]>
git.cameronkatri.com Git - mandoc.git/blob - html.c
1 /* $Id: html.c,v 1.126 2011/01/30 16:05:37 schwarze Exp $ */
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2011 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.
22 #include <sys/types.h>
42 #define HTML_CLRLINE (1 << 0)
43 #define HTML_NOSTACK (1 << 1)
44 #define HTML_AUTOCLOSE (1 << 2) /* Tag has auto-closure. */
47 static const struct htmldata htmltags
[TAG_MAX
] = {
48 {"html", HTML_CLRLINE
}, /* TAG_HTML */
49 {"head", HTML_CLRLINE
}, /* TAG_HEAD */
50 {"body", HTML_CLRLINE
}, /* TAG_BODY */
51 {"meta", HTML_CLRLINE
| HTML_NOSTACK
| HTML_AUTOCLOSE
}, /* TAG_META */
52 {"title", HTML_CLRLINE
}, /* TAG_TITLE */
53 {"div", HTML_CLRLINE
}, /* TAG_DIV */
54 {"h1", 0}, /* TAG_H1 */
55 {"h2", 0}, /* TAG_H2 */
56 {"span", 0}, /* TAG_SPAN */
57 {"link", HTML_CLRLINE
| HTML_NOSTACK
| HTML_AUTOCLOSE
}, /* TAG_LINK */
58 {"br", HTML_CLRLINE
| HTML_NOSTACK
| HTML_AUTOCLOSE
}, /* TAG_BR */
60 {"table", HTML_CLRLINE
}, /* TAG_TABLE */
61 {"tbody", HTML_CLRLINE
}, /* TAG_TBODY */
62 {"col", HTML_CLRLINE
| HTML_NOSTACK
| HTML_AUTOCLOSE
}, /* TAG_COL */
63 {"tr", HTML_CLRLINE
}, /* TAG_TR */
64 {"td", HTML_CLRLINE
}, /* TAG_TD */
65 {"li", HTML_CLRLINE
}, /* TAG_LI */
66 {"ul", HTML_CLRLINE
}, /* TAG_UL */
67 {"ol", HTML_CLRLINE
}, /* TAG_OL */
68 {"dl", HTML_CLRLINE
}, /* TAG_DL */
69 {"dt", HTML_CLRLINE
}, /* TAG_DT */
70 {"dd", HTML_CLRLINE
}, /* TAG_DD */
71 {"blockquote", HTML_CLRLINE
}, /* TAG_BLOCKQUOTE */
72 {"p", HTML_CLRLINE
| HTML_NOSTACK
| HTML_AUTOCLOSE
}, /* TAG_P */
73 {"pre", HTML_CLRLINE
}, /* TAG_PRE */
74 {"b", 0 }, /* TAG_B */
75 {"i", 0 }, /* TAG_I */
76 {"code", 0 }, /* TAG_CODE */
77 {"small", 0 }, /* TAG_SMALL */
80 static const char *const htmlattrs
[ATTR_MAX
] = {
81 "http-equiv", /* ATTR_HTTPEQUIV */
82 "content", /* ATTR_CONTENT */
83 "name", /* ATTR_NAME */
85 "href", /* ATTR_HREF */
86 "type", /* ATTR_TYPE */
87 "media", /* ATTR_MEDIA */
88 "class", /* ATTR_CLASS */
89 "style", /* ATTR_STYLE */
90 "width", /* ATTR_WIDTH */
92 "summary", /* ATTR_SUMMARY */
93 "align", /* ATTR_ALIGN */
94 "colspan", /* ATTR_COLSPAN */
97 static void print_num(struct html
*, const char *, size_t);
98 static void print_spec(struct html
*, enum roffdeco
,
99 const char *, size_t);
100 static void print_res(struct html
*, const char *, size_t);
101 static void print_ctag(struct html
*, enum htmltag
);
102 static void print_doctype(struct html
*);
103 static void print_xmltype(struct html
*);
104 static int print_encode(struct html
*, const char *, int);
105 static void print_metaf(struct html
*, enum roffdeco
);
106 static void print_attr(struct html
*,
107 const char *, const char *);
108 static void *ml_alloc(char *, enum htmltype
);
112 ml_alloc(char *outopts
, enum htmltype type
)
120 toks
[2] = "includes";
123 h
= calloc(1, sizeof(struct html
));
126 exit((int)MANDOCLEVEL_SYSERR
);
131 h
->symtab
= chars_init(CHARS_HTML
);
133 while (outopts
&& *outopts
)
134 switch (getsubopt(&outopts
, UNCONST(toks
), &v
)) {
142 h
->base_includes
= v
;
152 html_alloc(char *outopts
)
155 return(ml_alloc(outopts
, HTML_HTML_4_01_STRICT
));
160 xhtml_alloc(char *outopts
)
163 return(ml_alloc(outopts
, HTML_XHTML_1_0_STRICT
));
173 h
= (struct html
*)p
;
175 while ((tag
= h
->tags
.head
) != NULL
) {
176 h
->tags
.head
= tag
->next
;
181 chars_free(h
->symtab
);
188 print_gen_head(struct html
*h
)
190 struct htmlpair tag
[4];
192 tag
[0].key
