platform/upstream/v8.git
14 years agoFast string construct stub (ia32 only for now).
vitalyr@chromium.org [Thu, 26 Aug 2010 13:59:37 +0000 (13:59 +0000)]
Fast string construct stub (ia32 only for now).

Review URL: http://codereview.chromium.org/3211002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5357 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix indention in v8natives.js - no functionality change.
ricow@chromium.org [Thu, 26 Aug 2010 13:03:17 +0000 (13:03 +0000)]
Fix indention in v8natives.js - no functionality change.
Review URL: http://codereview.chromium.org/3110041

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5354 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoTemporarily disable test-heap-profiler/HeapSnapshotsDiff until the test is fixed...
ricow@chromium.org [Thu, 26 Aug 2010 09:27:14 +0000 (09:27 +0000)]
Temporarily disable test-heap-profiler/HeapSnapshotsDiff until the test is fixed (issue 848)

Review URL: http://codereview.chromium.org/3209005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5353 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix incorrect encoding of single and double precision registers for some VFP instruct...
erik.corry@gmail.com [Thu, 26 Aug 2010 08:53:00 +0000 (08:53 +0000)]
Fix incorrect encoding of single and double precision registers for some VFP instructions. Also fix incorrect disassembling of vldr/vstr.  This is a commit of codereview.chromium.org/3107027 for Rodolph Perfetta.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5352 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd inlining of binary smi operations in the full codegens on IA32
kasperl@chromium.org [Thu, 26 Aug 2010 08:50:38 +0000 (08:50 +0000)]
Add inlining of binary smi operations in the full codegens on IA32
and x64 (on IA32 we even fold constants into the instructions for
a more compact representation) and prepare the ARM full codegen for
the doing the same there.
Review URL: http://codereview.chromium.org/3195028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5351 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFixes bug in Object.freeze and Object.seal causing them to misbehave when Array.proto...
ricow@chromium.org [Thu, 26 Aug 2010 08:35:49 +0000 (08:35 +0000)]
Fixes bug in Object.freeze and Object.seal causing them to misbehave when Array.prototype has changed.

Object.freeze and Object.seal uses GetOwnPropertyNames which
returns an array with local property names. This array will also have
the additional properties defined on Array.prototype or
Object.prototype.

Note that, the implementation of GetOwnPropertyNames (when used as
Object.getOwnPropertyNames) is correct, since the spec says to create
a new Array (which would also have these properties).

Review URL: http://codereview.chromium.org/3137041

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5350 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoExplicitly set expected number of properties for instances of builtin types.
vitalyr@chromium.org [Wed, 25 Aug 2010 17:50:37 +0000 (17:50 +0000)]
Explicitly set expected number of properties for instances of builtin types.

We used to inherit an estimate on the expected number of properties in
%SetCode. It is not accurate and does not help for the types where
adding custom properties is rare (String, Date, etc.).

Review URL: http://codereview.chromium.org/3191026

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5348 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPrepare push of version 2.3.11 to trunk. Now working on version 2.4.0
ricow@chromium.org [Wed, 25 Aug 2010 15:02:07 +0000 (15:02 +0000)]
Prepare push of version 2.3.11 to trunk. Now working on version 2.4.0

This includes the ChangeLog part from the last push to trunk, since
the "prepare push to trunk" cl for that push was never committed.

Review URL: http://codereview.chromium.org/3111034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5344 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPrepare for optionally inlining smi cases in the code generated
kasperl@chromium.org [Wed, 25 Aug 2010 14:22:03 +0000 (14:22 +0000)]
Prepare for optionally inlining smi cases in the code generated
by the full codegens.
Review URL: http://codereview.chromium.org/3116042

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5343 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoReapplying r5147 (caching maps for slow case objects).
kaznacheev@chromium.org [Wed, 25 Aug 2010 13:25:54 +0000 (13:25 +0000)]
Reapplying r5147 (caching maps for slow case objects).

r5147 wrongly assumed that a code cache for a slow case map is always empty.
This patch solves this: whenever we attempt to add a stub to a map's code cache
we check that this map is cached. If it is we give the object its own copy
of the map and only then modify the map.

Review URL: http://codereview.chromium.org/3134027

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5342 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoChange code contains function to allow a return address just after the last instruction
sgjesse@chromium.org [Wed, 25 Aug 2010 12:51:27 +0000 (12:51 +0000)]
Change code contains function to allow a return address just after the last instruction

Then patching the return sequence for debugger break at return a call is placed at the very end of the instruction stream causing the return address to be just after the actual instruction in the code object.
Review URL: http://codereview.chromium.org/3151037

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5341 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoConvert this.length to uint32 in Array.prototype.[last]indexOf.
lrn@chromium.org [Wed, 25 Aug 2010 12:16:46 +0000 (12:16 +0000)]
Convert this.length to uint32 in Array.prototype.[last]indexOf.

