]>
git.cameronkatri.com Git - mandoc.git/blob - html.c
1 /* $Id: html.c,v 1.129 2011/03/17 09:16:38 kristaps 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
= mandoc_calloc(1, sizeof(struct html
));
127 h
->symtab
= chars_init(CHARS_HTML
);
129 while (outopts
&& *outopts
)
130 switch (getsubopt(&outopts
, UNCONST(toks
), &v
)) {
138 h
->base_includes
= v
;
148 html_alloc(char *outopts
)
151 return(ml_alloc(outopts
, HTML_HTML_4_01_STRICT
));
156 xhtml_alloc(char *outopts
)
159 return(ml_alloc(outopts
, HTML_XHTML_1_0_STRICT
));
169 h
= (struct html
*)p
;
171 while ((tag
= h
->tags
.head
) != NULL
) {
172 h
->tags
.head
= tag
->next
;
177 chars_free(h
->symtab
);
184 print_gen_head(struct html
*h
)
186 struct htmlpair tag
[4];
188 tag
[0].key
= ATTR_HTTPEQUIV
;
189 tag
[0].val
= "Content-Type";
190 tag
[1].key
= ATTR_CONTENT
;
191 tag
[1].val
= "text/html; charset=utf-8";
192 print_otag(h
, TAG_META
, 2, tag
);
194 tag
[0].key
= ATTR_NAME
;
195 tag
[0].val
= "resource-type";
196 tag
[1].key
= ATTR_CONTENT
;
197 tag
[1].val
= "document";
198 print_otag(h
, TAG_META
, 2, tag
);
201 tag
[0].key
= ATTR_REL
;
202 tag
[0].val
= "stylesheet";
203 tag
[1].key
= ATTR_HREF
;
204 tag
[1].val
= h
->style
;
205 tag
[2].key
= ATTR_TYPE
;
206 tag
[2].val
= "text/css";
207 tag
[3].key
= ATTR_MEDIA
;
209 print_otag(h
, TAG_LINK
, 4, tag
);
215 print_num(struct html
*h
, const char *p
, size_t len
)
219 rhs
= chars_num2char(p
, len
);
225 print_spec(struct html
*h
, enum roffdeco d
, const char *p
, size_t len
)
231 if ((cp
= chars_spec2cp(h
->symtab
, p
, len
)) > 0) {
234 } else if (-1 == cp
&& DECO_SSPECIAL
== d
) {
235 fwrite(p
, 1, len
, stdout
);
240 if (NULL
!= (rhs
= chars_spec2str(h
->symtab
, p
, len
, &sz
)))
241 fwrite(rhs
, 1, sz
, stdout
);
246 print_res(struct html
*h
, const char *p
, size_t len
)
252 if ((cp
= chars_res2cp(h
->symtab
, p
, len
)) > 0) {
258 if (NULL
!= (rhs
= chars_res2str(h
->symtab
, p
, len
, &sz
)))
259 fwrite(rhs
, 1, sz
, stdout
);
264 print_metaf(struct html
*h
, enum roffdeco deco
)
269 case (DECO_PREVIOUS
):
273 font
= HTMLFONT_ITALIC
;
276 font
= HTMLFONT_BOLD
;
279 font
= HTMLFONT_NONE
;
287 print_tagq(h
, h
->metaf
);
294 if (HTMLFONT_NONE
!= font
)
295 h
->metaf
= HTMLFONT_BOLD
== font
?
296 print_otag(h
, TAG_B
, 0, NULL
) :
297 print_otag(h
, TAG_I
, 0, NULL
);
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_NUMBERED
):
345 print_num(h
, seq
, sz
);
347 case (DECO_RESERVED
):
348 print_res(h
, seq
, sz
);
350 case (DECO_SSPECIAL
):
353 print_spec(h
, deco
, seq
, sz
);
355 case (DECO_PREVIOUS
):
364 print_metaf(h
, deco
);
372 if (DECO_NOSPACE
== deco
&& '\0' == *(p
+ 1))
381 print_attr(struct html
*h
, const char *key
, const char *val
)
383 printf(" %s=\"", key
);
384 (void)print_encode(h
, val
, 1);
390 print_otag(struct html
*h
, enum htmltag tag
,
391 int sz
, const struct htmlpair
*p
)
396 /* Push this tags onto the stack of open scopes. */
398 if ( ! (HTML_NOSTACK
& htmltags
[tag
].flags
)) {
399 t
= mandoc_malloc(sizeof(struct tag
));
401 t
->next
= h
->tags
.head
;
406 if ( ! (HTML_NOSPACE
& h
->flags
))
407 if ( ! (HTML_CLRLINE
& htmltags
[tag
].flags
)) {
409 if ( ! (HTML_KEEP
& h
->flags
)) {
410 if (HTML_PREKEEP
& h
->flags
)
411 h
->flags
|= HTML_KEEP
;
417 if ( ! (h
->flags
& HTML_NONOSPACE
))
418 h
->flags
&= ~HTML_NOSPACE
;
420 h
->flags
|= HTML_NOSPACE
;
422 /* Print out the tag name and attributes. */
424 printf("<%s", htmltags
[tag
].name
);
425 for (i
= 0; i
< sz
; i
++)
426 print_attr(h
, htmlattrs
[p
[i
].key
], p
[i
].val
);
428 /* Add non-overridable attributes. */
430 if (TAG_HTML
== tag
&& HTML_XHTML_1_0_STRICT
== h
->type
) {
431 print_attr(h
, "xmlns", "http://www.w3.org/1999/xhtml");
432 print_attr(h
, "xml:lang", "en");
