1 /* $Id: html.c,v 1.150 2011/10/05 21:35:17 kristaps Exp $ */
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2011 Ingo Schwarze <schwarze@openbsd.org>
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.
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.
22 #include <sys/types.h>
34 #include "libmandoc.h"
42 #define HTML_CLRLINE (1 << 0)
43 #define HTML_NOSTACK (1 << 1)
44 #define HTML_AUTOCLOSE (1 << 2) /* Tag has auto-closure. */
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 */
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 */
80 static const char *const htmlattrs[ATTR_MAX] = {
81 "http-equiv", /* ATTR_HTTPEQUIV */
82 "content", /* ATTR_CONTENT */
83 "name", /* ATTR_NAME */
85 "href", /* ATTR_HREF */
86 "type", /* ATTR_TYPE */
87 "media", /* ATTR_MEDIA */
88 "class", /* ATTR_CLASS */
89 "style", /* ATTR_STYLE */
90 "width", /* ATTR_WIDTH */
92 "summary", /* ATTR_SUMMARY */
93 "align", /* ATTR_ALIGN */
94 "colspan", /* ATTR_COLSPAN */
97 static const char *const roffscales[SCALE_MAX] = {
110 static void bufncat(struct html *, const char *, size_t);
111 static void print_ctag(struct html *, enum htmltag);
112 static int print_encode(struct html *, const char *, int);
113 static void print_metaf(struct html *, enum mandoc_esc);
114 static void print_attr(struct html *, const char *, const char *);
115 static void *ml_alloc(char *, enum htmltype);
118 ml_alloc(char *outopts, enum htmltype type)
126 toks[2] = "includes";
127 toks[3] = "fragment";
130 h = mandoc_calloc(1, sizeof(struct html));
134 h->symtab = mchars_alloc();
136 while (outopts && *outopts)
137 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
145 h->base_includes = v;
148 h->oflags |= HTML_FRAGMENT;
158 html_alloc(char *outopts)
161 return(ml_alloc(outopts, HTML_HTML_4_01_STRICT));
166 xhtml_alloc(char *outopts)
169 return(ml_alloc(outopts, HTML_XHTML_1_0_STRICT));
179 h = (struct html *)p;
181 while ((tag = h->tags.head) != NULL) {
182 h->tags.head = tag->next;
187 mchars_free(h->symtab);
194 print_gen_head(struct html *h)
196 struct htmlpair tag[4];
198 tag[0].key = ATTR_HTTPEQUIV;
199 tag[0].val = "Content-Type";
200 tag[1].key = ATTR_CONTENT;
201 tag[1].val = "text/html; charset=utf-8";
202 print_otag(h, TAG_META, 2, tag);
204 tag[0].key = ATTR_NAME;
205 tag[0].val = "resource-type";
206 tag[1].key = ATTR_CONTENT;
207 tag[1].val = "document";
208 print_otag(h, TAG_META, 2, tag);
211 tag[0].key = ATTR_REL;
212 tag[0].val = "stylesheet";
213 tag[1].key = ATTR_HREF;
214 tag[1].val = h->style;
215 tag[2].key = ATTR_TYPE;
216 tag[2].val = "text/css";
217 tag[3].key = ATTR_MEDIA;
219 print_otag(h, TAG_LINK, 4, tag);
224 print_metaf(struct html *h, enum mandoc_esc deco)
229 case (ESCAPE_FONTPREV):
232 case (ESCAPE_FONTITALIC):
233 font = HTMLFONT_ITALIC;
235 case (ESCAPE_FONTBOLD):
236 font = HTMLFONT_BOLD;
240 case (ESCAPE_FONTROMAN):
241 font = HTMLFONT_NONE;
249 print_tagq(h, h->metaf);
256 if (HTMLFONT_NONE != font)
257 h->metaf = HTMLFONT_BOLD == font ?
258 print_otag(h, TAG_B, 0, NULL) :
259 print_otag(h, TAG_I, 0, NULL);
263 html_strlen(const char *cp)
269 * Account for escaped sequences within string length
270 * calculations. This follows the logic in term_strlen() as we
271 * must calculate the width of produced strings.
272 * Assume that characters are always width of "1". This is
273 * hacky, but it gets the job done for approximation of widths.
