platform/upstream/v8.git
11 years agoMIPS: Tag smi-constants as smi. This also fixes code that copies holes into arrays.
palfia@homejinni.com [Mon, 27 May 2013 22:11:06 +0000 (22:11 +0000)]
MIPS: Tag smi-constants as smi. This also fixes code that copies holes into arrays.

Port r14814 (8a28ccfb)

BUG=

Review URL: https://codereview.chromium.org/16041005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14830 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix the hole loading optimization.
verwaest@chromium.org [Mon, 27 May 2013 17:33:14 +0000 (17:33 +0000)]
Fix the hole loading optimization.
- Holes are only ever loaded as double or tagged.
- Change to tagged has to deoptimize on undefined (no implicit
  conversions from double the hole NaN -> tagged undefined).

BUG=
R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/16099006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14829 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Fix unaligned address of double.
palfia@homejinni.com [Mon, 27 May 2013 17:12:48 +0000 (17:12 +0000)]
MIPS: Fix unaligned address of double.

TEST=mjsunit/debug-evaluate-locals-optimized-double

BUG=

Review URL: https://codereview.chromium.org/14631016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14828 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: make isolate accessible from returnvalue
palfia@homejinni.com [Mon, 27 May 2013 16:53:30 +0000 (16:53 +0000)]
MIPS: make isolate accessible from returnvalue

Port r14819 (1ee1db14)

BUG=

Review URL: https://codereview.chromium.org/16099003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14827 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoAdd template parameter to ReturnValue::Set.
dcarney@chromium.org [Mon, 27 May 2013 14:34:16 +0000 (14:34 +0000)]
Add template parameter to ReturnValue::Set.

E.g., v8-i18n wants to set the return value with a different type of a Persistent.

BUG=NONE
R=dcarney@chromium.org, svenpanne@chromium.org

Review URL: https://codereview.chromium.org/16102002

Patch from Marja Hölttä <marja@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14826 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoCollect type feedback in separate pass and store it in AST
rossberg@chromium.org [Mon, 27 May 2013 13:59:20 +0000 (13:59 +0000)]
Collect type feedback in separate pass and store it in AST

Notes:

- For now, just adds the missing type info fields to the AST nodes directly.
  I'd like to factor that out more nicely in a follow-up CL.

- All type feedback now is uniformly collected through AST nodes'
  RecordTypeFeedback functions. At some point, this logic should be moved
  out of ast.cc.

- The typing pass currently simulates the exact same conditions under
  which feedback was collected in Hydrogen before. That also should be
  made more generic in the future.

- Type information itself is unchanged. Making it more regular is
  yet more future work.

Some additional cleanups:

- Lifted out nested class ObjectLiteral::Property, to enable forward declaration.
- Moved around some auxiliary enums.

R=svenpanne@chromium.org
BUG=

Review URL: https://codereview.chromium.org/14990014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14825 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoExtract GlobalValueNumberer and helper classes from hydrogen.cc and move to hydrogen...
titzer@chromium.org [Mon, 27 May 2013 12:45:46 +0000 (12:45 +0000)]
Extract GlobalValueNumberer and helper classes from hydrogen.cc and move to hydrogen-gvn.cc.

BUG=

Review URL: https://codereview.chromium.org/16095004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14824 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Remove HCheckSmi, LCheckSmi and rename LCheckSmiAndReturn to LCheckSmi.
palfia@homejinni.com [Mon, 27 May 2013 12:39:01 +0000 (12:39 +0000)]
MIPS: Remove HCheckSmi, LCheckSmi and rename LCheckSmiAndReturn to LCheckSmi.

Port r14816 (0485153d)

BUG=

Review URL: https://codereview.chromium.org/16035008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14823 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoPrepare push to trunk. Now working on version 3.19.6.
svenpanne@chromium.org [Mon, 27 May 2013 12:19:49 +0000 (12:19 +0000)]
Prepare push to trunk.  Now working on version 3.19.6.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/15838010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14820 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agomake isolate accessible from returnvalue
dcarney@chromium.org [Mon, 27 May 2013 11:50:08 +0000 (11:50 +0000)]
make isolate accessible from returnvalue

R=svenpanne@chromium.org
BUG=

Review URL: https://codereview.chromium.org/16021010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14819 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix Object.freeze for objects with mixed accessors and data properties
verwaest@chromium.org [Mon, 27 May 2013 10:56:27 +0000 (10:56 +0000)]
Fix Object.freeze for objects with mixed accessors and data properties

The bug in the existing code was that it modified the |attributes|
local variable on its way through the loop in CopyUpToAddAttributes.
But that affected any properties updated after an accessor property.
The code now sets up a mask each time and applies that instead of
mutating |attributes|.

