1/*	$Id: html.c,v 1.150 2011/10/05 21:35:17 kristaps Exp $ */
2/*
3 * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2011 Ingo Schwarze <schwarze@openbsd.org>
5 *
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.
9 *
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.
17 */
18#ifdef HAVE_CONFIG_H
19#include "config.h"
20#endif
21
22#include <sys/types.h>
23
24#include <assert.h>
25#include <ctype.h>
26#include <stdarg.h>
27#include <stdio.h>
28#include <stdint.h>
29#include <stdlib.h>
30#include <string.h>
31#include <unistd.h>
32
33#include "mandoc.h"
34#include "libmandoc.h"
35#include "out.h"
36#include "html.h"
37#include "main.h"
38
39struct	htmldata {
40	const char	 *name;
41	int		  flags;
42#define	HTML_CLRLINE	 (1 << 0)
43#define	HTML_NOSTACK	 (1 << 1)
44#define	HTML_AUTOCLOSE	 (1 << 2) /* Tag has auto-closure. */
45};
46
47static	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 */
59	{"a",		0}, /* TAG_A */
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 */
78};
79
80static	const char	*const htmlattrs[ATTR_MAX] = {
81	"http-equiv", /* ATTR_HTTPEQUIV */
82	"content", /* ATTR_CONTENT */
83	"name", /* ATTR_NAME */
84	"rel", /* ATTR_REL */
85	"href", /* ATTR_HREF */
86	"type", /* ATTR_TYPE */
87	"media", /* ATTR_MEDIA */
88	"class", /* ATTR_CLASS */
89	"style", /* ATTR_STYLE */
90	"width", /* ATTR_WIDTH */
91	"id", /* ATTR_ID */
92	"summary", /* ATTR_SUMMARY */
93	"align", /* ATTR_ALIGN */
94	"colspan", /* ATTR_COLSPAN */
95};
96
97static	const char	*const roffscales[SCALE_MAX] = {
98	"cm", /* SCALE_CM */
99	"in", /* SCALE_IN */
100	"pc", /* SCALE_PC */
101	"pt", /* SCALE_PT */
102	"em", /* SCALE_EM */
103	"em", /* SCALE_MM */
104	"ex", /* SCALE_EN */
105	"ex", /* SCALE_BU */
106	"em", /* SCALE_VS */
107	"ex", /* SCALE_FS */
108};
109
110static	void	 bufncat(struct html *, const char *, size_t);
111static	void	 print_ctag(struct html *, enum htmltag);
112static	int	 print_encode(struct html *, const char *, int);
113static	void	 print_metaf(struct html *, enum mandoc_esc);
114static	void	 print_attr(struct html *, const char *, const char *);
115static	void	 *ml_alloc(char *, enum htmltype);
116
117static void *
118ml_alloc(char *outopts, enum htmltype type)
119{
120	struct html	*h;
121	const char	*toks[5];
122	char		*v;
123
124	toks[0] = "style";
125	toks[1] = "man";
126	toks[2] = "includes";
127	toks[3] = "fragment";
128	toks[4] = NULL;
129
130	h = mandoc_calloc(1, sizeof(struct html));
131
132	h->type = type;
133	h->tags.head = NULL;
134	h->symtab = mchars_alloc();
135
136	while (outopts && *outopts)
137		switch (getsubopt(&outopts, UNCONST(toks), &v)) {
138		case (0):
139			h->style = v;
140			break;
141		case (1):
142			h->base_man = v;
143			break;
144		case (2):
145			h->base_includes = v;
146			break;
147		case (3):
148			h->oflags |= HTML_FRAGMENT;
149			break;
150		default:
151			break;
152		}
153
154	return(h);
155}
156
157void *
158html_alloc(char *outopts)
159{
160
161	return(ml_alloc(outopts, HTML_HTML_4_01_STRICT));
162}
163
164
165void *
166xhtml_alloc(char *outopts)
167{
168
169	return(ml_alloc(outopts, HTML_XHTML_1_0_STRICT));
170}
171
172
173void
174html_free(void *p)
175{
176	struct tag	*tag;
177	struct html	*h;
178
179	h = (struct html *)p;
180
181	while ((tag = h->tags.head) != NULL) {
182		h->tags.head = tag->next;
183		free(tag);
184	}
185
186	if (h->symtab)
187		mchars_free(h->symtab);
188
189	free(h);
190}
191
192
193void
194print_gen_head(struct html *h)
195{
196	struct htmlpair	 tag[4];
197
198	tag[0].key = ATTR_HTTPEQUIV;
199	tag[0].val = "Content-Type";
