]>
git.cameronkatri.com Git - mandoc.git/blob - html.c
5a63ea5db78bd21369b64258255f44034252eedb
1 /* $Id: html.c,v 1.115 2010/12/15 15:59:23 kristaps 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 */
73 static const char *const htmlfonts
[HTMLFONT_MAX
] = {
79 static const char *const htmlattrs
[ATTR_MAX
] = {
96 static void print_spec(struct html
*, enum roffdeco
,
97 const char *, size_t);
98 static void print_res(struct html
*, const char *, size_t);
99 static void print_ctag(struct html
*, enum htmltag
);
100 static void print_doctype(struct html
*);
101 static void print_xmltype(struct html
*);
102 static int print_encode(struct html
*, const char *, int);
103 static void print_metaf(struct html
*, enum roffdeco
);
104 static void print_attr(struct html
*,
105 const char *, const char *);
106 static void *ml_alloc(char *, enum htmltype
);
110 ml_alloc(char *outopts
, enum htmltype type
)
118 toks
[2] = "includes";
121 h
= calloc(1, sizeof(struct html
));
124 exit((int)MANDOCLEVEL_SYSERR
);
129 h
->symtab
= chars_init(CHARS_HTML
);
131 while (outopts
&& *outopts
)
132 switch (getsubopt(&outopts
, UNCONST(toks
), &v
)) {
140 h
->base_includes
= v
;
150 html_alloc(char *outopts
)
153 return(ml_alloc(outopts
, HTML_HTML_4_01_STRICT
));
158 xhtml_alloc(char *outopts
)
161 return(ml_alloc(outopts
, HTML_XHTML_1_0_STRICT
));
171 h
= (struct html
*)p
;
173 while ((tag
= h
->tags
.head
) != NULL
) {
174 h
->tags
.head
= tag
->next
;
179 chars_free(h
->symtab
);
186 print_gen_head(struct html
*h
)
188 struct htmlpair tag
[4];
190 tag
[0].key
= ATTR_HTTPEQUIV
;
191 tag
[0].val
= "Content-Type";
192 tag
[1].key
= ATTR_CONTENT
;
193 tag
[1].val
= "text/html; charset=utf-8";
194 print_otag(h
, TAG_META
, 2, tag
);
196 tag
[0].key
= ATTR_NAME
;
197 tag
[0].val
= "resource-type";
198 tag
[1].key
= ATTR_CONTENT
;
199 tag
[1].val
= "document";
200 print_otag(h
, TAG_META
, 2, tag
);
203 tag
[0].key
= ATTR_REL
;
204 tag
[0].val
= "stylesheet";
205 tag
[1].key
= ATTR_HREF
;
206 tag
[1].val
= h
->style
;
207 tag
[2].key
= ATTR_TYPE
;
208 tag
[2].val
= "text/css";
209 tag
[3].key
= ATTR_MEDIA
;
211 print_otag(h
, TAG_LINK
, 4, tag
);
217 print_spec(struct html
*h
, enum roffdeco d
, const char *p
, size_t len
)
223 if ((cp
= chars_spec2cp(h
->symtab
, p
, len
)) > 0) {
226 } else if (-1 == cp
&& DECO_SSPECIAL
== d
) {
227 fwrite(p
, 1, len
, stdout
);
232 if (NULL
!= (rhs
= chars_spec2str(h
->symtab
, p
, len
, &sz
)))
233 fwrite(rhs
, 1, sz
, stdout
);
238 print_res(struct html
*h
, const char *p
, size_t len
)
244 if ((cp
= chars_res2cp(h
->symtab
, p
, len
)) > 0) {
250 if (NULL
!= (rhs
= chars_res2str(h
->symtab
, p
, len
, &sz
)))
251 fwrite(rhs
, 1, sz
, stdout
);
256 print_ofont(struct html
*h
, enum htmlfont font
)
263 /* FIXME: DECO_ROMAN should just close out preexisting. */
265 if (h
->metaf
&& h
->tags
.head
== h
->metaf
)
266 print_tagq(h
, h
->metaf
);
268 PAIR_CLASS_INIT(&tag
, htmlfonts
[font
]);
269 h
->metaf
= print_otag(h
, TAG_SPAN
, 1, &tag
);
275 print_metaf(struct html
*h
, enum roffdeco deco
)
280 case (DECO_PREVIOUS
):
284 font
= HTMLFONT_ITALIC
;
287 font
= HTMLFONT_BOLD
;
290 font
= HTMLFONT_NONE
;
297 (void)print_ofont(h
, font
);
302 print_encode(struct html
*h
, const char *p
, int norecurse
)
308 static const char rejs
[6] = { '\\', '<', '>', '&', ASCII_HYPH
, '\0' };
313 sz
= strcspn(p
, rejs
);
315 fwrite(p
, 1, sz
, stdout
);
322 } else if ('>' == *p
) {
325 } else if ('&' == *p
) {
328 } else if (ASCII_HYPH
== *p
) {
330 * Note: "soft hyphens" aren't graphically
331 * displayed when not breaking the text; we want
332 * them to be displayed.
