1 /* $Id: html.c,v 1.124 2010/12/27 21:41:05 schwarze Exp $ */
3 * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
5 * Permission to use, copy, modify, and distribute this software for any
6 * purpose with or without fee is hereby granted, provided that the above
7 * copyright notice and this permission notice appear in all copies.
9 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
10 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
11 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
12 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
13 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
14 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
15 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
21 #include <sys/types.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 {"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 */
79 static const char *const htmlattrs[ATTR_MAX] = {
80 "http-equiv", /* ATTR_HTTPEQUIV */
81 "content", /* ATTR_CONTENT */
82 "name", /* ATTR_NAME */
84 "href", /* ATTR_HREF */
85 "type", /* ATTR_TYPE */
86 "media", /* ATTR_MEDIA */
87 "class", /* ATTR_CLASS */
88 "style", /* ATTR_STYLE */
89 "width", /* ATTR_WIDTH */
91 "summary", /* ATTR_SUMMARY */
92 "align", /* ATTR_ALIGN */
95 static void print_spec(struct html *, enum roffdeco,
96 const char *, size_t);
97 static void print_res(struct html *, const char *, size_t);
98 static void print_ctag(struct html *, enum htmltag);
99 static void print_doctype(struct html *);
100 static void print_xmltype(struct html *);
101 static int print_encode(struct html *, const char *, int);
102 static void print_metaf(struct html *, enum roffdeco);
103 static void print_attr(struct html *,
104 const char *, const char *);
105 static void *ml_alloc(char *, enum htmltype);
109 ml_alloc(char *outopts, enum htmltype type)
117 toks[2] = "includes";
120 h = calloc(1, sizeof(struct html));
123 exit((int)MANDOCLEVEL_SYSERR);
128 h->symtab = chars_init(CHARS_HTML);
130 while (outopts && *outopts)
131 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
139 h->base_includes = v;
149 html_alloc(char *outopts)
152 return(ml_alloc(outopts, HTML_HTML_4_01_STRICT));
157 xhtml_alloc(char *outopts)
160 return(ml_alloc(outopts, HTML_XHTML_1_0_STRICT));
170 h = (struct html *)p;
172 while ((tag = h->tags.head) != NULL) {
173 h->tags.head = tag->next;
178 chars_free(h->symtab);
185 print_gen_head(struct html *h)
187 struct htmlpair tag[4];
189 tag[0].key = ATTR_HTTPEQUIV;
190 tag[0].val = "Content-Type";
191 tag[1].key = ATTR_CONTENT;
192 tag[1].val = "text/html; charset=utf-8";
193 print_otag(h, TAG_META, 2, tag);
195 tag[0].key = ATTR_NAME;
196 tag[0].val = "resource-type";
197 tag[1].key = ATTR_CONTENT;
198 tag[1].val = "document";
199 print_otag(h, TAG_META, 2, tag);
202 tag[0].key = ATTR_REL;
203 tag[0].val = "stylesheet";
204 tag[1].key = ATTR_HREF;
205 tag[1].val = h->style;
206 tag[2].key = ATTR_TYPE;
207 tag[2].val = "text/css";
208 tag[3].key = ATTR_MEDIA;
210 print_otag(h, TAG_LINK, 4, tag);
216 print_spec(struct html *h, enum roffdeco d, const char *p, size_t len)
222 if ((cp = chars_spec2cp(h->symtab, p, len)) > 0) {
225 } else if (-1 == cp && DECO_SSPECIAL == d) {
226 fwrite(p, 1, len, stdout);
231 if (NULL != (rhs = chars_spec2str(h->symtab, p, len, &sz)))
232 fwrite(rhs, 1, sz, stdout);
237 print_res(struct html *h, const char *p, size_t len)
243 if ((cp = chars_res2cp(h->symtab, p, len)) > 0) {
249 if (NULL != (rhs = chars_res2str(h->symtab, p, len, &sz)))
250 fwrite(rhs, 1, sz, stdout);
255 print_metaf(struct html *h, enum roffdeco deco)
260 case (DECO_PREVIOUS):
264 font = HTMLFONT_ITALIC;
267 font = HTMLFONT_BOLD;
270 font = HTMLFONT_NONE;
278 print_tagq(h, h->metaf);
285 if (HTMLFONT_NONE != font)
286 h->metaf = HTMLFONT_BOLD == font ?
287 print_otag(h, TAG_B, 0, NULL) :
288 print_otag(h, TAG_I, 0, NULL);
293 print_encode(struct html *h, const char *p, int norecurse)
299 static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' };
304 sz = strcspn(p, rejs);
306 fwrite(p, 1, sz, stdout);
313 } else if ('>' == *p) {
316 } else if ('&' == *p) {
319 } else if (ASCII_HYPH == *p) {
321 * Note: "soft hyphens" aren't graphically
322 * displayed when not breaking the text; we want
323 * them to be displayed.
