]> git.cameronkatri.com Git - mandoc.git/blob - html.c
Make mchars_num2char() return a char like it says.
[mandoc.git] / html.c
1 /* $Id: html.c,v 1.137 2011/04/30 22:24:31 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 "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 {"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_num(struct html *, const char *, size_t);
97 static void print_spec(struct html *, 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 mandoc_esc);
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 = mandoc_calloc(1, sizeof(struct html));
122
123 h->type = type;
124 h->tags.head = NULL;
125 h->symtab = mchars_alloc();
126
127 while (outopts && *outopts)
128 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
129 case (0):
130 h->style = v;
131 break;
132 case (1):
133 h->base_man = v;
134 break;
135 case (2):
136 h->base_includes = v;
137 break;
138 default:
139 break;
140 }
141
142 return(h);
143 }
144
145 void *
146 html_alloc(char *outopts)
147 {
148
149 return(ml_alloc(outopts, HTML_HTML_4_01_STRICT));
150 }
151
152
153 void *
154 xhtml_alloc(char *outopts)
155 {
156
157 return(ml_alloc(outopts, HTML_XHTML_1_0_STRICT));
158 }
159
160
161 void
162 html_free(void *p)
163 {
164 struct tag *tag;
165 struct html *h;
166
167 h = (struct html *)p;
168
169 while ((tag = h->tags.head) != NULL) {
170 h->tags.head = tag->next;
171 free(tag);
172 }
173
174 if (h->symtab)
175 mchars_free(h->symtab);
176
177 free(h);
178 }
179
180
181 void
182 print_gen_head(struct html *h)
183 {
184 struct htmlpair tag[4];
185
186 tag[0].key = ATTR_HTTPEQUIV;
187 tag[0].val = "Content-Type";
188 tag[1].key = ATTR_CONTENT;
189 tag[1].val = "text/html; charset=utf-8";
190 print_otag(h, TAG_META, 2, tag);
191
192 tag[0].key = ATTR_NAME;
193 tag[0].val = "resource-type";
194 tag[1].key = ATTR_CONTENT;
195 tag[1].val = "document";
196 print_otag(h, TAG_META, 2, tag);
197
198 if (h->style) {
199 tag[0].key = ATTR_REL;
200 tag[0].val = "stylesheet";
201 tag[1].key = ATTR_HREF;
202 tag[1].val = h->style;
203 tag[2].key = ATTR_TYPE;
204 tag[2].val = "text/css";
205 tag[3].key = ATTR_MEDIA;
206 tag[3].val = "all";
207 print_otag(h, TAG_LINK, 4, tag);
208 }
209 }
210
211 /* ARGSUSED */
212 static void
213 print_num(struct html *h, const char *p, size_t len)
214 {
215 char c;
216
217 if ('\0' != (c = mchars_num2char(p, len)))
218 putchar((int)c);
219 }
220
221 static void
222 print_spec(struct html *h, const char *p, size_t len)
223 {
224 int cp;
225 const char *rhs;
226 size_t sz;
227
228 if ((cp = mchars_spec2cp(h->symtab, p, len)) > 0) {
229 printf("&#%d;", cp);
230 return;
231 } else if (-1 == cp && 1 == len) {
232 fwrite(p, 1, len, stdout);
233 return;
234 } else if (-1 == cp)
235 return;
236
237 if (NULL != (rhs = mchars_spec2str(h->symtab, p, len, &sz)))
238 fwrite(rhs, 1, sz, stdout);
239 }
240
241
242 static void
243 print_res(struct html *h, const char *p, size_t len)
244 {
245 int cp;
246 const char *rhs;
247 size_t sz;
248
249 if ((cp = mchars_res2cp(h->symtab, p, len)) > 0) {
250 printf("&#%d;", cp);
251 return;
252 } else if (-1 == cp)
253 return;
254
255 if (NULL != (rhs = mchars_res2str(h->symtab, p, len, &sz)))
256 fwrite(rhs, 1, sz, stdout);
257 }
258
259
260 static void
261 print_metaf(struct html *h, enum mandoc_esc deco)
262 {
263 enum htmlfont font;
264
265 switch (deco) {
266 case (ESCAPE_FONTPREV):
267 font = h->metal;
268 break;
269 case (ESCAPE_FONTITALIC):
270 font = HTMLFONT_ITALIC;
271 break;
272 case (ESCAPE_FONTBOLD):
273 font = HTMLFONT_BOLD;
274 break;
275 case (ESCAPE_FONTROMAN):
276 font = HTMLFONT_NONE;
277 break;
278 default:
279 abort();
280 /* NOTREACHED */
281 }
282
283 if (h->metaf) {
284 print_tagq(h, h->metaf);
285 h->metaf = NULL;
286 }
287
288 h->metal = h->metac;
289 h->metac = font;
290
291 if (HTMLFONT_NONE != font)
292 h->metaf = HTMLFONT_BOLD == font ?
