From 59b3c105632a6067998ad36631b3eba9b2165219 Mon Sep 17 00:00:00 2001
From: George Koehler <xkernigh@netscape.net>
Date: Sun, 29 Oct 2017 17:45:10 -0400
Subject: [PATCH] Use (arith) 1 << ... when getting the sign bit.

This prevents an overflow reported by @hexcoder- in
https://github.com/davidgiven/ack/issues/56

lang/cem/cpp.ansi/LLlex.c used a plain 1 << ... and caused an overflow
on machines where sizeof(int) < sizeof(long).  Using 1L << ... would
work for now but might fail later if arith became long long.

C doesn't specify whether negative integers use 2's complement or some
other format.  Therefore, (arith) 1 << ... has an undefined value.  It
should still work because the value is some integer where the sign bit
is set and all other bits are clear.

(unsigned arith) 1 << ... would also get the sign bit, but casting it
from unsigned back to signed would make the same undefined value.

(arith) -1 << ... would assume 2's complement.
---
 lang/cem/cpp.ansi/LLlex.c  | 2 +-
 lang/cem/cpp.ansi/arith.h  | 4 ++++
 lang/cem/cpp.ansi/ch3bin.c | 2 --
 lang/m2/comp/cstoper.c     | 2 +-
 lang/m2/comp/type.c        | 2 --
 5 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/lang/cem/cpp.ansi/LLlex.c b/lang/cem/cpp.ansi/LLlex.c
index 7aa0a1eed..5b25a49d4 100644
--- a/lang/cem/cpp.ansi/LLlex.c
+++ b/lang/cem/cpp.ansi/LLlex.c
@@ -261,7 +261,7 @@ garbage:
 		register int base = 10, vch;
 		register arith val = 0;
 		int ovfl = 0;
-		arith ubound = ~(1<<(sizeof(arith)*8-1))/(base/2);
+		arith ubound = max_arith/(base/2);
 
 		/* Since the preprocessor only knows integers and has
 		 * nothing to do with ellipsis we just return when the
diff --git a/lang/cem/cpp.ansi/arith.h b/lang/cem/cpp.ansi/arith.h
index be74a1887..37a5f63fe 100644
--- a/lang/cem/cpp.ansi/arith.h
+++ b/lang/cem/cpp.ansi/arith.h
@@ -16,3 +16,7 @@
 /*	All preprocessor arithmetic should be done in longs.
 */
 #define	arith	long				/* dummy */
+
+#define	arith_size	(sizeof(arith))
+#define	arith_sign	((arith) 1 << (arith_size * 8 - 1))
+#define	max_arith	(~arith_sign)
diff --git a/lang/cem/cpp.ansi/ch3bin.c b/lang/cem/cpp.ansi/ch3bin.c
index f9f43cfbf..c458487a5 100644
--- a/lang/cem/cpp.ansi/ch3bin.c
+++ b/lang/cem/cpp.ansi/ch3bin.c
@@ -8,8 +8,6 @@
 #include	"Lpars.h"
 #include	"arith.h"
 
-#define arith_sign (1L << (sizeof(arith)*8-1))
-
 ch3bin(pval, pis_uns, oper, val, is_uns)
 	register arith *pval, val;
 	int oper, is_uns, *pis_uns;
diff --git a/lang/m2/comp/cstoper.c b/lang/m2/comp/cstoper.c
index 9d7fea099..53dfd0e7b 100644
--- a/lang/m2/comp/cstoper.c
+++ b/lang/m2/comp/cstoper.c
@@ -28,7 +28,7 @@
 
 extern char	*symbol2str();
 
-#define arith_sign	((arith) (1L << (sizeof(arith) * 8 - 1)))
+#define arith_sign	((arith) 1 << (sizeof(arith) * 8 - 1))
 
 #ifndef NOCROSS
 arith full_mask[MAXSIZE+1];/* full_mask[1] == 0xFF, full_mask[2] == 0xFFFF, .. */
diff --git a/lang/m2/comp/type.c b/lang/m2/comp/type.c
index bbf4841b4..5912771fe 100644
--- a/lang/m2/comp/type.c
+++ b/lang/m2/comp/type.c
@@ -51,8 +51,6 @@ arith
 	pointer_size = SZ_POINTER;
 #endif
 
-#define arith_sign	((arith) (1L << (sizeof(arith) * 8 - 1)))
-
 arith	ret_area_size;
 
 t_type