]>
git.cameronkatri.com Git - mandoc.git/blob - html.c
1 /* $Id: html.c,v 1.140 2011/05/17 10:48:06 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>
34 #include "libmandoc.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 const char *const roffscales
[SCALE_MAX
] = {
110 static void print_num(struct html
*, const char *, size_t);
111 static void print_spec(struct html
*, const char *, size_t);
112 static void print_res(struct html
*, const char *, size_t);
113 static void print_ctag(struct html
*, enum htmltag
);
114 static void print_doctype(struct html
*);
115 static void print_xmltype(struct html
*);
116 static int print_encode(struct html
*, const char *, int);
117 static void print_metaf(struct html
*, enum mandoc_esc
);
118 static void print_attr(struct html
*,
119 const char *, const char *);
120 static void *ml_alloc(char *, enum htmltype
);
124 ml_alloc(char *outopts
, enum htmltype type
)
132 toks
[2] = "includes";
135 h
= mandoc_calloc(1, sizeof(struct html
));
139 h
->symtab
= mchars_alloc();
141 while (outopts
&& *outopts
)
142 switch (getsubopt(&outopts
, UNCONST(toks
), &v
)) {
150 h
->base_includes
= v
;
160 html_alloc(char *outopts
)
163 return(ml_alloc(outopts
, HTML_HTML_4_01_STRICT
));
168 xhtml_alloc(char *outopts
)
171 return(ml_alloc(outopts
, HTML_XHTML_1_0_STRICT
));
181 h
= (struct html
*)p
;
183 while ((tag
= h
->tags
.head
) != NULL
) {
184 h
->tags
.head
= tag
->next
;
189 mchars_free(h
->symtab
);
196 print_gen_head(struct html
*h
)
198 struct htmlpair tag
[4];
200 tag
[0].key
= ATTR_HTTPEQUIV
;
201 tag
[0].val
= "Content-Type";
202 tag
[1].key
= ATTR_CONTENT
;
203 tag
[1].val
= "text/html; charset=utf-8";
204 print_otag(h
, TAG_META
, 2, tag
);
206 tag
[0].key
= ATTR_NAME
;
207 tag
[0].val
= "resource-type";
208 tag
[1].key
= ATTR_CONTENT
;
209 tag
[1].val
= "document";
210 print_otag(h
, TAG_META
, 2, tag
);
213 tag
[0].key
= ATTR_REL
;
214 tag
[0].val
= "stylesheet";
215 tag
[1].key
= ATTR_HREF
;
216 tag
[1].val
= h
->style
;
217 tag
[2].key
= ATTR_TYPE
;
218 tag
[2].val
= "text/css";
219 tag
[3].key
= ATTR_MEDIA
;
221 print_otag(h
, TAG_LINK
, 4, tag
);
227 print_num(struct html
*h
, const char *p
, size_t len
)
231 if ('\0' != (c
= mchars_num2char(p
, len
)))
236 print_spec(struct html
*h
, const char *p
, size_t len
)
242 if ((cp
= mchars_spec2cp(h
->symtab
, p
, len
)) > 0) {
245 } else if (-1 == cp
&& 1 == len
) {
246 fwrite(p
, 1, len
, stdout
);
251 if (NULL
!= (rhs
= mchars_spec2str(h
->symtab
, p
, len
, &sz
)))
252 fwrite(rhs
, 1, sz
, stdout
);
257 print_res(struct html
*h
, const char *p
, size_t len
)
263 if ((cp
= mchars_res2cp(h
->symtab
, p
, len
)) > 0) {
269 if (NULL
!= (rhs
= mchars_res2str(h
->symtab
, p
, len
, &sz
)))
270 fwrite(rhs
, 1, sz
, stdout
);
275 print_metaf(struct html
*h
, enum mandoc_esc deco
)
280 case (ESCAPE_FONTPREV
):
283 case (ESCAPE_FONTITALIC
):
284 font
= HTMLFONT_ITALIC
;
286 case (ESCAPE_FONTBOLD
):
287 font
= HTMLFONT_BOLD
;
289 case (ESCAPE_FONTROMAN
):
290 font
= HTMLFONT_NONE
;
298 print_tagq(h
, h
->metaf
);
305 if (HTMLFONT_NONE
!= font
)
306 h
->metaf
= HTMLFONT_BOLD
== font
?