325 /*printf("­");*/
328 } else if ('\0' == *p)
332 len = a2roffdeco(&deco, &seq, &sz);
335 case (DECO_RESERVED):
336 print_res(h, seq, sz);
338 case (DECO_SSPECIAL):
341 print_spec(h, deco, seq, sz);
343 case (DECO_PREVIOUS):
352 print_metaf(h, deco);
360 if (DECO_NOSPACE == deco && '\0' == *(p + 1))
369 print_attr(struct html *h, const char *key, const char *val)
371 printf(" %s=\"", key);
372 (void)print_encode(h, val, 1);
378 print_otag(struct html *h, enum htmltag tag,
379 int sz, const struct htmlpair *p)
384 /* Push this tags onto the stack of open scopes. */
386 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
387 t = malloc(sizeof(struct tag));
390 exit((int)MANDOCLEVEL_SYSERR);
393 t->next = h->tags.head;
398 if ( ! (HTML_NOSPACE & h->flags))
399 if ( ! (HTML_CLRLINE & htmltags[tag].flags)) {
401 if ( ! (HTML_KEEP & h->flags)) {
402 if (HTML_PREKEEP & h->flags)
403 h->flags |= HTML_KEEP;
409 if ( ! (h->flags & HTML_NONOSPACE))
410 h->flags &= ~HTML_NOSPACE;
412 h->flags |= HTML_NOSPACE;
414 /* Print out the tag name and attributes. */
416 printf("<%s", htmltags[tag].name);
417 for (i = 0; i < sz; i++)
418 print_attr(h, htmlattrs[p[i].key], p[i].val);
420 /* Add non-overridable attributes. */
422 if (TAG_HTML == tag && HTML_XHTML_1_0_STRICT == h->type) {
423 print_attr(h, "xmlns", "http://www.w3.org/1999/xhtml");
424 print_attr(h, "xml:lang", "en");
425 print_attr(h, "lang", "en");
428 /* Accomodate for XML "well-formed" singleton escaping. */
430 if (HTML_AUTOCLOSE & htmltags[tag].flags)
432 case (HTML_XHTML_1_0_STRICT):
441 h->flags |= HTML_NOSPACE;
443 if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags)
451 print_ctag(struct html *h, enum htmltag tag)
454 printf("</%s>", htmltags[tag].name);
455 if (HTML_CLRLINE & htmltags[tag].flags) {
456 h->flags |= HTML_NOSPACE;
463 print_gen_decls(struct html *h)
472 print_xmltype(struct html *h)
475 if (HTML_XHTML_1_0_STRICT == h->type)
476 puts("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
481 print_doctype(struct html *h)
488 case (HTML_HTML_4_01_STRICT):
490 doctype = "-//W3C//DTD HTML 4.01//EN";
491 dtd = "http://www.w3.org/TR/html4/strict.dtd";
495 doctype = "-//W3C//DTD XHTML 1.0 Strict//EN";
496 dtd = "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
500 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
506 print_text(struct html *h, const char *word)
509 if (word[0] && '\0' == word[1])
526 if ( ! (HTML_IGNDELIM & h->flags))
527 h->flags |= HTML_NOSPACE;
533 if ( ! (HTML_NOSPACE & h->flags)) {
535 if ( ! (HTML_KEEP & h->flags)) {
536 if (HTML_PREKEEP & h->flags)
537 h->flags |= HTML_KEEP;
543 assert(NULL == h->metaf);
544 if (HTMLFONT_NONE != h->metac)
545 h->metaf = HTMLFONT_BOLD == h->metac ?
546 print_otag(h, TAG_B, 0, NULL) :
547 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;
555 print_tagq(h, h->metaf);
559 h->flags &= ~HTML_IGNDELIM;
562 * Note that we don't process the pipe: the parser sees it as
563 * punctuation, but we don't in terms of typography.
565 if (word[0] && '\0' == word[1])
570 h->flags |= HTML_NOSPACE;
579 print_tagq(struct html *h, const struct tag *until)
583 while ((tag = h->tags.head) != NULL) {
586 print_ctag(h, tag->tag);
587 h->tags.head = tag->next;
589 if (until && tag == until)
596 print_stagq(struct html *h, const struct tag *suntil)
600 while ((tag = h->tags.head) != NULL) {
601 if (suntil && tag == suntil)
605 print_ctag(h, tag->tag);
606 h->tags.head = tag->next;
613 bufinit(struct html *h)
622 bufcat_style(struct html *h, const char *key, const char *val)
633 bufcat(struct html *h, const char *p)
636 bufncat(h, p, strlen(p));
641 buffmt(struct html *h, const char *fmt, ...)
646 (void)vsnprintf(h->buf + (int)h->buflen,
647 BUFSIZ - h->buflen - 1, fmt, ap);
649 h->buflen = strlen(h->buf);
654 bufncat(struct html *h, const char *p, size_t sz)
657 if (h->buflen + sz > BUFSIZ - 1)
658 sz = BUFSIZ - 1 - h->buflen;
660 (void)strncat(h->buf, p, sz);
666 buffmt_includes(struct html *h, const char *name)
670 pp = h->base_includes;
672 while (NULL != (p = strchr(pp, '%'))) {
673 bufncat(h, pp, (size_t)(p - pp));
690 buffmt_man(struct html *h,
691 const char *name, const char *sec)
698 while (NULL != (p = strchr(pp, '%'))) {
699 bufncat(h, pp, (size_t)(p - pp));
702 bufcat(h, sec ? sec : "1");
719 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
762 * XXX: the CSS spec isn't clear as to which types accept
763 * integer or real numbers, so we just make them all decimals.
765 buffmt(h, "%s: %.2f%s;", p, v, u);
770 html_idcat(char *dst, const char *src, int sz)
776 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
778 /* We can't start with a number (bah). */
790 for ( ; *dst != '\0' && sz; dst++, sz--)
793 for ( ; *src != '\0' && sz > 1; src++) {
794 ssz = snprintf(dst, (size_t)sz, "%.2x", *src);