293 print_otag(h, TAG_B, 0, NULL) :
294 print_otag(h, TAG_I, 0, NULL);
295 }
296
297
298 static int
299 print_encode(struct html *h, const char *p, int norecurse)
300 {
301 size_t sz;
302 int len, nospace;
303 const char *seq;
304 enum mandoc_esc esc;
305 static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' };
306
307 nospace = 0;
308
309 while ('\0' != *p) {
310 sz = strcspn(p, rejs);
311
312 fwrite(p, 1, sz, stdout);
313 p += (int)sz;
314
315 if ('\0' == *p)
316 break;
317
318 switch (*p++) {
319 case ('<'):
320 printf("&lt;");
321 continue;
322 case ('>'):
323 printf("&gt;");
324 continue;
325 case ('&'):
326 printf("&amp;");
327 continue;
328 case (ASCII_HYPH):
329 putchar('-');
330 continue;
331 default:
332 break;
333 }
334
335 esc = mandoc_escape(&p, &seq, &len);
336 if (ESCAPE_ERROR == esc)
337 break;
338
339 switch (esc) {
340 case (ESCAPE_NUMBERED):
341 print_num(h, seq, len);
342 break;
343 case (ESCAPE_PREDEF):
344 print_res(h, seq, len);
345 break;
346 case (ESCAPE_SPECIAL):
347 print_spec(h, seq, len);
348 break;
349 case (ESCAPE_FONTPREV):
350 /* FALLTHROUGH */
351 case (ESCAPE_FONTBOLD):
352 /* FALLTHROUGH */
353 case (ESCAPE_FONTITALIC):
354 /* FALLTHROUGH */
355 case (ESCAPE_FONTROMAN):
356 if (norecurse)
357 break;
358 print_metaf(h, esc);
359 break;
360 case (ESCAPE_NOSPACE):
361 if ('\0' == *p)
362 nospace = 1;
363 break;
364 default:
365 break;
366 }
367 }
368
369 return(nospace);
370 }
371
372
373 static void
374 print_attr(struct html *h, const char *key, const char *val)
375 {
376 printf(" %s=\"", key);
377 (void)print_encode(h, val, 1);
378 putchar('\"');
379 }
380
381
382 struct tag *
383 print_otag(struct html *h, enum htmltag tag,
384 int sz, const struct htmlpair *p)
385 {
386 int i;
387 struct tag *t;
388
389 /* Push this tags onto the stack of open scopes. */
390
391 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
392 t = mandoc_malloc(sizeof(struct tag));
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 /* Accommodate 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 void
506 print_text(struct html *h, const char *word)
507 {
508
509 if ( ! (HTML_NOSPACE & h->flags)) {
510 /* Manage keeps! */
511 if ( ! (HTML_KEEP & h->flags)) {
512 if (HTML_PREKEEP & h->flags)
513 h->flags |= HTML_KEEP;
514 putchar(' ');
515 } else
516 printf("&#160;");
517 }
518
519 assert(NULL == h->metaf);
520 if (HTMLFONT_NONE != h->metac)
521 h->metaf = HTMLFONT_BOLD == h->metac ?
522 print_otag(h, TAG_B, 0, NULL) :
523 print_otag(h, TAG_I, 0, NULL);
524
525 assert(word);
526 if ( ! print_encode(h, word, 0))
527 if ( ! (h->flags & HTML_NONOSPACE))
528 h->flags &= ~HTML_NOSPACE;
529
530 if (h->metaf) {
531 print_tagq(h, h->metaf);
532 h->metaf = NULL;
533 }
534
535 h->flags &= ~HTML_IGNDELIM;
536 }
537
538
539 void
540 print_tagq(struct html *h, const struct tag *until)
541 {
542 struct tag *tag;
543
544 while ((tag = h->tags.head) != NULL) {
545 /*
546 * Remember to close out and nullify the current
547 * meta-font and table, if applicable.
548 */
549 if (tag == h->metaf)
550 h->metaf = NULL;
551 if (tag == h->tblt)
552 h->tblt = NULL;
553 print_ctag(h, tag->tag);
554 h->tags.head = tag->next;
555 free(tag);
556 if (until && tag == until)
557 return;
558 }
559 }
560
561
562 void
563 print_stagq(struct html *h, const struct tag *suntil)
564 {
565 struct tag *tag;
566
567 while ((tag = h->tags.head) != NULL) {
568 if (suntil && tag == suntil)
569 return;
570 /*
571 * Remember to close out and nullify the current
572 * meta-font and table, if applicable.