200	tag[1].key = ATTR_CONTENT;
201	tag[1].val = "text/html; charset=utf-8";
202	print_otag(h, TAG_META, 2, tag);
203
204	tag[0].key = ATTR_NAME;
205	tag[0].val = "resource-type";
206	tag[1].key = ATTR_CONTENT;
207	tag[1].val = "document";
208	print_otag(h, TAG_META, 2, tag);
209
210	if (h->style) {
211		tag[0].key = ATTR_REL;
212		tag[0].val = "stylesheet";
213		tag[1].key = ATTR_HREF;
214		tag[1].val = h->style;
215		tag[2].key = ATTR_TYPE;
216		tag[2].val = "text/css";
217		tag[3].key = ATTR_MEDIA;
218		tag[3].val = "all";
219		print_otag(h, TAG_LINK, 4, tag);
220	}
221}
222
223static void
224print_metaf(struct html *h, enum mandoc_esc deco)
225{
226	enum htmlfont	 font;
227
228	switch (deco) {
229	case (ESCAPE_FONTPREV):
230		font = h->metal;
231		break;
232	case (ESCAPE_FONTITALIC):
233		font = HTMLFONT_ITALIC;
234		break;
235	case (ESCAPE_FONTBOLD):
236		font = HTMLFONT_BOLD;
237		break;
238	case (ESCAPE_FONT):
239		/* FALLTHROUGH */
240	case (ESCAPE_FONTROMAN):
241		font = HTMLFONT_NONE;
242		break;
243	default:
244		abort();
245		/* NOTREACHED */
246	}
247
248	if (h->metaf) {
249		print_tagq(h, h->metaf);
250		h->metaf = NULL;
251	}
252
253	h->metal = h->metac;
254	h->metac = font;
255
256	if (HTMLFONT_NONE != font)
257		h->metaf = HTMLFONT_BOLD == font ?
258			print_otag(h, TAG_B, 0, NULL) :
259			print_otag(h, TAG_I, 0, NULL);
260}
261
262int
263html_strlen(const char *cp)
264{
265	int		 ssz, sz;
266	const char	*seq, *p;
267
268	/*
269	 * Account for escaped sequences within string length
270	 * calculations.  This follows the logic in term_strlen() as we
271	 * must calculate the width of produced strings.
272	 * Assume that characters are always width of "1".  This is
273	 * hacky, but it gets the job done for approximation of widths.
274	 */
275
276	sz = 0;
277	while (NULL != (p = strchr(cp, '\\'))) {
278		sz += (int)(p - cp);
279		++cp;
280		switch (mandoc_escape(&cp, &seq, &ssz)) {
281		case (ESCAPE_ERROR):
282			return(sz);
283		case (ESCAPE_UNICODE):
284			/* FALLTHROUGH */
285		case (ESCAPE_NUMBERED):
286			/* FALLTHROUGH */
287		case (ESCAPE_SPECIAL):
288			sz++;
289			break;
290		default:
291			break;
292		}
293	}
294
295	assert(sz >= 0);
296	return(sz + strlen(cp));
297}
298
299static int
300print_encode(struct html *h, const char *p, int norecurse)
301{
302	size_t		 sz;
303	int		 c, len, nospace;
304	const char	*seq;
305	enum mandoc_esc	 esc;
306	static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' };
307
308	nospace = 0;
309
310	while ('\0' != *p) {
311		sz = strcspn(p, rejs);
312
313		fwrite(p, 1, sz, stdout);
314		p += (int)sz;
315
316		if ('\0' == *p)
317			break;
318
319		switch (*p++) {
320		case ('<'):
321			printf("&lt;");
322			continue;
323		case ('>'):
324			printf("&gt;");
325			continue;
326		case ('&'):
327			printf("&amp;");
328			continue;
329		case (ASCII_HYPH):
330			putchar('-');
331			continue;
332		default:
333			break;
334		}
335
336		esc = mandoc_escape(&p, &seq, &len);
337		if (ESCAPE_ERROR == esc)
338			break;
339
340		switch (esc) {
341		case (ESCAPE_UNICODE):
342			/* Skip passed "u" header. */
343			c = mchars_num2uc(seq + 1, len - 1);
344			if ('\0' != c)
345				printf("&#x%x;", c);
346			break;
347		case (ESCAPE_NUMBERED):
348			c = mchars_num2char(seq, len);
349			if ('\0' != c)
350				putchar(c);
351			break;
352		case (ESCAPE_SPECIAL):
353			c = mchars_spec2cp(h->symtab, seq, len);
354			if (c > 0)
355				printf("&#%d;", c);
356			else if (-1 == c && 1 == len)
357				putchar((int)*seq);
358			break;
359		case (ESCAPE_FONT):
