platform/upstream/v8.git
13 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

13 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

13 years agoBump the kPartialSnapshotCacheCapacity to 1400 to fix testing with snapshot on.
ricow@chromium.org [Wed, 11 Aug 2010 08:43:15 +0000 (08:43 +0000)]
Bump the kPartialSnapshotCacheCapacity to 1400 to fix testing with snapshot on.

Review URL: http://codereview.chromium.org/3172003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5234 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoPut direct code pointers into JSFunction objects. This is a first step
ricow@chromium.org [Wed, 11 Aug 2010 08:12:53 +0000 (08:12 +0000)]
Put direct code pointers into JSFunction objects. This is a first step
in allowing more flexible compilation and to simplify builtins lookup.

This changes a number of places where code objects are assigned to
SharedFunctionInfo objects to also assign this code object to the
JSFunction. In addition, the code flushing is changed slightly to
accomodate this (we need to flush the code from functions pointing to
SharedFunctionInfo objects that has already been flushed).

Review URL: http://codereview.chromium.org/3120006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5233 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoRemoved support for object literal get/set with number/string property name.
lrn@chromium.org [Tue, 10 Aug 2010 12:44:13 +0000 (12:44 +0000)]
Removed support for object literal get/set with number/string property name.
It doesn't work correctly for array indices.

Review URL: http://codereview.chromium.org/3109002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5232 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoFix fuzzer-found error where left and right were the same register in bitops.
erik.corry@gmail.com [Tue, 10 Aug 2010 12:30:14 +0000 (12:30 +0000)]
Fix fuzzer-found error where left and right were the same register in bitops.
Review URL: http://codereview.chromium.org/3115004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5231 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoFix CPU profiler crash in start / stop sequence when non-existent name is passed
mikhail.naganov@gmail.com [Tue, 10 Aug 2010 12:06:42 +0000 (12:06 +0000)]
Fix CPU profiler crash in start / stop sequence when non-existent name is passed

BUG=51594
TEST=test-cpu-profiler/CrashIfStoppingLastNonExistentProfile

Review URL: http://codereview.chromium.org/3108004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5230 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoRemove a couple of spill scopes
sgjesse@chromium.org [Tue, 10 Aug 2010 11:52:18 +0000 (11:52 +0000)]
Remove a couple of spill scopes
Review URL: http://codereview.chromium.org/3015066

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5229 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoFirst phase of migration to new indexed property query callbacks.
antonm@chromium.org [Tue, 10 Aug 2010 10:05:18 +0000 (10:05 +0000)]
First phase of migration to new indexed property query callbacks.

Eventually indexed property query callbacks will return attributes
(as an integer) or an empty handle if property is not intercepted.

To gradually migrate to this new API, USE_NEW_QUERY_CALLBACK
macro would control if old or new style API is used.

So the migration plan is:

1) introduce new API which should be explictily enabled;
2) switch to new API defining USE_NEW_QUERY_CALLBACK before
include of <v8.h> (that would require changes to client code as well)
3) remove old API from v8
4) remove #define USE_NEW_QUERY_CALLBACK from clients.

BUG=http://code.google.com/p/v8/issues/detail?id=816

Review URL: http://codereview.chromium.org/3101001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5228 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoAdd v8::Value::IsRegExp method.
antonm@chromium.org [Tue, 10 Aug 2010 09:53:56 +0000 (09:53 +0000)]
Add v8::Value::IsRegExp method.

Patch by Vlad Burlik.

Review URL: http://codereview.chromium.org/3064051

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5227 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoEnable static assertions in release mode.
mikhail.naganov@gmail.com [Tue, 10 Aug 2010 09:22:49 +0000 (09:22 +0000)]
Enable static assertions in release mode.

Review URL: http://codereview.chromium.org/3146004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5226 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoFix lint issue.
mikhail.naganov@gmail.com [Tue, 10 Aug 2010 07:37:59 +0000 (07:37 +0000)]
Fix lint issue.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5225 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoHeap profiler: Add static assertions for heap snapshot entities' sizes.
mikhail.naganov@gmail.com [Tue, 10 Aug 2010 07:30:14 +0000 (07:30 +0000)]
Heap profiler: Add static assertions for heap snapshot entities' sizes.

Review URL: http://codereview.chromium.org/2808112

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5224 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoA follow-up to r5211: fix a couple of issues detected on Windows.
mikhail.naganov@gmail.com [Mon, 9 Aug 2010 14:57:13 +0000 (14:57 +0000)]
A follow-up to r5211: fix a couple of issues detected on Windows.

