]>
git.cameronkatri.com Git - mandoc.git/blob - html.c
1 /* $Id: html.c,v 1.173 2014/09/28 11:33:15 kristaps Exp $ */
3 * Copyright (c) 2008, 2009, 2010, 2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2011, 2012, 2013, 2014 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.
20 #include <sys/types.h>
32 #include "mandoc_aux.h"
33 #include "libmandoc.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 {"pre", HTML_CLRLINE
}, /* TAG_PRE */
72 {"b", 0 }, /* TAG_B */
73 {"i", 0 }, /* TAG_I */
74 {"code", 0 }, /* TAG_CODE */
75 {"small", 0 }, /* TAG_SMALL */
76 {"style", HTML_CLRLINE
}, /* TAG_STYLE */
77 {"math", HTML_CLRLINE
}, /* TAG_MATH */
78 {"mrow", 0}, /* TAG_MROW */
79 {"mi", 0}, /* TAG_MI */
80 {"mo", 0}, /* TAG_MO */
81 {"msup", 0}, /* TAG_MSUP */
82 {"msub", 0}, /* TAG_MSUB */
83 {"msubsup", 0}, /* TAG_MSUBSUP */
84 {"mfrac", 0}, /* TAG_MFRAC */
85 {"msqrt", 0}, /* TAG_MSQRT */
86 {"mfenced", 0}, /* TAG_MFENCED */
87 {"mtable", 0}, /* TAG_MTABLE */
88 {"mtr", 0}, /* TAG_MTR */
89 {"mtd", 0}, /* TAG_MTD */
92 static const char *const htmlattrs
[ATTR_MAX
] = {
93 "name", /* ATTR_NAME */
95 "href", /* ATTR_HREF */
96 "type", /* ATTR_TYPE */
97 "media", /* ATTR_MEDIA */
98 "class", /* ATTR_CLASS */
99 "style", /* ATTR_STYLE */
101 "colspan", /* ATTR_COLSPAN */
102 "charset", /* ATTR_CHARSET */
103 "open", /* ATTR_OPEN */
104 "close", /* ATTR_CLOSE */
107 static const char *const roffscales
[SCALE_MAX
] = {
120 static void bufncat(struct html
*, const char *, size_t);
121 static void print_ctag(struct html
*, enum htmltag
);
122 static int print_escape(char);
123 static int print_encode(struct html
*, const char *, int);
124 static void print_metaf(struct html
*, enum mandoc_esc
);
125 static void print_attr(struct html
*, const char *, const char *);
126 static void *ml_alloc(char *);
130 ml_alloc(char *outopts
)
138 toks
[2] = "includes";
139 toks
[3] = "fragment";
142 h
= mandoc_calloc(1, sizeof(struct html
));
145 h
->symtab
= mchars_alloc();
147 while (outopts
&& *outopts
)
148 switch (getsubopt(&outopts
, UNCONST(toks
), &v
)) {
156 h
->base_includes
= v
;
159 h
->oflags
|= HTML_FRAGMENT
;
169 html_alloc(char *outopts
)
172 return(ml_alloc(outopts
));
176 xhtml_alloc(char *outopts
)
179 return(ml_alloc(outopts
));
188 h
= (struct html
*)p
;
190 while ((tag
= h
->tags
.head
) != NULL
) {
191 h
->tags
.head
= tag
->next
;
196 mchars_free(h
->symtab
);
202 print_gen_head(struct html
*h
)
204 struct htmlpair tag
[4];
207 tag
[0].key
= ATTR_CHARSET
;
208 tag
[0].val
= "utf-8";
209 print_otag(h
, TAG_META
, 1, tag
);
212 * Print a default style-sheet.