Review URL: http://codereview.chromium.org/3104033

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5340 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoStart using the overwrite mode from the full codegens to generate
kasperl@chromium.org [Wed, 25 Aug 2010 11:10:05 +0000 (11:10 +0000)]
Start using the overwrite mode from the full codegens to generate
slightly better code and allow passing arguments to binary op stubs
in registers on the platforms that support it.
Review URL: http://codereview.chromium.org/3203005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5339 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoMove code stub implementations from codegen-[platform].cc files to new code-stub...
whesse@chromium.org [Wed, 25 Aug 2010 09:44:44 +0000 (09:44 +0000)]
Move code stub implementations from codegen-[platform].cc files to new code-stub-[platform].cc files, and declarations to new code-stub-[platform].h files.
Remaining work is to do the same for platform-independent code stub declarations, and to remove all dependencies on codegen header files from code stub files.
Review URL: http://codereview.chromium.org/3195022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5338 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoImprove the code generated by the full codegen by keeping
kasperl@chromium.org [Wed, 25 Aug 2010 08:57:21 +0000 (08:57 +0000)]
Improve the code generated by the full codegen by keeping
track of the fall-through label and avoiding silly jumps.
Review URL: http://codereview.chromium.org/3166033

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5337 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoHide test failure be being less GC agressive in the test
sgjesse@chromium.org [Wed, 25 Aug 2010 07:26:12 +0000 (07:26 +0000)]
Hide test failure be being less GC agressive in the test

This is a temporary fix until the root cause is resolved.

TBR=kasperl@chromium.org
Review URL: http://codereview.chromium.org/3122038

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5336 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoSimplify the way constructors are called and avoid pushing
kasperl@chromium.org [Wed, 25 Aug 2010 07:18:32 +0000 (07:18 +0000)]
Simplify the way constructors are called and avoid pushing
a dummy receiver that we end up ignoring anyway.
Review URL: http://codereview.chromium.org/3136032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5335 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoUse collector for preparse data.
lrn@chromium.org [Wed, 25 Aug 2010 06:46:53 +0000 (06:46 +0000)]
Use collector for preparse data.

This is in preparation of adding more data to the preparse record.
Using a collector instead of a list avoids copying data when the backing
store grows.

Review URL: http://codereview.chromium.org/3153037

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5334 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix a bug in the handling of debug break in CallIC
sgjesse@chromium.org [Wed, 25 Aug 2010 05:57:02 +0000 (05:57 +0000)]
Fix a bug in the handling of debug break in CallIC

The change of calling convention in the CallIC was not reflected in the debug break code. Without the change to the debug break code the added test crashed.
Review URL: http://codereview.chromium.org/3167037

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5333 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd position information for compares, binary ops, and count
kasperl@chromium.org [Tue, 24 Aug 2010 13:51:23 +0000 (13:51 +0000)]
Add position information for compares, binary ops, and count
operations.
Review URL: http://codereview.chromium.org/3120027

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5332 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPorted string call IC-s to arm.
vitalyr@chromium.org [Tue, 24 Aug 2010 13:45:53 +0000 (13:45 +0000)]
Ported string call IC-s to arm.

Review URL: http://codereview.chromium.org/3143038

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5331 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPorted string call IC-s to x64.
vitalyr@chromium.org [Tue, 24 Aug 2010 13:39:29 +0000 (13:39 +0000)]
Ported string call IC-s to x64.

Review URL: http://codereview.chromium.org/3156045

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5330 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRefactor the tools/test.py script and related testcfg.py files.
ricow@chromium.org [Tue, 24 Aug 2010 13:34:59 +0000 (13:34 +0000)]
Refactor the tools/test.py script and related testcfg.py files.

This makes it possible to run several variants of the tests (with different flags that is) by adding extra lists to the VARIANT_FLAGS list. In addition, there is a number of smaller refactorings.

Review URL: http://codereview.chromium.org/3164023

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5329 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoIntroduce a new intermediate AST node for encapsulating the
kasperl@chromium.org [Tue, 24 Aug 2010 12:56:45 +0000 (12:56 +0000)]
Introduce a new intermediate AST node for encapsulating the
increment part of a count operation.
Review URL: http://codereview.chromium.org/3150032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5328 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoEnsure that scanner state is correctly reset when an error is encountered.
lrn@chromium.org [Tue, 24 Aug 2010 12:29:50 +0000 (12:29 +0000)]
Ensure that scanner state is correctly reset when an error is encountered.

Add a scope object to ensure that leaving a literal scanning prematurely
will clean up after itself.
Also reset the literal buffer if a scanner is reinitialized with a new
source code.

