]>
git.cameronkatri.com Git - mandoc.git/blob - html.c
1 /* $Id: html.c,v 1.187 2015/09/26 00:54:03 schwarze Exp $ */
3 * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2011-2015 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 AUTHORS DISCLAIM ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS 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"
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 */
90 {"munderover", 0}, /* TAG_MUNDEROVER */
91 {"munder", 0}, /* TAG_MUNDER*/
92 {"mover", 0}, /* TAG_MOVER*/
95 static const char *const htmlattrs
[ATTR_MAX
] = {
96 "name", /* ATTR_NAME */
98 "href", /* ATTR_HREF */
99 "type", /* ATTR_TYPE */
100 "media", /* ATTR_MEDIA */
101 "class", /* ATTR_CLASS */
102 "style", /* ATTR_STYLE */
104 "colspan", /* ATTR_COLSPAN */
105 "charset", /* ATTR_CHARSET */
106 "open", /* ATTR_OPEN */
107 "close", /* ATTR_CLOSE */
108 "mathvariant", /* ATTR_MATHVARIANT */
111 static const char *const roffscales
[SCALE_MAX
] = {
124 static void bufncat(struct html
*, const char *, size_t);
125 static void print_ctag(struct html
*, struct tag
*);
126 static int print_escape(char);
127 static int print_encode(struct html
*, const char *, int);
128 static void print_metaf(struct html
*, enum mandoc_esc
);
129 static void print_attr(struct html
*, const char *, const char *);
133 html_alloc(const struct mchars
*mchars
, const struct manoutput
*outopts
)
137 h
= mandoc_calloc(1, sizeof(struct html
));
142 h
->style
= outopts
->style
;
143 h
->base_man
= outopts
->man
;
144 h
->base_includes
= outopts
->includes
;
145 if (outopts
->fragment
)
146 h
->oflags
|= HTML_FRAGMENT
;
157 h
= (struct html
*)p
;
159 while ((tag
= h
->tags
.head
) != NULL
) {
160 h
->tags
.head
= tag
->next
;
168 print_gen_head(struct html
*h
)
170 struct htmlpair tag
[4];
173 tag
[0].key
= ATTR_CHARSET
;
174 tag
[0].val
= "utf-8";
175 print_otag(h
, TAG_META
, 1, tag
);
178 * Print a default style-sheet.
180 t
= print_otag(h
, TAG_STYLE
, 0, NULL
);
181 print_text(h
, "table.head, table.foot { width: 100%; }\n"
182 "td.head-rtitle, td.foot-os { text-align: right; }\n"
183 "td.head-vol { text-align: center; }\n"
184 "table.foot td { width: 50%; }\n"
185 "table.head td { width: 33%; }\n"
186 "div.spacer { margin: 1em 0; }\n");
190 tag
[0].key
= ATTR_REL
;
191 tag
[0].val
= "stylesheet";
192 tag
[1].key
= ATTR_HREF
;
193 tag
[1].val
= h
->style
;
194 tag
[2].key
= ATTR_TYPE
;
195 tag
[2].val
= "text/css";
196 tag
[3].key
= ATTR_MEDIA
;
198 print_otag(h
, TAG_LINK
, 4, tag
);
203 print_metaf(struct html
*h
, enum mandoc_esc deco
)
208 case ESCAPE_FONTPREV
:
211 case ESCAPE_FONTITALIC
:
212 font
= HTMLFONT_ITALIC
;
214 case ESCAPE_FONTBOLD
:
215 font
= HTMLFONT_BOLD
;
222 case ESCAPE_FONTROMAN
:
223 font
= HTMLFONT_NONE
;
230 print_tagq(h
, h
->metaf
);
238 case HTMLFONT_ITALIC
:
239 h
->metaf
= print_otag(h
, TAG_I
, 0, NULL
);
242 h
->metaf
= print_otag(h
, TAG_B
, 0, NULL
);
245 h
->metaf
= print_otag(h
, TAG_B
, 0, NULL
);
246 print_otag(h
, TAG_I
, 0, NULL
);
254 html_strlen(const char *cp
)
260 * Account for escaped sequences within string length
261 * calculations. This follows the logic in term_strlen() as we
262 * must calculate the width of produced strings.