 - storage of enums in bit fields;
 - removing dead entries from address -> id map in HeapObjectsMap;
 - layout of HeapEntry, to avoid class size increase on ia32 due to alignment;

Review URL: http://codereview.chromium.org/3096008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5217 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoFix bug in binary operations code. When loading a 32-bit integer as a
ager@chromium.org [Mon, 9 Aug 2010 14:54:23 +0000 (14:54 +0000)]
Fix bug in binary operations code. When loading a 32-bit integer as a
64-bit integer to the stack, the high and low parts of the 64-bit
value were pushed in the wrong order.

TBR=erik.corry@gmail.com
Review URL: http://codereview.chromium.org/3010060

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5216 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoIA32: Avoid going into stubs or runtime code for bitops even if the
erik.corry@gmail.com [Mon, 9 Aug 2010 13:12:02 +0000 (13:12 +0000)]
IA32: Avoid going into stubs or runtime code for bitops even if the
inputs are heap numbers or the result is a heap number (only with
SSE2).  Make it possible for a deferred code object to work without
spilling all registers.
Review URL: http://codereview.chromium.org/3054047

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5215 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoFix cast that fails to build on Windows.
lrn@chromium.org [Mon, 9 Aug 2010 12:44:48 +0000 (12:44 +0000)]
Fix cast that fails to build on Windows.

Review URL: http://codereview.chromium.org/3044055

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5213 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoChange dos line endings to unix line endings in a number of mjsunit test files.
ricow@chromium.org [Mon, 9 Aug 2010 12:21:50 +0000 (12:21 +0000)]
Change dos line endings to unix line endings in a number of mjsunit test files.

Review URL: http://codereview.chromium.org/3072031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5212 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoHeap profiler: reduce heap snapshots size.
mikhail.naganov@gmail.com [Mon, 9 Aug 2010 11:37:24 +0000 (11:37 +0000)]
Heap profiler: reduce heap snapshots size.

The size of a snapshot is now 65-80% of the JS heap size (tested on
GMail and Wave), previously it was >200%.

BUG=783

Review URL: http://codereview.chromium.org/3060008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5211 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoPrepare push to trunk. Now working on version 2.3.7.
lrn@chromium.org [Mon, 9 Aug 2010 09:20:35 +0000 (09:20 +0000)]
Prepare push to trunk.  Now working on version 2.3.7.

Review URL: http://codereview.chromium.org/3017063

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5207 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoChanged layout of object literal parser.
lrn@chromium.org [Mon, 9 Aug 2010 08:54:29 +0000 (08:54 +0000)]
Changed layout of object literal parser.

Review URL: http://codereview.chromium.org/3032064

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5206 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoCreate a new RegExp object for every evaluation of a RegExp literal.
lrn@chromium.org [Fri, 6 Aug 2010 13:04:27 +0000 (13:04 +0000)]
Create a new RegExp object for every evaluation of a RegExp literal.
Changes necessary to following ES5 semantics and matching Safari - in ES3
the same RegExp object was generated by each evaluation of the RegExp literal.
Fixes bug 704.

Review URL: http://codereview.chromium.org/3034060

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5198 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoFix bug 813, harmless assert triggered by ARM inside o3d.
erik.corry@gmail.com [Fri, 6 Aug 2010 09:04:17 +0000 (09:04 +0000)]
Fix bug 813, harmless assert triggered by ARM inside o3d.
Review URL: http://codereview.chromium.org/2868112

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5197 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoReapplying r5174 (Avoid GC when compiling CallIC stubs).
kaznacheev@chromium.org [Fri, 6 Aug 2010 08:49:59 +0000 (08:49 +0000)]
Reapplying r5174 (Avoid GC when compiling CallIC stubs).

I could not find a way to introduce simple yet useful assertions,
so I removed the assertions that broke tests.

The new patch just fixes the actual problem (GC triggered while raw pointers
are held in LookupResult).

Review URL: http://codereview.chromium.org/3030049

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5196 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoRe-apply r5165 (Added support for ES5's propertyname production)
sgjesse@chromium.org [Fri, 6 Aug 2010 08:03:44 +0000 (08:03 +0000)]
Re-apply r5165 (Added support for ES5's propertyname production)

TBR=lrn@chromium.org
Review URL: http://codereview.chromium.org/3073031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5192 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoCheck ASCIIness of lhs, not this.
antonm@chromium.org [Thu, 5 Aug 2010 18:12:39 +0000 (18:12 +0000)]
Check ASCIIness of lhs, not this.

In one pathalogical case it's possible to have this->IsAsciiRepresentation() &&
!this->TryFlattenGetString()->IsAsciiRepresentation()---if cons string has two byte
string which holds only ascii chars and second is an empty string.  In this case we
would return first which is not AsciiRepresentation(), however cons is.