Review URL: http://codereview.chromium.org/3137037

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5327 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoOptimize 'typeof <expr> == <string literal>' in the full codegen.
kasperl@chromium.org [Tue, 24 Aug 2010 11:41:26 +0000 (11:41 +0000)]
Optimize 'typeof <expr> == <string literal>' in the full codegen.
Review URL: http://codereview.chromium.org/3110034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5326 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoCreated collector class and used it to collect identifiers during scanning.
lrn@chromium.org [Tue, 24 Aug 2010 10:53:44 +0000 (10:53 +0000)]
Created collector class and used it to collect identifiers during scanning.

The collector class automatically expands to hold the values added to it,
like a List, but doesn't ensure that the backing store is contiguous, which
allows it to avoid copying back and forth as the buffer grows.

This is in preparation for identifyng identical symbols during preparsing.

Review URL: http://codereview.chromium.org/3181036

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5325 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRemove the full codegen syntax checker completely but be
kasperl@chromium.org [Tue, 24 Aug 2010 09:04:17 +0000 (09:04 +0000)]
Remove the full codegen syntax checker completely but be
careful to avoid making code with loops run too slowly.
Review URL: http://codereview.chromium.org/3107033

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5324 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoCleanup the AST code by removing unused parts and get rid of the
kasperl@chromium.org [Tue, 24 Aug 2010 07:26:49 +0000 (07:26 +0000)]
Cleanup the AST code by removing unused parts and get rid of the
flow graph code completely. Add new AST node for null compares
and use it to make the full codegens better at generating code
for null comparisons.
Review URL: http://codereview.chromium.org/3146037

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5323 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFixed remaining bugs in COW-regexp-results.
lrn@chromium.org [Mon, 23 Aug 2010 13:58:56 +0000 (13:58 +0000)]
Fixed remaining bugs in COW-regexp-results.

Review URL: http://codereview.chromium.org/3197011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5322 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoMove the function name inferrer code from the AstOptimizer to
kasperl@chromium.org [Mon, 23 Aug 2010 13:26:03 +0000 (13:26 +0000)]
Move the function name inferrer code from the AstOptimizer to
the parser in preparation for not using the optimizer when
using the full codegen. Code covered by existing tests.
Review URL: http://codereview.chromium.org/3141034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5321 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoSimplified the full codegens by removing the Expression::kTestValue
kasperl@chromium.org [Mon, 23 Aug 2010 12:55:29 +0000 (12:55 +0000)]
Simplified the full codegens by removing the Expression::kTestValue
and Expression::kValueTest. At the same time, prepare the code for
keeping track of fall through labels to avoid unnecessary jumps.
Review URL: http://codereview.chromium.org/3152042

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5320 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoDon't do a smi check twice in the TO_INTEGER macro.
sandholm@chromium.org [Mon, 23 Aug 2010 11:46:59 +0000 (11:46 +0000)]
Don't do a smi check twice in the TO_INTEGER macro.
Review URL: http://codereview.chromium.org/3115024

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5319 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix bug in 64 bit codegen causing us to always go slowcase in apply.
ricow@chromium.org [Mon, 23 Aug 2010 11:21:09 +0000 (11:21 +0000)]
Fix bug in 64 bit codegen causing us to always go slowcase in apply.

After introducing the direct instruction pointer in functions the
CallApplyLazy method was changed on all platforms. The change made to
the 64 bit version was wrong, not using the right operand for the code
comparisons always forcing us into slow case.

Review URL: http://codereview.chromium.org/3104027

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5317 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix an issue in the ARM port where a left shift was predicted to have a Smi result...
erik.corry@gmail.com [Mon, 23 Aug 2010 09:04:31 +0000 (09:04 +0000)]
Fix an issue in the ARM port where a left shift was predicted to have a Smi result when it had an int32 result.  This is a commit of codereview.chromium.org/3195004 for Rodolph Perfetta.  Part 2

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5316 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix an issue in the ARM port where a left shift was predicted to have a Smi result...
erik.corry@gmail.com [Mon, 23 Aug 2010 07:53:03 +0000 (07:53 +0000)]
Fix an issue in the ARM port where a left shift was predicted to have a Smi result when it had an int32 result.  This is a commit of codereview.chromium.org/3195004 for Rodolph Perfetta

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5315 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRemove the virtual declaration from a couple of non-virtual JumpTarget
ager@chromium.org [Mon, 23 Aug 2010 07:30:09 +0000 (07:30 +0000)]
Remove the virtual declaration from a couple of non-virtual JumpTarget
methods to avoid compiler warnings with some gcc versions.

