Rules that perform index->fixnum conversions must use arithmetic left

shift so that proper testing for overflow can occur.  The reason:
these rules also get invoked by generic arithmetic because the code
compressor sees them and combines instruction pairs to use them.

diff --git a/v7/src/compiler/machines/bobcat/make.scm-68040 b/v7/src/compiler/machines/bobcat/make.scm-68040
index dd00844e730caca39bb81a53b4ec08e035e658fb..62c2a1ec7cbee36bdf9b69445f219da22cbfe9a9 100644 (file)
--- a/v7/src/compiler/machines/bobcat/make.scm-68040
+++ b/v7/src/compiler/machines/bobcat/make.scm-68040
@@ -1,6 +1,6 @@
 #| -*-Scheme-*-
 
-$Header: /Users/cph/tmp/foo/mit-scheme/mit-scheme/v7/src/compiler/machines/bobcat/make.scm-68040,v 4.58 1989/11/02 08:08:54 cph Exp $
+$Header: /Users/cph/tmp/foo/mit-scheme/mit-scheme/v7/src/compiler/machines/bobcat/make.scm-68040,v 4.59 1989/11/15 02:48:07 cph Exp $
 
 Copyright (c) 1988, 1989 Massachusetts Institute of Technology
 
@@ -41,4 +41,4 @@ MIT in each case. |#
            ((package/reference (find-package name) 'INITIALIZE-PACKAGE!)))
          '((COMPILER MACROS)
            (COMPILER DECLARATIONS)))
-(add-system! (make-system "Liar (Motorola MC68020)" 4 58 '()))
\ No newline at end of file
+(add-system! (make-system "Liar (Motorola MC68020)" 4 59 '()))
\ No newline at end of file