307 print_otag(h
, TAG_B
, 0, NULL
) :
308 print_otag(h
, TAG_I
, 0, NULL
);
312 html_strlen(const char *cp
)
318 * Account for escaped sequences within string length
319 * calculations. This follows the logic in term_strlen() as we
320 * must calculate the width of produced strings.
321 * Assume that characters are always width of "1". This is
322 * hacky, but it gets the job done for approximation of widths.
326 while (NULL
!= (p
= strchr(cp
, '\\'))) {
329 switch (mandoc_escape(&cp
, &seq
, &ssz
)) {
332 case (ESCAPE_NUMBERED
):
334 case (ESCAPE_PREDEF
):
336 case (ESCAPE_SPECIAL
):
345 return(sz
+ strlen(cp
));
349 print_encode(struct html
*h
, const char *p
, int norecurse
)
355 static const char rejs
[6] = { '\\', '<', '>', '&', ASCII_HYPH
, '\0' };
360 sz
= strcspn(p
, rejs
);
362 fwrite(p
, 1, sz
, stdout
);
385 esc
= mandoc_escape(&p
, &seq
, &len
);
386 if (ESCAPE_ERROR
== esc
)
390 case (ESCAPE_NUMBERED
):
391 print_num(h
, seq
, len
);
393 case (ESCAPE_PREDEF
):
394 print_res(h
, seq
, len
);
396 case (ESCAPE_SPECIAL
):
397 print_spec(h
, seq
, len
);
399 case (ESCAPE_FONTPREV
):
401 case (ESCAPE_FONTBOLD
):
403 case (ESCAPE_FONTITALIC
):
405 case (ESCAPE_FONTROMAN
):
410 case (ESCAPE_NOSPACE
):
424 print_attr(struct html
*h
, const char *key
, const char *val
)
426 printf(" %s=\"", key
);
427 (void)print_encode(h
, val
, 1);
433 print_otag(struct html
*h
, enum htmltag tag
,
434 int sz
, const struct htmlpair
*p
)
439 /* Push this tags onto the stack of open scopes. */
441 if ( ! (HTML_NOSTACK
& htmltags
[tag
].flags
)) {
442 t
= mandoc_malloc(sizeof(struct tag
));
444 t
->next
= h
->tags
.head
;
449 if ( ! (HTML_NOSPACE
& h
->flags
))
450 if ( ! (HTML_CLRLINE
& htmltags
[tag
].flags
)) {
452 if ( ! (HTML_KEEP
& h
->flags
)) {
453 if (HTML_PREKEEP
& h
->flags
)
454 h
->flags
|= HTML_KEEP
;
460 if ( ! (h
->flags
& HTML_NONOSPACE
))
461 h
->flags
&= ~HTML_NOSPACE
;
463 h
->flags
|= HTML_NOSPACE
;
465 /* Print out the tag name and attributes. */
467 printf("<%s", htmltags
[tag
].name
);
468 for (i
= 0; i
< sz
; i
++)
469 print_attr(h
, htmlattrs
[p
[i
].key
], p
[i
].val
);
471 /* Add non-overridable attributes. */
473 if (TAG_HTML
== tag
&& HTML_XHTML_1_0_STRICT
== h
->type
) {
474 print_attr(h
, "xmlns", "http://www.w3.org/1999/xhtml");
475 print_attr(h
, "xml:lang", "en");
476 print_attr(h
, "lang", "en");
479 /* Accommodate for XML "well-formed" singleton escaping. */
481 if (HTML_AUTOCLOSE
& htmltags
[tag
].flags
)
483 case (HTML_XHTML_1_0_STRICT
):
492 h
->flags
|= HTML_NOSPACE
;
494 if ((HTML_AUTOCLOSE
| HTML_CLRLINE
) & htmltags
[tag
].flags
)
502 print_ctag(struct html
*h
, enum htmltag tag
)
505 printf("</%s>", htmltags
[tag
].name
);
506 if (HTML_CLRLINE
& htmltags
[tag
].flags
) {
507 h
->flags
|= HTML_NOSPACE
;
514 print_gen_decls(struct html
*h
)
523 print_xmltype(struct html
*h
)
526 if (HTML_XHTML_1_0_STRICT
== h
->type
)
527 puts("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
532 print_doctype(struct html
*h
)
539 case (HTML_HTML_4_01_STRICT
):
541 doctype
= "-//W3C//DTD HTML 4.01//EN";
542 dtd
= "http://www.w3.org/TR/html4/strict.dtd";
546 doctype
= "-//W3C//DTD XHTML 1.0 Strict//EN";
547 dtd
= "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
551 printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
556 print_text(struct html
*h
, const char *word
)
559 if ( ! (HTML_NOSPACE
& h
->flags
)) {
561 if ( ! (HTML_KEEP
& h
->flags
)) {
562 if (HTML_PREKEEP
& h
->flags
)
563 h
->flags
|= HTML_KEEP
;
569 assert(NULL
== h
->metaf
);
570 if (HTMLFONT_NONE
!= h
->metac
)
571 h
->metaf
= HTMLFONT_BOLD
== h
->metac
?
