===================================================================
RCS file: /cvs/mandoc/roff.c,v
retrieving revision 1.153
retrieving revision 1.155
diff -u -p -r1.153 -r1.155
--- mandoc/roff.c	2011/07/26 14:24:06	1.153
+++ mandoc/roff.c	2011/07/27 07:32:26	1.155
@@ -1,4 +1,4 @@
-/*	$Id: roff.c,v 1.153 2011/07/26 14:24:06 kristaps Exp $ */
+/*	$Id: roff.c,v 1.155 2011/07/27 07:32:26 kristaps Exp $ */
 /*
  * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010, 2011 Ingo Schwarze <schwarze@openbsd.org>
@@ -143,6 +143,12 @@ struct	predef {
 #define	PREDEF(__name, __str) \
 	{ (__name), (__str) },
 
+static	enum rofft	 roffhash_find(const char *, size_t);
+static	void		 roffhash_init(void);
+static	void		 roffnode_cleanscope(struct roff *);
+static	void		 roffnode_pop(struct roff *);
+static	void		 roffnode_push(struct roff *, enum rofft,
+				const char *, int, int);
 static	enum rofferr	 roff_block(ROFF_ARGS);
 static	enum rofferr	 roff_block_text(ROFF_ARGS);
 static	enum rofferr	 roff_block_sub(ROFF_ARGS);
@@ -153,13 +159,16 @@ static	enum rofferr	 roff_cond_text(ROFF_ARGS);
 static	enum rofferr	 roff_cond_sub(ROFF_ARGS);
 static	enum rofferr	 roff_ds(ROFF_ARGS);
 static	enum roffrule	 roff_evalcond(const char *, int *);
+static	void		 roff_free1(struct roff *);
 static	void		 roff_freestr(struct roff *);
 static	char		*roff_getname(struct roff *, char **, int, int);
 static	const char	*roff_getstrn(const struct roff *, 
 				const char *, size_t);
 static	enum rofferr	 roff_line_ignore(ROFF_ARGS);
 static	enum rofferr	 roff_nr(ROFF_ARGS);
-static	int		 roff_res(struct roff *, 
+static	enum rofft	 roff_parse(struct roff *, const char *, int *);
+static	enum rofferr	 roff_parsetext(char *);
+static	void		 roff_res(struct roff *, 
 				char **, size_t *, int, int);
 static	enum rofferr	 roff_rm(ROFF_ARGS);
 static	void		 roff_setstr(struct roff *,
@@ -172,7 +181,7 @@ static	enum rofferr	 roff_EN(ROFF_ARGS);
 static	enum rofferr	 roff_T_(ROFF_ARGS);
 static	enum rofferr	 roff_userdef(ROFF_ARGS);
 
-/* See roff_hash_find() */
+/* See roffhash_find() */
 
 #define	ASCII_HI	 126
 #define	ASCII_LO	 33
@@ -220,20 +229,11 @@ static	const struct predef predefs[PREDEFS_MAX] = {
 #include "predefs.in"
 };
 
-static	void		 roff_free1(struct roff *);
-static	enum rofft	 roff_hash_find(const char *, size_t);
-static	void		 roff_hash_init(void);
-static	void		 roffnode_cleanscope(struct roff *);
-static	void		 roffnode_push(struct roff *, enum rofft,
-				const char *, int, int);
-static	void		 roffnode_pop(struct roff *);
-static	enum rofft	 roff_parse(struct roff *, const char *, int *);
-
-/* See roff_hash_find() */
+/* See roffhash_find() */
 #define	ROFF_HASH(p)	(p[0] - ASCII_LO)
 
 static void
-roff_hash_init(void)
+roffhash_init(void)
 {
 	struct roffmac	 *n;
 	int		  buc, i;
@@ -258,7 +258,7 @@ roff_hash_init(void)
  * the nil-terminated string name could be found.
  */
 static enum rofft
-roff_hash_find(const char *p, size_t s)
+roffhash_find(const char *p, size_t s)
 {
 	int		 buc;
 	struct roffmac	*n;
@@ -387,7 +387,7 @@ roff_alloc(struct mparse *parse)
 	r->parse = parse;
 	r->rstackpos = -1;
 	
-	roff_hash_init();
+	roffhash_init();
 
 	for (i = 0; i < PREDEFS_MAX; i++) 
 		roff_setstr(r, predefs[i].name, predefs[i].str, 0);
@@ -400,8 +400,8 @@ roff_alloc(struct mparse *parse)
  * `\*', e.g., `\*(ab').  These must be handled before the actual line
  * is processed. 
  * This also checks the syntax of regular escapes.
-*/
-static int
+ */
+static void
 roff_res(struct roff *r, char **bufp, size_t *szp, int ln, int pos)
 {
 	enum mandoc_esc	 esc;
@@ -413,8 +413,7 @@ roff_res(struct roff *r, char **bufp, size_t *szp, int
 	size_t		 nsz;
 	char		*n;
 
