1 /* $Id: html.c,v 1.185 2015/01/21 20:33:25 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 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"
40 #define HTML_CLRLINE (1 << 0)
41 #define HTML_NOSTACK (1 << 1)
42 #define HTML_AUTOCLOSE (1 << 2) /* Tag has auto-closure. */
45 static const struct htmldata htmltags[TAG_MAX] = {
46 {"html", HTML_CLRLINE}, /* TAG_HTML */
47 {"head", HTML_CLRLINE}, /* TAG_HEAD */
48 {"body", HTML_CLRLINE}, /* TAG_BODY */
49 {"meta", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_META */
50 {"title", HTML_CLRLINE}, /* TAG_TITLE */
51 {"div", HTML_CLRLINE}, /* TAG_DIV */
52 {"h1", 0}, /* TAG_H1 */
53 {"h2", 0}, /* TAG_H2 */
54 {"span", 0}, /* TAG_SPAN */
55 {"link", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_LINK */
56 {"br", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_BR */
58 {"table", HTML_CLRLINE}, /* TAG_TABLE */
59 {"tbody", HTML_CLRLINE}, /* TAG_TBODY */
60 {"col", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_COL */
61 {"tr", HTML_CLRLINE}, /* TAG_TR */
62 {"td", HTML_CLRLINE}, /* TAG_TD */
63 {"li", HTML_CLRLINE}, /* TAG_LI */
64 {"ul", HTML_CLRLINE}, /* TAG_UL */
65 {"ol", HTML_CLRLINE}, /* TAG_OL */
66 {"dl", HTML_CLRLINE}, /* TAG_DL */
67 {"dt", HTML_CLRLINE}, /* TAG_DT */
68 {"dd", HTML_CLRLINE}, /* TAG_DD */
69 {"blockquote", HTML_CLRLINE}, /* TAG_BLOCKQUOTE */
70 {"pre", HTML_CLRLINE }, /* TAG_PRE */
71 {"b", 0 }, /* TAG_B */
72 {"i", 0 }, /* TAG_I */
73 {"code", 0 }, /* TAG_CODE */
74 {"small", 0 }, /* TAG_SMALL */
75 {"style", HTML_CLRLINE}, /* TAG_STYLE */
76 {"math", HTML_CLRLINE}, /* TAG_MATH */
77 {"mrow", 0}, /* TAG_MROW */
78 {"mi", 0}, /* TAG_MI */
79 {"mo", 0}, /* TAG_MO */
80 {"msup", 0}, /* TAG_MSUP */
81 {"msub", 0}, /* TAG_MSUB */
82 {"msubsup", 0}, /* TAG_MSUBSUP */
83 {"mfrac", 0}, /* TAG_MFRAC */
84 {"msqrt", 0}, /* TAG_MSQRT */
85 {"mfenced", 0}, /* TAG_MFENCED */
86 {"mtable", 0}, /* TAG_MTABLE */
87 {"mtr", 0}, /* TAG_MTR */
88 {"mtd", 0}, /* TAG_MTD */
89 {"munderover", 0}, /* TAG_MUNDEROVER */
90 {"munder", 0}, /* TAG_MUNDER*/
91 {"mover", 0}, /* TAG_MOVER*/
94 static const char *const htmlattrs[ATTR_MAX] = {
95 "name", /* ATTR_NAME */
97 "href", /* ATTR_HREF */
98 "type", /* ATTR_TYPE */
99 "media", /* ATTR_MEDIA */
100 "class", /* ATTR_CLASS */
101 "style", /* ATTR_STYLE */
103 "colspan", /* ATTR_COLSPAN */
104 "charset", /* ATTR_CHARSET */
105 "open", /* ATTR_OPEN */
106 "close", /* ATTR_CLOSE */
107 "mathvariant", /* ATTR_MATHVARIANT */
110 static const char *const roffscales[SCALE_MAX] = {
123 static void bufncat(struct html *, const char *, size_t);
124 static void print_ctag(struct html *, struct tag *);
125 static int print_escape(char);
126 static int print_encode(struct html *, const char *, int);
127 static void print_metaf(struct html *, enum mandoc_esc);
128 static void print_attr(struct html *, const char *, const char *);
132 html_alloc(const struct mchars *mchars, char *outopts)
140 toks[2] = "includes";
141 toks[3] = "fragment";
144 h = mandoc_calloc(1, sizeof(struct html));
149 while (outopts && *outopts)
150 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
158 h->base_includes = v;
161 h->oflags |= HTML_FRAGMENT;
176 h = (struct html *)p;
178 while ((tag = h->tags.head) != NULL) {
179 h->tags.head = tag->next;
187 print_gen_head(struct html *h)
189 struct htmlpair tag[4];
192 tag[0].key = ATTR_CHARSET;
193 tag[0].val = "utf-8";
194 print_otag(h, TAG_META, 1, tag);
197 * Print a default style-sheet.