= ATTR_HTTPEQUIV
;
193 tag
[0].val
= "Content-Type";
194 tag
[1].key
= ATTR_CONTENT
;
195 tag
[1].val
= "text/html; charset=utf-8";
196 print_otag(h
, TAG_META
, 2, tag
);
198 tag
[0].key
= ATTR_NAME
;
199 tag
[0].val
= "resource-type";
200 tag
[1].key
= ATTR_CONTENT
;
201 tag
[1].val
= "document";
202 print_otag(h
, TAG_META
, 2, tag
);
205 tag
[0].key
= ATTR_REL
;
206 tag
[0].val
= "stylesheet";
207 tag
[1].key
= ATTR_HREF
;
208 tag
[1].val
= h
->style
;
209 tag
[2].key
= ATTR_TYPE
;
210 tag
[2].val
= "text/css";
211 tag
[3].key
= ATTR_MEDIA
;
213 print_otag(h
, TAG_LINK
, 4, tag
);
219 print_num(struct html
*h
, const char *p
, size_t len
)
223 rhs
= chars_num2char(p
, len
);
230 print_spec(struct html
*h
, enum roffdeco d
, const char *p
, size_t len
)
236 if ((cp
= chars_spec2cp(h
->symtab
, p
, len
)) > 0) {
239 } else if (-1 == cp
&& DECO_SSPECIAL
== d
) {
240 fwrite(p
, 1, len
, stdout
);
245 if (NULL
!= (rhs
= chars_spec2str(h
->symtab
, p
, len
, &sz
)))
246 fwrite(rhs
, 1, sz
, stdout
);
251 print_res(struct html
*h
, const char *p
, size_t len
)
257 if ((cp
= chars_res2cp(h
->symtab
, p
, len
)) > 0) {
263 if (NULL
!= (rhs
= chars_res2str(h
->symtab
, p
, len
, &sz
)))
264 fwrite(rhs
, 1, sz
, stdout
);
269 print_metaf(struct html
*h
, enum roffdeco deco
)
274 case (DECO_PREVIOUS
):
278 font
= HTMLFONT_ITALIC
;
281 font
= HTMLFONT_BOLD
;
284 font
= HTMLFONT_NONE
;
292 print_tagq(h
, h
->metaf
);
299 if (HTMLFONT_NONE
!= font
)
300 h
->metaf
= HTMLFONT_BOLD
== font
?
301 print_otag(h
, TAG_B
, 0, NULL
) :
302 print_otag(h
, TAG_I
, 0, NULL
);
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_NUMBERED
):
350 print_num(h
, seq
, sz
);
352 case (DECO_RESERVED
):
353 print_res(h
, seq
, sz
);
355 case (DECO_SSPECIAL
):
358 print_spec(h
, deco
, seq
, sz
);
360 case (DECO_PREVIOUS
):
369 print_metaf(h
, deco
);
377 if (DECO_NOSPACE
== deco
&& '\0' == *(p
+ 1))
386 print_attr(struct html
*h
, const char *key
, const char *val
)
388 printf(" %s=\"", key
);
389 (void)print_encode(h
, val
, 1);
395 print_otag(struct html
*h
, enum htmltag tag
,
396 int sz
, const struct htmlpair
*p
)
401 /* Push this tags onto the stack of open scopes. */
403 if ( ! (HTML_NOSTACK
& htmltags
[tag
].flags
)) {
404 t
= malloc(sizeof(struct tag
));
407 exit((int)MANDOCLEVEL_SYSERR
);
410 t
->next
= h
->tags
.head
;
415 if ( ! (HTML_NOSPACE
& h
->flags
))
416 if ( ! (HTML_CLRLINE
& htmltags
[tag
].flags
)) {
418 if ( ! (HTML_KEEP
& h
->flags
)) {
419 if (HTML_PREKEEP
& h
->flags
)
420 h
->flags
|= HTML_KEEP
;
426 if ( ! (h
->flags
& HTML_NONOSPACE
))
427 h
->flags
&= ~HTML_NOSPACE
;
429 h
->flags
|= HTML_NOSPACE
;
431 /* Print out the tag name and attributes. */
433 printf("<%s", htmltags
[tag
].name
);
434 for (i
= 0; i
< sz
; i
++)
435 print_attr(h
, htmlattrs
[p
[i
].key
], p
[i
].val
);
437 /* Add non-overridable attributes. */
439 if (TAG_HTML
== tag
&& HTML_XHTML_1_0_STRICT
== h
->type
) {
440 print_attr(h
, "xmlns", "http://www.w3.org/1999/xhtml");
441 print_attr(h
, "xml:lang", "en");
442 print_attr(h
, "lang", "en");
445 /* Accomodate for XML "well-formed" singleton escaping. */
447 if (HTML_AUTOCLOSE
& htmltags
[tag
].flags
)
449 case (HTML_XHTML_1_0_STRICT
):
458 h
->flags
|= HTML_NOSPACE
;
460 if ((HTML_AUTOCLOSE
| HTML_CLRLINE
) & htmltags
[tag
].flags
)
468 print_ctag(struct html
*h
, enum htmltag tag
)
471 printf("</%s>", htmltags
[tag
].name
);
472 if (HTML_CLRLINE
& htmltags
[tag
].flags
