• finally, the code and manual page text to deprecate, and code to not

handle any more, octal 010 style constants, as promised
• overhaul the manpage re. arithmetic expressions, make the guarantees
  mksh code has explicitly, precisely, clear
• to reduce burden of the compiler, getint() now operates on mksh_uari_t
  internally; it already applied the sign after operation, anyway (C99
  guarantees wraparound on unsigned types, but for signed types we need
  specific compiler support; apparently, this comes from hardware limits)
• use const and shuffle order of locals around while here

check.t

@@ -1,4 +1,4 @@
-# $MirOS: src/bin/mksh/check.t,v 1.500 2011/12/09 20:40:02 tg Exp $
+# $MirOS: src/bin/mksh/check.t,v 1.501 2011/12/10 13:34:14 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 $
@@ -28,7 +28,7 @@
 # http://www.freebsd.org/cgi/cvsweb.cgi/src/tools/regression/bin/test/regress.sh?rev=HEAD
 
 expected-stdout:
-	@(#)MIRBSD KSH R40 2011/12/09
+	@(#)MIRBSD KSH R40 2011/12/10
 description:
 	Check version of shell.
 stdin:
@@ -3603,6 +3603,7 @@ expected-stdout:
 name: integer-base-check-flat-posix
 description:
 	Check behaviour of POSuX bases
+category: !nodeprecated
 stdin:
 	echo :$((10)).$((010)).$((0x10)).
 expected-stdout:
@@ -3612,6 +3613,7 @@ name: integer-base-check-flat-right
 description:
 	Check behaviour does not match POSuX, because a not type-safe
 	scripting language has *no* business interpreting "010" as octal
+category: nodeprecated
 stdin:
 	echo :$((10)).$((010)).$((0x10)).
 expected-stdout:

mksh.1

@@ -1,4 +1,4 @@
-.\" $MirOS: src/bin/mksh/mksh.1,v 1.280 2011/12/03 00:09:15 tg Exp $
+.\" $MirOS: src/bin/mksh/mksh.1,v 1.281 2011/12/10 13:34:16 tg Exp $
 .\" $OpenBSD: ksh.1,v 1.141 2011/09/03 22:59:08 jmc Exp $
 .\"-
 .\" Copyright © 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009,
@@ -72,7 +72,7 @@
 .\" with -mandoc, it might implement .Mx itself, but we want to
 .\" use our own definition. And .Dd must come *first*, always.
 .\"
-.Dd $Mdocdate: December 3 2011 $
+.Dd $Mdocdate: December 10 2011 $
 .\"
 .\" Check which macro package we use
 .\"
@@ -2465,14 +2465,8 @@ Grouping operators:
 ( )
 .Ed
 .Pp
-Integer constants and expressions are calculated using the
+Integer constants and expressions are calculated using an exactly 32-bit
-.Vt mksh_ari_t
+wide, signed or unsigned, type with silent wraparound on integer overflow.
-.Po if signed Pc
-or
-.Vt mksh_uari_t
-.Po if unsigned Pc
-type, and are limited to 32 bits.
-Overflows wrap silently.
 Integer constants may be specified with arbitrary bases using the notation
 .Ar base Ns # Ns Ar number ,
 where
@@ -2480,22 +2474,15 @@ where
 is a decimal integer specifying the base, and
 .Ar number
 is a number in the specified base.
-Additionally,
+Additionally, base-16 integers may be specified by prefixing them with
-integers may be prefixed with
+.Sq 0x Pq case-insensitive
-.Sq 0X
+in all forms of arithmetic expressions, except as numeric arguments to the
-or
-.Sq 0x
-(specifying base 16), similar to
-.At
-.Nm ksh ,
-or
-.Sq 0
-(base 8), as an
-.Nm
-extension, in all forms of arithmetic expressions,
-except as numeric arguments to the
 .Ic test
-command.
+built-in command.
+Base-8 integers may be specified by prefixing them with
+.Sq 0 ,
+but this is deprecated and will be removed in the next version of
+.Nm .
 As a special
 .Nm mksh
 extension, numbers to the base of one are treated as either (8-bit

sh.h

@@ -151,9 +151,9 @@
 #endif
 
 #ifdef EXTERN
-__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.509 2011/12/09 20:40:26 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.510 2011/12/10 13:34:18 tg Exp $");
 #endif
-#define MKSH_VERSION "R40 2011/12/09"
+#define MKSH_VERSION "R40 2011/12/10"
 
 /* arithmetics types */
 typedef int32_t mksh_ari_t;

var.c

@@ -26,7 +26,7 @@
 #include <sys/sysctl.h>
 #endif
 
-__RCSID("$MirOS: src/bin/mksh/var.c,v 1.136 2011/11/26 00:45:03 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/var.c,v 1.137 2011/12/10 13:34:19 tg Exp $");
 
 /*-
  * Variables
@@ -469,10 +469,10 @@ setint(struct tbl *vq, mksh_ari_t n)
 static int
 getint(struct tbl *vp, mksh_ari_t *nump, bool arith)
 {
-	char *s;
 	int c, base, neg;
+	mksh_uari_t num;
+	const char *s;
 	bool have_base = false;
-	mksh_ari_t num;
 
 	if (vp->flag&SPECIAL)
 		getspec(vp);
@@ -487,6 +487,13 @@ getint(struct tbl *vp, mksh_ari_t *nump, bool arith)
 	base = 10;
 	num = 0;
 	neg = 0;
+#ifdef MKSH_DISABLE_DEPRECATED
+	if (arith && s[0] == '0' && (s[1] | 0x20) == 'x') {
+		s += 2;
+		base = 16;
+		have_base = true;
+	}
+#else
 	if (arith && *s == '0' && *(s+1)) {
 		s++;
 		if (*s == 'x' || *s == 'X') {
@@ -499,19 +506,20 @@ getint(struct tbl *vp, mksh_ari_t *nump, bool arith)
 			base = 8;
 		have_base = true;
 	}
-	for (c = *s++; c ; c = *s++) {
+#endif
+	while ((c = *s++)) {
 		if (c == '-') {
 			neg++;
 			continue;
 		} else if (c == '#') {
-			base = (int)num;
+			if (have_base || num < 1 || num > 36)
-			if (have_base || base < 1 || base > 36)
 				return (-1);
+			base = (int)num;
 			if (base == 1) {
 				unsigned int wc;
 
 				if (!UTFMODE)
-					wc = *(unsigned char *)s;
+					wc = *(const unsigned char *)s;
 				else if (utf_mbtowc(&wc, s) == (size_t)-1)
 					/* OPTU-8 -> OPTU-16 */
 					/*
@@ -519,7 +527,7 @@ getint(struct tbl *vp, mksh_ari_t *nump, bool arith)
 					 * the same as 1#\x80 does, thus is
 					 * not round-tripping correctly XXX)
 					 */
-					wc = 0xEF00 + *(unsigned char *)s;
+					wc = 0xEF00 + *(const unsigned char *)s;
 				*nump = (mksh_ari_t)wc;
 				return (1);
 			}
@@ -538,9 +546,7 @@ getint(struct tbl *vp, mksh_ari_t *nump, bool arith)
 			return (-1);
 		num = num * base + c;
 	}
-	if (neg)
+	*nump = neg ? -((mksh_ari_t)num) : (mksh_ari_t)num;
-		num = -num;
-	*nump = num;
 	return (base);
 }