R=verwaest@chromium.org

Review URL: https://chromiumcodereview.appspot.com/16051002

Patch from Adam Klein <adamk@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14818 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoRegexp parser: reset flag after scanning ahead for capture groups.
yangguo@chromium.org [Mon, 27 May 2013 10:53:37 +0000 (10:53 +0000)]
Regexp parser: reset flag after scanning ahead for capture groups.

When the regexp pattern parser encounters an unbound reference to a
capturing group, it needs to scan ahead to decide whether it really
is a reference.  The scan advances to the end of the pattern string
and sets has_more_ to false, but fails to reset it to true so that
later on, parsing a character class wrongly fails.

R=ulan@chromium.org
BUG=v8:2690

Review URL: https://chromiumcodereview.appspot.com/15712006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14817 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoRemove HCheckSmi, LCheckSmi and rename LCheckSmiAndReturn to LCheckSmi.
verwaest@chromium.org [Mon, 27 May 2013 09:58:46 +0000 (09:58 +0000)]
Remove HCheckSmi, LCheckSmi and rename LCheckSmiAndReturn to LCheckSmi.

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15841007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14816 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix build
verwaest@chromium.org [Mon, 27 May 2013 08:50:52 +0000 (08:50 +0000)]
Fix build

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15881004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14815 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoTag smi-constants as smi. This also fixes code that copies holes into arrays.
verwaest@chromium.org [Mon, 27 May 2013 08:43:58 +0000 (08:43 +0000)]
Tag smi-constants as smi. This also fixes code that copies holes into arrays.

BUG=
R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15861009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14814 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Fix DIRECT_API_CALL_NEW and DIRECT_GETTER_CALL_NEW call.
palfia@homejinni.com [Sat, 25 May 2013 17:10:42 +0000 (17:10 +0000)]
MIPS: Fix DIRECT_API_CALL_NEW and DIRECT_GETTER_CALL_NEW call.

This commit fixes the register usage of DIRECT_API_CALL_NEW
and DIRECT_GETTER_CALL_NEW:

* These functions expect arguments in a0-a1 and not in a1-a2
  as the and DIRECT_API_CALL and DIRECT_GETTER_CALL do.
* Fixes the simulator to expect *_NEW arguments in a0-a1.
* Adds more comment to simulator to better explain the
  register usage.

TEST=cctest/test-api/LoadICFastApi_DirectCall_GCMoveStub,cctest/test-api/SimpleCallback

BUG=

Review URL: https://codereview.chromium.org/15813005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14813 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Fix hole handling, and ensure smi representation is handled properly
palfia@homejinni.com [Fri, 24 May 2013 20:20:08 +0000 (20:20 +0000)]
MIPS: Fix hole handling, and ensure smi representation is handled properly

Port r14807 (c26f44d6)

BUG=

Review URL: https://codereview.chromium.org/15821007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14812 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: DummyUse if Representation::Tagged()+Type::Smi() to Representation::Smi()
palfia@homejinni.com [Fri, 24 May 2013 20:19:18 +0000 (20:19 +0000)]
MIPS: DummyUse if Representation::Tagged()+Type::Smi() to Representation::Smi()

Port r14802 (d4a53e02)

BUG=

Review URL: https://codereview.chromium.org/16019009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14811 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Replace tagged keys for fast access by smi, and use smi in boundscheck.
palfia@homejinni.com [Fri, 24 May 2013 20:18:24 +0000 (20:18 +0000)]
MIPS: Replace tagged keys for fast access by smi, and use smi in boundscheck.

Port r14794 (05b08564)

BUG=

Review URL: https://codereview.chromium.org/15937009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14810 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoUn-flake test-cpu-profiler/SampleWhenFrameIsNotSetup
yurys@chromium.org [Fri, 24 May 2013 16:19:06 +0000 (16:19 +0000)]
Un-flake test-cpu-profiler/SampleWhenFrameIsNotSetup

It is OK for FindChild to return NULL. If the child must
exist GetChild should be used to force the assertion.