334 /*printf("­");*/
337 } else if ('\0' == *p
)
341 len
= a2roffdeco(&deco
, &seq
, &sz
);
344 case (DECO_RESERVED
):
345 print_res(h
, seq
, sz
);
347 case (DECO_SSPECIAL
):
350 print_spec(h
, deco
, seq
, sz
);
352 case (DECO_PREVIOUS
):
361 print_metaf(h
, deco
);
369 if (DECO_NOSPACE
== deco
&& '\0' == *(p
+ 1))
378 print_attr(struct html
*h
, const char *key
, const char *val
)
380 printf(" %s=\"", key
);
381 (void)print_encode(h
, val
, 1);
387 print_otag(struct html
*h
, enum htmltag tag
,
388 int sz
, const struct htmlpair
*p
)
393 /* Push this tags onto the stack of open scopes. */
395 if ( ! (HTML_NOSTACK
& htmltags
[tag
].flags
)) {
396 t
= malloc(sizeof(struct tag
));
399 exit((int)MANDOCLEVEL_SYSERR
);
402 t
->next
= h
->tags
.head
;
407 if ( ! (HTML_NOSPACE
& h
->flags
))
408 if ( ! (HTML_CLRLINE
& htmltags
[tag
].flags
)) {
410 if ( ! (HTML_KEEP
& h
->flags
)) {
411 if (HTML_PREKEEP
& h
->flags
)
412 h
->flags
|= HTML_KEEP
;
418 if ( ! (h
->flags
& HTML_NONOSPACE
))
419 h
->flags
&= ~HTML_NOSPACE
;
421 h
->flags
|= HTML_NOSPACE
;
423 /* Print out the tag name and attributes. */
425 printf("<%s", htmltags
[tag
].name
);
426 for (i
= 0; i
< sz
; i
++)
427 print_attr(h
, htmlattrs
[p
[i
].key
], p
[i
].val
);
429 /* Add non-overridable attributes. */
431 if (TAG_HTML
== tag
&& HTML_XHTML_1_0_STRICT
== h
->type
) {
432 print_attr(h
, "xmlns", "http://www.w3.org/1999/xhtml");
433 print_attr(h
, "xml:lang", "en");
434 print_attr(h
, "lang", "en");
437 /* Accomodate for XML "well-formed" singleton escaping. */
439 if (HTML_AUTOCLOSE
& htmltags
[tag
].flags
)
441 case (HTML_XHTML_1_0_STRICT
):
450 h
->flags
|= HTML_NOSPACE
;
456 print_ctag(struct html
*h
, enum htmltag tag
)
459 printf("</%s>", htmltags
[tag
].name
);
460 if (HTML_CLRLINE
& htmltags
[tag
].flags
) {
461 h
->flags
|= HTML_NOSPACE
;
468 print_gen_decls(struct html
*h
)
477 print_xmltype(struct html
*h
)
480 if (HTML_XHTML_1_0_STRICT
== h
->type
)
481 printf("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
486 print_doctype(struct html
*h
)
493 case (HTML_HTML_4_01_STRICT
):
495 doctype
= "-//W3C//DTD HTML 4.01//EN";
496 dtd
= "http://www.w3.org/TR/html4/strict.dtd";
500 doctype
= "-//W3C//DTD XHTML 1.0 Strict//EN";
501 dtd
= "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
505 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
511 print_text(struct html
*h
, const char *word
)
514 if (word