360			/* FALLTHROUGH */
361		case (ESCAPE_FONTPREV):
362			/* FALLTHROUGH */
363		case (ESCAPE_FONTBOLD):
364			/* FALLTHROUGH */
365		case (ESCAPE_FONTITALIC):
366			/* FALLTHROUGH */
367		case (ESCAPE_FONTROMAN):
368			if (norecurse)
369				break;
370			print_metaf(h, esc);
371			break;
372		case (ESCAPE_NOSPACE):
373			if ('\0' == *p)
374				nospace = 1;
375			break;
376		default:
377			break;
378		}
379	}
380
381	return(nospace);
382}
383
384
385static void
386print_attr(struct html *h, const char *key, const char *val)
387{
388	printf(" %s=\"", key);
389	(void)print_encode(h, val, 1);
390	putchar('\"');
391}
392
393
394struct tag *
395print_otag(struct html *h, enum htmltag tag,
396		int sz, const struct htmlpair *p)
397{
398	int		 i;
399	struct tag	*t;
400
401	/* Push this tags onto the stack of open scopes. */
402
403	if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
404		t = mandoc_malloc(sizeof(struct tag));
405		t->tag = tag;
406		t->next = h->tags.head;
407		h->tags.head = t;
408	} else
409		t = NULL;
410
411	if ( ! (HTML_NOSPACE & h->flags))
412		if ( ! (HTML_CLRLINE & htmltags[tag].flags)) {
413			/* Manage keeps! */
414			if ( ! (HTML_KEEP & h->flags)) {
415				if (HTML_PREKEEP & h->flags)
416					h->flags |= HTML_KEEP;
417				putchar(' ');
418			} else
419				printf("&#160;");
420		}
421
422	if ( ! (h->flags & HTML_NONOSPACE))
423		h->flags &= ~HTML_NOSPACE;
424	else
425		h->flags |= HTML_NOSPACE;
426
427	/* Print out the tag name and attributes. */
428
429	printf("<%s", htmltags[tag].name);
430	for (i = 0; i < sz; i++)
431		print_attr(h, htmlattrs[p[i].key], p[i].val);
432
433	/* Add non-overridable attributes. */
434
435	if (TAG_HTML == tag && HTML_XHTML_1_0_STRICT == h->type) {
436		print_attr(h, "xmlns", "http://www.w3.org/1999/xhtml");
437		print_attr(h, "xml:lang", "en");
438		print_attr(h, "lang", "en");
439	}
440
441	/* Accommodate for XML "well-formed" singleton escaping. */
442
443	if (HTML_AUTOCLOSE & htmltags[tag].flags)
444		switch (h->type) {
445		case (HTML_XHTML_1_0_STRICT):
446			putchar('/');
447			break;
448		default:
449			break;
450		}
451
452	putchar('>');
453
454	h->flags |= HTML_NOSPACE;
455
456	if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags)
457		putchar('\n');
458
459	return(t);
460}
461
462
463static void
464print_ctag(struct html *h, enum htmltag tag)
465{
466
467	printf("</%s>", htmltags[tag].name);
468	if (HTML_CLRLINE & htmltags[tag].flags) {
469		h->flags |= HTML_NOSPACE;
470		putchar('\n');
471	}
472}
473
474void
475print_gen_decls(struct html *h)
476{
477	const char	*doctype;
478	const char	*dtd;
479	const char	*name;
480
481	switch (h->type) {
482	case (HTML_HTML_4_01_STRICT):
483		name = "HTML";
484		doctype = "-//W3C//DTD HTML 4.01//EN";
485		dtd = "http://www.w3.org/TR/html4/strict.dtd";
486		break;
487	default:
488		puts("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
489		name = "html";
490		doctype = "-//W3C//DTD XHTML 1.0 Strict//EN";
491		dtd = "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
492		break;
493	}
494
495	printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n",
496			name, doctype, dtd);
497}
498
499void
500print_text(struct html *h, const char *word)
501{
502
503	if ( ! (HTML_NOSPACE & h->flags)) {
504		/* Manage keeps! */
505		if ( ! (HTML_KEEP & h->flags)) {
506			if (HTML_PREKEEP & h->flags)
507				h->flags |= HTML_KEEP;
508			putchar(' ');
509		} else
510			printf("&#160;");
511	}
512
513	assert(NULL == h->metaf);
514	if (HTMLFONT_NONE != h->metac)
515		h->metaf = HTMLFONT_BOLD == h->metac ?