Review URL: http://codereview.chromium.org/3170036

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5314 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix mismatched new[]/delete for external string resources only used
ager@chromium.org [Mon, 23 Aug 2010 06:30:00 +0000 (06:30 +0000)]
Fix mismatched new[]/delete for external string resources only used
for testing.

Review URL: http://codereview.chromium.org/3165051

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5313 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRe-enable test-api/RegExpInterruption on arm.
ricow@chromium.org [Fri, 20 Aug 2010 13:00:22 +0000 (13:00 +0000)]
Re-enable test-api/RegExpInterruption on arm.

This test has been disabled for a very long time, and the fix in
http://code.google.com/p/v8/source/detail?r=5308 should make this go
away.

I will watch the buildbot continuously over the next few days to make
sure it does not appear again. If it does not I will close bug 240.

Review URL: http://codereview.chromium.org/3127023

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5312 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoWe can use the array trim trick in old paged space as well as
erik.corry@gmail.com [Fri, 20 Aug 2010 12:38:33 +0000 (12:38 +0000)]
We can use the array trim trick in old paged space as well as
new space.
Review URL: http://codereview.chromium.org/3143032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5311 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoUse Copy-on-write arrays for cached regexp results.
lrn@chromium.org [Fri, 20 Aug 2010 09:37:22 +0000 (09:37 +0000)]
Use Copy-on-write arrays for cached regexp results.

Review URL: http://codereview.chromium.org/3158020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5310 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoChange code pointer in function objects to a pointer to the first
ricow@chromium.org [Fri, 20 Aug 2010 07:10:18 +0000 (07:10 +0000)]
Change code pointer in function objects to a pointer to the first
instruction.

By changing the pointer to the code object to a pointer to the first
instruction we can call directly this instruction directly instead of
looking up the address through the code object.

Review URL: http://codereview.chromium.org/3156028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5309 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix local handle leak in RegExpStringModificationTest.
vitalyr@chromium.org [Thu, 19 Aug 2010 14:00:58 +0000 (14:00 +0000)]
Fix local handle leak in RegExpStringModificationTest.

Review URL: http://codereview.chromium.org/3133027

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5308 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRemove unused function.
vitalyr@chromium.org [Thu, 19 Aug 2010 13:28:28 +0000 (13:28 +0000)]
Remove unused function.

Review URL: http://codereview.chromium.org/3107022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5307 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix indentation in utils.h.
ager@chromium.org [Thu, 19 Aug 2010 08:49:26 +0000 (08:49 +0000)]
Fix indentation in utils.h.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5306 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoLanding for Ryan Dahl.
ager@chromium.org [Thu, 19 Aug 2010 08:14:29 +0000 (08:14 +0000)]
Landing for Ryan Dahl.

Remove const from BitCast.

Fixes build for old GCCs.

Review URL: http://codereview.chromium.org/3130033/show

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5305 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix issue 833: eliminate race condition in ProfilerEventsProcessor.
mikhail.naganov@gmail.com [Wed, 18 Aug 2010 15:36:00 +0000 (15:36 +0000)]
Fix issue 833: eliminate race condition in ProfilerEventsProcessor.

When under a load, the ProfilerEventsProcessor thread can start
(enter Run) after Stop has been called, thus resetting running_
data member back to 'true', and starting an infinite loop.

BUG=833

Review URL: http://codereview.chromium.org/3157022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5304 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix two-phase lookup problem in FlexibleBodyVisitor and FixedBodyVisitor.
vegorov@chromium.org [Wed, 18 Aug 2010 15:24:52 +0000 (15:24 +0000)]
Fix two-phase lookup problem in FlexibleBodyVisitor and FixedBodyVisitor.

In FlexibleBodyVisitor and FixedBodyVisitor, the calls to IteratePointers, which lives in the base class BodyVisitorBase<StaticVisitor>, needs to be qualified with the dependent base class to be able to be bound during two-phase lookup.

Review URL: http://codereview.chromium.org/3119026

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5303 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRemove unused typedef.
antonm@chromium.org [Wed, 18 Aug 2010 13:29:20 +0000 (13:29 +0000)]
Remove unused typedef.

Review URL: http://codereview.chromium.org/3164022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5302 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoMake instance_size immediately useful for all fixed size objects.
vitalyr@chromium.org [Wed, 18 Aug 2010 13:00:38 +0000 (13:00 +0000)]
Make instance_size immediately useful for all fixed size objects.

For variable sized objects this field doesn't really make any sense so
by putting a special value there we can improve SizeFromMap().

