]>
git.cameronkatri.com Git - mandoc.git/blob - html.c
1 /* $Id: html.c,v 1.127 2011/03/15 16:23:51 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
= 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
);
229 print_spec(struct html
*h
, enum roffdeco d
, const char *p
, size_t len
)
235 if ((cp
= chars_spec2cp(h
->symtab
, p
, len
)) > 0) {
238 } else if (-1 == cp
&& DECO_SSPECIAL
== d
) {
239 fwrite(p
, 1, len
, stdout
);
244 if (NULL
!= (rhs
= chars_spec2str(h
->symtab
, p
, len
, &sz
)))
245 fwrite(rhs
, 1, sz
, stdout
);
250 print_res(struct html
*h
, const char *p
, size_t len
)
256 if ((cp
= chars_res2cp(h
->symtab
, p
, len
)) > 0) {
262 if (NULL
!= (rhs
= chars_res2str(h
->symtab
, p
, len
, &sz
)))
263 fwrite(rhs
, 1, sz
, stdout
);
268 print_metaf(struct html
*h
, enum roffdeco deco
)
273 case (DECO_PREVIOUS
):
277 font
= HTMLFONT_ITALIC
;
280 font
= HTMLFONT_BOLD
;
283 font
= HTMLFONT_NONE
;
291 print_tagq(h
, h
->metaf
);
298 if (HTMLFONT_NONE
!= font
)
299 h
->metaf
= HTMLFONT_BOLD
== font
?
300 print_otag(h
, TAG_B
, 0, NULL
) :
301 print_otag(h
, TAG_I
, 0, NULL
);
306 print_encode(struct html
*h
, const char *p
, int norecurse
)
312 static const char rejs
[6] = { '\\', '<', '>', '&', ASCII_HYPH
, '\0' };
317 sz
= strcspn(p
, rejs
);
319 fwrite(p
, 1, sz
, stdout
);
326 } else if ('>' == *p
) {
329 } else if ('&' == *p
) {
332 } else if (ASCII_HYPH
== *p
) {
334 * Note: "soft hyphens" aren't graphically
335 * displayed when not breaking the text; we want
336 * them to be displayed.
338 /*printf("­");*/
341 } else if ('\0' == *p
)
345 len
= a2roffdeco(&deco
, &seq
, &sz
);
348 case (DECO_NUMBERED
):
349 print_num(h
, seq
, sz
);
351 case (DECO_RESERVED
):
352 print_res(h
, seq
, sz
);
354 case (DECO_SSPECIAL
):
357 print_spec(h
, deco
, seq
, sz
);
359 case (DECO_PREVIOUS
):
368 print_metaf(h
, deco
);
376 if (DECO_NOSPACE
== deco
&& '\0' == *(p
+ 1))
385 print_attr(struct html
*h
, const char *key
, const char *val
)
387 printf(" %s=\"", key
);
388 (void)print_encode(h
, val
, 1);
394 print_otag(struct html
*h
, enum htmltag tag
,
395 int sz
, const struct htmlpair
*p
)
400 /* Push this tags onto the stack of open scopes. */
402 if ( ! (HTML_NOSTACK
& htmltags
[tag
].flags
)) {
403 t
= malloc(sizeof(struct tag
));
406 exit((int)MANDOCLEVEL_SYSERR
);
409 t
->next
= h
->tags
.head
;
414 if ( ! (HTML_NOSPACE
& h
->flags
))
415 if ( ! (HTML_CLRLINE
& htmltags
[tag
].flags
)) {
417 if ( ! (HTML_KEEP
& h
->flags
)) {
418 if (HTML_PREKEEP
& h
->flags
)
419 h
->flags
|= HTML_KEEP
;
425 if ( ! (h
->flags
& HTML_NONOSPACE
))
426 h
->flags
&= ~HTML_NOSPACE
;
428 h
->flags
|= HTML_NOSPACE
;
430 /* Print out the tag name and attributes. */
432 printf("<%s", htmltags
[tag
].name
);
433 for (i
= 0; i
< sz
; i
++)
434 print_attr(h
, htmlattrs
[p
[i
].key
], p
[i
].val
);
436 /* Add non-overridable attributes. */
438 if (TAG_HTML
== tag
&& HTML_XHTML_1_0_STRICT
== h
->type
) {
439 print_attr(h
, "xmlns", "http://www.w3.org/1999/xhtml");
440 print_attr(h
, "xml:lang", "en");
441 print_attr(h
, "lang", "en");
444 /* Accomodate for XML "well-formed" singleton escaping. */
446 if (HTML_AUTOCLOSE
& htmltags
[tag
].flags
)
448 case (HTML_XHTML_1_0_STRICT
):
457 h
->flags
|= HTML_NOSPACE
;
459 if ((HTML_AUTOCLOSE
| HTML_CLRLINE
) & htmltags
[tag
].flags
)
467 print_ctag(struct html
*h
, enum htmltag tag
)
470 printf("</%s>", htmltags
[tag
].name
);
471 if (HTML_CLRLINE
& htmltags
[tag
].flags
