]>
git.cameronkatri.com Git - mandoc.git/blob - html.c
1 /* $Id: html.c,v 1.122 2010/12/24 00:46:49 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 */
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 */
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_metaf(struct html
*h
, enum roffdeco deco
)
261 case (DECO_PREVIOUS
):
265 font
= HTMLFONT_ITALIC
;
268 font
= HTMLFONT_BOLD
;
271 font
= HTMLFONT_NONE
;
279 print_tagq(h
, h
->metaf
);
286 if (HTMLFONT_NONE
!= font
)
287 h
->metaf
= HTMLFONT_BOLD
== font
?
288 print_otag(h
, TAG_B
, 0, NULL
) :
289 print_otag(h
, TAG_I
, 0, NULL
);
294 print_encode(struct html
*h
, const char *p
, int norecurse
)
300 static const char rejs
[6] = { '\\', '<', '>', '&', ASCII_HYPH
, '\0' };
305 sz
= strcspn(p
, rejs
);
307 fwrite(p
, 1, sz
, stdout
);
314 } else if ('>' == *p
) {
317 } else if ('&' == *p
) {
320 } else if (ASCII_HYPH
== *p
) {
322 * Note: "soft hyphens" aren't graphically
323 * displayed when not breaking the text; we want
324 * them to be displayed.
326 /*printf("­");*/
329 } else if ('\0' == *p
)
333 len
= a2roffdeco(&deco
, &seq
, &sz
);
336 case (DECO_RESERVED
):
337 print_res(h
, seq
, sz
);
339 case (DECO_SSPECIAL
):
342 print_spec(h
, deco
, seq
, sz
);
344 case (DECO_PREVIOUS
):
353 print_metaf(h
, deco
);
361 if (DECO_NOSPACE
== deco
&& '\0' == *(p
+ 1))
370 print_attr(struct html
*h
, const char *key
, const char *val
)
372 printf(" %s=\"", key
);
373 (void)print_encode(h
, val
, 1);
379 print_otag(struct html
*h
, enum htmltag tag
,
380 int sz
, const struct htmlpair
*p
)
385 /* Push this tags onto the stack of open scopes. */
387 if ( ! (HTML_NOSTACK
& htmltags
[tag
].flags
)) {
388 t
= malloc(sizeof(struct tag
));
391 exit((int)MANDOCLEVEL_SYSERR
);
394 t
->next
= h
->tags
.head
;
399 if ( ! (HTML_NOSPACE
& h
->flags
))
400 if ( ! (HTML_CLRLINE
& htmltags
[tag
].flags
)) {
402 if ( ! (HTML_KEEP
& h
->flags
)) {
403 if (HTML_PREKEEP
& h
->flags
)
404 h
->flags
|= HTML_KEEP
;
410 if ( ! (h
->flags
& HTML_NONOSPACE
))
411 h
->flags
&= ~HTML_NOSPACE
;
413 h
->flags
|= HTML_NOSPACE
;
415 /* Print out the tag name and attributes. */
417 printf("<%s", htmltags
[tag
].name
);
418 for (i
= 0; i
< sz
; i
++)
419 print_attr(h
, htmlattrs
[p
[i
].key
], p
[i
].val
);
421 /* Add non-overridable attributes. */
423 if (TAG_HTML
== tag
&& HTML_XHTML_1_0_STRICT
== h
->type
) {
424 print_attr(h
, "xmlns", "http://www.w3.org/1999/xhtml");
425 print_attr(h
, "xml:lang", "en");
426 print_attr(h
, "lang", "en");
429 /* Accomodate for XML "well-formed" singleton escaping. */
431 if (HTML_AUTOCLOSE
& htmltags
[tag
].flags
)
433 case (HTML_XHTML_1_0_STRICT
):
442 h
->flags
|= HTML_NOSPACE
;
444 if ((HTML_AUTOCLOSE
| HTML_CLRLINE
) & htmltags
[tag
].flags
)
452 print_ctag(struct html
*h
, enum htmltag tag
)
455 printf("</%s>", htmltags
[tag
].name
);
456 if (HTML_CLRLINE
& htmltags
[tag
].flags
) {
457 h
->flags
|= HTML_NOSPACE
;
464 print_gen_decls(struct html
*h
)
473 print_xmltype(struct html
*h
)
476 if (HTML_XHTML_1_0_STRICT
== h
->type
)
477 puts("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
482 print_doctype(struct html
*h
)
489 case (HTML_HTML_4_01_STRICT
):
491 doctype
= "-//W3C//DTD HTML 4.01//EN";
492 dtd
= "http://www.w3.org/TR/html4/strict.dtd";
496 doctype
= "-//W3C//DTD XHTML 1.0 Strict//EN";
497 dtd
= "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
501 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
507 print_text(struct html
*h
, const char *word
)
510 if (word
[0] && '\0' == word
[1])
527 if ( ! (HTML_IGNDELIM
& h
->flags
))
528 h
->flags
|= HTML_NOSPACE
;
534 if ( ! (HTML_NOSPACE
& h
->flags
)) {
536 if ( ! (HTML_KEEP
& h
->flags
)) {
537 if (HTML_PREKEEP
& h
->flags
)
538 h
->flags
|= HTML_KEEP
;
544 assert(NULL
== h
->metaf
);
545 if (HTMLFONT_NONE
!= h
->metac
)
546 h
->metaf
= HTMLFONT_BOLD
== h
->metac
?