516			print_otag(h, TAG_B, 0, NULL) :
517			print_otag(h, TAG_I, 0, NULL);
518
519	assert(word);
520	if ( ! print_encode(h, word, 0)) {
521		if ( ! (h->flags & HTML_NONOSPACE))
522			h->flags &= ~HTML_NOSPACE;
523	} else
524		h->flags |= HTML_NOSPACE;
525
526	if (h->metaf) {
527		print_tagq(h, h->metaf);
528		h->metaf = NULL;
529	}
530
531	h->flags &= ~HTML_IGNDELIM;
532}
533
534
535void
536print_tagq(struct html *h, const struct tag *until)
537{
538	struct tag	*tag;
539
540	while ((tag = h->tags.head) != NULL) {
541		/*
542		 * Remember to close out and nullify the current
543		 * meta-font and table, if applicable.
544		 */
545		if (tag == h->metaf)
546			h->metaf = NULL;
547		if (tag == h->tblt)
548			h->tblt = NULL;
549		print_ctag(h, tag->tag);
550		h->tags.head = tag->next;
551		free(tag);
552		if (until && tag == until)
553			return;
554	}
555}
556
557
558void
559print_stagq(struct html *h, const struct tag *suntil)
560{
561	struct tag	*tag;
562
563	while ((tag = h->tags.head) != NULL) {
564		if (suntil && tag == suntil)
565			return;
566		/*
567		 * Remember to close out and nullify the current
568		 * meta-font and table, if applicable.
569		 */
570		if (tag == h->metaf)
571			h->metaf = NULL;
572		if (tag == h->tblt)
573			h->tblt = NULL;
574		print_ctag(h, tag->tag);
575		h->tags.head = tag->next;
576		free(tag);
577	}
578}
579
580void
581bufinit(struct html *h)
582{
583
584	h->buf[0] = '\0';
585	h->buflen = 0;
586}
587
588void
589bufcat_style(struct html *h, const char *key, const char *val)
590{
591
592	bufcat(h, key);
593	bufcat(h, ":");
594	bufcat(h, val);
595	bufcat(h, ";");
596}
597
598void
599bufcat(struct html *h, const char *p)
600{
601
602	h->buflen = strlcat(h->buf, p, BUFSIZ);
603	assert(h->buflen < BUFSIZ);
604}
605
606void
607bufcat_fmt(struct html *h, const char *fmt, ...)
608{
609	va_list		 ap;
610
611	va_start(ap, fmt);
612	(void)vsnprintf(h->buf + (int)h->buflen,
613			BUFSIZ - h->buflen - 1, fmt, ap);
614	va_end(ap);
615	h->buflen = strlen(h->buf);
616}
617
618static void
619bufncat(struct html *h, const char *p, size_t sz)
620{
621
622	assert(h->buflen + sz + 1 < BUFSIZ);
623	strncat(h->buf, p, sz);
624	h->buflen += sz;
625}
626
627void
628buffmt_includes(struct html *h, const char *name)
629{
630	const char	*p, *pp;
631
632	pp = h->base_includes;
633
634	bufinit(h);
635	while (NULL != (p = strchr(pp, '%'))) {
636		bufncat(h, pp, (size_t)(p - pp));
637		switch (*(p + 1)) {
638		case('I'):
639			bufcat(h, name);
640			break;
641		default:
642			bufncat(h, p, 2);
643			break;
644		}
645		pp = p + 2;
646	}
647	if (pp)
648		bufcat(h, pp);
649}
650
651void
652buffmt_man(struct html *h,
653		const char *name, const char *sec)
654{
655	const char	*p, *pp;
656
657	pp = h->base_man;
658
659	bufinit(h);
660	while (NULL != (p = strchr(pp, '%'))) {
661		bufncat(h, pp, (size_t)(p - pp));
662		switch (*(p + 1)) {
663		case('S'):
664			bufcat(h, sec ? sec : "1");
665			break;
666		case('N'):
667			bufcat_fmt(h, name);
668			break;
669		default:
670			bufncat(h, p, 2);
671			break;
672		}
673		pp = p + 2;
674	}
675	if (pp)
676		bufcat(h, pp);
677}
678
679void
680bufcat_su(struct html *h, const char *p, const struct roffsu *su)
681{
682	double		 v;
683
684	v = su->scale;
685	if (SCALE_MM == su->unit && 0.0 == (v /= 100.0))
686		v = 1.0;
687
688	bufcat_fmt(h, "%s: %.2f%s;", p, v, roffscales[su->unit]);
689}
690
691void
692bufcat_id(struct html *h, const char *src)
693{
694
695	/* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
696
697	while ('\0' != *src)
698		bufcat_fmt(h, "%.2x", *src++);
699}
700