1 /* $Id: html.c,v 1.192 2016/01/04 12:45:29 schwarze Exp $ */
3 * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2011-2015 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 AUTHORS DISCLAIM ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS 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"
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 *, struct tag *);
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 manoutput *outopts)
137 h = mandoc_calloc(1, sizeof(struct html));
140 h->style = outopts->style;
141 h->base_man = outopts->man;
142 h->base_includes = outopts->includes;
143 if (outopts->fragment)
144 h->oflags |= HTML_FRAGMENT;
155 h = (struct html *)p;
157 while ((tag = h->tags.head) != NULL) {
158 h->tags.head = tag->next;
166 print_gen_head(struct html *h)
168 struct htmlpair tag[4];
171 tag[0].key = ATTR_CHARSET;
172 tag[0].val = "utf-8";
173 print_otag(h, TAG_META, 1, tag);
176 * Print a default style-sheet.
178 t = print_otag(h, TAG_STYLE, 0, NULL);
179 print_text(h, "table.head, table.foot { width: 100%; }\n"
180 "td.head-rtitle, td.foot-os { text-align: right; }\n"
181 "td.head-vol { text-align: center; }\n"
182 "table.foot td { width: 50%; }\n"
183 "table.head td { width: 33%; }\n"
184 "div.spacer { margin: 1em 0; }\n");
188 tag[0].key = ATTR_REL;
189 tag[0].val = "stylesheet";
190 tag[1].key = ATTR_HREF;
191 tag[1].val = h->style;
192 tag[2].key = ATTR_TYPE;
193 tag[2].val = "text/css";
194 tag[3].key = ATTR_MEDIA;
196 print_otag(h, TAG_LINK, 4, tag);
201 print_metaf(struct html *h, enum mandoc_esc deco)
206 case ESCAPE_FONTPREV:
209 case ESCAPE_FONTITALIC:
210 font = HTMLFONT_ITALIC;
212 case ESCAPE_FONTBOLD:
213 font = HTMLFONT_BOLD;
219 case ESCAPE_FONTROMAN:
220 font = HTMLFONT_NONE;
227 print_tagq(h, h->metaf);
235 case HTMLFONT_ITALIC:
236 h->metaf = print_otag(h, TAG_I, 0, NULL);
239 h->metaf = print_otag(h, TAG_B, 0, NULL);
242 h->metaf = print_otag(h, TAG_B, 0, NULL);
243 print_otag(h, TAG_I, 0, NULL);
251 html_strlen(const char *cp)
257 * Account for escaped sequences within string length
258 * calculations. This follows the logic in term_strlen() as we
259 * must calculate the width of produced strings.
260 * Assume that characters are always width of "1". This is
261 * hacky, but it gets the job done for approximation of widths.
267 rsz = strcspn(cp, "\\");
279 switch (mandoc_escape(&cp, NULL, NULL)) {
283 case ESCAPE_NUMBERED:
285 case ESCAPE_OVERSTRIKE:
291 case ESCAPE_SKIPCHAR:
333 print_encode(struct html *h, const char *p, int norecurse)
339 static const char rejs[9] = { '\\', '<', '>', '&', '"',
340 ASCII_NBRSP, ASCII_HYPH, ASCII_BREAK, '\0' };
345 if (HTML_SKIPCHAR & h->flags && '\\' != *p) {
346 h->flags &= ~HTML_SKIPCHAR;
351 sz = strcspn(p, rejs);
353 fwrite(p, 1, sz, stdout);
359 if (print_escape(*p++))
362 esc = mandoc_escape(&p, &seq, &len);
363 if (ESCAPE_ERROR == esc)
368 case ESCAPE_FONTPREV:
369 case ESCAPE_FONTBOLD:
370 case ESCAPE_FONTITALIC:
372 case ESCAPE_FONTROMAN:
376 case ESCAPE_SKIPCHAR:
377 h->flags |= HTML_SKIPCHAR;
383 if (h->flags & HTML_SKIPCHAR) {
384 h->flags &= ~HTML_SKIPCHAR;
390 /* Skip past "u" header. */
391 c = mchars_num2uc(seq + 1, len - 1);
393 case ESCAPE_NUMBERED:
394 c = mchars_num2char(seq, len);
399 c = mchars_spec2cp(seq, len);
407 case ESCAPE_OVERSTRIKE:
415 if ((c < 0x20 && c != 0x09) ||
416 (c > 0x7E && c < 0xA0))
420 else if ( ! print_escape(c))
428 print_attr(struct html *h, const char *key, const char *val)
430 printf(" %s=\"", key);
431 (void)print_encode(h, val, 1);
436 print_otag(struct html *h, enum htmltag tag,
437 int sz, const struct htmlpair *p)
442 /* Push this tags onto the stack of open scopes. */
444 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
445 t = mandoc_malloc(sizeof(struct tag));
447 t->next = h->tags.head;
452 if ( ! (HTML_NOSPACE & h->flags))
453 if ( ! (HTML_CLRLINE & htmltags[tag].flags)) {
455 if ( ! (HTML_KEEP & h->flags)) {
456 if (HTML_PREKEEP & h->flags)
457 h->flags |= HTML_KEEP;
463 if ( ! (h->flags & HTML_NONOSPACE))
464 h->flags &= ~HTML_NOSPACE;
466 h->flags |= HTML_NOSPACE;
468 /* Print out the tag name and attributes. */
470 printf("<%s", htmltags[tag].name);
471 for (i = 0; i < sz; i++)
472 print_attr(h, htmlattrs[p[i].key], p[i].val);
474 /* Accommodate for "well-formed" singleton escaping. */
476 if (HTML_AUTOCLOSE & htmltags[tag].flags)
481 h->flags |= HTML_NOSPACE;
483 if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags)
490 print_ctag(struct html *h, struct tag *tag)
494 * Remember to close out and nullify the current
495 * meta-font and table, if applicable.
