===================================================================
RCS file: /cvs/mandoc/html.c,v
retrieving revision 1.68
retrieving revision 1.81
diff -u -p -r1.68 -r1.81
--- mandoc/html.c	2009/10/28 05:08:17	1.68
+++ mandoc/html.c	2009/11/05 10:16:01	1.81
@@ -1,4 +1,4 @@
-/*	$Id: html.c,v 1.68 2009/10/28 05:08:17 kristaps Exp $ */
+/*	$Id: html.c,v 1.81 2009/11/05 10:16:01 kristaps Exp $ */
 /*
  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
  *
@@ -18,9 +18,8 @@
 
 #include <assert.h>
 #include <ctype.h>
-#include <err.h>
-#include <stdio.h>
 #include <stdarg.h>
+#include <stdio.h>
 #include <stdint.h>
 #include <stdlib.h>
 #include <string.h>
@@ -100,17 +99,16 @@ html_alloc(char *outopts)
 	toks[2] = "includes";
 	toks[3] = NULL;
 
-	if (NULL == (h = calloc(1, sizeof(struct html))))
-		return(NULL);
+	h = calloc(1, sizeof(struct html));
+	if (NULL == h) {
+		perror(NULL);
+		exit(EXIT_FAILURE);
+	}
 
 	h->tags.head = NULL;
 	h->ords.head = NULL;
+	h->symtab = chars_init(CHARS_HTML);
 
-	if (NULL == (h->symtab = chars_init(CHARS_HTML))) {
-		free(h);
-		return(NULL);
-	}
-
 	while (outopts && *outopts)
 		switch (getsubopt(&outopts, UNCONST(toks), &v)) {
 		case (0):
@@ -191,15 +189,13 @@ static void
 print_spec(struct html *h, const char *p, int len)
 {
 	const char	*rhs;
-	int		 i;
 	size_t		 sz;
 
 	rhs = chars_a2ascii(h->symtab, p, (size_t)len, &sz);
 
 	if (NULL == rhs) 
 		return;
-	for (i = 0; i < (int)sz; i++) 
-		putchar(rhs[i]);
+	fwrite(rhs, 1, sz, stdout);
 }
 
 
@@ -207,15 +203,13 @@ static void
 print_res(struct html *h, const char *p, int len)
 {
 	const char	*rhs;
-	int		 i;
 	size_t		 sz;
 
 	rhs = chars_a2res(h->symtab, p, (size_t)len, &sz);
 
 	if (NULL == rhs)
 		return;
-	for (i = 0; i < (int)sz; i++) 
-		putchar(rhs[i]);
+	fwrite(rhs, 1, sz, stdout);
 }
 
 
@@ -276,21 +270,19 @@ print_escape(struct html *h, const char **p)
 			return;
 		}
 
-		switch (*wp) {
-		case ('B'):
-			/* TODO */
-			break;
-		case ('I'):
-			/* TODO */
-			break;
-		case ('P'):
-			/* FALLTHROUGH */
-		case ('R'):
-			/* TODO */
-			break;
-		default:
-			break;
-		}
+		/* 
+		 * These aren't supported, as they're symmetry-breaking
+		 * constructs that don't play well with hierarchical
+		 * mark-up.  Consider:
+		 *
+		 * \fBHello.
+		 * .PP
+		 * World.
+		 *
+		 * The style started before "Hello" wouldn't be able to
+		 * propogate into the next `PP' because we'd exit the
+		 * current paragraph's scope.
+		 */
 
 		*p = wp;
 		return;
@@ -322,26 +314,27 @@ print_escape(struct html *h, const char **p)
 static void
 print_encode(struct html *h, const char *p)
 {
+	size_t		 sz;
 
 	for (; *p; p++) {
+		sz = strcspn(p, "\\<>&");
+
+		fwrite(p, 1, sz, stdout);
+		p += /* LINTED */
+			sz;
+
 		if ('\\' == *p) {
 			print_escape(h, &p);
 			continue;
-		}
-		switch (*p) {
-		case ('<'):
-			printf("&lt;");
+		} else if ('\0' == *p)
 			break;
-		case ('>'):
+
+		if ('<' == *p)
+			printf("&lt;");
+		else if ('>' == *p)
 			printf("&gt;");
-			break;
-		case ('&'):
+		else if ('&' == *p)
 			printf("&amp;");
-			break;
-		default:
-			putchar(*p);
-			break;
-		}
 	}
 }
 