Review URL: http://codereview.chromium.org/3036046

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5188 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoRevert r5165 (Added support for ES5's propertyname production)
sgjesse@chromium.org [Thu, 5 Aug 2010 14:46:34 +0000 (14:46 +0000)]
Revert r5165 (Added support for ES5's propertyname production)

This is to test in Chromium without this change.

TBR=lrn@chromium.org
Review URL: http://codereview.chromium.org/3027043

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5182 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoAdd information about memory allocator's size and capacity and heap snapshot into...
antonm@chromium.org [Thu, 5 Aug 2010 14:12:50 +0000 (14:12 +0000)]
Add information about memory allocator's size and capacity and heap snapshot into heap stats.

That might help us debug OOM crashes in V8.

Review URL: http://codereview.chromium.org/3046049

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5181 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoDo not apply JS_RETURN and DEBUG_BREAK_SLOT relocations on x64.
vegorov@chromium.org [Thu, 5 Aug 2010 13:38:27 +0000 (13:38 +0000)]
Do not apply JS_RETURN and DEBUG_BREAK_SLOT relocations on x64.

Improve tests for debug API to check behavior of JS_RETURN and DEBUG_BREAK_SLOT relocations.

Review URL: http://codereview.chromium.org/3058049

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5180 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoFix GCC 4.4.4 build
sgjesse@chromium.org [Thu, 5 Aug 2010 12:34:53 +0000 (12:34 +0000)]
Fix GCC 4.4.4 build

Force a number functions to not inline in order th work around a bug in GCC 4.4.4.

See http://codereview.chromium.org/3083016 for more info.
Review URL: http://codereview.chromium.org/3034056

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5179 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoAdd RelocInfo::DEBUG_BREAK_SLOT to RelocInfo::kApplyMask on ia32/x64 to ensure that...
vegorov@chromium.org [Thu, 5 Aug 2010 11:39:01 +0000 (11:39 +0000)]
Add RelocInfo::DEBUG_BREAK_SLOT to RelocInfo::kApplyMask on ia32/x64 to ensure that debug break slots get relocated correctly during compacting GC.

Review URL: http://codereview.chromium.org/3058048

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5178 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoRevert r5174.
vegorov@chromium.org [Thu, 5 Aug 2010 11:19:43 +0000 (11:19 +0000)]
Revert r5174.
Review URL: http://codereview.chromium.org/2868108

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5177 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoObject.seal(obj) and Object.freeze(obj) should return the input obj.
sgjesse@chromium.org [Thu, 5 Aug 2010 10:56:15 +0000 (10:56 +0000)]
Object.seal(obj) and Object.freeze(obj) should return the input obj.

BUG=http://code.google.com/p/v8/issues/detail?id=809
TEST=Seal/freeze an object and check if Object.seal and Object.freeze returns the given object.

Burcu Dogan <burcujdogan@gmail.com>

Review URL: http://codereview.chromium.org/3056049

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5176 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoModify polymorphic keyed load IC stub to load cached properties from the property...
whesse@chromium.org [Thu, 5 Aug 2010 10:45:07 +0000 (10:45 +0000)]
Modify polymorphic keyed load IC stub to load cached properties from the property array of a fast-case JSObject.
Review URL: http://codereview.chromium.org/3053042

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5175 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoAvoid GC when compiling CallIC stubs.
kaznacheev@chromium.org [Thu, 5 Aug 2010 08:37:12 +0000 (08:37 +0000)]
Avoid GC when compiling CallIC stubs.

In rare cases GC could be called from ComputeCallMiss function thus
breaking CallIC::LoadFunction.

Review URL: http://codereview.chromium.org/3047027

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5174 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoPrepare push to trunk. We are now working on 2.3.6.
sgjesse@chromium.org [Wed, 4 Aug 2010 09:30:55 +0000 (09:30 +0000)]
Prepare push to trunk. We are now working on 2.3.6.
Review URL: http://codereview.chromium.org/3078024

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5171 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoVirtualAlloc on Windows 7 does not currently provide sufficient randomization to...
sgjesse@chromium.org [Mon, 2 Aug 2010 15:27:25 +0000 (15:27 +0000)]
VirtualAlloc on Windows 7 does not currently provide sufficient randomization to protect JIT code from being aligned in large regions at a predictable location.

This patch manually randomizes the allocation address for PAGE_EXECUTE_READWRITE regions between kAllocationRandomAddressMin and kAllocationRandomAddressMax.