BUG=v8:2628
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/15786004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14809 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agotrace compare_nil_ic state in --trace-ic output
olivf@chromium.org [Fri, 24 May 2013 15:20:48 +0000 (15:20 +0000)]
trace compare_nil_ic state in --trace-ic output

BUG=
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/15735020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14808 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix hole handling, and ensure smi representation is handled properly
verwaest@chromium.org [Fri, 24 May 2013 13:40:02 +0000 (13:40 +0000)]
Fix hole handling, and ensure smi representation is handled properly

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/16013003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14807 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoRemove offset() and is_in_object() from hydrogen and lithium LoadNamedField and Store...
titzer@chromium.org [Fri, 24 May 2013 12:40:08 +0000 (12:40 +0000)]
Remove offset() and is_in_object() from hydrogen and lithium LoadNamedField and StoreNamedField and use the appropriate methods on HObjectAccess instead

BUG=

Review URL: https://codereview.chromium.org/15881003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14806 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoPrepare push to trunk. Now working on version 3.19.5.
svenpanne@chromium.org [Fri, 24 May 2013 12:34:41 +0000 (12:34 +0000)]
Prepare push to trunk.  Now working on version 3.19.5.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/16002003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14803 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoDummyUse if Representation::Tagged()+Type::Smi() to Representation::Smi()
verwaest@chromium.org [Fri, 24 May 2013 12:30:24 +0000 (12:30 +0000)]
DummyUse if Representation::Tagged()+Type::Smi() to Representation::Smi()

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15778006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14802 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoUpdate representation-from-uses to support smi.
verwaest@chromium.org [Fri, 24 May 2013 12:29:37 +0000 (12:29 +0000)]
Update representation-from-uses to support smi.

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15692004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14801 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agofix some names in compare nil ic
olivf@chromium.org [Fri, 24 May 2013 11:44:55 +0000 (11:44 +0000)]
fix some names in compare nil ic

The compare nil ic introduces a mechanism called lightweight miss. The
idea is to insert a direct call to the miss handler instead of going
through the deopt routine. This can be use for uninitialized stubs, to
directly jump to the runtime.

This cl
1. Cleans up naming
2. localizes the lightweight miss functionality on the only user (compare nil ic)
3. fixes DoCodegen, to actually call the correct method (uninitialized vs initialized)

BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/15806005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14798 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix edge case in stack trace formatting.
yangguo@chromium.org [Fri, 24 May 2013 11:33:46 +0000 (11:33 +0000)]
Fix edge case in stack trace formatting.

Bug description: in strict mode, null as receiver is not implicitly converted
to the global object, so that when formatting the stack trace, the receiver of
the stack frame is null. The IS_OBJECT check returns true for null, but
%GetDataProperty expected a JSObject, which results in a failed RUNTIME_ASSERT.

R=mvstanton@chromium.org
BUG=237617

Review URL: https://chromiumcodereview.appspot.com/15670003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14797 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMake (most of) --trace-codegen available in release mode. Better output.
svenpanne@chromium.org [Fri, 24 May 2013 10:57:59 +0000 (10:57 +0000)]
Make (most of) --trace-codegen available in release mode. Better output.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/15941004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14796 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agofix thread safety issue in FunctionTemplate test
svenpanne@chromium.org [Fri, 24 May 2013 09:32:10 +0000 (09:32 +0000)]
fix thread safety issue in FunctionTemplate test

R=svenpanne@chromium.org
BUG=

Review URL: https://codereview.chromium.org/15791005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14795 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoReplace tagged keys for fast access by smi, and use smi in boundscheck.
verwaest@chromium.org [Fri, 24 May 2013 08:52:35 +0000 (08:52 +0000)]
Replace tagged keys for fast access by smi, and use smi in boundscheck.
BUG=
R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15763004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14794 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoIntroduce ObjectAccess, which is used by LoadNamedField and StoreNamedField to denote...
titzer@chromium.org [Fri, 24 May 2013 08:38:21 +0000 (08:38 +0000)]
Introduce ObjectAccess, which is used by LoadNamedField and StoreNamedField to denote what parts of an object are referred to by a given load or store. Refactor HGraphBuilder to use ObjectAccess, which removes the need to manually set GVN flags and simplifies the code as well.

Review URL: https://codereview.chromium.org/14284010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14791 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoAdd belated test for the SeqStringSetChar bug.
yangguo@chromium.org [Fri, 24 May 2013 08:37:27 +0000 (08:37 +0000)]
Add belated test for the SeqStringSetChar bug.

