Changeset 94622 in webkit for trunk/Source/JavaScriptCore/ChangeLog

Timestamp:

Sep 6, 2011, 7:00:16 PM (14 years ago)

Author:

[email protected]

Message:

Source/JavaScriptCore: ​https://bugs.webkit.org/show_bug.cgi?id=67486
This reverts r65993 which gives wrong results for rshift
in some corner cases (see the test).

Patch by Yong Li <​[email protected]> on 2011-09-06
Reviewed by Gavin Barraclough.

New test: fast/js/floating-point-truncate-rshift.html

• assembler/ARMAssembler.h:
• assembler/MacroAssemblerARM.h:

(JSC::MacroAssemblerARM::supportsFloatingPointTruncate):
(JSC::MacroAssemblerARM::branchTruncateDoubleToInt32):

LayoutTests: ​https://bugs.webkit.org/show_bug.cgi?id=67486
Add a test case for branchTruncateDoubleToInt32() which is
used in right shift optimization when the assembler
"supportsFloatingPointTruncate".

Patch by Yong Li <​[email protected]> on 2011-09-06
Reviewed by Gavin Barraclough.

• fast/js/floating-point-truncate-rshift-expected.txt: Added.
• fast/js/floating-point-truncate-rshift.html: Added.

File:

• trunk/Source/JavaScriptCore/ChangeLog (modified) (1 diff)

trunk/Source/JavaScriptCore/ChangeLog

@@ +1 @@
+2011-09-06  Yong Li  <[email protected]>
+
+        https://bugs.webkit.org/show_bug.cgi?id=67486
+        This reverts r65993 which gives wrong results for rshift
+        in some corner cases (see the test).
+
+        Reviewed by Gavin Barraclough.
+
+        New test: fast/js/floating-point-truncate-rshift.html
+
+        * assembler/ARMAssembler.h:
+        * assembler/MacroAssemblerARM.h:
+        (JSC::MacroAssemblerARM::supportsFloatingPointTruncate):
+        (JSC::MacroAssemblerARM::branchTruncateDoubleToInt32):
+
 2011-09-06  Filip Pizlo  <[email protected]>