]>
git.cameronkatri.com Git - mandoc.git/blob - html.c
d99bd58fbdc28e083b30567fa0481cddda98c3db
1 /* $Id: html.c,v 1.121 2010/12/22 11:15:16 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 */
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 {"u", 0 }, /* TAG_U */
76 {"code", 0 }, /* TAG_CODE */
79 static const char *const htmlfonts
[HTMLFONT_MAX
] = {
85 static const char *const htmlattrs
[ATTR_MAX
] = {
86 "http-equiv", /* ATTR_HTTPEQUIV */
87 "content", /* ATTR_CONTENT */
88 "name", /* ATTR_NAME */
90 "href", /* ATTR_HREF */
91 "type", /* ATTR_TYPE */
92 "media", /* ATTR_MEDIA */
93 "class", /* ATTR_CLASS */
94 "style", /* ATTR_STYLE */
95 "width", /* ATTR_WIDTH */
97 "summary", /* ATTR_SUMMARY */
98 "align", /* ATTR_ALIGN */
101 static void print_spec(struct html
*, enum roffdeco
,
102 const char *, size_t);
103 static void print_res(struct html
*, const char *, size_t);
104 static void print_ctag(struct html
*, enum htmltag
);
105 static void print_doctype(struct html
*);
106 static void print_xmltype(struct html
*);
107 static int print_encode(struct html
*, const char *, int);
108 static void print_metaf(struct html
*, enum roffdeco
);
109 static void print_attr(struct html
*,
110 const char *, const char *);
111 static void *ml_alloc(char *, enum htmltype
);
115 ml_alloc(char *outopts
, enum htmltype type
)
123 toks
[2] = "includes";
126 h
= calloc(1, sizeof(struct html
));
129 exit((int)MANDOCLEVEL_SYSERR
);
134 h
->symtab
= chars_init(CHARS_HTML
);
136 while (outopts
&& *outopts
)
137 switch (getsubopt(&outopts
, UNCONST(toks
), &v
)) {
145 h
->base_includes
= v
;
155 html_alloc(char *outopts
)
158 return(ml_alloc(outopts
, HTML_HTML_4_01_STRICT
));
163 xhtml_alloc(char *outopts
)
166 return(ml_alloc(outopts
, HTML_XHTML_1_0_STRICT
));
176 h
= (struct html
*)p
;
178 while ((tag
= h
->tags
.head
) != NULL
) {
179 h
->tags
.head
= tag
->next
;
184 chars_free(h
->symtab
);
191 print_gen_head(struct html
*h
)
193 struct htmlpair tag
[4];
195 tag
[0].key
= ATTR_HTTPEQUIV
;
196 tag
[0].val
= "Content-Type";
197 tag
[1].key
= ATTR_CONTENT
;
198 tag
[1].val
= "text/html; charset=utf-8";
199 print_otag(h
, TAG_META
, 2, tag
);
201 tag
[0].key
= ATTR_NAME
;
202 tag
[0].val
= "resource-type";
203 tag
[1].key
= ATTR_CONTENT
;
204 tag
[1].val
= "document";
205 print_otag(h
, TAG_META
, 2, tag
);
208 tag
[0].key
= ATTR_REL
;
209 tag
[0].val
= "stylesheet";
210 tag
[1].key
= ATTR_HREF
;
211 tag
[1].val
= h
->style
;
212 tag
[2].key
= ATTR_TYPE
;
213 tag
[2].val
= "text/css";
214 tag
[3].key
= ATTR_MEDIA
;
216 print_otag(h
, TAG_LINK
, 4, tag
);
222 print_spec(struct html
*h
, enum roffdeco d
, const char *p
, size_t len
)
228 if ((cp
= chars_spec2cp(h
->symtab
, p
, len
)) > 0) {
231 } else if (-1 == cp
&& DECO_SSPECIAL
== d
) {
232 fwrite(p
, 1, len
, stdout
);
237 if (NULL
!= (rhs
= chars_spec2str(h
->symtab
, p
, len
, &sz
)))
238 fwrite(rhs
, 1, sz
, stdout
);
243 print_res(struct html
*h
, const char *p
, size_t len
)
249 if ((cp