-	/* Search for a leading backslash and save a pointer to it. */
-
+again:
 	cp = *bufp + pos;
 	while (NULL != (cp = strchr(cp, '\\'))) {
 		stesc = cp++;
@@ -426,7 +425,7 @@ roff_res(struct roff *r, char **bufp, size_t *szp, int
 		 */
 
 		if ('\0' == *cp)
-			return(1);
+			return;
 
 		if ('*' != *cp) {
 			res = cp;
@@ -437,7 +436,7 @@ roff_res(struct roff *r, char **bufp, size_t *szp, int
 			mandoc_msg
 				(MANDOCERR_BADESCAPE, r->parse, 
 				 ln, (int)(stesc - *bufp), NULL);
-			continue;
+			return;
 		}
 
 		cp++;
@@ -450,7 +449,7 @@ roff_res(struct roff *r, char **bufp, size_t *szp, int
 
 		switch (*cp) {
 		case ('\0'):
-			return(1);
+			return;
 		case ('('):
 			cp++;
 			maxl = 2;
@@ -473,7 +472,7 @@ roff_res(struct roff *r, char **bufp, size_t *szp, int
 					(MANDOCERR_BADESCAPE, 
 					 r->parse, ln, 
 					 (int)(stesc - *bufp), NULL);
-				return(1); 
+				return;
 			}
 			if (0 == maxl && ']' == *cp)
 				break;
@@ -495,6 +494,8 @@ roff_res(struct roff *r, char **bufp, size_t *szp, int
 
 		/* Replace the escape sequence by the string. */
 
+		pos += (stesc - *bufp);
+
 		nsz = *szp + strlen(res) + 1;
 		n = mandoc_malloc(nsz);
 
@@ -506,10 +507,52 @@ roff_res(struct roff *r, char **bufp, size_t *szp, int
 
 		*bufp = n;
 		*szp = nsz;
-		return(0);
+		goto again;
 	}
+}
 
-	return(1);
+/*
+ * Process text streams: convert all breakable hyphens into ASCII_HYPH.
+ */
+static enum rofferr
+roff_parsetext(char *p)
+{
+	char		 l, r;
+	size_t		 sz;
+	const char	*start;
+	enum mandoc_esc	 esc;
+
+	start = p;
+
+	while ('\0' != *p) {
+		sz = strcspn(p, "-\\");
+		p += sz;
+
+		if ('\\' == *p) {
+			/* Skip over escapes. */
+			p++;
+			esc = mandoc_escape
+				((const char **)&p, NULL, NULL);
+			if (ESCAPE_ERROR == esc)
+				break;
+			continue;
+		} else if ('-' != *p || p == start)
+			continue;
+
+		l = *(p - 1);
+		r = *(p + 1);
+
+		if ('\\' != l &&
+				'\t' != r && '\t' != l &&
+				' ' != r && ' ' != l &&
+				'-' != r && '-' != l &&
+				! isdigit((unsigned char)l) &&
+			       	! isdigit((unsigned char)r))
+			*p = ASCII_HYPH;
+		p++;
+	}
+
+	return(ROFF_CONT);
 }
 
 enum rofferr
@@ -525,8 +568,7 @@ roff_parseln(struct roff *r, int ln, char **bufp, 
 	 * words to fill in.
 	 */
 
-	if ( ! roff_res(r, bufp, szp, ln, pos))
-		return(ROFF_REPARSE);
+	roff_res(r, bufp, szp, ln, pos);
 
 	ppos = pos;
 	ctl = mandoc_getcontrol(*bufp, &pos);
@@ -551,13 +593,13 @@ roff_parseln(struct roff *r, int ln, char **bufp, 
 			return(eqn_read(&r->eqn, ln, *bufp, pos, offs));
 		if (r->tbl)
 			return(tbl_read(r->tbl, ln, *bufp, pos));
-		return(ROFF_CONT);
+		return(roff_parsetext(*bufp + pos));
 	} else if ( ! ctl) {
 		if (r->eqn)
 			return(eqn_read(&r->eqn, ln, *bufp, pos, offs));
 		if (r->tbl)
 			return(tbl_read(r->tbl, ln, *bufp, pos));
-		return(ROFF_CONT);
+		return(roff_parsetext(*bufp + pos));
 	} else if (r->eqn)
 		return(eqn_read(&r->eqn, ln, *bufp, ppos, offs));
 
@@ -637,7 +679,7 @@ roff_parse(struct roff *r, const char *buf, int *pos)
 	maclen = strcspn(mac + 1, " \\\t\0") + 1;
 
 	t = (r->current_string = roff_getstrn(r, mac, maclen))
-	    ? ROFF_USERDEF : roff_hash_find(mac, maclen);
+	    ? ROFF_USERDEF : roffhash_find(mac, maclen);
 
 	*pos += (int)maclen;