R=titzer@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/15849003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14790 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agocallback handler map not correctly populated by direct use of SetCallHandler
dcarney@chromium.org [Fri, 24 May 2013 07:59:33 +0000 (07:59 +0000)]
callback handler map not correctly populated by direct use of SetCallHandler

R=svenpanne@chromium.org
BUG=

Review URL: https://codereview.chromium.org/15814005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14789 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Fix assert on ARM
palfia@homejinni.com [Fri, 24 May 2013 01:11:33 +0000 (01:11 +0000)]
MIPS: Fix assert on ARM

Port r14781 (a3491cb7)

BUG=

Review URL: https://codereview.chromium.org/15952002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14788 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Require smi for keyed store to smi.
palfia@homejinni.com [Fri, 24 May 2013 01:10:23 +0000 (01:10 +0000)]
MIPS: Require smi for keyed store to smi.

Port r14779 (33d24080)

BUG=

Review URL: https://codereview.chromium.org/15889007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14787 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Tag length of FixedArrayBase and smi-array[x] as smi representation
palfia@homejinni.com [Fri, 24 May 2013 01:09:23 +0000 (01:09 +0000)]
MIPS: Tag length of FixedArrayBase and smi-array[x] as smi representation

Port r14778 (c7fdf61b)

BUG=

Review URL: https://codereview.chromium.org/15927005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14786 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Improve SeqStringSetChar implementation.
palfia@homejinni.com [Fri, 24 May 2013 01:06:57 +0000 (01:06 +0000)]
MIPS: Improve SeqStringSetChar implementation.

Port r14769 (f0000b20)

BUG=

Review URL: https://codereview.chromium.org/15949002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14785 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Always require exact maps.
palfia@homejinni.com [Fri, 24 May 2013 01:05:42 +0000 (01:05 +0000)]
MIPS: Always require exact maps.

Port r14768 (5dec309f)

BUG=

Review URL: https://codereview.chromium.org/15948003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14784 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Omit smi-check for write-barrier unless tagged.
palfia@homejinni.com [Fri, 24 May 2013 01:04:08 +0000 (01:04 +0000)]
MIPS: Omit smi-check for write-barrier unless tagged.

Port r14767 (40dd85ec)

BUG=

Review URL: https://codereview.chromium.org/15745021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14783 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Implement HChange support for Smis and use it in Load/StoreNameField
palfia@homejinni.com [Fri, 24 May 2013 01:02:27 +0000 (01:02 +0000)]
MIPS: Implement HChange support for Smis and use it in Load/StoreNameField

Port r14765 (22625125)

BUG=

Review URL: https://codereview.chromium.org/15781006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14782 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix assert on ARM
verwaest@chromium.org [Thu, 23 May 2013 16:54:12 +0000 (16:54 +0000)]
Fix assert on ARM

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15886004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14781 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoSimplifying GC heuristics, deleted old generation allocation limit.
hpayer@chromium.org [Thu, 23 May 2013 15:11:43 +0000 (15:11 +0000)]
Simplifying GC heuristics, deleted old generation allocation limit.

BUG=

Review URL: https://codereview.chromium.org/14700006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14780 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoRequire smi for keyed store to smi.
verwaest@chromium.org [Thu, 23 May 2013 15:01:17 +0000 (15:01 +0000)]
Require smi for keyed store to smi.

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15884002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14779 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoTag length of FixedArrayBase and smi-array[x] as smi representation
verwaest@chromium.org [Thu, 23 May 2013 14:38:39 +0000 (14:38 +0000)]
Tag length of FixedArrayBase and smi-array[x] as smi representation

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15858006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14778 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix embedded new-space pointer in LCmpObjectEqAndBranch.
mstarzinger@chromium.org [Thu, 23 May 2013 14:06:28 +0000 (14:06 +0000)]
Fix embedded new-space pointer in LCmpObjectEqAndBranch.

R=mvstanton@chromium.org
BUG=chromium:240032
TEST=mjsunit/regress/regress-crbug-240032

Review URL: https://codereview.chromium.org/15779004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14777 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoAdd asserts to String::GetFlatContent.
yangguo@chromium.org [Thu, 23 May 2013 13:53:49 +0000 (13:53 +0000)]
Add asserts to String::GetFlatContent.

R=verwaest@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/13841012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14776 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMissing type cell on ia32 from bindings.
mvstanton@chromium.org [Thu, 23 May 2013 13:45:33 +0000 (13:45 +0000)]
Missing type cell on ia32 from bindings.