BUG=none
TEST=allocate lots of javascript code and check for contiguous allocations

Patch by Paul Mehta <pmehta@chromium.org>

Review URL: http://codereview.chromium.org/2832095

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5169 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoRevert r5147 due to failing assert, with no simple solution. Issue 808.
whesse@chromium.org [Mon, 2 Aug 2010 15:08:17 +0000 (15:08 +0000)]
Revert r5147 due to failing assert, with no simple solution.  Issue 808.
Review URL: http://codereview.chromium.org/3087001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5167 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoAdded support for ES5's propertyname production.
lrn@chromium.org [Mon, 2 Aug 2010 12:09:10 +0000 (12:09 +0000)]
Added support for ES5's propertyname production.
Object initialisers and dot-notation property access allows keywords in ES5.
Also allowed non-identifiers after "get" or "set" in an object initialiser.

Review URL: http://codereview.chromium.org/3047038

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5165 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoPrepare push to trunk. Now working on version 2.3.5.
whesse@chromium.org [Mon, 2 Aug 2010 11:02:38 +0000 (11:02 +0000)]
Prepare push to trunk.  Now working on version 2.3.5.
Review URL: http://codereview.chromium.org/3040030

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5163 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

13 years agoFix issue 806.
vegorov@chromium.org [Mon, 2 Aug 2010 09:14:44 +0000 (09:14 +0000)]
Fix issue 806.

Ensure that we are not using r12 as a receiver in inlined NamedStore code.

Review URL: http://codereview.chromium.org/3081007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5162 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoUnicode: Reduced size of tables.
lrn@chromium.org [Fri, 30 Jul 2010 12:59:57 +0000 (12:59 +0000)]
Unicode: Reduced size of tables.

Review URL: http://codereview.chromium.org/3043032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5161 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRoll over again fixed 3029033 Fix 'step in' after live edit stack manipulation
peter.rybin@gmail.com [Fri, 30 Jul 2010 11:58:43 +0000 (11:58 +0000)]
Roll over again fixed 3029033 Fix 'step in' after live edit stack manipulation

Review URL: http://codereview.chromium.org/3066018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5160 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoUpdated unicode library.
lrn@chromium.org [Fri, 30 Jul 2010 07:10:22 +0000 (07:10 +0000)]
Updated unicode library.
Added Nl category to letters predicate (as requried for JS identifiers).
Changed/simplified representation of canonicalization ranges.
Truncated tables to code points in the BMP (all that is used by JS).
Reformatted tables to avoid excessively long lines.
Removed duplicate entries from multi-character mapping result tables.

Review URL: http://codereview.chromium.org/3030026

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5155 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRoll back all step-in-fix-related changes
peter.rybin@gmail.com [Thu, 29 Jul 2010 21:46:19 +0000 (21:46 +0000)]
Roll back all step-in-fix-related changes

Review URL: http://codereview.chromium.org/3012035

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5154 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix failing tests and rollback two from Testing change for failing unit test
peter.rybin@gmail.com [Thu, 29 Jul 2010 21:29:45 +0000 (21:29 +0000)]
Fix failing tests and rollback two from Testing change for failing unit test

Review URL: http://codereview.chromium.org/3052026

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5153 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRollback one from Testing change for failing unit test
peter.rybin@gmail.com [Thu, 29 Jul 2010 20:32:14 +0000 (20:32 +0000)]
Rollback one from Testing change for failing unit test

Review URL: http://codereview.chromium.org/2870083

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5152 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoTesting change for failing unit test
peter.rybin@gmail.com [Thu, 29 Jul 2010 20:23:19 +0000 (20:23 +0000)]
Testing change for failing unit test

Review URL: http://codereview.chromium.org/3031032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5151 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix 'step in' after live edit stack manipulation
peter.rybin@gmail.com [Thu, 29 Jul 2010 16:40:14 +0000 (16:40 +0000)]
Fix 'step in' after live edit stack manipulation

Review URL: http://codereview.chromium.org/3029033

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5150 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix XCode build.
vegorov@chromium.org [Thu, 29 Jul 2010 08:31:36 +0000 (08:31 +0000)]
Fix XCode build.

Review URL: http://codereview.chromium.org/3056028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5149 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd debugger protocol request for setting global flags.
podivilov@chromium.org [Wed, 28 Jul 2010 15:50:05 +0000 (15:50 +0000)]
Add debugger protocol request for setting global flags.

Review URL: http://codereview.chromium.org/2880011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5148 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoCache maps for slow case objects.
kaznacheev@chromium.org [Wed, 28 Jul 2010 15:08:32 +0000 (15:08 +0000)]
Cache maps for slow case objects.

Review URL: http://codereview.chromium.org/3032028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5147 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoBreakpoint position should be inside function body.
podivilov@chromium.org [Wed, 28 Jul 2010 13:02:03 +0000 (13:02 +0000)]
Breakpoint position should be inside function body.

