (su->unit) {
737 case (SCALE_CM):
738 u = "cm";
739 break;
740 case (SCALE_IN):
741 u = "in";
742 break;
743 case (SCALE_PC):
744 u = "pc";
745 break;
746 case (SCALE_PT):
747 u = "pt";
748 break;
749 case (SCALE_EM):
750 u = "em";
751 break;
752 case (SCALE_MM):
753 if (0 == (v /= 100))
754 v = 1;
755 u = "em";
756 break;
757 case (SCALE_EN):
758 u = "ex";
759 break;
760 case (SCALE_BU):
761 u = "ex";
762 break;
763 case (SCALE_VS):
764 u = "em";
765 break;
766 default:
767 u = "ex";
768 break;
769 }
770
771 /*
772 * XXX: the CSS spec isn't clear as to which types accept
773 * integer or real numbers, so we just make them all decimals.
774 */
775 buffmt(h, "%s: %.2f%s;", p, v, u);
776 }
777
778
779 void
780 html_idcat(char *dst, const char *src, int sz)
781 {
782 int ssz;
783
784 assert(sz > 2);
785
786 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
787
788 /* We can't start with a number (bah). */
789
790 if ('#' == *dst) {
791 dst++;
792 sz--;
793 }
794 if ('\0' == *dst) {
795 *dst++ = 'x';
796 *dst = '\0';
797 sz--;
798 }
799
800 for ( ; *dst != '\0' && sz; dst++, sz--)
801 /* Jump to end. */ ;
802
803 for ( ; *src != '\0' && sz > 1; src++) {
804 ssz = snprintf(dst, (size_t)sz, "%.2x", *src);
805 sz -= ssz;
806 dst += ssz;
807 }
808 }