) {
473 h
->flags
|= HTML_NOSPACE
;
480 print_gen_decls(struct html
*h
)
489 print_xmltype(struct html
*h
)
492 if (HTML_XHTML_1_0_STRICT
== h
->type
)
493 puts("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
498 print_doctype(struct html
*h
)
505 case (HTML_HTML_4_01_STRICT
):
507 doctype
= "-//W3C//DTD HTML 4.01//EN";
508 dtd
= "http://www.w3.org/TR/html4/strict.dtd";
512 doctype
= "-//W3C//DTD XHTML 1.0 Strict//EN";
513 dtd
= "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
517 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
523 print_text(struct html
*h
, const char *word
)
526 if (word
[0] && '\0' == word
[1])
543 if ( ! (HTML_IGNDELIM
& h
->flags
))
544 h
->flags
|= HTML_NOSPACE
;
550 if ( ! (HTML_NOSPACE
& h
->flags
)) {
552 if ( ! (HTML_KEEP
& h
->flags
)) {
553 if (HTML_PREKEEP
& h
->flags
)
554 h
->flags
|= HTML_KEEP
;
560 assert(NULL
== h
->metaf
);
561 if (HTMLFONT_NONE
!= h
->metac
)
562 h
->metaf
= HTMLFONT_BOLD
== h
->metac
?
563 print_otag(h
, TAG_B
, 0, NULL
) :
564 print_otag(h
, TAG_I
, 0, NULL
);
567 if ( ! print_encode(h
, word
, 0))
568 if ( ! (h
->flags
& HTML_NONOSPACE
))
569 h
->flags
&= ~HTML_NOSPACE
;
572 print_tagq(h
, h
->metaf
);
576 h
->flags
&= ~HTML_IGNDELIM
;
579 * Note that we don't process the pipe: the parser sees it as
580 * punctuation, but we don't in terms of typography.
582 if (word
[0] && '\0' == word
[1])
587 h
->flags
|= HTML_NOSPACE
;
596 print_tagq(struct html
*h
, const struct tag
*until
)
600 while ((tag
= h
->tags
.head
) != NULL
) {
602 * Remember to close out and nullify the current
603 * meta-font and table, if applicable.
609 print_ctag(h
, tag
->tag
);
610 h
->tags
.head
= tag
->next
;
612 if (until
&& tag
== until
)
619 print_stagq(struct html
*h
, const struct tag
*suntil
)
623 while ((tag
= h
->tags
.head
) != NULL
) {
624 if (suntil
&& tag
== suntil
)
627 * Remember to close out and nullify the current
628 * meta-font and table, if applicable.
634 print_ctag(h
, tag
->tag
);
635 h
->tags
.head
= tag
->next
;
642 bufinit(struct html
*h
)
651 bufcat_style(struct html
*h
, const char *key
, const char *val
)
662 bufcat(struct html
*h
, const char *p
)
665 bufncat(h
, p
, strlen(p
));
670 buffmt(struct html
*h
, const char *fmt
, ...)
675 (void)vsnprintf(h
->buf
+ (int)h
->buflen
,
676 BUFSIZ
- h
->buflen
- 1, fmt
, ap
);
678 h
->buflen
= strlen(h
->buf
);
683 bufncat(struct html
*h
, const char *p
, size_t sz
)
686 if (h
->buflen
+ sz
> BUFSIZ
- 1)
687 sz
= BUFSIZ
- 1 - h
->buflen
;
689 (void)strncat(h
->buf
, p
, sz
);
695 buffmt_includes(struct html
*h
, const char *name
)
699 pp
= h
->base_includes
;
701 while (NULL
!= (p
= strchr(pp
, '%'))) {
702 bufncat(h
, pp
, (size_t)(p
- pp
));
719 buffmt_man(struct html
*h
,
720 const char *name
, const char *sec
)
727 while (NULL
!= (p
= strchr(pp
, '%'))) {
728 bufncat(h
, pp
, (size_t)(p
- pp
));
731 bufcat(h
, sec
? sec
: "1");
748 bufcat_su(struct html
*h
, const char *p
, const struct roffsu
*su
)
791 * XXX: the CSS spec isn't clear as to which types accept
792 * integer or real numbers, so we just make them all decimals.
794 buffmt(h
, "%s: %.2f%s;", p
, v
, u
);
799 html_idcat(char *dst
, const char *src
, int sz
)
805 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
807 /* We can't start with a number (bah). */
819 for ( ; *dst
!= '\0' && sz
; dst
++, sz
--)
822 for ( ; *src
!= '\0' && sz
> 1; src
++) {
823 ssz
= snprintf(dst
, (size_t)sz
, "%.2x", *src
);