===================================================================
RCS file: /cvs/mandoc/html.c,v
retrieving revision 1.93
retrieving revision 1.101
diff -u -p -r1.93 -r1.101
--- mandoc/html.c	2010/01/29 14:39:37	1.93
+++ mandoc/html.c	2010/06/07 10:52:44	1.101
@@ -1,4 +1,4 @@
-/*	$Id: html.c,v 1.93 2010/01/29 14:39:37 kristaps Exp $ */
+/*	$Id: html.c,v 1.101 2010/06/07 10:52:44 kristaps Exp $ */
 /*
  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
  *
@@ -29,13 +29,12 @@
 #include <string.h>
 #include <unistd.h>
 
+#include "mandoc.h"
 #include "out.h"
 #include "chars.h"
 #include "html.h"
 #include "main.h"
 
-#define	UNCONST(a)	((void *)(uintptr_t)(const void *)(a))
-
 struct	htmldata {
 	const char	 *name;
 	int		  flags;
@@ -54,7 +53,7 @@ static	const struct htmldata htmltags[TAG_MAX] = {
 	{"h1",		0}, /* TAG_H1 */
 	{"h2",		0}, /* TAG_H2 */
 	{"span",	0}, /* TAG_SPAN */
-	{"link",	HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
+	{"link",	HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_LINK */
 	{"br",		HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_BR */
 	{"a",		0}, /* TAG_A */
 	{"table",	HTML_CLRLINE}, /* TAG_TABLE */
@@ -96,6 +95,8 @@ 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);
 
 
@@ -294,11 +295,12 @@ print_encode(struct html *h, const char *p, int norecu
 	int		 len, nospace;
 	const char	*seq;
 	enum roffdeco	 deco;
+	static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' };
 
 	nospace = 0;
 
 	for (; *p; p++) {
-		sz = strcspn(p, "\\<>&");
+		sz = strcspn(p, rejs);
 
 		fwrite(p, 1, sz, stdout);
 		p += /* LINTED */
@@ -313,6 +315,15 @@ print_encode(struct html *h, const char *p, int norecu
 		} else if ('&' == *p) {
 			printf("&amp;");
 			continue;
+		} else if (ASCII_HYPH == *p) {
+			/*
+			 * Note: "soft hyphens" aren't graphically
+			 * displayed when not breaking the text; we want
+			 * them to be displayed.
+			 */
+			/*printf("&#173;");*/
+			putchar('-');
+			continue;
 		} else if ('\0' == *p)
 			break;
 
@@ -351,6 +362,15 @@ print_encode(struct html *h, const char *p, int norecu
 }
 
 
+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)
@@ -358,6 +378,8 @@ 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) {
@@ -374,14 +396,22 @@ 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, 1);
-		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):
@@ -422,21 +452,9 @@ print_gen_decls(struct html *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);
+	if (HTML_XHTML_1_0_STRICT == h->type)
+		printf("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
 }
 
 
@@ -445,19 +463,23 @@ 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 HTML PUBLIC \"%s\" \"%s\">\n", doctype, dtd);
+	printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n", 
+			name, doctype, dtd);
 }
 
 
@@ -482,8 +504,6 @@ print_text(struct html *h, const char *p)
 		case(')'):
 			/* FALLTHROUGH */
 		case(']'):
-			/* FALLTHROUGH */
-		case('}'):
 			if ( ! (HTML_IGNDELIM & h->flags))
 				h->flags |= HTML_NOSPACE;
 			break;
@@ -498,13 +518,15 @@ print_text(struct html *h, const char *p)
 	if ( ! print_encode(h, p, 0))
 		h->flags &= ~HTML_NOSPACE;
 
+	/* 
+	 * Note that we don't process the pipe: the parser sees it as
+	 * punctuation, but we don't in terms of typography.
+	 */
 	if (*p && 0 == *(p + 1))
 		switch (*p) {
 		case('('):
 			/* FALLTHROUGH */
 		case('['):
-			/* FALLTHROUGH */
-		case('{'):
 			h->flags |= HTML_NOSPACE;
 			break;
 		default: