===================================================================
RCS file: /cvs/mandoc/html.c,v
retrieving revision 1.135
retrieving revision 1.140
diff -u -p -r1.135 -r1.140
--- mandoc/html.c	2011/04/30 22:14:02	1.135
+++ mandoc/html.c	2011/05/17 10:48:06	1.140
@@ -1,4 +1,4 @@
-/*	$Id: html.c,v 1.135 2011/04/30 22:14:02 kristaps Exp $ */
+/*	$Id: html.c,v 1.140 2011/05/17 10:48:06 kristaps Exp $ */
 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2011 Ingo Schwarze <schwarze@openbsd.org>
@@ -31,6 +31,7 @@
 #include <unistd.h>
 
 #include "mandoc.h"
+#include "libmandoc.h"
 #include "out.h"
 #include "html.h"
 #include "main.h"
@@ -93,6 +94,19 @@ static	const char	*const htmlattrs[ATTR_MAX] = {
 	"colspan", /* ATTR_COLSPAN */
 };
 
+static	const char	*const roffscales[SCALE_MAX] = {
+	"cm", /* SCALE_CM */
+	"in", /* SCALE_IN */
+	"pc", /* SCALE_PC */
+	"pt", /* SCALE_PT */
+	"em", /* SCALE_EM */
+	"em", /* SCALE_MM */
+	"ex", /* SCALE_EN */
+	"ex", /* SCALE_BU */
+	"em", /* SCALE_VS */
+	"ex", /* SCALE_FS */
+};
+
 static	void		  print_num(struct html *, const char *, size_t);
 static	void		  print_spec(struct html *, const char *, size_t);
 static	void		  print_res(struct html *, const char *, size_t);
@@ -122,7 +136,7 @@ ml_alloc(char *outopts, enum htmltype type)
 
 	h->type = type;
 	h->tags.head = NULL;
-	h->symtab = mchars_init();
+	h->symtab = mchars_alloc();
 
 	while (outopts && *outopts)
 		switch (getsubopt(&outopts, UNCONST(toks), &v)) {
@@ -212,11 +226,10 @@ print_gen_head(struct html *h)
 static void
 print_num(struct html *h, const char *p, size_t len)
 {
-	const char	*rhs;
+	char		 c;
 
-	rhs = mchars_num2char(p, len);
-	if (rhs)
-		putchar((int)*rhs);
+	if ('\0' != (c = mchars_num2char(p, len)))
+		putchar((int)c);
 }
 
 static void
@@ -295,7 +308,43 @@ print_metaf(struct html *h, enum mandoc_esc deco)
 			print_otag(h, TAG_I, 0, NULL);
 }
 
+int
+html_strlen(const char *cp)
+{
+	int		 ssz, sz;
+	const char	*seq, *p;
 
+	/*
+	 * Account for escaped sequences within string length
+	 * calculations.  This follows the logic in term_strlen() as we
+	 * must calculate the width of produced strings.
+	 * Assume that characters are always width of "1".  This is
+	 * hacky, but it gets the job done for approximation of widths.
+	 */
+
+	sz = 0;
+	while (NULL != (p = strchr(cp, '\\'))) {
+		sz += (int)(p - cp);
+		++cp;
+		switch (mandoc_escape(&cp, &seq, &ssz)) {
+		case (ESCAPE_ERROR):
+			return(sz);
+		case (ESCAPE_NUMBERED):
+			/* FALLTHROUGH */
+		case (ESCAPE_PREDEF):
+			/* FALLTHROUGH */
+		case (ESCAPE_SPECIAL):
+			sz++;
+			break;
+		default:
+			break;
+		}
+	}
+
+	assert(sz >= 0);
+	return(sz + strlen(cp));
+}
+
 static int
 print_encode(struct html *h, const char *p, int norecurse)
 {
@@ -693,50 +742,12 @@ void
 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
 {
 	double		 v;
-	const char	*u;
 
 	v = su->scale;
+	if (SCALE_MM == su->unit && 0.0 == (v /= 100.0))
+		v = 1.0;
 
-	switch (su->unit) {
-	case (SCALE_CM):
-		u = "cm";
-		break;
-	case (SCALE_IN):
-		u = "in";
-		break;
-	case (SCALE_PC):
-		u = "pc";
-		break;
-	case (SCALE_PT):
-		u = "pt";
-		break;
-	case (SCALE_EM):
-		u = "em";
-		break;
-	case (SCALE_MM):
-		if (0 == (v /= 100))
-			v = 1;
-		u = "em";
-		break;
-	case (SCALE_EN):
-		u = "ex";
-		break;
-	case (SCALE_BU):
-		u = "ex";
-		break;
-	case (SCALE_VS):
-		u = "em";
-		break;
-	default:
-		u = "ex";
-		break;
-	}
-
-	/* 
-	 * XXX: the CSS spec isn't clear as to which types accept
-	 * integer or real numbers, so we just make them all decimals.
-	 */
-	buffmt(h, "%s: %.2f%s;", p, v, u);
+	buffmt(h, "%s: %.2f%s;", p, v, roffscales[su->unit]);
 }