199 t = print_otag(h, TAG_STYLE, 0, NULL);
200 print_text(h, "table.head, table.foot { width: 100%; }\n"
201 "td.head-rtitle, td.foot-os { text-align: right; }\n"
202 "td.head-vol { text-align: center; }\n"
203 "table.foot td { width: 50%; }\n"
204 "table.head td { width: 33%; }\n"
205 "div.spacer { margin: 1em 0; }\n");
209 tag[0].key = ATTR_REL;
210 tag[0].val = "stylesheet";
211 tag[1].key = ATTR_HREF;
212 tag[1].val = h->style;
213 tag[2].key = ATTR_TYPE;
214 tag[2].val = "text/css";
215 tag[3].key = ATTR_MEDIA;
217 print_otag(h, TAG_LINK, 4, tag);
222 print_metaf(struct html *h, enum mandoc_esc deco)
227 case ESCAPE_FONTPREV:
230 case ESCAPE_FONTITALIC:
231 font = HTMLFONT_ITALIC;
233 case ESCAPE_FONTBOLD:
234 font = HTMLFONT_BOLD;
241 case ESCAPE_FONTROMAN:
242 font = HTMLFONT_NONE;
250 print_tagq(h, h->metaf);
258 case HTMLFONT_ITALIC:
259 h->metaf = print_otag(h, TAG_I, 0, NULL);
262 h->metaf = print_otag(h, TAG_B, 0, NULL);
265 h->metaf = print_otag(h, TAG_B, 0, NULL);
266 print_otag(h, TAG_I, 0, NULL);
274 html_strlen(const char *cp)
280 * Account for escaped sequences within string length
281 * calculations. This follows the logic in term_strlen() as we
282 * must calculate the width of produced strings.
283 * Assume that characters are always width of "1". This is
284 * hacky, but it gets the job done for approximation of widths.
290 rsz = strcspn(cp, "\\");
302 switch (mandoc_escape(&cp, NULL, NULL)) {
307 case ESCAPE_NUMBERED:
311 case ESCAPE_OVERSTRIKE:
317 case ESCAPE_SKIPCHAR:
359 print_encode(struct html *h, const char *p, int norecurse)
365 static const char rejs[9] = { '\\', '<', '>', '&', '"',
366 ASCII_NBRSP, ASCII_HYPH, ASCII_BREAK, '\0' };
371 if (HTML_SKIPCHAR & h->flags && '\\' != *p) {
372 h->flags &= ~HTML_SKIPCHAR;
377 sz = strcspn(p, rejs);
379 fwrite(p, 1, sz, stdout);
385 if (print_escape(*p++))
388 esc = mandoc_escape(&p, &seq, &len);
389 if (ESCAPE_ERROR == esc)
395 case ESCAPE_FONTPREV:
397 case ESCAPE_FONTBOLD:
399 case ESCAPE_FONTITALIC:
403 case ESCAPE_FONTROMAN:
407 case ESCAPE_SKIPCHAR:
408 h->flags |= HTML_SKIPCHAR;
414 if (h->flags & HTML_SKIPCHAR) {
415 h->flags &= ~HTML_SKIPCHAR;
421 /* Skip past "u" header. */
422 c = mchars_num2uc(seq + 1, len - 1);
424 case ESCAPE_NUMBERED:
425 c = mchars_num2char(seq, len);
430 c = mchars_spec2cp(h->symtab, seq, len);
438 case ESCAPE_OVERSTRIKE:
446 if ((c < 0x20 && c != 0x09) ||
447 (c > 0x7E && c < 0xA0))
451 else if ( ! print_escape(c))
459 print_attr(struct html *h, const char *key, const char *val)
461 printf(" %s=\"", key);
462 (void)print_encode(h, val, 1);
467 print_otag(struct html *h, enum htmltag tag,
468 int sz, const struct htmlpair *p)
473 /* Push this tags onto the stack of open scopes. */
475 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
476 t = mandoc_malloc(sizeof(struct tag));
478 t->next = h->tags.head;
483 if ( ! (HTML_NOSPACE & h->flags))
484 if ( ! (HTML_CLRLINE & htmltags[tag].flags)) {
486 if ( ! (HTML_KEEP & h->flags)) {
487 if (HTML_PREKEEP & h->flags)
488 h->flags |= HTML_KEEP;
494 if ( ! (h->flags & HTML_NONOSPACE))
495 h->flags &= ~HTML_NOSPACE;
497 h->flags |= HTML_NOSPACE;
499 /* Print out the tag name and attributes. */
501 printf("<%s", htmltags[tag].name);
502 for (i = 0; i < sz; i++)
503 print_attr(h, htmlattrs[p[i].key], p[i].val);
505 /* Accommodate for "well-formed" singleton escaping. */
507 if (HTML_AUTOCLOSE & htmltags[tag].flags)
512 h->flags |= HTML_NOSPACE;
514 if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags)
521 print_ctag(struct html *h, struct tag *tag)
525 * Remember to close out and nullify the current
526 * meta-font and table, if applicable.