Review URL: http://codereview.chromium.org/3127016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5301 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix Windows x64 build broken by r5299.
sgjesse@chromium.org [Wed, 18 Aug 2010 11:49:40 +0000 (11:49 +0000)]
Fix Windows x64 build broken by r5299.
Review URL: http://codereview.chromium.org/3158019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5300 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoTracks the maximum usage of executable memory allocated by V8 and allows the histogra...
sgjesse@chromium.org [Wed, 18 Aug 2010 10:45:15 +0000 (10:45 +0000)]
Tracks the maximum usage of executable memory allocated by V8 and allows the histogram data to be gathered and reported.

This patch is contains only the usage tracking logic from 3030048 (already LGTM'd). It does not implement the RWX Limit.

BUG=52122
TEST=Check the V8.ExecutableMemoryMax histogram in the Chrome about:histograms page

Review URL: http://codereview.chromium.org/3161015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5299 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoUse consistent types when calling RegExpImpl::IrregexpExecOnce
sgjesse@chromium.org [Wed, 18 Aug 2010 08:54:25 +0000 (08:54 +0000)]
Use consistent types when calling RegExpImpl::IrregexpExecOnce
Review URL: http://codereview.chromium.org/3167021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5298 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoHeap profiler: allow returning aggregated snapshots via the new API.
mikhail.naganov@gmail.com [Wed, 18 Aug 2010 08:19:29 +0000 (08:19 +0000)]
Heap profiler: allow returning aggregated snapshots via the new API.

This is intended for smoother migration to the new API in Chromium.
Also, aggregated heap snapshots can be used for cheaply obtaining
heap statistics, e.g. in tests.

Review URL: http://codereview.chromium.org/3124024

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5297 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPrepare push of version 2.3.9 to trunk. Now working on versin 2.3.10.
ricow@chromium.org [Wed, 18 Aug 2010 07:14:52 +0000 (07:14 +0000)]
Prepare push of version 2.3.9 to trunk. Now working on versin 2.3.10.

Review URL: http://codereview.chromium.org/3186010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5290 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoDecrease the DST cache increment to take into account the suspension
ager@chromium.org [Wed, 18 Aug 2010 07:01:07 +0000 (07:01 +0000)]
Decrease the DST cache increment to take into account the suspension
of DST in Egypt during Ramadan 2010.

Review URL: http://codereview.chromium.org/3122025

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5289 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix uninitialized field in Debug::ThreadInit().
vitalyr@chromium.org [Tue, 17 Aug 2010 17:46:23 +0000 (17:46 +0000)]
Fix uninitialized field in Debug::ThreadInit().

Originally reported by Maxim.Mossienko on the isolates branch.

Review URL: http://codereview.chromium.org/3164020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5288 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoVarious improvements to oom_dump and instance type lists.
antonm@chromium.org [Tue, 17 Aug 2010 13:48:03 +0000 (13:48 +0000)]
Various improvements to oom_dump and instance type lists.

1) addresses various Mark's concern;
2) adds some missing instance types and rearranges existing ones to follow
  InstanceType order;
3) various minor cleanups.

Review URL: http://codereview.chromium.org/3119023

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5286 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix presubmit and check failures introduced by r5284.
vegorov@chromium.org [Tue, 17 Aug 2010 12:10:27 +0000 (12:10 +0000)]
Fix presubmit and check failures introduced by r5284.

Review URL: http://codereview.chromium.org/3141021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5285 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoMerge flush code phase into marking phase.
vegorov@chromium.org [Tue, 17 Aug 2010 11:44:01 +0000 (11:44 +0000)]
Merge flush code phase into marking phase.

Review URL: http://codereview.chromium.org/3135026

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5284 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix breakpoints on inlined named stores in code from the optimizing compiler
sgjesse@chromium.org [Tue, 17 Aug 2010 11:06:12 +0000 (11:06 +0000)]
Fix breakpoints on inlined named stores in code from the optimizing compiler
Review URL: http://codereview.chromium.org/3164018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5283 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd classname to a static function call.
sgjesse@chromium.org [Tue, 17 Aug 2010 09:27:51 +0000 (09:27 +0000)]
Add classname to a static function call.
Review URL: http://codereview.chromium.org/3125014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5282 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd support for vstr for single precision VFP register. This is a commit of http...
erik.corry@gmail.com [Tue, 17 Aug 2010 08:43:45 +0000 (08:43 +0000)]
Add support for vstr for single precision VFP register.  This is a commit of codereview.chromium.org/3064045 for Rodolph Perfetta

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5281 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix compilation for ARMv4 for OpenBSD/FreeBSD
sgjesse@chromium.org [Tue, 17 Aug 2010 07:17:20 +0000 (07:17 +0000)]
Fix compilation for ARMv4 for OpenBSD/FreeBSD

This is a port of r5260.
Review URL: http://codereview.chromium.org/3165021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5280 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoArray builtins: add inline modifier to help gcc.
vitalyr@chromium.org [Mon, 16 Aug 2010 17:19:38 +0000 (17:19 +0000)]
Array builtins: add inline modifier to help gcc.