Javascript constructors called from C++ code didn't have a type cell
properly filled in on ia32. This showed up as a bug in webkit bindings.
Re-enabled flag optimize-constructed-arrays.

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/15870002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14775 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoDon't use fast literal if the boilerplate map is still deprecated.
verwaest@chromium.org [Thu, 23 May 2013 11:30:24 +0000 (11:30 +0000)]
Don't use fast literal if the boilerplate map is still deprecated.

R=mvstanton@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15660005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14774 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix Windows build after r14770
dslomov@chromium.org [Thu, 23 May 2013 10:29:28 +0000 (10:29 +0000)]
Fix Windows build after r14770

TBR=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/15848003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14771 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoExternalization API for ArrayBuffer
dslomov@chromium.org [Thu, 23 May 2013 10:01:42 +0000 (10:01 +0000)]
Externalization API for ArrayBuffer

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/15001041

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14770 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoImprove SeqStringSetChar implementation.
yangguo@chromium.org [Thu, 23 May 2013 09:51:06 +0000 (09:51 +0000)]
Improve SeqStringSetChar implementation.

R=jkummerow@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/15743006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14769 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoAlways require exact maps.
verwaest@chromium.org [Thu, 23 May 2013 09:19:18 +0000 (09:19 +0000)]
Always require exact maps.

R=danno@chromium.org

Review URL: https://chromiumcodereview.appspot.com/13923003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14768 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoOmit smi-check for write-barrier unless tagged.
verwaest@chromium.org [Thu, 23 May 2013 09:18:52 +0000 (09:18 +0000)]
Omit smi-check for write-barrier unless tagged.

R=mvstanton@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15848002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14767 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoAvoid creating duplicate entries for a value when merging HSimulates
jkummerow@chromium.org [Thu, 23 May 2013 09:17:01 +0000 (09:17 +0000)]
Avoid creating duplicate entries for a value when merging HSimulates

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/15660003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14766 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoImplement HChange support for Smis and use it in Load/StoreNameField
verwaest@chromium.org [Thu, 23 May 2013 08:32:07 +0000 (08:32 +0000)]
Implement HChange support for Smis and use it in Load/StoreNameField

BUG=
R=verwaest@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15303004

Patch from Daniel Clifford <danno@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14765 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agode-isolate Persistent::Dispose
dcarney@chromium.org [Thu, 23 May 2013 08:19:27 +0000 (08:19 +0000)]
de-isolate Persistent::Dispose

R=svenpanne@chromium.org
BUG=

Review URL: https://codereview.chromium.org/15648008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14764 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMove global pretenuring flag check to ShouldGloballyPretenure().
hpayer@chromium.org [Thu, 23 May 2013 08:17:03 +0000 (08:17 +0000)]
Move global pretenuring flag check to ShouldGloballyPretenure().

BUG=

Review URL: https://codereview.chromium.org/15734007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14763 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMake Object.freeze fast
verwaest@chromium.org [Thu, 23 May 2013 07:05:58 +0000 (07:05 +0000)]
Make Object.freeze fast

This patch both speeds up the freeze operation itself, but also
allows properties to remain in fast mode. Objects with non-empty
elements backing stores still end up with slow elements.

Relanding r14758 and r14759 with fix for Test262: only mark properties
and elements READ_ONLY if they are not JS setter/getters. Tightened up
tests to assert frozen-ness, and added targeted tests for the new code
(covering accessors).

BUG=v8:1858, 115960
R=verwaest@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15691007

Patch from Adam Klein <adamk@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14762 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Handle holes in smi-untag from LoadKeyed requiring hole handling.
palfia@homejinni.com [Wed, 22 May 2013 23:09:22 +0000 (23:09 +0000)]
MIPS: Handle holes in smi-untag from LoadKeyed requiring hole handling.

Port r14756 (fda1989f)

BUG=

Review URL: https://codereview.chromium.org/15690003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14761 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoRevert "Make Object.freeze fast"
adamk@chromium.org [Wed, 22 May 2013 21:27:00 +0000 (21:27 +0000)]
Revert "Make Object.freeze fast"
and "Fix Object.freeze on dictionary-backed arrays to properly freeze elements"

This reverts r14758 and r14759 due to introducing failures in Test262

TBR=verwaest@chromium.org

Review URL: https://codereview.chromium.org/15681004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14760 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix Object.freeze on dictionary-backed arrays to properly freeze elements
adamk@chromium.org [Wed, 22 May 2013 20:40:04 +0000 (20:40 +0000)]
Fix Object.freeze on dictionary-backed arrays to properly freeze elements

Follow-up to r14758: slightly rearranges JSObject::Freeze() to avoid duplicating
code while still retaining proper dictionary elements storage behavior.