Review URL: http://codereview.chromium.org/2883042

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5145 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix error in optimized x.apply(y, arguments) code generation on ARM. Fixes issue...
whesse@chromium.org [Wed, 28 Jul 2010 12:50:27 +0000 (12:50 +0000)]
Fix error in optimized x.apply(y, arguments) code generation on ARM.  Fixes issue 784.  Adds regression test.
Review URL: http://codereview.chromium.org/3048035

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5144 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoUse hashlib module instead of md5 if it is present.
antonm@chromium.org [Wed, 28 Jul 2010 12:35:46 +0000 (12:35 +0000)]
Use hashlib module instead of md5 if it is present.

md5 module is deprecated since Python 2.5 and we'd better off using hashlib instead.

Review URL: http://codereview.chromium.org/3026030

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5143 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoReleasing allocated chunks in Pool's destructor.
antonm@chromium.org [Wed, 28 Jul 2010 12:34:41 +0000 (12:34 +0000)]
Releasing allocated chunks in Pool's destructor.

That should make it not leak if v8 wasn't intialized at all.

Review URL: http://codereview.chromium.org/3048034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5142 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPort faster callbacks invocation to x64.
antonm@chromium.org [Wed, 28 Jul 2010 09:36:53 +0000 (09:36 +0000)]
Port faster callbacks invocation to x64.

It's a port of http://code.google.com/p/v8/source/detail?r=3209 to x64 platform.
That allows invocation of callbacks without going into runtime.

Review URL: http://codereview.chromium.org/2801008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5141 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoX64: Fix order of operands on movq(reg, xmm).
lrn@chromium.org [Tue, 27 Jul 2010 12:02:21 +0000 (12:02 +0000)]
X64: Fix order of operands on movq(reg, xmm).
Fixes bug 792, and the Transcendental cache is the only use of this operation.

Review URL: http://codereview.chromium.org/2806080

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5138 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix some bugs in Function.prototype.bind implementation.
lrn@chromium.org [Tue, 27 Jul 2010 09:20:21 +0000 (09:20 +0000)]
Fix some bugs in Function.prototype.bind implementation.
Correctly handle not passing thisArg.
Fixes to NewObjectFromBound to use correct argument count, not leak memory,
and handle constructors that throw exceptions.

Review URL: http://codereview.chromium.org/2878057

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5137 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRemove global.print from v8natives.js (fixes issue 791)
ricow@chromium.org [Tue, 27 Jul 2010 06:18:32 +0000 (06:18 +0000)]
Remove global.print from v8natives.js (fixes issue 791)

Review URL: http://codereview.chromium.org/3045016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5135 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPrepare push to trunk. Now working on version 2.3.4.
whesse@chromium.org [Mon, 26 Jul 2010 10:45:46 +0000 (10:45 +0000)]
Prepare push to trunk.  Now working on version 2.3.4.
Review URL: http://codereview.chromium.org/3058009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5132 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix break position not to be outside of the script
peter.rybin@gmail.com [Fri, 23 Jul 2010 17:21:55 +0000 (17:21 +0000)]
Fix break position not to be outside of the script

Review URL: http://codereview.chromium.org/3017021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5131 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix aliasing problem in inlined stores on x64 and ia32. The receiver
ager@chromium.org [Fri, 23 Jul 2010 11:55:03 +0000 (11:55 +0000)]
Fix aliasing problem in inlined stores on x64 and ia32. The receiver
and the value can share a register. We need to remove this aliasing
before modifying the registers.

I haven't managed to generate a stand-alon test case for this
yet. I'll do that as a separate change. This was found while loading
Wave.

Review URL: http://codereview.chromium.org/3039025

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5127 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix bug in r5123, Comparison(), by unusing results before unconditional jump to smi...
whesse@chromium.org [Fri, 23 Jul 2010 11:20:59 +0000 (11:20 +0000)]
Fix bug in r5123, Comparison(), by unusing results before unconditional jump to smi comparison JumpTarget.
Review URL: http://codereview.chromium.org/3026019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5126 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix presubmit errors.
ricow@chromium.org [Fri, 23 Jul 2010 10:15:21 +0000 (10:15 +0000)]
Fix presubmit errors.

Review URL: http://codereview.chromium.org/3009017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5125 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoImplement Function.prototype.bind (ES5 15.3.4.5).
ricow@chromium.org [Fri, 23 Jul 2010 10:08:55 +0000 (10:08 +0000)]
Implement Function.prototype.bind (ES5 15.3.4.5).

