1 /* $Id: html.c,v 1.181 2014/10/29 00:17:43 schwarze Exp $ */
3 * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2011, 2012, 2013, 2014 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.
20 #include <sys/types.h>
32 #include "mandoc_aux.h"
33 #include "libmandoc.h"
41 #define HTML_CLRLINE (1 << 0)
42 #define HTML_NOSTACK (1 << 1)
43 #define HTML_AUTOCLOSE (1 << 2) /* Tag has auto-closure. */
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 */
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 {"pre", HTML_CLRLINE }, /* TAG_PRE */
72 {"b", 0 }, /* TAG_B */
73 {"i", 0 }, /* TAG_I */
74 {"code", 0 }, /* TAG_CODE */
75 {"small", 0 }, /* TAG_SMALL */
76 {"style", HTML_CLRLINE}, /* TAG_STYLE */
77 {"math", HTML_CLRLINE}, /* TAG_MATH */
78 {"mrow", 0}, /* TAG_MROW */
79 {"mi", 0}, /* TAG_MI */
80 {"mo", 0}, /* TAG_MO */
81 {"msup", 0}, /* TAG_MSUP */
82 {"msub", 0}, /* TAG_MSUB */
83 {"msubsup", 0}, /* TAG_MSUBSUP */
84 {"mfrac", 0}, /* TAG_MFRAC */
85 {"msqrt", 0}, /* TAG_MSQRT */
86 {"mfenced", 0}, /* TAG_MFENCED */
87 {"mtable", 0}, /* TAG_MTABLE */
88 {"mtr", 0}, /* TAG_MTR */
89 {"mtd", 0}, /* TAG_MTD */
90 {"munderover", 0}, /* TAG_MUNDEROVER */
91 {"munder", 0}, /* TAG_MUNDER*/
92 {"mover", 0}, /* TAG_MOVER*/
95 static const char *const htmlattrs[ATTR_MAX] = {
96 "name", /* ATTR_NAME */
98 "href", /* ATTR_HREF */
99 "type", /* ATTR_TYPE */
100 "media", /* ATTR_MEDIA */
101 "class", /* ATTR_CLASS */
102 "style", /* ATTR_STYLE */
104 "colspan", /* ATTR_COLSPAN */
105 "charset", /* ATTR_CHARSET */
106 "open", /* ATTR_OPEN */
107 "close", /* ATTR_CLOSE */
108 "mathvariant", /* ATTR_MATHVARIANT */
111 static const char *const roffscales[SCALE_MAX] = {
124 static void bufncat(struct html *, const char *, size_t);
125 static void print_ctag(struct html *, enum htmltag);
126 static int print_escape(char);
127 static int print_encode(struct html *, const char *, int);
128 static void print_metaf(struct html *, enum mandoc_esc);
129 static void print_attr(struct html *, const char *, const char *);
133 html_alloc(const struct mchars *mchars, char *outopts)
141 toks[2] = "includes";
142 toks[3] = "fragment";
145 h = mandoc_calloc(1, sizeof(struct html));
150 while (outopts && *outopts)
151 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
159 h->base_includes = v;
162 h->oflags |= HTML_FRAGMENT;
177 h = (struct html *)p;
179 while ((tag = h->tags.head) != NULL) {
180 h->tags.head = tag->next;
188 print_gen_head(struct html *h)
190 struct htmlpair tag[4];
193 tag[0].key = ATTR_CHARSET;
194 tag[0].val = "utf-8";
195 print_otag(h, TAG_META, 1, tag);
198 * Print a default style-sheet.