214 t
= print_otag(h
, TAG_STYLE
, 0, NULL
);
215 print_text(h
, "table.head, table.foot { width: 100%; }\n"
216 "td.head-rtitle, td.foot-os { text-align: right; }\n"
217 "td.head-vol { text-align: center; }\n"
218 "table.foot td { width: 50%; }\n"
219 "table.head td { width: 33%; }\n"
220 "div.spacer { margin: 1em 0; }\n");
224 tag
[0].key
= ATTR_REL
;
225 tag
[0].val
= "stylesheet";
226 tag
[1].key
= ATTR_HREF
;
227 tag
[1].val
= h
->style
;
228 tag
[2].key
= ATTR_TYPE
;
229 tag
[2].val
= "text/css";
230 tag
[3].key
= ATTR_MEDIA
;
232 print_otag(h
, TAG_LINK
, 4, tag
);
237 print_metaf(struct html
*h
, enum mandoc_esc deco
)
242 case ESCAPE_FONTPREV
:
245 case ESCAPE_FONTITALIC
:
246 font
= HTMLFONT_ITALIC
;
248 case ESCAPE_FONTBOLD
:
249 font
= HTMLFONT_BOLD
;
256 case ESCAPE_FONTROMAN
:
257 font
= HTMLFONT_NONE
;
265 print_tagq(h
, h
->metaf
);
273 case HTMLFONT_ITALIC
:
274 h
->metaf
= print_otag(h
, TAG_I
, 0, NULL
);
277 h
->metaf
= print_otag(h
, TAG_B
, 0, NULL
);
280 h
->metaf
= print_otag(h
, TAG_B
, 0, NULL
);
281 print_otag(h
, TAG_I
, 0, NULL
);
289 html_strlen(const char *cp
)
295 * Account for escaped sequences within string length
296 * calculations. This follows the logic in term_strlen() as we
297 * must calculate the width of produced strings.
298 * Assume that characters are always width of "1". This is
299 * hacky, but it gets the job done for approximation of widths.
305 rsz
= strcspn(cp
, "\\");
317 switch (mandoc_escape(&cp
, NULL
, NULL
)) {
322 case ESCAPE_NUMBERED
:
330 case ESCAPE_SKIPCHAR
:
372 print_encode(struct html
*h
, const char *p
, int norecurse
)
378 static const char rejs
[9] = { '\\', '<', '>', '&', '"',
379 ASCII_NBRSP
, ASCII_HYPH
, ASCII_BREAK
, '\0' };
384 if (HTML_SKIPCHAR
& h
->flags
&& '\\' != *p
) {
385 h
->flags
&= ~HTML_SKIPCHAR
;
390 sz
= strcspn(p
, rejs
);
392 fwrite(p
, 1, sz
, stdout
);
398 if (print_escape(*p
++))
401 esc
= mandoc_escape(&p
, &seq
, &len
);
402 if (ESCAPE_ERROR
== esc
)
408 case ESCAPE_FONTPREV
:
410 case ESCAPE_FONTBOLD
:
412 case ESCAPE_FONTITALIC
:
416 case ESCAPE_FONTROMAN
:
420 case ESCAPE_SKIPCHAR
:
421 h
->flags
|= HTML_SKIPCHAR
;
427 if (h
->flags
& HTML_SKIPCHAR
) {
428 h
->flags
&= ~HTML_SKIPCHAR
;
434 /* Skip past "u" header. */
435 c
= mchars_num2uc(seq
+ 1, len
- 1);
439 case ESCAPE_NUMBERED
:
440 c
= mchars_num2char(seq
, len
);
441 if ( ! ('\0' == c
|| print_escape(c
)))
445 c
= mchars_spec2cp(h
->symtab
, seq
, len
);
448 else if (-1 == c
&& 1 == len
&&
465 print_attr(struct html
*h
, const char *key
, const char *val
)
467 printf(" %s=\"", key
);
468 (void)print_encode(h
, val
, 1);
473 print_otag(struct html
*h
, enum htmltag tag
,
474 int sz
, const struct htmlpair
*p
)
479 /* Push this tags onto the stack of open scopes. */
481 if ( ! (HTML_NOSTACK
& htmltags
[tag
].flags
)) {
482 t
= mandoc_malloc(sizeof(struct tag
));
484 t
->next
= h
->tags
.head
;
489 if ( ! (HTML_NOSPACE
& h
->flags
))
490 if ( ! (HTML_CLRLINE
& htmltags
[tag
].flags
)) {
492 if ( ! (HTML_KEEP
& h
->flags
)) {
493 if (HTML_PREKEEP
& h
->flags
)
494 h
->flags
|= HTML_KEEP
;
500 if ( ! (h
->flags
& HTML_NONOSPACE
))
501 h
->flags
&= ~HTML_NOSPACE
;
503 h
->flags
|= HTML_NOSPACE
;