Please note that we do not implement correctly the setting of caller
and arguments on the returned objects, since we already have these
properties on function objects (and they are non-configurable).

Also corrects indention in DefineOwnProperty.

Review URL: http://codereview.chromium.org/3046010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5124 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoUse static type information when creating the inlined code for comparison operators.
whesse@chromium.org [Fri, 23 Jul 2010 09:05:46 +0000 (09:05 +0000)]
Use static type information when creating the inlined code for comparison operators.
Review URL: http://codereview.chromium.org/3017025

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5123 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFixed a couple of issues with store inlining on arm.
ager@chromium.org [Fri, 23 Jul 2010 08:47:15 +0000 (08:47 +0000)]
Fixed a couple of issues with store inlining on arm.

Spill and merge virtual frames explicitly in the deferred code.

Account for the fact that the inlined write barrier size depends on
the size of the new space masks.

Review URL: http://codereview.chromium.org/3018015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5122 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoChange asserts to STATIC_ASSERT if they can be checked at compilation time. Just...
whesse@chromium.org [Fri, 23 Jul 2010 08:25:48 +0000 (08:25 +0000)]
Change asserts to STATIC_ASSERT if they can be checked at compilation time.  Just in the codegen-xxx.cc files on all platforms.
Review URL: http://codereview.chromium.org/3017018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5121 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoProperly propagate failures from helper methods.
antonm@chromium.org [Thu, 22 Jul 2010 13:39:18 +0000 (13:39 +0000)]
Properly propagate failures from helper methods.

Otherwise failures are not reported and callback may fail to setup.

Review URL: http://codereview.chromium.org/3035017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5120 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd ARM_LINK_FLAGS when build d8 with arch=arm to make
ager@chromium.org [Thu, 22 Jul 2010 12:57:48 +0000 (12:57 +0000)]
Add ARM_LINK_FLAGS when build d8 with arch=arm to make
cross-compiling d8 easier.

Review URL: http://codereview.chromium.org/3062002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5119 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix d8 build error on a fresh checkout.
serya@chromium.org [Thu, 22 Jul 2010 12:55:14 +0000 (12:55 +0000)]
Fix d8 build error on a fresh checkout.

Review URL: http://codereview.chromium.org/2809069

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5118 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix Win64 build.
ager@chromium.org [Thu, 22 Jul 2010 09:01:00 +0000 (09:01 +0000)]
Fix Win64 build.

TBR=ricow@chromium.org
Review URL: http://codereview.chromium.org/3017024

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5117 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPort inlined in-object property stores to ARM.
ager@chromium.org [Thu, 22 Jul 2010 08:17:40 +0000 (08:17 +0000)]
Port inlined in-object property stores to ARM.
Review URL: http://codereview.chromium.org/2878043

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5116 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPort inline in-object property stores from ia32 to x64.
ager@chromium.org [Thu, 22 Jul 2010 06:31:49 +0000 (06:31 +0000)]
Port inline in-object property stores from ia32 to x64.
Review URL: http://codereview.chromium.org/3008017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5115 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoSupport object literal map caching for literals with keys that are array indices.
kaznacheev@chromium.org [Thu, 22 Jul 2010 06:28:47 +0000 (06:28 +0000)]
Support object literal map caching for literals with keys that are array indices.

The caching code used to require that all keys in the literal were symbol strings
in order to be able to cache object literal maps.

Relaxed this requirement to allow array indices as well (as in { "1": "one", two: "two" }).

Changed the way 'number_of_properties' is calculated so that redundant property slots are not
allocated for array indices.

Review URL: http://codereview.chromium.org/2858064

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5114 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRemove VirtualFrame::CallStoreIC(void) and CallCommonStoreIC from virtual-frame-x64...
whesse@chromium.org [Wed, 21 Jul 2010 15:16:01 +0000 (15:16 +0000)]
Remove VirtualFrame::CallStoreIC(void) and CallCommonStoreIC from virtual-frame-x64.cc.  Make implementations of CallStoreIC(key, is_contextual) and CallKeyedStoreIC() the same on x64 and ia32.  Make ia32 and x64 implementations more alike by moving some functions and changing comments.
Review URL: http://codereview.chromium.org/3060002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5113 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd EmitKeyedStore and EmitKeyedPropertyAssignment to x64. Removal of Reference...
whesse@chromium.org [Wed, 21 Jul 2010 11:52:57 +0000 (11:52 +0000)]
Add EmitKeyedStore and EmitKeyedPropertyAssignment to x64.  Removal of Reference class from VisitAssignment complete on all platforms.
Review URL: http://codereview.chromium.org/3028015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5112 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPrepare push to trunk. Now working on version 2.3.3.
fschneider@chromium.org [Wed, 21 Jul 2010 10:53:38 +0000 (10:53 +0000)]
Prepare push to trunk. Now working on version 2.3.3.

