Support Dave Korn’s alternative 「'a'」 (or 「'…'」) form for base-one

integers in addition to my 「1#a」 (or 「1#…」), which also allows for
finer end-of-character checking. Note that this is locale-dependent in
ksh93, set ±U dependent in mksh, and mksh’s OPTU-16 encoding is used.

check.t

@@ -1,4 +1,4 @@
-# $MirOS: src/bin/mksh/check.t,v 1.303 2009/09/06 17:42:11 tg Exp $
+# $MirOS: src/bin/mksh/check.t,v 1.304 2009/09/06 17:55:53 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 2009/08/30
+	@(#)MIRBSD KSH R39 2009/09/06
 description:
 	Check version of shell.
 stdin:
@@ -5520,6 +5520,26 @@ expected-stdout:
 	00000120  000A EFE0 EF80 EF80 - 000A FFFD EFEF EFBF  |.<EFBFBD><EFBFBD><EFBFBD>.<EFBFBD><EFBFBD><EFBFBD>|
 	00000128  EFBE EFEF EFBF EFBF - 000A                 |<EFBFBD><EFBFBD><EFBFBD><EFBFBD>.|
 ---
+name: integer-base-one-4
+description:
+	Check if ksh93-style base-one integers work
+stdin:
+	set -U
+	print 1 $(('a'))
+	(print 2f $(('aa'))) 2>&1 | sed "s/^[^']*'/2p '/"
+	print 3 $(('…'))
+	x="'a'"
+	print "4 <$x>"
+	print 5 $(($x))
+	print 6 $((x))
+expected-stdout:
+	1 97
+	2p 'aa': multi-character character constant
+	3 8230
+	4 <'a'>
+	5 97
+	6 97
+---
 name: ulimit-1
 description:
 	Check if we can use a specific syntax idiom for ulimit

expr.c

@@ -22,7 +22,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/expr.c,v 1.33 2009/08/28 21:01:25 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/expr.c,v 1.34 2009/09/06 17:55:54 tg Exp $");
 
 /* The order of these enums is constrained by the order of opinfo[] */
 enum token {
@@ -528,6 +528,20 @@ exprtoken(Expr_state *es)
 		cp += utf_ptradj(cp);
 		strndupx(tvar, es->tokp, cp - es->tokp, ATEMP);
 		goto process_tvar;
+	} else if (c == '\'') {
+		++cp;
+		cp += utf_ptradj(cp);
+		if (*cp++ != '\'')
+			evalerr(es, ET_STR,
+			    "multi-character character constant");
+		/* 'x' -> 1#x (x = one multibyte character) */
+		c = cp - es->tokp;
+		tvar = alloc(c + /* NUL */ 1, ATEMP);
+		tvar[0] = '1';
+		tvar[1] = '#';
+		memcpy(tvar + 2, es->tokp + 1, c - 2);
+		tvar[c] = '\0';
+		goto process_tvar;
 	} else if (ksh_isdigit(c)) {
 		while (c != '_' && (ksh_isalnux(c) || c == '#'))
 			c = *cp++;

mksh.1

@@ -1,4 +1,4 @@
-.\" $MirOS: src/bin/mksh/mksh.1,v 1.181 2009/09/06 17:42:13 tg Exp $
+.\" $MirOS: src/bin/mksh/mksh.1,v 1.182 2009/09/06 17:55:54 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
@@ -2230,6 +2230,14 @@ extension, numbers to the base of one are treated as either (8-bit
 transparent) ASCII or Unicode codepoints, depending on the shell's
 .Ic utf8\-mode
 flag (current setting).
+The
+.At
+.Nm ksh93
+syntax of
+.Dq \*(aqx\*(aq
+instead of
+.Dq 1#x
+is also supported.
 Note that NUL bytes (integral value of zero) cannot be used.
 In Unicode mode, raw octets are mapped into the range EF80..EFFF as in
 OPTU-8, which is in the PUA and has been assigned by CSUR for this use.

sh.h

@@ -134,9 +134,9 @@
 #endif
 
 #ifdef EXTERN
-__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.338 2009/09/06 17:42:14 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.339 2009/09/06 17:55:55 tg Exp $");
 #endif
-#define MKSH_VERSION "R39 2009/08/30"
+#define MKSH_VERSION "R39 2009/09/06"
 
 #ifndef MKSH_INCLUDES_ONLY