Review URL: http://codereview.chromium.org/3181013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5279 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoGive correct value ranges to enumeration types used as opaque types.
vegorov@chromium.org [Mon, 16 Aug 2010 17:14:34 +0000 (17:14 +0000)]
Give correct value ranges to enumeration types used as opaque types.

This allows to remove special handling of GCC 4.4 (disabling of Value Range Propagation) from SConstruct.

BUG=http://code.google.com/p/v8/issues/detail?id=830

Review URL: http://codereview.chromium.org/3135022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5278 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoCopy-on-write arrays.
vitalyr@chromium.org [Mon, 16 Aug 2010 16:06:46 +0000 (16:06 +0000)]
Copy-on-write arrays.

Object model changes
----------------------------------------
New fixed_cow_array_map is used for the elements array of a JSObject
to mark it as COW. The JSObject's map and other fields are not
affected. The JSObject's map still has the "fast elements" bit set. It
means we can do only the receiver map check in keyed loads and the
receiver and the elements map checks in keyed stores. So introducing
COW arrays doesn't hurt performance of these operations. But note that
the elements map check is necessary in all mutating operations because
the "has fast elements" bit now means "has fast elements for reading".
EnsureWritableFastElements can be used in runtime functions to perform
the necessary lazy copying.

Generated code changes
----------------------------------------
Generic keyed load is updated to only do the receiver map check (this
could have been done earlier). FastCloneShallowArrayStub now has two
modes: clone elements and use COW elements. AssertFastElements macro
is added to check the elements when necessary. The custom call IC
generators for Array.prototype.{push,pop} are updated to avoid going
to the slow case (and patching the IC) when calling the builtin should
work.

COW enablement
----------------------------------------
Currently we only put shallow and simple literal arrays in the COW
mode. This is done by the parser.

Review URL: http://codereview.chromium.org/3144002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5275 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPrepare push of version 2.3.8 to trunk. Now working on version 2.3.9.
ager@chromium.org [Mon, 16 Aug 2010 15:49:02 +0000 (15:49 +0000)]
Prepare push of version 2.3.8 to trunk. Now working on version 2.3.9.
Review URL: http://codereview.chromium.org/3117018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5274 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoMake the Integer32 type info only cover the signed 32 bit integers.
erik.corry@gmail.com [Mon, 16 Aug 2010 15:08:32 +0000 (15:08 +0000)]
Make the Integer32 type info only cover the signed 32 bit integers.
Fix some bit op bugs introduced last week on IA32:
http://code.google.com/p/chromium/issues/detail?id=52096
Review URL: http://codereview.chromium.org/3151017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5273 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoInitial implementation of oom_dump utility.
antonm@chromium.org [Mon, 16 Aug 2010 12:36:02 +0000 (12:36 +0000)]
Initial implementation of oom_dump utility.

Note: it intentionally lags behind current HeapStats (which now has os_error field)
as these new minidumps could not be produced by current versions of Chrome yet.

Review URL: http://codereview.chromium.org/3170015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5272 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRemove temporary support for two indexed property query APIs.
antonm@chromium.org [Mon, 16 Aug 2010 11:53:52 +0000 (11:53 +0000)]
Remove temporary support for two indexed property query APIs.

Review URL: http://codereview.chromium.org/3143015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5271 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoARM: Ensure that we are not in a spilled scope when calling
erik.corry@gmail.com [Mon, 16 Aug 2010 11:43:30 +0000 (11:43 +0000)]
ARM: Ensure that we are not in a spilled scope when calling
Load() or constructing a reference.
Review URL: http://codereview.chromium.org/3125011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5270 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd rcr to disassembler.
erik.corry@gmail.com [Mon, 16 Aug 2010 11:18:42 +0000 (11:18 +0000)]
Add rcr to disassembler.
Review URL: http://codereview.chromium.org/3108023

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5268 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoARM: backend opt for ToBoolean: JIT code generation for ToBool
sgjesse@chromium.org [Mon, 16 Aug 2010 07:52:49 +0000 (07:52 +0000)]
ARM: backend opt for ToBoolean: JIT code generation for ToBool

Upgraded the CodeGenerator::ToBoolean() function in the ARM backend to use complete JIT code generation and not make runtime calls to ToBool (when VFP is enabled).

This change also includes the vcmp VFP instruction that supports a constant 0.0 as the second operand.

