We don't do hyphenation, but we allow breaking the line at hyphens that are

already there in the middle of words.  So far, we only allowed this on text
lines.  Now it turns out some macros allow this for their arguments, too,
in particular .Nd and most of the .%? citation macros.

Issue found by Franco Fichtner <franco at lastsummer dot de> while doing
systematic groff-mandoc comparisons in the DragonFly base system, THANKS!

While here, garbage collect two empty prevalidator function pointer lists
and sort a couple of function declarations.

1 files changed, 62 insertions, 20 deletions

diff --git a/mdoc_validate.c b/mdoc_validate.c
index c13422f2..415454db 100644
--- a/mdoc_validate.c
+++ b/mdoc_validate.c
@@ -1,4 +1,4 @@
-/*	$Id: mdoc_validate.c,v 1.195 2013/10/06 13:32:46 schwarze Exp $ */
+/*	$Id: mdoc_validate.c,v 1.196 2013/10/06 22:46:15 schwarze Exp $ */
 /*
  * Copyright (c) 2008-2012 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010, 2011, 2012, 2013 Ingo Schwarze <schwarze@openbsd.org>
@@ -19,7 +19,7 @@
 #include "config.h"
 #endif
 
-#ifndef	OSNAME
+#ifndef OSNAME
 #include <sys/utsname.h>
 #endif
 
@@ -97,18 +97,19 @@ static	int	 post_bl_block_width(POST_ARGS);
 static	int	 post_bl_block_tag(POST_ARGS);
 static	int	 post_bl_head(POST_ARGS);
 static	int	 post_bx(POST_ARGS);
+static	int	 post_defaults(POST_ARGS);
 static	int	 post_dd(POST_ARGS);
 static	int	 post_dt(POST_ARGS);
-static	int	 post_defaults(POST_ARGS);
-static	int	 post_literal(POST_ARGS);
 static	int	 post_eoln(POST_ARGS);
+static	int	 post_hyph(POST_ARGS);
+static	int	 post_ignpar(POST_ARGS);
 static	int	 post_it(POST_ARGS);
 static	int	 post_lb(POST_ARGS);
+static	int	 post_literal(POST_ARGS);
 static	int	 post_nm(POST_ARGS);
 static	int	 post_ns(POST_ARGS);
 static	int	 post_os(POST_ARGS);
 static	int	 post_par(POST_ARGS);
-static	int	 post_ignpar(POST_ARGS);
 static	int	 post_prol(POST_ARGS);
 static	int	 post_root(POST_ARGS);
 static	int	 post_rs(POST_ARGS);
@@ -142,28 +143,30 @@ static	v_post	 posts_bx[] = { post_bx, NULL };
 static	v_post	 posts_bool[] = { ebool, NULL };
 static	v_post	 posts_eoln[] = { post_eoln, NULL };
 static	v_post	 posts_defaults[] = { post_defaults, NULL };
+static	v_post	 posts_d1[] = { bwarn_ge1, post_hyph, NULL };
 static	v_post	 posts_dd[] = { post_dd, post_prol, NULL };
 static	v_post	 posts_dl[] = { post_literal, bwarn_ge1, NULL };
 static	v_post	 posts_dt[] = { post_dt, post_prol, NULL };
 static	v_post	 posts_fo[] = { hwarn_eq1, bwarn_ge1, NULL };
+static	v_post	 posts_hyph[] = { post_hyph, NULL };
+static	v_post	 posts_hyphtext[] = { ewarn_ge1, post_hyph, NULL };
 static	v_post	 posts_it[] = { post_it, NULL };
 static	v_post	 posts_lb[] = { post_lb, NULL };
-static	v_post	 posts_nd[] = { berr_ge1, NULL };
+static	v_post	 posts_nd[] = { berr_ge1, post_hyph, NULL };
 static	v_post	 posts_nm[] = { post_nm, NULL };
 static	v_post	 posts_notext[] = { ewarn_eq0, NULL };
 static	v_post	 posts_ns[] = { post_ns, NULL };
 static	v_post	 posts_os[] = { post_os, post_prol, NULL };
 static	v_post	 posts_pp[] = { post_par, ewarn_eq0, NULL };
 static	v_post	 posts_rs[] = { post_rs, NULL };
-static	v_post	 posts_sh[] = { post_ignpar, hwarn_ge1, post_sh, NULL };
+static	v_post	 posts_sh[] = { post_ignpar,hwarn_ge1,post_sh,post_hyph,NULL };
 static	v_post	 posts_sp[] = { post_par, ewarn_le1, NULL };
-static	v_post	 posts_ss[] = { post_ignpar, hwarn_ge1, NULL };
+static	v_post	 posts_ss[] = { post_ignpar, hwarn_ge1, post_hyph, NULL };
 static	v_post	 posts_st[] = { post_st, NULL };
 static	v_post	 posts_std[] = { post_std, NULL };
 static	v_post	 posts_text[] = { ewarn_ge1, NULL };
 static	v_post	 posts_text1[] = { ewarn_eq1, NULL };
