since we have wcwidth anyway, expose it as ${%strvar} to the user, iff

utf8-mode is enabled (otherwise it'll be a synonym for ${#strvar} aka
the number of octets in it)

check.t

@@ -1,4 +1,4 @@
-# $MirOS: src/bin/mksh/check.t,v 1.340 2009/11/21 22:32:05 tg Exp $
+# $MirOS: src/bin/mksh/check.t,v 1.341 2009/11/21 23:23:16 tg Exp $
 # $OpenBSD: bksl-nl.t,v 1.2 2001/01/28 23:04:56 niklas Exp $
 # $OpenBSD: history.t,v 1.5 2001/01/28 23:04:56 niklas Exp $
 # $OpenBSD: read.t,v 1.3 2003/03/10 03:48:16 david Exp $
@@ -25,7 +25,7 @@
 # http://www.research.att.com/~gsf/public/ifs.sh
 
 expected-stdout:
-	@(#)MIRBSD KSH R39 2009/11/09
+	@(#)MIRBSD KSH R39 2009/11/21
 description:
 	Check version of shell.
 stdin:
@@ -6680,3 +6680,62 @@ expected-stdout:
 	done
 expected-stderr-pattern: /.*-x.*option/
 ---
+name: wcswidth-1
+description:
+	Check the new wcswidth feature
+stdin:
+	s=何
+	set +U
+	print octets: ${#s} ${%s} .
+	set -U
+	print characters: ${#s} .
+	print columns: ${%s} .
+expected-stdout:
+	octets: 3 3 .
+	characters: 1 .
+	columns: 2 .
+---
+name: wcswidth-2
+description:
+	Check some corner cases
+stdin:
+	print % $% .
+expected-stdout:
+	% $% .
+---
+name: wcswidth-3
+description:
+	Check some corner cases
+stdin:
+	print ${%} .
+expected-stderr-pattern:
+	/bad substitution/
+expected-exit: 1
+---
+name: wcswidth-4a
+description:
+	Check some corner cases
+stdin:
+	print ${%*} .
+expected-stderr-pattern:
+	/bad substitution/
+expected-exit: 1
+---
+name: wcswidth-4b
+description:
+	Check some corner cases
+stdin:
+	print ${%@} .
+expected-stderr-pattern:
+	/bad substitution/
+expected-exit: 1
+---
+name: wcswidth-4c
+description:
+	Check some corner cases
+stdin:
+	:
+	print ${%?} .
+expected-stdout:
+	1 .
+---

eval.c

@@ -22,7 +22,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/eval.c,v 1.72 2009/11/21 22:32:08 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/eval.c,v 1.73 2009/11/21 23:23:17 tg Exp $");
 
 /*
  * string expansion
@@ -911,14 +911,20 @@ varsub(Expand *xp, const char *sp, const char *word,
 	struct tbl *vp;
 	bool zero_ok = false;
 
-	if (sp[0] == '\0')	/* Bad variable name */
+	if ((stype = sp[0]) == '\0')	/* Bad variable name */
 		return (-1);
 
 	xp->var = NULL;
 
-	/* ${#var}, string length or array size */
+	/*-
-	if (sp[0] == '#' && (c = sp[1]) != '\0') {
+	 * ${#var}, string length (-U: characters, +U: octets) or array size
-		/* Can't have any modifiers for ${#...} */
+	 * ${%var}, string width (-U: screen columns, +U: octets)
+	 */
+	c = sp[1];
+	if (stype == '%' && c == '\0')
+		return (-1);
+	if ((stype == '#' || stype == '%') && c != '\0') {
+		/* Can't have any modifiers for ${#...} or ${%...} */
 		if (*word != CSUBST)
 			return (-1);
 		sp++;
@@ -927,6 +933,8 @@ varsub(Expand *xp, const char *sp, const char *word,
 		    p[2] == ']') {
 			int n = 0;
 
+			if (stype != '#')
+				return (-1);
 			vp = global(arrayname(sp));
 			if (vp->flag & (ISSET|ARRAY))
 				zero_ok = true;
@@ -934,17 +942,19 @@ varsub(Expand *xp, const char *sp, const char *word,
 				if (vp->flag & ISSET)
 					n++;
 			c = n;
-		} else if (c == '*' || c == '@')
+		} else if (c == '*' || c == '@') {
+			if (stype != '#')
+				return (-1);
 			c = e->loc->argc;
-		else {
+		} else {
 			p = str_val(global(sp));
 			zero_ok = p != null;
-			c = utflen(p);
+			c = stype == '#' ? (int)utflen(p) : utf_mbswidth(p);
 		}
 		if (Flag(FNOUNSET) && c == 0 && !zero_ok)
 			errorf("%s: parameter not set", sp);
 		*stypep = 0; /* unqualified variable/string substitution */
-		xp->str = shf_smprintf("%lu", (unsigned long)c);
+		xp->str = shf_smprintf("%u", (unsigned int)c);
 		return (XSUB);
 	}
 

lex.c

@@ -22,7 +22,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/lex.c,v 1.100 2009/10/04 12:45:22 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/lex.c,v 1.101 2009/11/21 23:23:18 tg Exp $");
 
 /*
  * states while lexing word
@@ -1548,7 +1548,7 @@ get_brace_var(XString *wsp, char *wp)
 		/* State machine to figure out where the variable part ends. */
 		switch (state) {
 		case PS_INITIAL:
-			if (c == '#' || c == '!') {
+			if (c == '#' || c == '!' || c == '%') {
 				state = PS_SAW_HASH;
 				break;
 			}

mksh.1

@@ -1,4 +1,4 @@
-.\" $MirOS: src/bin/mksh/mksh.1,v 1.197 2009/11/21 23:21:23 tg Exp $
+.\" $MirOS: src/bin/mksh/mksh.1,v 1.198 2009/11/21 23:23:18 tg Exp $
 .\" $OpenBSD: ksh.1,v 1.129 2009/05/28 06:09:06 jmc Exp $
 .\"-
 .\" Copyright © 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009
@@ -1449,6 +1449,12 @@ of the string value of parameter
 The number of elements in the array
 .Ar name .
 .Pp
+.It Pf ${% Ns Ar name Ns \&}
+The width
+.Pq in screen columns
+of the string value of parameter
+.Ar name .
+.Pp
 .It Pf ${! Ns Ar name Ns }
 The name of the variable referred to by
 .Ar name .

sh.h

@@ -134,9 +134,9 @@
 #endif
 
 #ifdef EXTERN
-__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.360 2009/11/09 23:35:11 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.361 2009/11/21 23:23:20 tg Exp $");
 #endif
-#define MKSH_VERSION "R39 2009/11/09"
+#define MKSH_VERSION "R39 2009/11/21"
 
 #ifndef MKSH_INCLUDES_ONLY