Patch by Subrato K De <subratokde@codeaurora.org>

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5267 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoLiveEdit: implement stack manipulations for x64
peter.rybin@gmail.com [Fri, 13 Aug 2010 13:54:28 +0000 (13:54 +0000)]
LiveEdit: implement stack manipulations for x64

Review URL: http://codereview.chromium.org/3120011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5264 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoARM: Fix another bug in r5252
sgjesse@chromium.org [Fri, 13 Aug 2010 12:36:59 +0000 (12:36 +0000)]
ARM: Fix another bug in r5252

Also hoist the load of the valueOf symbol out of the loop.
Review URL: http://codereview.chromium.org/3170014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5263 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoARM: Fix bug in r5252
sgjesse@chromium.org [Fri, 13 Aug 2010 11:18:16 +0000 (11:18 +0000)]
ARM: Fix bug in r5252
Review URL: http://codereview.chromium.org/3123013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5262 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd last OS error into heap stats.
antonm@chromium.org [Fri, 13 Aug 2010 11:11:36 +0000 (11:11 +0000)]
Add last OS error into heap stats.

That could allow us to understand why commit of from space sometimes fails.
Another option would be start a separate structure with OS-related info, but
as it's a single field, let's put it into HeapStats, at least for now.

Review URL: http://codereview.chromium.org/3118013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5261 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix compilation for ARMv4.
erik.corry@gmail.com [Fri, 13 Aug 2010 10:19:45 +0000 (10:19 +0000)]
Fix compilation for ARMv4.
Review URL: http://codereview.chromium.org/3161009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5260 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRemove experimental fast-codegen. We are no longer working on this
ager@chromium.org [Fri, 13 Aug 2010 09:07:09 +0000 (09:07 +0000)]
Remove experimental fast-codegen. We are no longer working on this
approach.

Review URL: http://codereview.chromium.org/3152016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5259 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoHandle accessors when generating Error.stack
sgjesse@chromium.org [Fri, 13 Aug 2010 08:31:52 +0000 (08:31 +0000)]
Handle accessors when generating Error.stack

BUG=http://code.google.com/p/v8/issues/detail?id=798
TEST=test/mjsunit/regress/regress-798.js
Review URL: http://codereview.chromium.org/3082012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5258 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAllow allocation in FailedAccessCheckCallback to allow embedders to
ager@chromium.org [Fri, 13 Aug 2010 06:55:44 +0000 (06:55 +0000)]
Allow allocation in FailedAccessCheckCallback to allow embedders to
for instance throw exceptions in case of failed access checks.

Review URL: http://codereview.chromium.org/3165016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5257 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoReduce size of function restarter frame.
peter.rybin@gmail.com [Thu, 12 Aug 2010 17:27:07 +0000 (17:27 +0000)]
Reduce size of function restarter frame.

Review URL: http://codereview.chromium.org/3133007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5256 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoLiveEdit: reflect scope_info moving out of Code
peter.rybin@gmail.com [Thu, 12 Aug 2010 16:01:56 +0000 (16:01 +0000)]
LiveEdit: reflect scope_info moving out of Code

Review URL: http://codereview.chromium.org/3131008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5255 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPreserve constant function transition when adding the same function.
vitalyr@chromium.org [Thu, 12 Aug 2010 14:51:59 +0000 (14:51 +0000)]
Preserve constant function transition when adding the same function.

This should help in cases like:
function Constructor() {
  this.foo = constFunction;
  this.bar = "baz";
}

for (...) {
  o = new Constructor();
  // Constant call IC will work.
  o.foo();
  // Inlined property load will see the same map.
  use(o.bar);
}

This change also fixes a latent bug in custom call IC-s for strings
exposed by string-charcodeat.js.

Review URL: http://codereview.chromium.org/3160006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5254 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRemove runtime function from fuzzing
sgjesse@chromium.org [Thu, 12 Aug 2010 14:38:50 +0000 (14:38 +0000)]
Remove runtime function from fuzzing

Don't do fuzzing of _IsStringWrapperSafeForDefaultValueOf as it expects a non-smi as its first argument.
Review URL: http://codereview.chromium.org/3109010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5253 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoHandle overwriting valueOf on String objects correctly when adding
sgjesse@chromium.org [Thu, 12 Aug 2010 13:43:08 +0000 (13:43 +0000)]
Handle overwriting valueOf on String objects correctly when adding

This adds a check to the fast case string add to ensure that the String object still have the default valueOf function. The default valueOf is sitting on a hidden prototype of String.prototype.

Before using the fast case valueOf the object is checked for a local valueOf property. For slow case objects this check always reports true (the dictionary is not probed, so valueOf might be there) and for fast case objects the descriptor array is checked for the valueOf symbol (just liniar scan). After that the prototype is checked for beeing the initial value of String.prototype. If this all pass (that is the default valueOf is still in place) this result is cached on the map making the check fast the next time.