572 print_otag(h
, TAG_B
, 0, NULL
) :
573 print_otag(h
, TAG_I
, 0, NULL
);
576 if ( ! print_encode(h
, word
, 0))
577 if ( ! (h
->flags
& HTML_NONOSPACE
))
578 h
->flags
&= ~HTML_NOSPACE
;
581 print_tagq(h
, h
->metaf
);
585 h
->flags
&= ~HTML_IGNDELIM
;
590 print_tagq(struct html
*h
, const struct tag
*until
)
594 while ((tag
= h
->tags
.head
) != NULL
) {
596 * Remember to close out and nullify the current
597 * meta-font and table, if applicable.
603 print_ctag(h
, tag
->tag
);
604 h
->tags
.head
= tag
->next
;
606 if (until
&& tag
== until
)
613 print_stagq(struct html
*h
, const struct tag
*suntil
)
617 while ((tag
= h
->tags
.head
) != NULL
) {
618 if (suntil
&& tag
== suntil
)
621 * Remember to close out and nullify the current
622 * meta-font and table, if applicable.
628 print_ctag(h
, tag
->tag
);
629 h
->tags
.head
= tag
->next
;
636 bufinit(struct html
*h
)
645 bufcat_style(struct html
*h
, const char *key
, const char *val
)
656 bufcat(struct html
*h
, const char *p
)
659 bufncat(h
, p
, strlen(p
));
664 buffmt(struct html
*h
, const char *fmt
, ...)
669 (void)vsnprintf(h
->buf
+ (int)h
->buflen
,
670 BUFSIZ
- h
->buflen
- 1, fmt
, ap
);
672 h
->buflen
= strlen(h
->buf
);
677 bufncat(struct html
*h
, const char *p
, size_t sz
)
680 if (h
->buflen
+ sz
> BUFSIZ
- 1)
681 sz
= BUFSIZ
- 1 - h
->buflen
;
683 (void)strncat(h
->buf
, p
, sz
);
689 buffmt_includes(struct html
*h
, const char *name
)
693 pp
= h
->base_includes
;
695 while (NULL
!= (p
= strchr(pp
, '%'))) {
696 bufncat(h
, pp
, (size_t)(p
- pp
));
713 buffmt_man(struct html
*h
,
714 const char *name
, const char *sec
)
721 while (NULL
!= (p
= strchr(pp
, '%'))) {
722 bufncat(h
, pp
, (size_t)(p
- pp
));
725 bufcat(h
, sec
? sec
: "1");
742 bufcat_su(struct html
*h
, const char *p
, const struct roffsu
*su
)
747 if (SCALE_MM
== su
->unit
&& 0.0 == (v
/= 100.0))
750 buffmt(h
, "%s: %.2f%s;", p
, v
, roffscales
[su
->unit
]);
755 html_idcat(char *dst
, const char *src
, int sz
)
761 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
763 /* We can't start with a number (bah). */
775 for ( ; *dst
!= '\0' && sz
; dst
++, sz
--)
778 for ( ; *src
!= '\0' && sz
> 1; src
++) {
779 ssz
= snprintf(dst
, (size_t)sz
, "%.2x", *src
);