200 t = print_otag(h, TAG_STYLE, 0, NULL);
201 print_text(h, "table.head, table.foot { width: 100%; }\n"
202 "td.head-rtitle, td.foot-os { text-align: right; }\n"
203 "td.head-vol { text-align: center; }\n"
204 "table.foot td { width: 50%; }\n"
205 "table.head td { width: 33%; }\n"
206 "div.spacer { margin: 1em 0; }\n");
210 tag[0].key = ATTR_REL;
211 tag[0].val = "stylesheet";
212 tag[1].key = ATTR_HREF;
213 tag[1].val = h->style;
214 tag[2].key = ATTR_TYPE;
215 tag[2].val = "text/css";
216 tag[3].key = ATTR_MEDIA;
218 print_otag(h, TAG_LINK, 4, tag);
223 print_metaf(struct html *h, enum mandoc_esc deco)
228 case ESCAPE_FONTPREV:
231 case ESCAPE_FONTITALIC:
232 font = HTMLFONT_ITALIC;
234 case ESCAPE_FONTBOLD:
235 font = HTMLFONT_BOLD;
242 case ESCAPE_FONTROMAN:
243 font = HTMLFONT_NONE;
251 print_tagq(h, h->metaf);
259 case HTMLFONT_ITALIC:
260 h->metaf = print_otag(h, TAG_I, 0, NULL);
263 h->metaf = print_otag(h, TAG_B, 0, NULL);
266 h->metaf = print_otag(h, TAG_B, 0, NULL);
267 print_otag(h, TAG_I, 0, NULL);
275 html_strlen(const char *cp)
281 * Account for escaped sequences within string length
282 * calculations. This follows the logic in term_strlen() as we
283 * must calculate the width of produced strings.
284 * Assume that characters are always width of "1". This is
285 * hacky, but it gets the job done for approximation of widths.
291 rsz = strcspn(cp, "\\");
303 switch (mandoc_escape(&cp, NULL, NULL)) {
308 case ESCAPE_NUMBERED:
316 case ESCAPE_SKIPCHAR:
358 print_encode(struct html *h, const char *p, int norecurse)
364 static const char rejs[9] = { '\\', '<', '>', '&', '"',
365 ASCII_NBRSP, ASCII_HYPH, ASCII_BREAK, '\0' };
370 if (HTML_SKIPCHAR & h->flags && '\\' != *p) {
371 h->flags &= ~HTML_SKIPCHAR;
376 sz = strcspn(p, rejs);
378 fwrite(p, 1, sz, stdout);
384 if (print_escape(*p++))
387 esc = mandoc_escape(&p, &seq, &len);
388 if (ESCAPE_ERROR == esc)
394 case ESCAPE_FONTPREV:
396 case ESCAPE_FONTBOLD:
398 case ESCAPE_FONTITALIC:
402 case ESCAPE_FONTROMAN:
406 case ESCAPE_SKIPCHAR:
407 h->flags |= HTML_SKIPCHAR;
413 if (h->flags & HTML_SKIPCHAR) {
414 h->flags &= ~HTML_SKIPCHAR;
420 /* Skip past "u" header. */
421 c = mchars_num2uc(seq + 1, len - 1);
423 case ESCAPE_NUMBERED:
424 c = mchars_num2char(seq, len);
429 c = mchars_spec2cp(h->symtab, seq, len);
440 if ((c < 0x20 && c != 0x09) ||
441 (c > 0x7E && c < 0xA0))
445 else if ( ! print_escape(c))
453 print_attr(struct html *h, const char *key, const char *val)
455 printf(" %s=\"", key);
456 (void)print_encode(h, val, 1);
461 print_otag(struct html *h, enum htmltag tag,
462 int sz, const struct htmlpair *p)
467 /* Push this tags onto the stack of open scopes. */
469 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
470 t = mandoc_malloc(sizeof(struct tag));
472 t->next = h->tags.head;
477 if ( ! (HTML_NOSPACE & h->flags))
478 if ( ! (HTML_CLRLINE & htmltags[tag].flags)) {
480 if ( ! (HTML_KEEP & h->flags)) {
481 if (HTML_PREKEEP & h->flags)
482 h->flags |= HTML_KEEP;
488 if ( ! (h->flags & HTML_NONOSPACE))
489 h->flags &= ~HTML_NOSPACE;
491 h->flags |= HTML_NOSPACE;
493 /* Print out the tag name and attributes. */
495 printf("<%s", htmltags[tag].name);