[0] && '\0' == word
[1])
531 if ( ! (HTML_IGNDELIM
& h
->flags
))
532 h
->flags
|= HTML_NOSPACE
;
538 if ( ! (HTML_NOSPACE
& h
->flags
)) {
540 if ( ! (HTML_KEEP
& h
->flags
)) {
541 if (HTML_PREKEEP
& h
->flags
)
542 h
->flags
|= HTML_KEEP
;
549 if ( ! print_encode(h
, word
, 0))
550 if ( ! (h
->flags
& HTML_NONOSPACE
))
551 h
->flags
&= ~HTML_NOSPACE
;
553 h
->flags
&= ~HTML_IGNDELIM
;
556 * Note that we don't process the pipe: the parser sees it as
557 * punctuation, but we don't in terms of typography.
559 if (word
[0] && '\0' == word
[1])
564 h
->flags
|= HTML_NOSPACE
;
573 print_tagq(struct html
*h
, const struct tag
*until
)
577 while ((tag
= h
->tags
.head
) != NULL
) {
580 print_ctag(h
, tag
->tag
);
581 h
->tags
.head
= tag
->next
;
583 if (until
&& tag
== until
)
590 print_stagq(struct html
*h
, const struct tag
*suntil
)
594 while ((tag
= h
->tags
.head
) != NULL
) {
595 if (suntil
&& tag
== suntil
)
599 print_ctag(h
, tag
->tag
);
600 h
->tags
.head
= tag
->next
;
607 bufinit(struct html
*h
)
616 bufcat_style(struct html
*h
, const char *key
, const char *val
)
627 bufcat(struct html
*h
, const char *p
)
630 bufncat(h
, p
, strlen(p
));
635 buffmt(struct html
*h
, const char *fmt
, ...)
640 (void)vsnprintf(h
->buf
+ (int)h
->buflen
,
641 BUFSIZ
- h
->buflen
- 1, fmt
, ap
);
643 h
->buflen
= strlen(h
->buf
);
648 bufncat(struct html
*h
, const char *p
, size_t sz
)
651 if (h
->buflen
+ sz
> BUFSIZ
- 1)
652 sz
= BUFSIZ
- 1 - h
->buflen
;
654 (void)strncat(h
->buf
, p
, sz
);
660 buffmt_includes(struct html
*h
, const char *name
)
664 pp
= h
->base_includes
;
666 while (NULL
!= (p
= strchr(pp
, '%'))) {
667 bufncat(h
, pp
, (size_t)(p
- pp
));
684 buffmt_man(struct html
*h
,
685 const char *name
, const char *sec
)
692 while (NULL
!= (p
= strchr(pp
, '%'))) {
693 bufncat(h
, pp
, (size_t)(p
- pp
));
696 bufcat(h
, sec
? sec
: "1");
713 bufcat_su(struct html
*h
, const char *p
, const struct roffsu
*su
)
756 * XXX: the CSS spec isn't clear as to which types accept
757 * integer or real numbers, so we just make them all decimals.
759 buffmt(h
, "%s: %.2f%s;", p
, v
, u
);
764 html_idcat(char *dst
, const char *src
, int sz
)
770 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
772 for ( ; *dst
!= '\0' && sz
; dst
++, sz
--)
777 /* We can't start with a number (bah). */
783 for ( ; *src
!= '\0' && sz
> 1; src
++) {
784 ssz
= snprintf(dst
, (size_t)sz
, "%.2x", *src
);