= chars_res2cp(h
->symtab
, p
, len
)) > 0) {
255 if (NULL
!= (rhs
= chars_res2str(h
->symtab
, p
, len
, &sz
)))
256 fwrite(rhs
, 1, sz
, stdout
);
261 print_ofont(struct html
*h
, enum htmlfont font
)
268 /* FIXME: DECO_ROMAN should just close out preexisting. */
270 if (h
->metaf
&& h
->tags
.head
== h
->metaf
)
271 print_tagq(h
, h
->metaf
);
273 PAIR_CLASS_INIT(&tag
, htmlfonts
[font
]);
274 h
->metaf
= print_otag(h
, TAG_SPAN
, 1, &tag
);
280 print_metaf(struct html
*h
, enum roffdeco deco
)
285 case (DECO_PREVIOUS
):
289 font
= HTMLFONT_ITALIC
;
292 font
= HTMLFONT_BOLD
;
295 font
= HTMLFONT_NONE
;
302 (void)print_ofont(h
, font
);
307 print_encode(struct html
*h
, const char *p
, int norecurse
)
313 static const char rejs
[6] = { '\\', '<', '>', '&', ASCII_HYPH
, '\0' };
318 sz
= strcspn(p
, rejs
);
320 fwrite(p
, 1, sz
, stdout
);
327 } else if ('>' == *p
) {
330 } else if ('&' == *p
) {
333 } else if (ASCII_HYPH
== *p
) {
335 * Note: "soft hyphens" aren't graphically
336 * displayed when not breaking the text; we want
337 * them to be displayed.
339 /*printf("­");*/
342 } else if ('\0' == *p
)
346 len
= a2roffdeco(&deco
, &seq
, &sz
);
349 case (DECO_RESERVED
):
350 print_res(h
, seq
, sz
);
352 case (DECO_SSPECIAL
):
355 print_spec(h
, deco
, seq
, sz
);
357 case (DECO_PREVIOUS
):
366 print_metaf(h
, deco
);
374 if (DECO_NOSPACE
== deco
&& '\0' == *(p
+ 1))
383 print_attr(struct html
*h
, const char *key
, const char *val
)
385 printf(" %s=\"", key
);
386 (void)print_encode(h
, val
, 1);
392 print_otag(struct html
*h
, enum htmltag tag
,
393 int sz
, const struct htmlpair
*p
)
398 /* Push this tags onto the stack of open scopes. */
400 if ( ! (HTML_NOSTACK
& htmltags
[tag
].flags
)) {
401 t
= malloc(sizeof(struct tag
));
404 exit((int)MANDOCLEVEL_SYSERR
);
407 t
->next
= h
->tags
.head
;
412 if ( ! (HTML_NOSPACE
& h
->flags
))
413 if ( ! (HTML_CLRLINE
& htmltags
[tag
].flags
)) {
415 if ( ! (HTML_KEEP
& h
->flags
)) {
416 if (HTML_PREKEEP
& h
->flags
)
417 h
->flags
|= HTML_KEEP
;
423 if ( ! (h
->flags
& HTML_NONOSPACE
))
424 h
->flags
&= ~HTML_NOSPACE
;
426 h
->flags
|= HTML_NOSPACE
;
428 /* Print out the tag name and attributes. */
430 printf("<%s", htmltags
[tag
].name
);
431 for (i
= 0; i
< sz
; i
++)
432 print_attr(h
, htmlattrs
[p
[i
].key
], p
[i
].val
);
434 /* Add non-overridable attributes. */
436 if (TAG_HTML
== tag
&& HTML_XHTML_1_0_STRICT
== h
->type
) {
437 print_attr(h
, "xmlns", "http://www.w3.org/1999/xhtml");
438 print_attr(h
, "xml:lang", "en");
439 print_attr(h
, "lang", "en");
442 /* Accomodate for XML "well-formed" singleton escaping. */
444 if (HTML_AUTOCLOSE
& htmltags
[tag
].flags
)
446 case (HTML_XHTML_1_0_STRICT
):
455 h
->flags
|= HTML_NOSPACE
;
457 if ((HTML_AUTOCLOSE
| HTML_CLRLINE
) & htmltags
[tag
].flags
)
465 print_ctag(struct html
*h
, enum htmltag tag
)
468 printf("</%s>", htmltags
[tag
].name
);
469 if (HTML_CLRLINE
& htmltags
[tag
].flags
) {
470 h
->flags
|= HTML_NOSPACE
;