263 * Assume that characters are always width of "1". This is
264 * hacky, but it gets the job done for approximation of widths.
270 rsz
= strcspn(cp
, "\\");
282 switch (mandoc_escape(&cp
, NULL
, NULL
)) {
287 case ESCAPE_NUMBERED
:
291 case ESCAPE_OVERSTRIKE
:
297 case ESCAPE_SKIPCHAR
:
339 print_encode(struct html
*h
, const char *p
, int norecurse
)
345 static const char rejs
[9] = { '\\', '<', '>', '&', '"',
346 ASCII_NBRSP
, ASCII_HYPH
, ASCII_BREAK
, '\0' };
351 if (HTML_SKIPCHAR
& h
->flags
&& '\\' != *p
) {
352 h
->flags
&= ~HTML_SKIPCHAR
;
357 sz
= strcspn(p
, rejs
);
359 fwrite(p
, 1, sz
, stdout
);
365 if (print_escape(*p
++))
368 esc
= mandoc_escape(&p
, &seq
, &len
);
369 if (ESCAPE_ERROR
== esc
)
375 case ESCAPE_FONTPREV
:
377 case ESCAPE_FONTBOLD
:
379 case ESCAPE_FONTITALIC
:
383 case ESCAPE_FONTROMAN
:
387 case ESCAPE_SKIPCHAR
:
388 h
->flags
|= HTML_SKIPCHAR
;
394 if (h
->flags
& HTML_SKIPCHAR
) {
395 h
->flags
&= ~HTML_SKIPCHAR
;
401 /* Skip past "u" header. */
402 c
= mchars_num2uc(seq
+ 1, len
- 1);
404 case ESCAPE_NUMBERED
:
405 c
= mchars_num2char(seq
, len
);
410 c
= mchars_spec2cp(h
->symtab
, seq
, len
);
418 case ESCAPE_OVERSTRIKE
:
426 if ((c
< 0x20 && c
!= 0x09) ||
427 (c
> 0x7E && c
< 0xA0))
431 else if ( ! print_escape(c
))
439 print_attr(struct html
*h
, const char *key
, const char *val
)
441 printf(" %s=\"", key
);
442 (void)print_encode(h
, val
, 1);
447 print_otag(struct html
*h
, enum htmltag tag
,
448 int sz
, const struct htmlpair
*p
)
453 /* Push this tags onto the stack of open scopes. */
455 if ( ! (HTML_NOSTACK
& htmltags
[tag
].flags
)) {
456 t
= mandoc_malloc(sizeof(struct tag
));
458 t
->next
= h
->tags
.head
;
463 if ( ! (HTML_NOSPACE
& h
->flags
))
464 if ( ! (HTML_CLRLINE
& htmltags
[tag
].flags
)) {
466 if ( ! (HTML_KEEP
& h
->flags
)) {
467 if (HTML_PREKEEP
& h
->flags
)
468 h
->flags
|= HTML_KEEP
;
474 if ( ! (h
->flags
& HTML_NONOSPACE
))
475 h
->flags
&= ~HTML_NOSPACE
;
477 h
->flags
|= HTML_NOSPACE
;
479 /* Print out the tag name and attributes. */
481 printf("<%s", htmltags
[tag
].name
);
482 for (i
= 0; i
< sz
; i
++)
483 print_attr(h
, htmlattrs
[p
[i
].key
], p
[i
].val
);
485 /* Accommodate for "well-formed" singleton escaping. */
487 if (HTML_AUTOCLOSE
& htmltags
[tag
].flags
)
492 h
->flags
|= HTML_NOSPACE
;
494 if ((HTML_AUTOCLOSE
| HTML_CLRLINE
) & htmltags
[tag
].flags
)
501 print_ctag(struct html
*h
, struct tag
*tag
)
505 * Remember to close out and nullify the current
506 * meta-font and table, if applicable.