Review URL: http://codereview.chromium.org/3053006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5109 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd EmitNamedStore and CallStoreIC(name, is_contextual) to x64 platform.
whesse@chromium.org [Wed, 21 Jul 2010 10:31:52 +0000 (10:31 +0000)]
Add EmitNamedStore and CallStoreIC(name, is_contextual) to x64 platform.
Review URL: http://codereview.chromium.org/3050008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5108 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd EmitSlotAssignment to x64 platform.
whesse@chromium.org [Wed, 21 Jul 2010 08:50:57 +0000 (08:50 +0000)]
Add EmitSlotAssignment to x64 platform.
Review URL: http://codereview.chromium.org/2878042

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5107 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoLanding for Rodolph Perfetta.
ager@chromium.org [Wed, 21 Jul 2010 07:42:51 +0000 (07:42 +0000)]
Landing for Rodolph Perfetta.

Add support for saturation instruction (ARMv6 or above).
The byte array clamping code has been updated accordingly.

Review URL: http://codereview.chromium.org/3036008/show

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5106 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoInline in-object property stores on ia32 when in loop and not in top-level code.
ager@chromium.org [Wed, 21 Jul 2010 06:59:34 +0000 (06:59 +0000)]
Inline in-object property stores on ia32 when in loop and not in top-level code.
Review URL: http://codereview.chromium.org/3046006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5105 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoStop pushing arguments onto the stack in CompareStub until just before calling runtime.
whesse@chromium.org [Tue, 20 Jul 2010 12:41:43 +0000 (12:41 +0000)]
Stop pushing arguments onto the stack in CompareStub until just before calling runtime.
This is a fixed version of change 5097, which had the problem that LoadFloatOperands tried to load the arguments from the stack.
Review URL: http://codereview.chromium.org/3040010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5103 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoOutput time spent in code flushing in GC NVP trace.
vegorov@chromium.org [Tue, 20 Jul 2010 10:59:00 +0000 (10:59 +0000)]
Output time spent in code flushing in GC NVP trace.

Add support for flushcode scope and cumulative stats into gc-nvp-trace-processor.

Review URL: http://codereview.chromium.org/3054003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5102 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoUpdate inheritance hierarchy in Objects.h to better reflect the actual hierarchy.
ricow@chromium.org [Tue, 20 Jul 2010 07:14:19 +0000 (07:14 +0000)]
Update inheritance hierarchy in Objects.h to better reflect the actual hierarchy.

We do not have a GlobalContext class.

We do, however, have a SerializedScopeInfo class that inherits FixedArray.

Starting comment is also slightly misleading, since not all objects in the hierarchy is described in objects.h (e.g., Context is described in contexts.h).

Review URL: http://codereview.chromium.org/2811057

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5101 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoUpdate Code::NUMBER_OF_KINDS to include binary op ICs.
ager@chromium.org [Tue, 20 Jul 2010 06:23:36 +0000 (06:23 +0000)]
Update Code::NUMBER_OF_KINDS to include binary op ICs.

Review URL: http://codereview.chromium.org/3027013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5100 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoLand for Victor Wang.
ager@chromium.org [Tue, 20 Jul 2010 05:51:10 +0000 (05:51 +0000)]
Land for Victor Wang.

Add direct_dependent_settings for v8 dll so targets that depend on v8
has USING_V8_SHARED defined.

Review URL: http://codereview.chromium.org/2946014/show

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5099 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRevert 5097 on ia32 platform until problem is diagnosed.
whesse@chromium.org [Mon, 19 Jul 2010 14:26:51 +0000 (14:26 +0000)]
Revert 5097 on ia32 platform until problem is diagnosed.
Review URL: http://codereview.chromium.org/3018009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5098 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoDelay the pushing of arguments until just before calling runtime in CompareStub.
whesse@chromium.org [Mon, 19 Jul 2010 13:35:41 +0000 (13:35 +0000)]
Delay the pushing of arguments until just before calling runtime in CompareStub.
This will allow jumps to runtime that skip useless tests.
Review URL: http://codereview.chromium.org/3016004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5097 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoAdd a check that weak object handle is not in NEAR_DEATH state after weak callback...
antonm@chromium.org [Mon, 19 Jul 2010 13:26:25 +0000 (13:26 +0000)]
Add a check that weak object handle is not in NEAR_DEATH state after weak callback invocation.