@@ -354,8 +347,11 @@ print_otag(struct html *h, enum htmltag tag, 
 	struct tag	*t;
 
 	if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
-		if (NULL == (t = malloc(sizeof(struct tag))))
-			err(EXIT_FAILURE, "malloc");
+		t = malloc(sizeof(struct tag));
+		if (NULL == t) {
+			perror(NULL);
+			exit(EXIT_FAILURE);
+		}
 		t->tag = tag;
 		t->next = h->tags.head;
 		h->tags.head = t;
@@ -364,16 +360,16 @@ print_otag(struct html *h, enum htmltag tag, 
 
 	if ( ! (HTML_NOSPACE & h->flags))
 		if ( ! (HTML_CLRLINE & htmltags[tag].flags))
-			printf(" ");
+			putchar(' ');
 
 	printf("<%s", htmltags[tag].name);
 	for (i = 0; i < sz; i++) {
 		printf(" %s=\"", htmlattrs[p[i].key]);
 		assert(p->val);
 		print_encode(h, p[i].val);
-		printf("\"");
+		putchar('\"');
 	}
-	printf(">");
+	putchar('>');
 
 	h->flags |= HTML_NOSPACE;
 	if (HTML_CLRLINE & htmltags[tag].flags)
@@ -391,11 +387,11 @@ print_ctag(struct html *h, enum htmltag tag)
 {
 	
 	printf("</%s>", htmltags[tag].name);
-	if (HTML_CLRLINE & htmltags[tag].flags)
+	if (HTML_CLRLINE & htmltags[tag].flags) {
 		h->flags |= HTML_NOSPACE;
-	if (HTML_CLRLINE & htmltags[tag].flags)
 		h->flags |= HTML_NEWLINE;
-	else
+		putchar('\n');
+	} else
 		h->flags &= ~HTML_NEWLINE;
 }
 
@@ -440,7 +436,7 @@ print_text(struct html *h, const char *p)
 		}
 
 	if ( ! (h->flags & HTML_NOSPACE))
-		printf(" ");
+		putchar(' ');
 
 	h->flags &= ~HTML_NOSPACE;
 	h->flags &= ~HTML_NEWLINE;
@@ -651,48 +647,28 @@ bufcat_su(struct html *h, const char *p, const struct 
 
 
 void
-html_idcpy(char *dst, const char *src, int sz)
+html_idcat(char *dst, const char *src, int sz)
 {
+	int		 ssz;
 
 	assert(sz);
-	dst[0] = '\0';
-	html_idcat(dst, src, sz);
-}
 
-
-void
-html_idcat(char *dst, const char *src, int sz)
-{
-	int		 i;
-
 	/* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
 
-	for (i = 0; *dst != '\0' && i < sz - 1; dst++, i++)
+	for ( ; *dst != '\0' && sz; dst++, sz--)
 		/* Jump to end. */ ;
 
-	for ( ; *src != '\0' && i < sz - 1; src++, i++) {
-		if (isalnum((u_char)*src)) {
-			*dst++ = *src;
-			continue;
-		}
+	assert(sz > 2);
 
-		switch (*src) {
-		case (';'):
-			*dst++ = ';';
-			break;
-		case ('-'):
-			*dst++ = '-';
-			break;
-		case (':'):
-			*dst++ = ':';
-			break;
-		case ('_'):
-			/* FALLTHROUGH */
-		default:
-			*dst++ = '_';
-			break;
-		}
-	}
+	/* We can't start with a number (bah). */
 
+	*dst++ = 'x';
 	*dst = '\0';
+	sz--;
+
+	for ( ; *src != '\0' && sz > 1; src++) {
+		ssz = snprintf(dst, (size_t)sz, "%.2x", *src);
+		sz -= ssz;
+		dst += ssz;
+	}
 }