1 /* $Id: html.c,v 1.80 2009/11/02 06:22:44 kristaps Exp $ */
3 * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
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.
17 #include <sys/types.h>
33 #define UNCONST(a) ((void *)(uintptr_t)(const void *)(a))
35 #define DOCTYPE "-//W3C//DTD HTML 4.01//EN"
36 #define DTD "http://www.w3.org/TR/html4/strict.dtd"
41 #define HTML_CLRLINE (1 << 0)
42 #define HTML_NOSTACK (1 << 1)
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}, /* 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 {"p", HTML_CLRLINE}, /* TAG_P */
55 {"span", 0}, /* TAG_SPAN */
56 {"link", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
57 {"br", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
59 {"table", HTML_CLRLINE}, /* TAG_TABLE */
60 {"col", HTML_CLRLINE | HTML_NOSTACK}, /* 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 {"base", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_BASE */
69 static const char *const htmlattrs[ATTR_MAX] = {
87 extern int getsubopt(char **, char * const *, char **);
91 html_alloc(char *outopts)
102 h = calloc(1, sizeof(struct html));
110 h->symtab = chars_init(CHARS_HTML);
112 while (outopts && *outopts)
113 switch (getsubopt(&outopts, UNCONST(toks), &v)) {
121 h->base_includes = v;
138 h = (struct html *)p;
140 while ((ord = h->ords.head) != NULL) {
141 h->ords.head = ord->next;
145 while ((tag = h->tags.head) != NULL) {
146 h->tags.head = tag->next;
151 chars_free(h->symtab);
158 print_gen_head(struct html *h)
160 struct htmlpair tag[4];
162 tag[0].key = ATTR_HTTPEQUIV;
163 tag[0].val = "Content-Type";
164 tag[1].key = ATTR_CONTENT;
165 tag[1].val = "text/html; charset=utf-8";
166 print_otag(h, TAG_META, 2, tag);
168 tag[0].key = ATTR_NAME;
169 tag[0].val = "resource-type";
170 tag[1].key = ATTR_CONTENT;
171 tag[1].val = "document";
172 print_otag(h, TAG_META, 2, tag);
175 tag[0].key = ATTR_REL;
176 tag[0].val = "stylesheet";
177 tag[1].key = ATTR_HREF;
178 tag[1].val = h->style;
179 tag[2].key = ATTR_TYPE;
180 tag[2].val = "text/css";
181 tag[3].key = ATTR_MEDIA;
183 print_otag(h, TAG_LINK, 4, tag);
189 print_spec(struct html *h, const char *p, int len)
194 rhs = chars_a2ascii(h->symtab, p, (size_t)len, &sz);
198 fwrite(rhs, 1, sz, stdout);
203 print_res(struct html *h, const char *p, int len)
208 rhs = chars_a2res(h->symtab, p, (size_t)len, &sz);
212 fwrite(rhs, 1, sz, stdout);
217 print_escape(struct html *h, const char **p)
232 if (0 == *wp || 0 == *(wp + 1)) {
233 *p = 0 == *wp ? wp : wp + 1;
237 print_spec(h, wp, 2);
241 } else if ('*' == *wp) {
250 if (0 == *wp || 0 == *(wp + 1)) {
251 *p = 0 == *wp ? wp : wp + 1;
267 } else if ('f' == *wp) {
292 } else if ('[' != *wp) {
293 print_spec(h, wp, 1);
299 for (j = 0; *wp && ']' != *wp; wp++, j++)
308 print_spec(h, wp - j, j);
310 print_res(h, wp - j, j);
317 print_encode(struct html *h, const char *p)
322 sz = strcspn(p, "\\<>&");
324 fwrite(p, 1, sz, stdout);
331 } else if ('\0' == *p)
345 print_otag(struct html *h, enum htmltag tag,
346 int sz, const struct htmlpair *p)
351 if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
352 t = malloc(sizeof(struct tag));
358 t->next = h->tags.head;
363 if ( ! (HTML_NOSPACE & h->flags))
364 if ( ! (HTML_CLRLINE & htmltags[tag].flags))
367 printf("<%s", htmltags[tag].name);
368 for (i = 0; i < sz; i++) {
369 printf(" %s=\"", htmlattrs[p[i].key]);
371 print_encode(h, p[i].val);
376 h->flags |= HTML_NOSPACE;
377 if (HTML_CLRLINE & htmltags[tag].flags)
378 h->flags |= HTML_NEWLINE;
380 h->flags &= ~HTML_NEWLINE;
388 print_ctag(struct html *h, enum htmltag tag)
391 printf("</%s>", htmltags[tag].name);
392 if (HTML_CLRLINE & htmltags[tag].flags) {
393 h->flags |= HTML_NOSPACE;
394 h->flags |= HTML_NEWLINE;
397 h->flags &= ~HTML_NEWLINE;
403 print_gen_doctype(struct html *h)
406 printf("<!DOCTYPE HTML PUBLIC \"%s\" \"%s\">", DOCTYPE, DTD);
411 print_text(struct html *h, const char *p)
414 if (*p && 0 == *(p + 1))
433 if ( ! (HTML_IGNDELIM & h->flags))
434 h->flags |= HTML_NOSPACE;
440 if ( ! (h->flags & HTML_NOSPACE))
443 h->flags &= ~HTML_NOSPACE;
444 h->flags &= ~HTML_NEWLINE;
449 if (*p && 0 == *(p + 1))
456 h->flags |= HTML_NOSPACE;
465 print_tagq(struct html *h, const struct tag *until)
469 while ((tag = h->tags.head) != NULL) {
470 print_ctag(h, tag->tag);
471 h->tags.head = tag->next;
473 if (until && tag == until)
480 print_stagq(struct html *h, const struct tag *suntil)
484 while ((tag = h->tags.head) != NULL) {
485 if (suntil && tag == suntil)
487 print_ctag(h, tag->tag);
488 h->tags.head = tag->next;
495 bufinit(struct html *h)
504 bufcat_style(struct html *h, const char *key, const char *val)
515 bufcat(struct html *h, const char *p)
518 bufncat(h, p, strlen(p));
523 buffmt(struct html *h, const char *fmt, ...)
528 (void)vsnprintf(h->buf + (int)h->buflen,
529 BUFSIZ - h->buflen - 1, fmt, ap);
531 h->buflen = strlen(h->buf);
536 bufncat(struct html *h, const char *p, size_t sz)
539 if (h->buflen + sz > BUFSIZ - 1)
540 sz = BUFSIZ - 1 - h->buflen;
542 (void)strncat(h->buf, p, sz);
548 buffmt_includes(struct html *h, const char *name)
552 pp = h->base_includes;
554 while (NULL != (p = strchr(pp, '%'))) {
555 bufncat(h, pp, (size_t)(p - pp));
572 buffmt_man(struct html *h,
573 const char *name, const char *sec)
580 while (NULL != (p = strchr(pp, '%'))) {
581 bufncat(h, pp, (size_t)(p - pp));
584 bufcat(h, sec ? sec : "1");
601 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
644 buffmt(h, "%s: %f%s;", p, v, u);
647 buffmt(h, "%s: %d%s;", p, (int)v, u);
652 html_idcat(char *dst, const char *src, int sz)
658 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
660 for ( ; *dst != '\0' && sz; dst++, sz--)
665 /* We can't start with a number (bah). */
671 for ( ; *src != '\0' && sz > 1; src++) {
672 ssz = snprintf(dst, (size_t)sz, "%.2x", *src);