533 printf("</%s>", htmltags[tag->tag].name);
534 if (HTML_CLRLINE & htmltags[tag->tag].flags) {
535 h->flags |= HTML_NOSPACE;
539 h->tags.head = tag->next;
544 print_gen_decls(struct html *h)
547 puts("<!DOCTYPE html>");
551 print_text(struct html *h, const char *word)
554 if ( ! (HTML_NOSPACE & h->flags)) {
556 if ( ! (HTML_KEEP & h->flags)) {
557 if (HTML_PREKEEP & h->flags)
558 h->flags |= HTML_KEEP;
564 assert(NULL == h->metaf);
566 case HTMLFONT_ITALIC:
567 h->metaf = print_otag(h, TAG_I, 0, NULL);
570 h->metaf = print_otag(h, TAG_B, 0, NULL);
573 h->metaf = print_otag(h, TAG_B, 0, NULL);
574 print_otag(h, TAG_I, 0, NULL);
581 if ( ! print_encode(h, word, 0)) {
582 if ( ! (h->flags & HTML_NONOSPACE))
583 h->flags &= ~HTML_NOSPACE;
584 h->flags &= ~HTML_NONEWLINE;
586 h->flags |= HTML_NOSPACE | HTML_NONEWLINE;
589 print_tagq(h, h->metaf);
593 h->flags &= ~HTML_IGNDELIM;
597 print_tagq(struct html *h, const struct tag *until)
601 while ((tag = h->tags.head) != NULL) {
603 if (until && tag == until)
609 print_stagq(struct html *h, const struct tag *suntil)
613 while ((tag = h->tags.head) != NULL) {
614 if (suntil && tag == suntil)
621 print_paragraph(struct html *h)
626 PAIR_CLASS_INIT(&tag, "spacer");
627 t = print_otag(h, TAG_DIV, 1, &tag);
633 bufinit(struct html *h)
641 bufcat_style(struct html *h, const char *key, const char *val)
651 bufcat(struct html *h, const char *p)
655 * XXX This is broken and not easy to fix.
656 * When using the -Oincludes option, buffmt_includes()
657 * may pass in strings overrunning BUFSIZ, causing a crash.
660 h->buflen = strlcat(h->buf, p, BUFSIZ);
661 assert(h->buflen < BUFSIZ);
665 bufcat_fmt(struct html *h, const char *fmt, ...)
670 (void)vsnprintf(h->buf + (int)h->buflen,
671 BUFSIZ - h->buflen - 1, fmt, ap);
673 h->buflen = strlen(h->buf);
677 bufncat(struct html *h, const char *p, size_t sz)
680 assert(h->buflen + sz + 1 < BUFSIZ);
681 strncat(h->buf, p, sz);
686 buffmt_includes(struct html *h, const char *name)
690 pp = h->base_includes;
693 while (NULL != (p = strchr(pp, '%'))) {
694 bufncat(h, pp, (size_t)(p - pp));
710 buffmt_man(struct html *h, const char *name, const char *sec)
717 while (NULL != (p = strchr(pp, '%'))) {
718 bufncat(h, pp, (size_t)(p - pp));
721 bufcat(h, sec ? sec : "1");
724 bufcat_fmt(h, "%s", name);
737 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
742 if (SCALE_MM == su->unit && 0.0 == (v /= 100.0))
744 else if (SCALE_BU == su->unit)
747 bufcat_fmt(h, "%s: %.2f%s;", p, v, roffscales[su->unit]);
751 bufcat_id(struct html *h, const char *src)
754 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
757 bufcat_fmt(h, "%.2x", *src++);