277 while (NULL != (p = strchr(cp, '\\'))) {
280 switch (mandoc_escape(&cp, &seq, &ssz)) {
283 case (ESCAPE_UNICODE):
285 case (ESCAPE_NUMBERED):
287 case (ESCAPE_SPECIAL):
296 return(sz + strlen(cp));
300 print_encode(struct html *h, const char *p, int norecurse)
306 static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' };
311 sz = strcspn(p, rejs);
313 fwrite(p, 1, sz, stdout);
336 esc = mandoc_escape(&p, &seq, &len);
337 if (ESCAPE_ERROR == esc)
341 case (ESCAPE_UNICODE):
342 /* Skip passed "u" header. */
343 c = mchars_num2uc(seq + 1, len - 1);
347 case (ESCAPE_NUMBERED):
348 c = mchars_num2char(seq, len);
352 case (ESCAPE_SPECIAL):
353 c = mchars_spec2cp(h->symtab, seq, len);
356 else if (-1 == c && 1 == len)
361 case (ESCAPE_FONTPREV):
363 case (ESCAPE_FONTBOLD):
365 case (ESCAPE_FONTITALIC):
367 case (ESCAPE_FONTROMAN):
372 case (ESCAPE_NOSPACE):
386 print_attr(struct html *h, const char *key, const char *val)
388 printf(" %s=\"", key);
389 (void)print_encode(h, val, 1);
395 print_otag(struct html *h, enum htmltag tag,
396 int sz, const struct htmlpair *p)
401 /* Push this tags onto the stack of open scopes. */
403 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
404 t = mandoc_malloc(sizeof(struct tag));
406 t->next = h->tags.head;
411 if ( ! (HTML_NOSPACE & h->flags))
412 if ( ! (HTML_CLRLINE & htmltags[tag].flags)) {
414 if ( ! (HTML_KEEP & h->flags)) {
415 if (HTML_PREKEEP & h->flags)
416 h->flags |= HTML_KEEP;
422 if ( ! (h->flags & HTML_NONOSPACE))
423 h->flags &= ~HTML_NOSPACE;
425 h->flags |= HTML_NOSPACE;
427 /* Print out the tag name and attributes. */
429 printf("<%s", htmltags[tag].name);
430 for (i = 0; i < sz; i++)
431 print_attr(h, htmlattrs[p[i].key], p[i].val);
433 /* Add non-overridable attributes. */
435 if (TAG_HTML == tag && HTML_XHTML_1_0_STRICT == h->type) {
436 print_attr(h, "xmlns", "http://www.w3.org/1999/xhtml");
437 print_attr(h, "xml:lang", "en");
438 print_attr(h, "lang", "en");
441 /* Accommodate for XML "well-formed" singleton escaping. */
443 if (HTML_AUTOCLOSE & htmltags[tag].flags)
445 case (HTML_XHTML_1_0_STRICT):
454 h->flags |= HTML_NOSPACE;
456 if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags)
464 print_ctag(struct html *h, enum htmltag tag)
467 printf("</%s>", htmltags[tag].name);
468 if (HTML_CLRLINE & htmltags[tag].flags) {
469 h->flags |= HTML_NOSPACE;
475 print_gen_decls(struct html *h)
482 case (HTML_HTML_4_01_STRICT):
484 doctype = "-//W3C//DTD HTML 4.01//EN";
485 dtd = "http://www.w3.org/TR/html4/strict.dtd";
488 puts("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
490 doctype = "-//W3C//DTD XHTML 1.0 Strict//EN";
491 dtd = "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
495 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
500 print_text(struct html *h, const char *word)
503 if ( ! (HTML_NOSPACE & h->flags)) {
505 if ( ! (HTML_KEEP & h->flags)) {
506 if (HTML_PREKEEP & h->flags)
507 h->flags |= HTML_KEEP;
513 assert(NULL == h->metaf);
514 if (HTMLFONT_NONE != h->metac)
515 h->metaf = HTMLFONT_BOLD == h->metac ?
516 print_otag(h, TAG_B, 0, NULL) :
517 print_otag(h, TAG_I, 0, NULL);
520 if ( ! print_encode(h, word, 0)) {
521 if ( ! (h->flags & HTML_NONOSPACE))
522 h->flags &= ~HTML_NOSPACE;
524 h->flags |= HTML_NOSPACE;
527 print_tagq(h, h->metaf);
531 h->flags &= ~HTML_IGNDELIM;
536 print_tagq(struct html *h, const struct tag *until)
540 while ((tag = h->tags.head) != NULL) {
542 * Remember to close out and nullify the current
543 * meta-font and table, if applicable.
549 print_ctag(h, tag->tag);
550 h->tags.head = tag->next;
552 if (until && tag == until)
559 print_stagq(struct html *h, const struct tag *suntil)
563 while ((tag = h->tags.head) != NULL) {
564 if (suntil && tag == suntil)
567 * Remember to close out and nullify the current
568 * meta-font and table, if applicable.
574 print_ctag(h, tag->tag);
575 h->tags.head = tag->next;
581 bufinit(struct html *h)
589 bufcat_style(struct html *h, const char *key, const char *val)
599 bufcat(struct html *h, const char *p)
602 h->buflen = strlcat(h->buf, p, BUFSIZ);
603 assert(h->buflen < BUFSIZ);
607 bufcat_fmt(struct html *h, const char *fmt, ...)
612 (void)vsnprintf(h->buf + (int)h->buflen,
613 BUFSIZ - h->buflen - 1, fmt, ap);
615 h->buflen = strlen(h->buf);
619 bufncat(struct html *h, const char *p, size_t sz)
622 assert(h->buflen + sz + 1 < BUFSIZ);
623 strncat(h->buf, p, sz);
628 buffmt_includes(struct html *h, const char *name)
632 pp = h->base_includes;
635 while (NULL != (p = strchr(pp, '%'))) {
636 bufncat(h, pp, (size_t)(p - pp));
652 buffmt_man(struct html *h,
653 const char *name, const char *sec)
660 while (NULL != (p = strchr(pp, '%'))) {
661 bufncat(h, pp, (size_t)(p - pp));
664 bufcat(h, sec ? sec : "1");
680 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
685 if (SCALE_MM == su->unit && 0.0 == (v /= 100.0))
688 bufcat_fmt(h, "%s: %.2f%s;", p, v, roffscales[su->unit]);
692 bufcat_id(struct html *h, const char *src)
695 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
698 bufcat_fmt(h, "%.2x", *src++);