===================================================================
RCS file: /cvs/mandoc/html.c,v
retrieving revision 1.86
retrieving revision 1.97
diff -u -p -r1.86 -r1.97
--- mandoc/html.c	2009/11/14 12:00:24	1.86
+++ mandoc/html.c	2010/04/03 12:46:35	1.97
@@ -1,4 +1,4 @@
-/*	$Id: html.c,v 1.86 2009/11/14 12:00:24 kristaps Exp $ */
+/*	$Id: html.c,v 1.97 2010/04/03 12:46:35 kristaps Exp $ */
 /*
  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
  *
@@ -14,6 +14,10 @@
  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  */
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif
+
 #include <sys/types.h>
 
 #include <assert.h>
@@ -32,40 +36,42 @@
 
 #define	UNCONST(a)	((void *)(uintptr_t)(const void *)(a))
 
-#define	DOCTYPE		"-//W3C//DTD HTML 4.01//EN"
-#define	DTD		"http://www.w3.org/TR/html4/strict.dtd"
-
 struct	htmldata {
 	const char	 *name;
 	int		  flags;
 #define	HTML_CLRLINE	 (1 << 0)
 #define	HTML_NOSTACK	 (1 << 1)
+#define	HTML_AUTOCLOSE	 (1 << 2) /* Tag has auto-closure. */
 };
 
 static	const struct htmldata htmltags[TAG_MAX] = {
 	{"html",	HTML_CLRLINE}, /* TAG_HTML */
 	{"head",	HTML_CLRLINE}, /* TAG_HEAD */
 	{"body",	HTML_CLRLINE}, /* TAG_BODY */
-	{"meta",	HTML_CLRLINE | HTML_NOSTACK}, /* TAG_META */
+	{"meta",	HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_META */
 	{"title",	HTML_CLRLINE}, /* TAG_TITLE */
 	{"div",		HTML_CLRLINE}, /* TAG_DIV */
 	{"h1",		0}, /* TAG_H1 */
 	{"h2",		0}, /* TAG_H2 */
-	{"p",		HTML_CLRLINE}, /* TAG_P */
 	{"span",	0}, /* TAG_SPAN */
 	{"link",	HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
-	{"br",		HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
+	{"br",		HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_BR */
 	{"a",		0}, /* TAG_A */
 	{"table",	HTML_CLRLINE}, /* TAG_TABLE */
-	{"col",		HTML_CLRLINE | HTML_NOSTACK}, /* TAG_COL */
+	{"col",		HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_COL */
 	{"tr",		HTML_CLRLINE}, /* TAG_TR */
 	{"td",		HTML_CLRLINE}, /* TAG_TD */
 	{"li",		HTML_CLRLINE}, /* TAG_LI */
 	{"ul",		HTML_CLRLINE}, /* TAG_UL */
 	{"ol",		HTML_CLRLINE}, /* TAG_OL */
-	{"base",	HTML_CLRLINE | HTML_NOSTACK}, /* TAG_BASE */
 };
 
+static	const char	*const htmlfonts[HTMLFONT_MAX] = {
+	"roman",
+	"bold",
+	"italic"
+};
+
 static	const char	*const htmlattrs[ATTR_MAX] = {
 	"http-equiv",
 	"content",
@@ -83,19 +89,20 @@ static	const char	*const htmlattrs[ATTR_MAX] = {
 	"summary",
 };
 
-#ifdef __linux__
-extern	int		  getsubopt(char **, char * const *, char **);
-#endif
-
-
 static	void		  print_spec(struct html *, const char *, size_t);
 static	void		  print_res(struct html *, const char *, size_t);
 static	void		  print_ctag(struct html *, enum htmltag);
-static	int		  print_encode(struct html *, const char *);
+static	void		  print_doctype(struct html *);
+static	void		  print_xmltype(struct html *);
+static	int		  print_encode(struct html *, const char *, int);
+static	void		  print_metaf(struct html *, enum roffdeco);
+static	void		  print_attr(struct html *, 
+				const char *, const char *);
+static	void		 *ml_alloc(char *, enum htmltype);
 
 
-void *
-html_alloc(char *outopts)
+static void *
+ml_alloc(char *outopts, enum htmltype type)
 {
 	struct html	*h;
 	const char	*toks[4];
@@ -112,6 +119,7 @@ html_alloc(char *outopts)
 		exit(EXIT_FAILURE);
 	}
 