433 print_attr(h
, "lang", "en");
436 /* Accomodate for XML "well-formed" singleton escaping. */
438 if (HTML_AUTOCLOSE
& htmltags
[tag
].flags
)
440 case (HTML_XHTML_1_0_STRICT
):
449 h
->flags
|= HTML_NOSPACE
;
451 if ((HTML_AUTOCLOSE
| HTML_CLRLINE
) & htmltags
[tag
].flags
)
459 print_ctag(struct html
*h
, enum htmltag tag
)
462 printf("</%s>", htmltags
[tag
].name
);
463 if (HTML_CLRLINE
& htmltags
[tag
].flags
) {
464 h
->flags
|= HTML_NOSPACE
;
471 print_gen_decls(struct html
*h
)
480 print_xmltype(struct html
*h
)
483 if (HTML_XHTML_1_0_STRICT
== h
->type
)
484 puts("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
489 print_doctype(struct html
*h
)
496 case (HTML_HTML_4_01_STRICT
):
498 doctype
= "-//W3C//DTD HTML 4.01//EN";
499 dtd
= "http://www.w3.org/TR/html4/strict.dtd";
503 doctype
= "-//W3C//DTD XHTML 1.0 Strict//EN";
504 dtd
= "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
508 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
514 print_text(struct html
*h
, const char *word
)
517 if (DELIM_CLOSE
== mandoc_isdelim(word
))
518 if ( ! (HTML_IGNDELIM
& h
->flags
))
519 h
->flags
|= HTML_NOSPACE
;
521 if ( ! (HTML_NOSPACE
& h
->flags
)) {
523 if ( ! (HTML_KEEP
& h
->flags
)) {
524 if (HTML_PREKEEP
& h
->flags
)
525 h
->flags
|= HTML_KEEP
;
531 assert(NULL
== h
->metaf
);
532 if (HTMLFONT_NONE
!= h
->metac
)
533 h
->metaf
= HTMLFONT_BOLD
== h
->metac
?
534 print_otag(h
, TAG_B
, 0, NULL
) :
535 print_otag(h
, TAG_I
, 0, NULL
);
538 if ( ! print_encode(h
, word
, 0))
539 if ( ! (h
->flags
& HTML_NONOSPACE
))
540 h
->flags
&= ~HTML_NOSPACE
;
543 print_tagq(h
, h
->metaf
);
547 h
->flags
&= ~HTML_IGNDELIM
;
549 if (DELIM_OPEN
== mandoc_isdelim(word
))
550 h
->flags
|= HTML_NOSPACE
;
555 print_tagq(struct html
*h
, const struct tag
*until
)
559 while ((tag
= h
->tags
.head
) != NULL
) {
561 * Remember to close out and nullify the current
562 * meta-font and table, if applicable.
568 print_ctag(h
, tag
->tag
);
569 h
->tags
.head
= tag
->next
;
571 if (until
&& tag
== until
)
578 print_stagq(struct html
*h
, const struct tag
*suntil
)
582 while ((tag
= h
->tags
.head
) != NULL
) {
583 if (suntil
&& tag
== suntil
)
586 * Remember to close out and nullify the current
587 * meta-font and table, if applicable.
593 print_ctag(h
, tag
->tag
);
594 h
->tags
.head
= tag
->next
;
601 bufinit(struct html
*h
)
610 bufcat_style(struct html
*h
, const char *key
, const char *val
)
621 bufcat(struct html
*h
, const char *p
)
624 bufncat(h
, p
, strlen(p
));
629 buffmt(struct html
*h
, const char *fmt
, ...)
634 (void)vsnprintf(h
->buf
+ (int)h
->buflen
,
635 BUFSIZ
- h
->buflen
- 1, fmt
, ap
);
637 h
->buflen
= strlen(h
->buf
);
642 bufncat(struct html
*h
, const char *p
, size_t sz
)
645 if (h
->buflen
+ sz
> BUFSIZ
- 1)
646 sz
= BUFSIZ
- 1 - h
->buflen
;
648 (void)strncat(h
->buf
, p
, sz
);
654 buffmt_includes(struct html
*h
, const char *name
)
658 pp
= h
->base_includes
;
660 while (NULL
!= (p
= strchr(pp
, '%'))) {
661 bufncat(h
, pp
, (size_t)(p
- pp
));
678 buffmt_man(struct html
*h
,
679 const char *name
, const char *sec
)
686 while (NULL
!= (p
= strchr(pp
, '%'))) {
687 bufncat(h
, pp
, (size_t)(p
- pp
));
690 bufcat(h
, sec
? sec
: "1");
707 bufcat_su(struct html
*h
, const char *p
, const struct roffsu
*su
)
750 * XXX: the CSS spec isn't clear as to which types accept
751 * integer or real numbers, so we just make them all decimals.
753 buffmt(h
, "%s: %.2f%s;", p
, v
, u
);
758 html_idcat(char *dst
, const char *src
, int sz
)
764 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
766 /* We can't start with a number (bah). */
778 for ( ; *dst
!= '\0' && sz
; dst
++, sz
--)
781 for ( ; *src
!= '\0' && sz
> 1; src
++) {
782 ssz
= snprintf(dst
, (size_t)sz
, "%.2x", *src
);