505 /* Print out the tag name and attributes. */
507 printf("<%s", htmltags
[tag
].name
);
508 for (i
= 0; i
< sz
; i
++)
509 print_attr(h
, htmlattrs
[p
[i
].key
], p
[i
].val
);
511 /* Accommodate for "well-formed" singleton escaping. */
513 if (HTML_AUTOCLOSE
& htmltags
[tag
].flags
)
518 h
->flags
|= HTML_NOSPACE
;
520 if ((HTML_AUTOCLOSE
| HTML_CLRLINE
) & htmltags
[tag
].flags
)
527 print_ctag(struct html
*h
, enum htmltag tag
)
530 printf("</%s>", htmltags
[tag
].name
);
531 if (HTML_CLRLINE
& htmltags
[tag
].flags
) {
532 h
->flags
|= HTML_NOSPACE
;
538 print_gen_decls(struct html
*h
)
541 puts("<!DOCTYPE html>");
545 print_text(struct html
*h
, const char *word
)
548 if ( ! (HTML_NOSPACE
& h
->flags
)) {
550 if ( ! (HTML_KEEP
& h
->flags
)) {
551 if (HTML_PREKEEP
& h
->flags
)
552 h
->flags
|= HTML_KEEP
;
558 assert(NULL
== h
->metaf
);
560 case HTMLFONT_ITALIC
:
561 h
->metaf
= print_otag(h
, TAG_I
, 0, NULL
);
564 h
->metaf
= print_otag(h
, TAG_B
, 0, NULL
);
567 h
->metaf
= print_otag(h
, TAG_B
, 0, NULL
);
568 print_otag(h
, TAG_I
, 0, NULL
);
575 if ( ! print_encode(h
, word
, 0)) {
576 if ( ! (h
->flags
& HTML_NONOSPACE
))
577 h
->flags
&= ~HTML_NOSPACE
;
579 h
->flags
|= HTML_NOSPACE
;
582 print_tagq(h
, h
->metaf
);
586 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
)
612 print_stagq(struct html
*h
, const struct tag
*suntil
)
616 while ((tag
= h
->tags
.head
) != NULL
) {
617 if (suntil
&& tag
== suntil
)
620 * Remember to close out and nullify the current
621 * meta-font and table, if applicable.
627 print_ctag(h
, tag
->tag
);
628 h
->tags
.head
= tag
->next
;
634 print_paragraph(struct html
*h
)
639 PAIR_CLASS_INIT(&tag
, "spacer");
640 t
= print_otag(h
, TAG_DIV
, 1, &tag
);
646 bufinit(struct html
*h
)
654 bufcat_style(struct html
*h
, const char *key
, const char *val
)
664 bufcat(struct html
*h
, const char *p
)
668 * XXX This is broken and not easy to fix.
669 * When using the -Oincludes option, buffmt_includes()
670 * may pass in strings overrunning BUFSIZ, causing a crash.
673 h
->buflen
= strlcat(h
->buf
, p
, BUFSIZ
);
674 assert(h
->buflen
< BUFSIZ
);
678 bufcat_fmt(struct html
*h
, const char *fmt
, ...)
683 (void)vsnprintf(h
->buf
+ (int)h
->buflen
,
684 BUFSIZ
- h
->buflen
- 1, fmt
, ap
);
686 h
->buflen
= strlen(h
->buf
);
690 bufncat(struct html
*h
, const char *p
, size_t sz
)
693 assert(h
->buflen
+ sz
+ 1 < BUFSIZ
);
694 strncat(h
->buf
, p
, sz
);
699 buffmt_includes(struct html
*h
, const char *name
)
703 pp
= h
->base_includes
;
706 while (NULL
!= (p
= strchr(pp
, '%'))) {
707 bufncat(h
, pp
, (size_t)(p
- pp
));
723 buffmt_man(struct html
*h
, const char *name
, const char *sec
)
730 while (NULL
!= (p
= strchr(pp
, '%'))) {
731 bufncat(h
, pp
, (size_t)(p
- pp
));
734 bufcat(h
, sec
? sec
: "1");
737 bufcat_fmt(h
, "%s", name
);
750 bufcat_su(struct html
*h
, const char *p
, const struct roffsu
*su
)
755 if (SCALE_MM
== su
->unit
&& 0.0 == (v
/= 100.0))
757 else if (SCALE_BU
== su
->unit
)
760 bufcat_fmt(h
, "%s: %.2f%s;", p
, v
, roffscales
[su
->unit
]);
764 bufcat_id(struct html
*h
, const char *src
)
767 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
770 bufcat_fmt(h
, "%.2x", *src
++);