+	h->type = type;
 	h->tags.head = NULL;
 	h->ords.head = NULL;
 	h->symtab = chars_init(CHARS_HTML);
@@ -134,7 +142,22 @@ html_alloc(char *outopts)
 	return(h);
 }
 
+void *
+html_alloc(char *outopts)
+{
 
+	return(ml_alloc(outopts, HTML_HTML_4_01_STRICT));
+}
+
+
+void *
+xhtml_alloc(char *outopts)
+{
+
+	return(ml_alloc(outopts, HTML_XHTML_1_0_STRICT));
+}
+
+
 void
 html_free(void *p)
 {
@@ -220,8 +243,54 @@ print_res(struct html *h, const char *p, size_t len)
 }
 
 
+struct tag *
+print_ofont(struct html *h, enum htmlfont font)
+{
+	struct htmlpair	 tag;
+
+	h->metal = h->metac;
+	h->metac = font;
+
+	/* FIXME: DECO_ROMAN should just close out preexisting. */
+
+	if (h->metaf && h->tags.head == h->metaf)
+		print_tagq(h, h->metaf);
+
+	PAIR_CLASS_INIT(&tag, htmlfonts[font]);
+	h->metaf = print_otag(h, TAG_SPAN, 1, &tag);
+	return(h->metaf);
+}
+
+
+static void
+print_metaf(struct html *h, enum roffdeco deco)
+{
+	enum htmlfont	 font;
+
+	switch (deco) {
+	case (DECO_PREVIOUS):
+		font = h->metal;
+		break;
+	case (DECO_ITALIC):
+		font = HTMLFONT_ITALIC;
+		break;
+	case (DECO_BOLD):
+		font = HTMLFONT_BOLD;
+		break;
+	case (DECO_ROMAN):
+		font = HTMLFONT_NONE;
+		break;
+	default:
+		abort();
+		/* NOTREACHED */
+	}
+
+	(void)print_ofont(h, font);
+}
+
+
 static int
-print_encode(struct html *h, const char *p)
+print_encode(struct html *h, const char *p, int norecurse)
 {
 	size_t		 sz;
 	int		 len, nospace;
@@ -259,6 +328,17 @@ print_encode(struct html *h, const char *p)
 		case (DECO_SPECIAL):
 			print_spec(h, seq, sz);
 			break;
+		case (DECO_PREVIOUS):
+			/* FALLTHROUGH */
+		case (DECO_BOLD):
+			/* FALLTHROUGH */
+		case (DECO_ITALIC):
+			/* FALLTHROUGH */
+		case (DECO_ROMAN):
+			if (norecurse)
+				break;
+			print_metaf(h, deco);
+			break;
 		default:
 			break;
 		}
@@ -273,6 +353,15 @@ print_encode(struct html *h, const char *p)
 }
 
 
+static void
+print_attr(struct html *h, const char *key, const char *val)
+{
+	printf(" %s=\"", key);
+	(void)print_encode(h, val, 1);
+	putchar('\"');
+}
+
+
 struct tag *
 print_otag(struct html *h, enum htmltag tag, 
 		int sz, const struct htmlpair *p)
@@ -280,6 +369,8 @@ print_otag(struct html *h, enum htmltag tag, 
 	int		 i;
 	struct tag	*t;
 
+	/* Push this tags onto the stack of open scopes. */
+
 	if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
 		t = malloc(sizeof(struct tag));
 		if (NULL == t) {
@@ -296,26 +387,38 @@ print_otag(struct html *h, enum htmltag tag, 
 		if ( ! (HTML_CLRLINE & htmltags[tag].flags))
 			putchar(' ');
 
+	/* Print out the tag name and attributes. */
+
 	printf("<%s", htmltags[tag].name);
-	for (i = 0; i < sz; i++) {
-		printf(" %s=\"", htmlattrs[p[i].key]);
-		assert(p->val);
-		(void)print_encode(h, p[i].val);
-		putchar('\"');
+	for (i = 0; i < sz; i++)
+		print_attr(h, htmlattrs[p[i].key], p[i].val);
+
+	/* Add non-overridable attributes. */
+
+	if (TAG_HTML == tag && HTML_XHTML_1_0_STRICT == h->type) {
+		print_attr(h, "xmlns", "http://www.w3.org/1999/xhtml");
+		print_attr(h, "xml:lang", "en");
+		print_attr(h, "lang", "en");
 	}
+
+	/* Accomodate for XML "well-formed" singleton escaping. */
+
+	if (HTML_AUTOCLOSE & htmltags[tag].flags)
+		switch (h->type) {
+		case (HTML_XHTML_1_0_STRICT):
+			putchar('/');
+			break;
+		default:
+			break;
+		}
+
 	putchar('>');
 
 	h->flags |= HTML_NOSPACE;
-	if (HTML_CLRLINE & htmltags[tag].flags)
-		h->flags |= HTML_NEWLINE;
-	else
-		h->flags &= ~HTML_NEWLINE;
-
 	return(t);
 }
 