If object enters NEAR_DEATH state, it must be explicitly cleared and/or disposed, otherwise
it would retain JS object forever.  Note as well that parameter is reset to NULL on first
invocation so weak handle callback would be in hard situation.

Review URL: http://codereview.chromium.org/3011009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5096 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix issue 785. For-in now works on strings: for (var i in "asdf") now works
whesse@chromium.org [Mon, 19 Jul 2010 12:45:21 +0000 (12:45 +0000)]
Fix issue 785.  For-in now works on strings: for (var i in "asdf") now works
all the time, not just the first time it is run.
Review URL: http://codereview.chromium.org/3037008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5095 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoLanding for pmehta.
ager@chromium.org [Mon, 19 Jul 2010 10:02:11 +0000 (10:02 +0000)]
Landing for pmehta.

Changed a static cast from static_cast<int> to static_cast<size_t>
that previously introduced a signed/unsigned comparison issue in the
main allocator for V8 (MemoryAllocator::AllocateRawMemory) that could
be used to bypass the V8 allocation limitations or trigger integer
overflows.

Review URL: http://codereview.chromium.org/3027006/show

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5094 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoLanding for hans@chromium.org.
ager@chromium.org [Mon, 19 Jul 2010 09:51:33 +0000 (09:51 +0000)]
Landing for hans@chromium.org.

TYPE_CHECK in v8.h should assign to volatile qualified null-pointer.

Unless the pointer is volatile qualified, Clang will warn that LLVM
removes the assignment during optimization. This is not a problem as
that code should never execute, but the warning is treated as an error
when building Chromium, and thus stops the build.

Code review URL: http://codereview.chromium.org/3005010/show

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5093 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoPrepare push to trunk. Now working on version 2.3.2.
ricow@chromium.org [Mon, 19 Jul 2010 07:33:05 +0000 (07:33 +0000)]
Prepare push to trunk. Now working on version 2.3.2.

Review URL: http://codereview.chromium.org/2811055

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5090 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoReorder SharedFunctionInfo setup in lazy compile, quick fix for crbug.com/49099
ricow@chromium.org [Mon, 19 Jul 2010 06:29:30 +0000 (06:29 +0000)]
Reorder SharedFunctionInfo setup in lazy compile, quick fix for crbug.com/49099

This fixes issue 49099 (http://crbug.com/49099) that makes chromium
crash if a gc is triggered by setting the scopeinfo in CompileLazy. If
this gc triggers code flushing the compiled function could be flushed,
causing the ASSERT(shared->is_compiled()) to fail. By reordering the
two methods we make sure that no allocation happens before the end of
the method.

Review URL: http://codereview.chromium.org/3035006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5089 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoRefactor type checks in v8natives.js and runtime.js.
ricow@chromium.org [Fri, 16 Jul 2010 11:21:08 +0000 (11:21 +0000)]
Refactor type checks in v8natives.js and runtime.js.

This includes adding a new inline IsSpecObject method to the code
generator.  The old approach was somehow ineffecient since we would
call both IsObject, IsUndetectable and IsFunction to determine if
something was an object according to the spec. This change introduces
a new macro that determines if something is an object according to the
spec (and this does not include null).

This change also corrects a few places where undetectable objects was
not allowed even when they should be (priorly they would use only
IS_SPEC_OBJECT_OR_NULL, which would return false on an undetectable
object, the new IS_SPEC_OBJECT returns true on an undetectable object.

Review URL: http://codereview.chromium.org/2877018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5087 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoStringDictionary::FindEntry optimized for symbol strings.
serya@chromium.org [Fri, 16 Jul 2010 10:07:57 +0000 (10:07 +0000)]
StringDictionary::FindEntry optimized for symbol strings.

Review URL: http://codereview.chromium.org/3020003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5086 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix compilation errors on X64. Error from r5080 is constructor Operand(register)...
whesse@chromium.org [Fri, 16 Jul 2010 08:20:39 +0000 (08:20 +0000)]
Fix compilation errors on X64. Error from r5080 is constructor Operand(register) with too few arguments, while constructing a dummy return value in unreachable code to avoid compiler errors. Also added explicit cast from intptr_t to int in profile-generator.cc to avoid compiler warnings, introduced in r5078.
Review URL: http://codereview.chromium.org/3030002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5084 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

14 years agoFix TestHeapEntryIdsAndGC test. The assert was wrong.
mikhail.naganov@gmail.com [Thu, 15 Jul 2010 15:41:05 +0000 (15:41 +0000)]
Fix TestHeapEntryIdsAndGC test. The assert was wrong.

TBR=ager@chromium.org

Review URL: http://codereview.chromium.org/2861051

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5082 ce2b1a6d-e550-0410-aec6-3dcde31c8c00