Also fix a lint error.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/15737018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14759 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMake Object.freeze fast
adamk@chromium.org [Wed, 22 May 2013 18:53:58 +0000 (18:53 +0000)]
Make Object.freeze fast

This patch both speeds up the freeze operation itself, but also
allows properties to remain in fast mode. Objects with non-empty
elements backing stores still end up with slow elements.

BUG=v8:1858, 115960
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/14888005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14758 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix bogus deopt in BuildEmitDeepCopy for holey arrays.
mstarzinger@chromium.org [Wed, 22 May 2013 17:58:21 +0000 (17:58 +0000)]
Fix bogus deopt in BuildEmitDeepCopy for holey arrays.

R=verwaest@chromium.org
BUG=chromium:242924
TEST=mjsunit/regress/regress-crbug-242924

Review URL: https://codereview.chromium.org/15735012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14757 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoHandle holes in smi-untag from LoadKeyed requiring hole handling.
verwaest@chromium.org [Wed, 22 May 2013 16:32:33 +0000 (16:32 +0000)]
Handle holes in smi-untag from LoadKeyed requiring hole handling.

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15737003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14756 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoDon't allow copying holes to fields.
verwaest@chromium.org [Wed, 22 May 2013 15:33:53 +0000 (15:33 +0000)]
Don't allow copying holes to fields.

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15745006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14753 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix corner case in optimized code map zapping.
mstarzinger@chromium.org [Wed, 22 May 2013 15:29:02 +0000 (15:29 +0000)]
Fix corner case in optimized code map zapping.

R=jkummerow@chromium.org
TEST=mjsunit/math-floor-part2

Review URL: https://codereview.chromium.org/15743005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14752 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix Windows shared library build.
ulan@chromium.org [Wed, 22 May 2013 15:12:07 +0000 (15:12 +0000)]
Fix Windows shared library build.

We cannot dll-export templates.

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15697006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14751 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoPrepare push to trunk. Now working on version 3.19.4.
ulan@chromium.org [Wed, 22 May 2013 14:03:32 +0000 (14:03 +0000)]
Prepare push to trunk.  Now working on version 3.19.4.

R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15746005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14748 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix VisitLogicalExpression for empty blocks on RHS.
mstarzinger@chromium.org [Wed, 22 May 2013 13:27:00 +0000 (13:27 +0000)]
Fix VisitLogicalExpression for empty blocks on RHS.

R=jkummerow@chromium.org
BUG=chromium:242870
TEST=mjsunit/regress/regress-crbug-242870

Review URL: https://codereview.chromium.org/15744002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14747 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix unexpected elements transition in JSON.parse
yangguo@chromium.org [Wed, 22 May 2013 13:24:18 +0000 (13:24 +0000)]
Fix unexpected elements transition in JSON.parse

R=verwaest@chromium.org
BUG=241344

Review URL: https://chromiumcodereview.appspot.com/15739003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14746 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoKeep representations while overwriting transitions.
verwaest@chromium.org [Wed, 22 May 2013 10:46:33 +0000 (10:46 +0000)]
Keep representations while overwriting transitions.

BUG=chromium:241477
R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15718002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14745 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoUse explicit type feedback clearing in some tests.
mstarzinger@chromium.org [Wed, 22 May 2013 09:17:27 +0000 (09:17 +0000)]
Use explicit type feedback clearing in some tests.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/15711004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14744 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoForce GC before executing unbox double arrays test to avoid timeouts.
hpayer@chromium.org [Wed, 22 May 2013 09:05:22 +0000 (09:05 +0000)]
Force GC before executing unbox double arrays test to avoid timeouts.