 static	v_post	 posts_vt[] = { post_vt, NULL };
-static	v_post	 posts_wline[] = { bwarn_ge1, NULL };
 static	v_pre	 pres_an[] = { pre_an, NULL };
 static	v_pre	 pres_bd[] = { pre_display, pre_bd, pre_literal, pre_par, NULL };
 static	v_pre	 pres_bl[] = { pre_bl, pre_par, NULL };
@@ -171,8 +174,6 @@ static	v_pre	 pres_d1[] = { pre_display, NULL };
 static	v_pre	 pres_dl[] = { pre_literal, pre_display, NULL };
 static	v_pre	 pres_dd[] = { pre_dd, NULL };
 static	v_pre	 pres_dt[] = { pre_dt, NULL };
-static	v_pre	 pres_er[] = { NULL, NULL };
-static	v_pre	 pres_fd[] = { NULL, NULL };
 static	v_pre	 pres_it[] = { pre_it, pre_par, NULL };
 static	v_pre	 pres_os[] = { pre_os, NULL };
 static	v_pre	 pres_pp[] = { pre_par, NULL };
@@ -188,7 +189,7 @@ static	const struct valids mdoc_valids[MDOC_MAX] = {
 	{ pres_sh, posts_sh },			/* Sh */ 
 	{ pres_ss, posts_ss },			/* Ss */ 
 	{ pres_pp, posts_pp },			/* Pp */ 
-	{ pres_d1, posts_wline },		/* D1 */
+	{ pres_d1, posts_d1 },			/* D1 */
 	{ pres_dl, posts_dl },			/* Dl */
 	{ pres_bd, posts_bd },			/* Bd */
 	{ NULL, NULL },				/* Ed */
@@ -201,11 +202,11 @@ static	const struct valids mdoc_valids[MDOC_MAX] = {
 	{ NULL, NULL },				/* Cd */ 
 	{ NULL, NULL },				/* Cm */
 	{ NULL, NULL },				/* Dv */ 
-	{ pres_er, NULL },			/* Er */ 
+	{ NULL, NULL },				/* Er */ 
 	{ NULL, NULL },				/* Ev */ 
 	{ pres_std, posts_std },		/* Ex */ 
 	{ NULL, NULL },				/* Fa */ 
-	{ pres_fd, posts_text },		/* Fd */
+	{ NULL, posts_text },			/* Fd */
 	{ NULL, NULL },				/* Fl */
 	{ NULL, NULL },				/* Fn */ 
 	{ NULL, NULL },				/* Ft */ 
@@ -223,15 +224,15 @@ static	const struct valids mdoc_valids[MDOC_MAX] = {
 	{ NULL, posts_vt },			/* Vt */ 
 	{ NULL, posts_text },			/* Xr */ 
 	{ NULL, posts_text },			/* %A */
-	{ NULL, posts_text },			/* %B */ /* FIXME: can be used outside Rs/Re. */
+	{ NULL, posts_hyphtext },		/* %B */ /* FIXME: can be used outside Rs/Re. */
 	{ NULL, posts_text },			/* %D */
 	{ NULL, posts_text },			/* %I */
 	{ NULL, posts_text },			/* %J */
-	{ NULL, posts_text },			/* %N */
-	{ NULL, posts_text },			/* %O */
+	{ NULL, posts_hyphtext },		/* %N */
+	{ NULL, posts_hyphtext },		/* %O */
 	{ NULL, posts_text },			/* %P */
-	{ NULL, posts_text },			/* %R */
-	{ NULL, posts_text },			/* %T */ /* FIXME: can be used outside Rs/Re. */
+	{ NULL, posts_hyphtext },		/* %R */
+	{ NULL, posts_hyphtext },		/* %T */ /* FIXME: can be used outside Rs/Re. */
 	{ NULL, posts_text },			/* %V */
 	{ NULL, NULL },				/* Ac */
 	{ NULL, NULL },				/* Ao */
@@ -271,7 +272,7 @@ static	const struct valids mdoc_valids[MDOC_MAX] = {
 	{ NULL, NULL },				/* So */
 	{ NULL, NULL },				/* Sq */
 	{ NULL, posts_bool },			/* Sm */ 
-	{ NULL, NULL },				/* Sx */
+	{ NULL, posts_hyph },			/* Sx */
 	{ NULL, NULL },				/* Sy */
 	{ NULL, NULL },				/* Tn */
 	{ NULL, NULL },				/* Ux */
@@ -1856,6 +1857,47 @@ post_rs(POST_ARGS)
 	return(1);
 }
 
+/*
+ * For some arguments of some macros,
+ * convert all breakable hyphens into ASCII_HYPH.
+ */
+static int
+post_hyph(POST_ARGS)
+{
+	struct mdoc_node	*n, *nch;
+	char			*cp;
+
+	n = mdoc->last;
+	switch (n->type) {
+	case (MDOC_HEAD):
+		if (MDOC_Sh == n->tok || MDOC_Ss == n->tok)
+			break;
+		return(1);
+	case (MDOC_BODY):
+		if (MDOC_D1 == n->tok || MDOC_Nd == n->tok)
+			break;
+		return(1);
+	case (MDOC_ELEM):
+		break;
+	default:
+		return(1);
+	}
+
+	for (nch = n->child; nch; nch = nch->next) {
+		if (MDOC_TEXT != nch->type)
+			continue;
+		cp = nch->string;
+		if (3 > strnlen(cp, 3))
+			continue;
+		while ('\0' != *(++cp))
+			if ('-' == *cp &&
+			    isalpha((unsigned char)cp[-1]) &&
+			    isalpha((unsigned char)cp[1]))
+				*cp = ASCII_HYPH;
+	}
+	return(1);
+}
+
 static int
 post_ns(POST_ARGS)
 {