From: Taylor R Campbell <campbell@mumble.net>
Date: Sun, 23 Feb 2014 17:08:55 +0000 (+0000)
Subject: Don't rewrite (flo:- 0. x) as (flo:negate x) on amd64.
X-Git-Tag: release-9.2.0~36
X-Git-Url: https://birchwood-abbey.net/git?a=commitdiff_plain;h=97b43235a9af3d5d3f339ce0189d06d2e07b87e9;p=mit-scheme.git

Don't rewrite (flo:- 0. x) as (flo:negate x) on amd64.

This is wrong because (flo:- 0. 0.) is 0. but (flo:negate 0.) is -0.

This caused some broken code to be miscompiled and yield the correct
result accidentally!
---

diff --git a/src/compiler/machines/x86-64/rulrew.scm b/src/compiler/machines/x86-64/rulrew.scm
index f785b2ce9..3838549ef 100644
--- a/src/compiler/machines/x86-64/rulrew.scm
+++ b/src/compiler/machines/x86-64/rulrew.scm
@@ -237,14 +237,6 @@ USA.
   (QUALIFIER (rtl:constant-flonum-test operand flo:nonzero?))
   (rtl:make-object->float operand))
 
-(define-rule rewriting
-  (FLONUM-2-ARGS FLONUM-SUBTRACT
-                 (REGISTER (? operand1 register-known-value))
-                 (? operand2)
-                 (? overflow?))
-  (QUALIFIER (rtl:constant-flonum-test operand1 flo:zero?))
-  (rtl:make-flonum-1-arg 'FLONUM-NEGATE operand2 overflow?))
-
 (define (rtl:constant-flonum-test expression predicate)
   (and (rtl:object->float? expression)
        (let ((expression (rtl:object->float-expression expression)))