547 print_otag(h
, TAG_B
, 0, NULL
) :
548 print_otag(h
, TAG_I
, 0, NULL
);
551 if ( ! print_encode(h
, word
, 0))
552 if ( ! (h
->flags
& HTML_NONOSPACE
))
553 h
->flags
&= ~HTML_NOSPACE
;
556 print_tagq(h
, h
->metaf
);
560 h
->flags
&= ~HTML_IGNDELIM
;
563 * Note that we don't process the pipe: the parser sees it as
564 * punctuation, but we don't in terms of typography.
566 if (word
[0] && '\0' == word
[1])
571 h
->flags
|= HTML_NOSPACE
;
580 print_tagq(struct html
*h
, const struct tag
*until
)
584 while ((tag
= h
->tags
.head
) != NULL
) {
587 print_ctag(h
, tag
->tag
);
588 h
->tags
.head
= tag
->next
;
590 if (until
&& tag
== until
)
597 print_stagq(struct html
*h
, const struct tag
*suntil
)
601 while ((tag
= h
->tags
.head
) != NULL
) {
602 if (suntil
&& tag
== suntil
)
606 print_ctag(h
, tag
->tag
);
607 h
->tags
.head
= tag
->next
;
614 bufinit(struct html
*h
)
623 bufcat_style(struct html
*h
, const char *key
, const char *val
)
634 bufcat(struct html
*h
, const char *p
)
637 bufncat(h
, p
, strlen(p
));
642 buffmt(struct html
*h
, const char *fmt
, ...)
647 (void)vsnprintf(h
->buf
+ (int)h
->buflen
,
648 BUFSIZ
- h
->buflen
- 1, fmt
, ap
);
650 h
->buflen
= strlen(h
->buf
);
655 bufncat(struct html
*h
, const char *p
, size_t sz
)
658 if (h
->buflen
+ sz
> BUFSIZ
- 1)
659 sz
= BUFSIZ
- 1 - h
->buflen
;
661 (void)strncat(h
->buf
, p
, sz
);
667 buffmt_includes(struct html
*h
, const char *name
)
671 pp
= h
->base_includes
;
673 while (NULL
!= (p
= strchr(pp
, '%'))) {
674 bufncat(h
, pp
, (size_t)(p
- pp
));
691 buffmt_man(struct html
*h
,
692 const char *name
, const char *sec
)
699 while (NULL
!= (p
= strchr(pp
, '%'))) {
700 bufncat(h
, pp
, (size_t)(p
- pp
));
703 bufcat(h
, sec
? sec
: "1");
720 bufcat_su(struct html
*h
, const char *p
, const struct roffsu
*su
)
763 * XXX: the CSS spec isn't clear as to which types accept
764 * integer or real numbers, so we just make them all decimals.
766 buffmt(h
, "%s: %.2f%s;", p
, v
, u
);
771 html_idcat(char *dst
, const char *src
, int sz
)
777 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
779 for ( ; *dst
!= '\0' && sz
; dst
++, sz
--)
784 /* We can't start with a number (bah). */
790 for ( ; *src
!= '\0' && sz
> 1; src
++) {
791 ssz
= snprintf(dst
, (size_t)sz
, "%.2x", *src
);