513 printf("</%s>", htmltags
[tag
->tag
].name
);
514 if (HTML_CLRLINE
& htmltags
[tag
->tag
].flags
) {
515 h
->flags
|= HTML_NOSPACE
;
519 h
->tags
.head
= tag
->next
;
524 print_gen_decls(struct html
*h
)
527 puts("<!DOCTYPE html>");
531 print_text(struct html
*h
, const char *word
)
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
);
546 case HTMLFONT_ITALIC
:
547 h
->metaf
= print_otag(h
, TAG_I
, 0, NULL
);
550 h
->metaf
= print_otag(h
, TAG_B
, 0, NULL
);
553 h
->metaf
= print_otag(h
, TAG_B
, 0, NULL
);
554 print_otag(h
, TAG_I
, 0, NULL
);
561 if ( ! print_encode(h
, word
, 0)) {
562 if ( ! (h
->flags
& HTML_NONOSPACE
))
563 h
->flags
&= ~HTML_NOSPACE
;
564 h
->flags
&= ~HTML_NONEWLINE
;
566 h
->flags
|= HTML_NOSPACE
| HTML_NONEWLINE
;
569 print_tagq(h
, h
->metaf
);
573 h
->flags
&= ~HTML_IGNDELIM
;
577 print_tagq(struct html
*h
, const struct tag
*until
)
581 while ((tag
= h
->tags
.head
) != NULL
) {
583 if (until
&& tag
== until
)
589 print_stagq(struct html
*h
, const struct tag
*suntil
)
593 while ((tag
= h
->tags
.head
) != NULL
) {
594 if (suntil
&& tag
== suntil
)
601 print_paragraph(struct html
*h
)
606 PAIR_CLASS_INIT(&tag
, "spacer");
607 t
= print_otag(h
, TAG_DIV
, 1, &tag
);
613 bufinit(struct html
*h
)
621 bufcat_style(struct html
*h
, const char *key
, const char *val
)
631 bufcat(struct html
*h
, const char *p
)
635 * XXX This is broken and not easy to fix.
636 * When using the -Oincludes option, buffmt_includes()
637 * may pass in strings overrunning BUFSIZ, causing a crash.
640 h
->buflen
= strlcat(h
->buf
, p
, BUFSIZ
);
641 assert(h
->buflen
< BUFSIZ
);
645 bufcat_fmt(struct html
*h
, const char *fmt
, ...)
650 (void)vsnprintf(h
->buf
+ (int)h
->buflen
,
651 BUFSIZ
- h
->buflen
- 1, fmt
, ap
);
653 h
->buflen
= strlen(h
->buf
);
657 bufncat(struct html
*h
, const char *p
, size_t sz
)
660 assert(h
->buflen
+ sz
+ 1 < BUFSIZ
);
661 strncat(h
->buf
, p
, sz
);
666 buffmt_includes(struct html
*h
, const char *name
)
670 pp
= h
->base_includes
;
673 while (NULL
!= (p
= strchr(pp
, '%'))) {
674 bufncat(h
, pp
, (size_t)(p
- pp
));
690 buffmt_man(struct html
*h
, const char *name
, const char *sec
)
697 while (NULL
!= (p
= strchr(pp
, '%'))) {
698 bufncat(h
, pp
, (size_t)(p
- pp
));
701 bufcat(h
, sec
? sec
: "1");
704 bufcat_fmt(h
, "%s", name
);
717 bufcat_su(struct html
*h
, const char *p
, const struct roffsu
*su
)
722 if (SCALE_MM
== su
->unit
&& 0.0 == (v
/= 100.0))
724 else if (SCALE_BU
== su
->unit
)
727 bufcat_fmt(h
, "%s: %.2f%s;", p
, v
, roffscales
[su
->unit
]);
731 bufcat_id(struct html
*h
, const char *src
)
734 /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
737 bufcat_fmt(h
, "%.2x", *src
++);