BUG=

Review URL: https://codereview.chromium.org/15292002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14743 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agobuild fix for 14738
dcarney@chromium.org [Wed, 22 May 2013 07:32:36 +0000 (07:32 +0000)]
build fix for 14738

BUG=
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/15688003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14739 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoimplement fast ReturnValue setters
dcarney@chromium.org [Wed, 22 May 2013 06:35:38 +0000 (06:35 +0000)]
implement fast ReturnValue setters

R=svenpanne@chromium.org
BUG=

Review URL: https://codereview.chromium.org/15398008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14738 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: fix arm simulator after 14725
palfia@homejinni.com [Tue, 21 May 2013 23:09:04 +0000 (23:09 +0000)]
MIPS: fix arm simulator after 14725

Port r14731 (046b5d8d)

BUG=

Review URL: https://codereview.chromium.org/15606002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14737 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: new style of property/function callbacks
palfia@homejinni.com [Tue, 21 May 2013 21:09:58 +0000 (21:09 +0000)]
MIPS: new style of property/function callbacks

Port r14725 (d393d88)

BUG=

Review URL: https://codereview.chromium.org/15562007
Patch from Balazs Kilvady <kilvadyb@homejinni.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14736 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoDon't track representations in context extensions.
verwaest@chromium.org [Tue, 21 May 2013 17:21:27 +0000 (17:21 +0000)]
Don't track representations in context extensions.
This also enables verification of representations.

BUG=
R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15411003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14735 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoAdd regression test for fix from r14732.
mstarzinger@chromium.org [Tue, 21 May 2013 14:20:42 +0000 (14:20 +0000)]
Add regression test for fix from r14732.

R=verwaest@chromium.org
BUG=chromium:242502
TEST=mjsunit/regress/regress-crbug-242502

Review URL: https://codereview.chromium.org/15288008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14734 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoa tiny bug in DehoistArrayIndex
mmassi@chromium.org [Tue, 21 May 2013 14:09:28 +0000 (14:09 +0000)]
a tiny bug in DehoistArrayIndex

R=mmassi@chromium.org

Review URL: https://codereview.chromium.org/15268009

Patch from Weiliang Lin <weiliang.lin2@gmail.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14733 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoZap map value in the stub-cache on clear.
verwaest@chromium.org [Tue, 21 May 2013 12:59:48 +0000 (12:59 +0000)]
Zap map value in the stub-cache on clear.
The empty_string is a valid key for keyed loads and stores.

BUG=
R=mstarzinger@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15565002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14732 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agofix arm simulator after 14725
dcarney@chromium.org [Tue, 21 May 2013 12:17:04 +0000 (12:17 +0000)]
fix arm simulator after 14725

BUG=

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/15484006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14731 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMake sure we register extensions only once. Removed unused member variable.
svenpanne@chromium.org [Tue, 21 May 2013 12:03:49 +0000 (12:03 +0000)]
Make sure we register extensions only once. Removed unused member variable.

V8::Initialize() can be called from various threads simultaneously, so
everything should be done only once per process or within the Isolate. For
registering extensions, we do the former now.

R=dslomov@chromium.org

Review URL: https://codereview.chromium.org/15564002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14730 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoDisable flaky test after r14723.
mstarzinger@chromium.org [Tue, 21 May 2013 11:24:20 +0000 (11:24 +0000)]
Disable flaky test after r14723.

R=svenpanne@chromium.org
BUG=v8:2628
TEST=cctest/test-cpu-profiler/SampleWhenFrameIsNotSetup

Review URL: https://codereview.chromium.org/15415007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14729 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoDon't create new maps in CurrentMapForDeprecated.
verwaest@chromium.org [Tue, 21 May 2013 11:20:24 +0000 (11:20 +0000)]
Don't create new maps in CurrentMapForDeprecated.

R=yangguo@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15358005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14728 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoRemove unneeded argument from Parser::GetSymbol
wingo@igalia.com [Tue, 21 May 2013 10:45:58 +0000 (10:45 +0000)]
Remove unneeded argument from Parser::GetSymbol

Parser::GetSymbol can't actually fail, so no need for the bool* ok
argument or the CHECK_OK in callers.

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/15421007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14727 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMake Runtime_Abort print to stderr instead of stdout.
mstarzinger@chromium.org [Tue, 21 May 2013 09:25:57 +0000 (09:25 +0000)]
Make Runtime_Abort print to stderr instead of stdout.

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/15298005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14726 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agonew style of property/function callbacks
dcarney@chromium.org [Tue, 21 May 2013 06:36:24 +0000 (06:36 +0000)]
new style of property/function callbacks

R=svenpanne@chromium.org
BUG=

Review URL: https://codereview.chromium.org/12494012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14725 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoARM: Smi refactoring and improvements.
rodolph.perfetta@gmail.com [Fri, 17 May 2013 15:38:14 +0000 (15:38 +0000)]
ARM: Smi refactoring and improvements.