573 */
574 if (tag == h->metaf)
575 h->metaf = NULL;
576 if (tag == h->tblt)
577 h->tblt = NULL;
578 print_ctag(h, tag->tag);
579 h->tags.head = tag->next;
580 free(tag);
581 }
582 }
583
584
585 void
586 bufinit(struct html *h)
587 {
588
589 h->buf[0] = '\0';
590 h->buflen = 0;
591 }
592
593
594 void
595 bufcat_style(struct html *h, const char *key, const char *val)
596 {
597
598 bufcat(h, key);
599 bufncat(h, ":", 1);
600 bufcat(h, val);
601 bufncat(h, ";", 1);
602 }
603
604
605 void
606 bufcat(struct html *h, const char *p)
607 {
608
609 bufncat(h, p, strlen(p));
610 }
611
612
613 void
614 buffmt(struct html *h, const char *fmt, ...)
615 {
616 va_list ap;
617
618 va_start(ap, fmt);
619 (void)vsnprintf(h->buf + (int)h->buflen,
620 BUFSIZ - h->buflen - 1, fmt, ap);
621 va_end(ap);
622 h->buflen = strlen(h->buf);
623 }
624
625
626 void
627 bufncat(struct html *h, const char *p, size_t sz)
628 {
629
630 if (h->buflen + sz > BUFSIZ - 1)
631 sz = BUFSIZ - 1 - h->buflen;
632
633 (void)strncat(h->buf, p, sz);
634 h->buflen += sz;
635 }
636
637
638 void
639 buffmt_includes(struct html *h, const char *name)
640 {
641 const char *p, *pp;
642
643 pp = h->base_includes;
644
645 while (NULL != (p = strchr(pp, '%'))) {
646 bufncat(h, pp, (size_t)(p - pp));
647 switch (*(p + 1)) {
648 case('I'):
649 bufcat(h, name);
650 break;
651 default:
652 bufncat(h, p, 2);
653 break;
654 }
655 pp = p + 2;
656 }
657 if (pp)
658 bufcat(h, pp);
659 }
660
661
662 void
663 buffmt_man(struct html *h,
664 const char *name, const char *sec)
665 {
666 const char *p, *pp;
667
668 pp = h->base_man;
669
670 /* LINTED */
671 while (NULL != (p = strchr(pp, '%'))) {
672 bufncat(h, pp, (size_t)(p - pp));
673 switch (*(p + 1)) {
674 case('S'):
675 bufcat(h, sec ? sec : "1");
676 break;
677 case('N'):
678 buffmt(h, name);
679 break;
680 default:
681 bufncat(h, p, 2);
682 break;
683 }
684 pp = p + 2;
685 }
686 if (pp)
687 bufcat(h, pp);
688 }
689
690
691 void
692 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
693 {
694 double v;
695 const char *u;
696
697 v = su->scale;
698
699 switch (su->unit) {
700 case (SCALE_CM):
701 u = "cm";
702 break;
703 case (SCALE_IN):
704 u = "in";
705 break;
706 case (SCALE_PC):
707 u = "pc";
708 break;
709 case (SCALE_PT):
710 u = "pt";
711 break;
712 case (SCALE_EM):
713 u = "em";
714 break;
715 case (SCALE_MM):
716 if (0 == (v /= 100))
717 v = 1;
718 u = "em";
719 break;
720 case (SCALE_EN):
721 u = "ex";
722 break;
723 case (SCALE_BU):
724 u = "ex";
725 break;
726 case (SCALE_VS):
727 u = "em";
728 break;
729 default:
730 u = "ex";
731 break;
732 }
733
734 /*
735 * XXX: the CSS spec isn't clear as to which types accept
736 * integer or real numbers, so we just make them all decimals.
737 */
738 buffmt(h, "%s: %.2f%s;", p, v, u);
739 }
740
741
742 void
743 html_idcat(char *dst, const char *src, int sz)
744 {
745 int ssz;
746
747 assert(sz > 2);
748
749 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
750
751 /* We can't start with a number (bah). */
752
753 if ('#' == *dst) {
754 dst++;
755 sz--;
756 }
757 if ('\0' == *dst) {
758 *dst++ = 'x';
759 *dst = '\0';
760 sz--;
761 }
762
763 for ( ; *dst != '\0' && sz; dst++, sz--)
764 /* Jump to end. */ ;
765
766 for ( ; *src != '\0' && sz > 1; src++) {
767 ssz = snprintf(dst, (size_t)sz, "%.2x", *src);
768 sz -= ssz;
769 dst += ssz;
770 }
771 }