 
-/* ARGSUSED */
 static void
 print_ctag(struct html *h, enum htmltag tag)
 {
@@ -323,22 +426,66 @@ print_ctag(struct html *h, enum htmltag tag)
 	printf("</%s>", htmltags[tag].name);
 	if (HTML_CLRLINE & htmltags[tag].flags) {
 		h->flags |= HTML_NOSPACE;
-		h->flags |= HTML_NEWLINE;
 		putchar('\n');
-	} else
-		h->flags &= ~HTML_NEWLINE;
+	} 
 }
 
 
-/* ARGSUSED */
 void
-print_gen_doctype(struct html *h)
+print_gen_decls(struct html *h)
 {
-	
-	printf("<!DOCTYPE HTML PUBLIC \"%s\" \"%s\">", DOCTYPE, DTD);
+
+	print_xmltype(h);
+	print_doctype(h);
 }
 
 
+static void
+print_xmltype(struct html *h)
+{
+	const char	*decl;
+
+	switch (h->type) {
+	case (HTML_XHTML_1_0_STRICT):
+		decl = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>";
+		break;
+	default:
+		decl = NULL;
+		break;
+	}
+
+	if (NULL == decl)
+		return;
+
+	printf("%s\n", decl);
+}
+
+
+static void
+print_doctype(struct html *h)
+{
+	const char	*doctype;
+	const char	*dtd;
+	const char	*name;
+
+	switch (h->type) {
+	case (HTML_HTML_4_01_STRICT):
+		name = "HTML";
+		doctype = "-//W3C//DTD HTML 4.01//EN";
+		dtd = "http://www.w3.org/TR/html4/strict.dtd";
+		break;
+	default:
+		name = "html";
+		doctype = "-//W3C//DTD XHTML 1.0 Strict//EN";
+		dtd = "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
+		break;
+	}
+
+	printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n", 
+			name, doctype, dtd);
+}
+
+
 void
 print_text(struct html *h, const char *p)
 {
@@ -360,8 +507,6 @@ print_text(struct html *h, const char *p)
 		case(')'):
 			/* FALLTHROUGH */
 		case(']'):
-			/* FALLTHROUGH */
-		case('}'):
 			if ( ! (HTML_IGNDELIM & h->flags))
 				h->flags |= HTML_NOSPACE;
 			break;
@@ -372,19 +517,17 @@ print_text(struct html *h, const char *p)
 	if ( ! (h->flags & HTML_NOSPACE))
 		putchar(' ');
 
-	h->flags &= ~HTML_NEWLINE;
-
 	assert(p);
-	if ( ! print_encode(h, p))
+	if ( ! print_encode(h, p, 0))
 		h->flags &= ~HTML_NOSPACE;
 
 	if (*p && 0 == *(p + 1))
 		switch (*p) {
+		case('|'):
+			/* FALLTHROUGH */
 		case('('):
 			/* FALLTHROUGH */
 		case('['):
-			/* FALLTHROUGH */
-		case('{'):
 			h->flags |= HTML_NOSPACE;
 			break;
 		default:
@@ -399,6 +542,8 @@ print_tagq(struct html *h, const struct tag *until)
 	struct tag	*tag;
 
 	while ((tag = h->tags.head) != NULL) {
+		if (tag == h->metaf)
+			h->metaf = NULL;
 		print_ctag(h, tag->tag);
 		h->tags.head = tag->next;
 		free(tag);
@@ -416,6 +561,8 @@ print_stagq(struct html *h, const struct tag *suntil)
 	while ((tag = h->tags.head) != NULL) {
 		if (suntil && tag == suntil)
 			return;
+		if (tag == h->metaf)
+			h->metaf = NULL;
 		print_ctag(h, tag->tag);
 		h->tags.head = tag->next;
 		free(tag);