496 for (i = 0; i < sz; i++)
497 print_attr(h, htmlattrs[p[i].key], p[i].val);
499 /* Accommodate for "well-formed" singleton escaping. */
501 if (HTML_AUTOCLOSE & htmltags[tag].flags)
506 h->flags |= HTML_NOSPACE;
508 if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags)
515 print_ctag(struct html *h, enum htmltag tag)
518 printf("</%s>", htmltags[tag].name);
519 if (HTML_CLRLINE & htmltags[tag].flags) {
520 h->flags |= HTML_NOSPACE;
526 print_gen_decls(struct html *h)
529 puts("<!DOCTYPE html>");
533 print_text(struct html *h, const char *word)
536 if ( ! (HTML_NOSPACE & h->flags)) {
538 if ( ! (HTML_KEEP & h->flags)) {
539 if (HTML_PREKEEP & h->flags)
540 h->flags |= HTML_KEEP;
546 assert(NULL == h->metaf);
548 case HTMLFONT_ITALIC:
549 h->metaf = print_otag(h, TAG_I, 0, NULL);
552 h->metaf = print_otag(h, TAG_B, 0, NULL);
555 h->metaf = print_otag(h, TAG_B, 0, NULL);
556 print_otag(h, TAG_I, 0, NULL);
563 if ( ! print_encode(h, word, 0)) {
564 if ( ! (h->flags & HTML_NONOSPACE))
565 h->flags &= ~HTML_NOSPACE;
567 h->flags |= HTML_NOSPACE;
570 print_tagq(h, h->metaf);
574 h->flags &= ~HTML_IGNDELIM;
578 print_tagq(struct html *h, const struct tag *until)
582 while ((tag = h->tags.head) != NULL) {
584 * Remember to close out and nullify the current
585 * meta-font and table, if applicable.
591 print_ctag(h, tag->tag);
592 h->tags.head = tag->next;
594 if (until && tag == until)
600 print_stagq(struct html *h, const struct tag *suntil)
604 while ((tag = h->tags.head) != NULL) {
605 if (suntil && tag == suntil)
608 * Remember to close out and nullify the current
609 * meta-font and table, if applicable.
615 print_ctag(h, tag->tag);
616 h->tags.head = tag->next;
622 print_paragraph(struct html *h)
627 PAIR_CLASS_INIT(&tag, "spacer");
628 t = print_otag(h, TAG_DIV, 1, &tag);
634 bufinit(struct html *h)
642 bufcat_style(struct html *h, const char *key, const char *val)
652 bufcat(struct html *h, const char *p)
656 * XXX This is broken and not easy to fix.
657 * When using the -Oincludes option, buffmt_includes()
658 * may pass in strings overrunning BUFSIZ, causing a crash.
661 h->buflen = strlcat(h->buf, p, BUFSIZ);
662 assert(h->buflen < BUFSIZ);
666 bufcat_fmt(struct html *h, const char *fmt, ...)
671 (void)vsnprintf(h->buf + (int)h->buflen,
672 BUFSIZ - h->buflen - 1, fmt, ap);
674 h->buflen = strlen(h->buf);
678 bufncat(struct html *h, const char *p, size_t sz)
681 assert(h->buflen + sz + 1 < BUFSIZ);
682 strncat(h->buf, p, sz);
687 buffmt_includes(struct html *h, const char *name)
691 pp = h->base_includes;
694 while (NULL != (p = strchr(pp, '%'))) {
695 bufncat(h, pp, (size_t)(p - pp));
711 buffmt_man(struct html *h, const char *name, const char *sec)
718 while (NULL != (p = strchr(pp, '%'))) {
719 bufncat(h, pp, (size_t)(p - pp));
722 bufcat(h, sec ? sec : "1");
725 bufcat_fmt(h, "%s", name);
738 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
743 if (SCALE_MM == su->unit && 0.0 == (v /= 100.0))
745 else if (SCALE_BU == su->unit)
748 bufcat_fmt(h, "%s: %.2f%s;", p, v, roffscales[su->unit]);
752 bufcat_id(struct html *h, const char *src)
755 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
758 bufcat_fmt(h, "%.2x", *src++);