This is only implemented in the optimizing compiler, as the two usages of %_IsStringWrapperSafeForDefaultValueOf is never hit by the full compiler.

I will port to x64 and ARM when this has been reviewed for ia32.

I will remove the performance counters prior to final commit.

BUG=http://code.google.com/p/v8/issues/detail?id=760
TEST=test/mjsunit/regress/regress-760-1.js
TEST=test/mjsunit/regress/regress-760-2.js

Review URL: http://codereview.chromium.org/3117006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5252 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix strict aliasing rule violations in dtoa.c.
vegorov@chromium.org [Thu, 12 Aug 2010 11:43:10 +0000 (11:43 +0000)]
Fix strict aliasing rule violations in dtoa.c.

Review URL: http://codereview.chromium.org/3172008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5251 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd object-visiting.cc into GYP/XCode/MSVC project files.
vegorov@chromium.org [Wed, 11 Aug 2010 15:18:01 +0000 (15:18 +0000)]
Add object-visiting.cc into GYP/XCode/MSVC project files.

Review URL: http://codereview.chromium.org/3166008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5248 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix compilation on Mac.
vegorov@chromium.org [Wed, 11 Aug 2010 14:51:02 +0000 (14:51 +0000)]
Fix compilation on Mac.

Review URL: http://codereview.chromium.org/3180001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5247 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoGeneralize virtually dispatched scavenger to virtually dispatched specialized visitors.
vegorov@chromium.org [Wed, 11 Aug 2010 14:30:14 +0000 (14:30 +0000)]
Generalize virtually dispatched scavenger to virtually dispatched specialized visitors.

Review URL: http://codereview.chromium.org/3066044

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5246 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoX64: Make SmiCompare use the Smi-register when possible.
lrn@chromium.org [Wed, 11 Aug 2010 14:23:12 +0000 (14:23 +0000)]
X64: Make SmiCompare use the Smi-register when possible.

Review URL: http://codereview.chromium.org/3104007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5245 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoChange lazy compilation stub to a builtin.
ricow@chromium.org [Wed, 11 Aug 2010 13:48:58 +0000 (13:48 +0000)]
Change lazy compilation stub to a builtin.

This change changes the lazy compilation stub to a builtin and
eliminates the argc (argument count for the function for which to
create a lazy stub) parameter.

Review URL: http://codereview.chromium.org/3146008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5244 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoMinor change to for-in
sgjesse@chromium.org [Wed, 11 Aug 2010 13:46:10 +0000 (13:46 +0000)]
Minor change to for-in

Return (smi) 0 instead of object null from the FILTER_KEY builtin.

Add a test which tests keys being deleted during for-in.
Review URL: http://codereview.chromium.org/3170004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5243 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPrepare push to trunk. Now working on version 2.3.8.
vegorov@chromium.org [Wed, 11 Aug 2010 13:12:56 +0000 (13:12 +0000)]
Prepare push to trunk. Now working on version 2.3.8.
Review URL: http://codereview.chromium.org/3166005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5240 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix compilation with strict aliasing enabled on GCC 4.4.1
vegorov@chromium.org [Wed, 11 Aug 2010 13:01:28 +0000 (13:01 +0000)]
Fix compilation with strict aliasing enabled on GCC 4.4.1

Review URL: http://codereview.chromium.org/3166006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5239 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoARM: Remove some spill scopes
sgjesse@chromium.org [Wed, 11 Aug 2010 10:55:44 +0000 (10:55 +0000)]
ARM: Remove some spill scopes
Review URL: http://codereview.chromium.org/3119005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5238 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAllow compiling with strict aliasing enabled on GCC 4.4.
lrn@chromium.org [Wed, 11 Aug 2010 10:52:34 +0000 (10:52 +0000)]
Allow compiling with strict aliasing enabled on GCC 4.4.
Patch by Jay Freeman.
Fixes issue 463.

Review URL: http://codereview.chromium.org/3117007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5237 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoHeap profiler: enforce compaction before taking a snapshot.
mikhail.naganov@gmail.com [Wed, 11 Aug 2010 09:54:30 +0000 (09:54 +0000)]
Heap profiler: enforce compaction before taking a snapshot.

Review URL: http://codereview.chromium.org/3118007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5236 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix issue 822: handling of JSObject::elements in CalculateNetworkSize.
mikhail.naganov@gmail.com [Wed, 11 Aug 2010 09:03:23 +0000 (09:03 +0000)]
Fix issue 822: handling of JSObject::elements in CalculateNetworkSize.

BUG=822
TEST=test-heap-profiler/Issue822

Review URL: http://codereview.chromium.org/3128004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5235 ce2b1a6d-e550-0410-aec6-3dcde31c8c00