502 printf("</%s>", htmltags[tag->tag].name);
503 if (HTML_CLRLINE & htmltags[tag->tag].flags) {
504 h->flags |= HTML_NOSPACE;
508 h->tags.head = tag->next;
513 print_gen_decls(struct html *h)
516 puts("<!DOCTYPE html>");
520 print_text(struct html *h, const char *word)
523 if ( ! (HTML_NOSPACE & h->flags)) {
525 if ( ! (HTML_KEEP & h->flags)) {
526 if (HTML_PREKEEP & h->flags)
527 h->flags |= HTML_KEEP;
533 assert(NULL == h->metaf);
535 case HTMLFONT_ITALIC:
536 h->metaf = print_otag(h, TAG_I, 0, NULL);
539 h->metaf = print_otag(h, TAG_B, 0, NULL);
542 h->metaf = print_otag(h, TAG_B, 0, NULL);
543 print_otag(h, TAG_I, 0, NULL);
550 if ( ! print_encode(h, word, 0)) {
551 if ( ! (h->flags & HTML_NONOSPACE))
552 h->flags &= ~HTML_NOSPACE;
553 h->flags &= ~HTML_NONEWLINE;
555 h->flags |= HTML_NOSPACE | HTML_NONEWLINE;
558 print_tagq(h, h->metaf);
562 h->flags &= ~HTML_IGNDELIM;
566 print_tagq(struct html *h, const struct tag *until)
570 while ((tag = h->tags.head) != NULL) {
572 if (until && tag == until)
578 print_stagq(struct html *h, const struct tag *suntil)
582 while ((tag = h->tags.head) != NULL) {
583 if (suntil && tag == suntil)
590 print_paragraph(struct html *h)
595 PAIR_CLASS_INIT(&tag, "spacer");
596 t = print_otag(h, TAG_DIV, 1, &tag);
602 bufinit(struct html *h)
610 bufcat_style(struct html *h, const char *key, const char *val)
620 bufcat(struct html *h, const char *p)
624 * XXX This is broken and not easy to fix.
625 * When using the -Oincludes option, buffmt_includes()
626 * may pass in strings overrunning BUFSIZ, causing a crash.
629 h->buflen = strlcat(h->buf, p, BUFSIZ);
630 assert(h->buflen < BUFSIZ);
634 bufcat_fmt(struct html *h, const char *fmt, ...)
639 (void)vsnprintf(h->buf + (int)h->buflen,
640 BUFSIZ - h->buflen - 1, fmt, ap);
642 h->buflen = strlen(h->buf);
646 bufncat(struct html *h, const char *p, size_t sz)
649 assert(h->buflen + sz + 1 < BUFSIZ);
650 strncat(h->buf, p, sz);
655 buffmt_includes(struct html *h, const char *name)
659 pp = h->base_includes;
662 while (NULL != (p = strchr(pp, '%'))) {
663 bufncat(h, pp, (size_t)(p - pp));
679 buffmt_man(struct html *h, const char *name, const char *sec)
686 while (NULL != (p = strchr(pp, '%'))) {
687 bufncat(h, pp, (size_t)(p - pp));
690 bufcat(h, sec ? sec : "1");
693 bufcat_fmt(h, "%s", name);
706 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
711 if (SCALE_MM == su->unit && 0.0 == (v /= 100.0))
713 else if (SCALE_BU == su->unit)
716 bufcat_fmt(h, "%s: %.2f%s;", p, v, roffscales[su->unit]);
720 bufcat_id(struct html *h, const char *src)
723 /* Cf. <http://www.w3.org/TR/html5/dom.html#the-id-attribute>. */
725 for (; '\0' != *src; src++)
726 bufncat(h, *src == ' ' ? "_" : src, 1);