) {
472 h
->flags
|= HTML_NOSPACE
;
479 print_gen_decls(struct html
*h
)
488 print_xmltype(struct html
*h
)
491 if (HTML_XHTML_1_0_STRICT
== h
->type
)
492 puts("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
497 print_doctype(struct html
*h
)
504 case (HTML_HTML_4_01_STRICT
):
506 doctype
= "-//W3C//DTD HTML 4.01//EN";
507 dtd
= "http://www.w3.org/TR/html4/strict.dtd";
511 doctype
= "-//W3C//DTD XHTML 1.0 Strict//EN";
512 dtd
= "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
516 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
522 print_text(struct html
*h
, const char *word
)
525 if (word
[0] && '\0' == word
[1])
542 if ( ! (HTML_IGNDELIM
& h
->flags
))
543 h
->flags
|= HTML_NOSPACE
;
549 if ( ! (HTML_NOSPACE
& h
->flags
)) {
551 if ( ! (HTML_KEEP
& h
->flags
)) {
552 if (HTML_PREKEEP
& h
->flags
)
553 h
->flags
|= HTML_KEEP
;
559 assert(NULL
== h
->metaf
);
560 if (HTMLFONT_NONE
!= h
->metac
)
561 h
->metaf
= HTMLFONT_BOLD
== h
->metac
?
562 print_otag(h
, TAG_B
, 0, NULL
) :
563 print_otag(h
, TAG_I
, 0, NULL
);
566 if ( ! print_encode(h
, word
, 0))
567 if ( ! (h
->flags
& HTML_NONOSPACE
))
568 h
->flags
&= ~HTML_NOSPACE
;
571 print_tagq(h
, h
->metaf
);
575 h
->flags
&= ~HTML_IGNDELIM
;
578 * Note that we don't process the pipe: the parser sees it as
579 * punctuation, but we don't in terms of typography.
581 if (word
[0] && '\0' == word
[1])
586 h
->flags
|= HTML_NOSPACE
;
595 print_tagq(struct html
*h
, const struct tag
*until
)
599 while ((tag
= h
->tags
.head
) != NULL
) {
601 * Remember to close out and nullify the current
602 * meta-font and table, if applicable.
608 print_ctag(h
, tag
->tag
);
609 h
->tags
.head
= tag
->next
;
611 if (until
&& tag
== until
)
618 print_stagq(struct html
*h
, const struct tag
*suntil
)
622 while ((tag
= h
->tags
.head
) != NULL
) {
623 if (suntil
&& tag
== suntil
)
626 * Remember to close out and nullify the current
627 * meta-font and table, if applicable.
633 print_ctag(h
, tag
->tag
);
634 h
->tags
.head
= tag
->next
;
641 bufinit(struct html
*h
)
650 bufcat_style(struct html
*h
, const char *key
, const char *val
)
661 bufcat(struct html
*h
, const char *p
)
664 bufncat(h
, p
, strlen(p
));
669 buffmt(struct html
*h
, const char *fmt
, ...)
674 (void)vsnprintf(h
->buf
+ (int)h
->buflen
,
675 BUFSIZ
- h
->buflen
- 1, fmt
, ap
);
677 h
->buflen
= strlen(h
->buf
);
682 bufncat(struct html
*h
, const char *p
, size_t sz
)
685 if (h
->buflen
+ sz
> BUFSIZ
- 1)
686 sz
= BUFSIZ
- 1 - h
->buflen
;
688 (void)strncat(h
->buf
, p
, sz
);
694 buffmt_includes(struct html
*h
, const char *name
)
698 pp
= h
->base_includes
;
700 while (NULL
!= (p
= strchr(pp
, '%'))) {
701 bufncat(h
, pp
, (size_t)(p
- pp
));
718 buffmt_man(struct html
*h
,
719 const char *name
, const char *sec
)
726 while (NULL
!= (p
= strchr(pp
, '%'))) {
727 bufncat(h
, pp
, (size_t)(p
- pp
));
730 bufcat(h
, sec
? sec
: "1");
747 bufcat_su(struct html
*h
, const char *p
, const struct roffsu
*su
)
790 * XXX: the CSS spec isn't clear as to which types accept
791 * integer or real numbers, so we just make them all decimals.
793 buffmt(h
, "%s: %.2f%s;", p
, v
, u
);
798 html_idcat(char *dst
, const char *src
, int sz
)
804 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
806 /* We can't start with a number (bah). */
818 for ( ; *dst
!= '\0' && sz
; dst
++, sz
--)
821 for ( ; *src
!= '\0' && sz
> 1; src
++) {
822 ssz
= snprintf(dst
, (size_t)sz
, "%.2x", *src
);