477 print_gen_decls(struct html
*h
)
486 print_xmltype(struct html
*h
)
489 if (HTML_XHTML_1_0_STRICT
== h
->type
)
490 puts("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
495 print_doctype(struct html
*h
)
502 case (HTML_HTML_4_01_STRICT
):
504 doctype
= "-//W3C//DTD HTML 4.01//EN";
505 dtd
= "http://www.w3.org/TR/html4/strict.dtd";
509 doctype
= "-//W3C//DTD XHTML 1.0 Strict//EN";
510 dtd
= "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
514 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
520 print_text(struct html
*h
, const char *word
)
523 if (word
[0] && '\0' == word
[1])
540 if ( ! (HTML_IGNDELIM
& h
->flags
))
541 h
->flags
|= HTML_NOSPACE
;
547 if ( ! (HTML_NOSPACE
& h
->flags
)) {
549 if ( ! (HTML_KEEP
& h
->flags
)) {
550 if (HTML_PREKEEP
& h
->flags
)
551 h
->flags
|= HTML_KEEP
;
558 if ( ! print_encode(h
, word
, 0))
559 if ( ! (h
->flags
& HTML_NONOSPACE
))
560 h
->flags
&= ~HTML_NOSPACE
;
562 h
->flags
&= ~HTML_IGNDELIM
;
565 * Note that we don't process the pipe: the parser sees it as
566 * punctuation, but we don't in terms of typography.
568 if (word
[0] && '\0' == word
[1])
573 h
->flags
|= HTML_NOSPACE
;
582 print_tagq(struct html
*h
, const struct tag
*until
)
586 while ((tag
= h
->tags
.head
) != NULL
) {
589 print_ctag(h
, tag
->tag
);
590 h
->tags
.head
= tag
->next
;
592 if (until
&& tag
== until
)
599 print_stagq(struct html
*h
, const struct tag
*suntil
)
603 while ((tag
= h
->tags
.head
) != NULL
) {
604 if (suntil
&& tag
== suntil
)
608 print_ctag(h
, tag
->tag
);
609 h
->tags
.head
= tag
->next
;
616 bufinit(struct html
*h
)
625 bufcat_style(struct html
*h
, const char *key
, const char *val
)
636 bufcat(struct html
*h
, const char *p
)
639 bufncat(h
, p
, strlen(p
));
644 buffmt(struct html
*h
, const char *fmt
, ...)
649 (void)vsnprintf(h
->buf
+ (int)h
->buflen
,
650 BUFSIZ
- h
->buflen
- 1, fmt
, ap
);
652 h
->buflen
= strlen(h
->buf
);
657 bufncat(struct html
*h
, const char *p
, size_t sz
)
660 if (h
->buflen
+ sz
> BUFSIZ
- 1)
661 sz
= BUFSIZ
- 1 - h
->buflen
;
663 (void)strncat(h
->buf
, p
, sz
);
669 buffmt_includes(struct html
*h
, const char *name
)
673 pp
= h
->base_includes
;
675 while (NULL
!= (p
= strchr(pp
, '%'))) {
676 bufncat(h
, pp
, (size_t)(p
- pp
));
693 buffmt_man(struct html
*h
,
694 const char *name
, const char *sec
)
701 while (NULL
!= (p
= strchr(pp
, '%'))) {
702 bufncat(h
, pp
, (size_t)(p
- pp
));
705 bufcat(h
, sec
? sec
: "1");
722 bufcat_su(struct html
*h
, const char *p
, const struct roffsu
*su
)
765 * XXX: the CSS spec isn't clear as to which types accept
766 * integer or real numbers, so we just make them all decimals.
768 buffmt(h
, "%s: %.2f%s;", p
, v
, u
);
773 html_idcat(char *dst
, const char *src
, int sz
)
779 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
781 for ( ; *dst
!= '\0' && sz
; dst
++, sz
--)
786 /* We can't start with a number (bah). */
792 for ( ; *src
!= '\0' && sz
> 1; src
++) {
793 ssz
= snprintf(dst
, (size_t)sz
, "%.2x", *src
);