Refactoring:
 * consistent use of SmiTag/Untag
 * added a few Smi macros and helpers
Improvements
 * small optimisations (e.g. merging untag and cmp #0)
 * added fixed point to double conversion instructions for simpler conversions

More on the last point: a Smi can be seen as a fixed point number with the
a one bit fractional part. Fixed to double instructions allow us to convert
a Smi to a double without untagging.

BUG=none
TEST=none

Review URL: https://chromiumcodereview.appspot.com/15085026

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14724 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoAllow for no samples in test-cpu-profiler/SampleWhenFrameIsNotSetup
yurys@chromium.org [Fri, 17 May 2013 15:10:25 +0000 (15:10 +0000)]
Allow for no samples in test-cpu-profiler/SampleWhenFrameIsNotSetup

The test should only check that there are no sample stacks that never possible in the JS code being profiled.

BUG=v8:2628
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/14845018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14723 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoReduce register pressure in some runtime functions
rossberg@chromium.org [Fri, 17 May 2013 13:54:12 +0000 (13:54 +0000)]
Reduce register pressure in some runtime functions

Hoped to cure the spurious 3% regression we saw on beloved chromium-rel-win7-dual/sunspider after http://code.google.com/p/v8/source/detail?r=14236

R=mstarzinger@chromium.org
BUG=chromium:233441

Review URL: https://codereview.chromium.org/15297004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14722 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoPrepare push to trunk. Now working on version 3.19.3.
jkummerow@chromium.org [Fri, 17 May 2013 13:37:26 +0000 (13:37 +0000)]
Prepare push to trunk.  Now working on version 3.19.3.

R=danno@chromium.org

Review URL: https://codereview.chromium.org/15295007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14719 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoTurning off optimize-constructed-arrays to investigate a WebKit/bindings issue.
mvstanton@chromium.org [Fri, 17 May 2013 12:33:48 +0000 (12:33 +0000)]
Turning off optimize-constructed-arrays to investigate a WebKit/bindings issue.

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/15303002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14718 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Unify deoptimizer for JavaScript frames.
palfia@homejinni.com [Fri, 17 May 2013 10:46:42 +0000 (10:46 +0000)]
MIPS: Unify deoptimizer for JavaScript frames.

Port r14715 (84633474)

Original commit message:
This unifies the translation of an optimized frame to a full JavaScript
frame. Only the frame's context and fp register as well as alignment
padding are different on each architecture and can be factored out.

BUG=

Review URL: https://codereview.chromium.org/15291002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14717 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix transition test to support allocation site info.
hpayer@chromium.org [Fri, 17 May 2013 08:56:45 +0000 (08:56 +0000)]
Fix transition test to support allocation site info.

BUG=

Review URL: https://codereview.chromium.org/15270002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14716 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoUnify deoptimizer for JavaScript frames.
mstarzinger@chromium.org [Fri, 17 May 2013 08:27:56 +0000 (08:27 +0000)]
Unify deoptimizer for JavaScript frames.

This unifies the translation of an optimized frame to a full JavaScript
frame. Only the frame's context and fp register as well as alignment
padding are different on each architecture and can be factored out.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/14843020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14715 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFill in one-word-fillers for the unused property fields.
danno@chromium.org [Fri, 17 May 2013 06:58:06 +0000 (06:58 +0000)]
Fill in one-word-fillers for the unused property fields.

BUG=chromium:240056

Review URL: https://chromiumcodereview.appspot.com/15247003
Patch from Toon Verwaest <verwaest@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14714 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoFix bugs in rewriting combined with attributes and accessors
verwaest@chromium.org [Fri, 17 May 2013 03:16:20 +0000 (03:16 +0000)]
Fix bugs in rewriting combined with attributes and accessors

R=danno@chromium.org

Review URL: https://chromiumcodereview.appspot.com/14843023

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14713 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

11 years agoMIPS: Implement yield* (delegating yield)
palfia@homejinni.com [Fri, 17 May 2013 01:10:09 +0000 (01:10 +0000)]
MIPS: Implement yield* (delegating yield)

Port r14669 (b64ce42a)

Original commit message:
Ideally this would have been implemented via desugaring at parse-time,
but yield* is an expression, and its desugaring includes statements like
while and try/catch.  We'd have to have BlockExpression in the AST to
support that, and it's not worth it for this feature.

So instead we implement all of the logic in
FullCodeGenerator::VisitYield.  Delegating yield AST nodes now have a
try handler index, for the try/catch.  Otherwise the implementation is
straightforward.

BUG=

Review URL: https://codereview.chromium.org/15210004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14712 ce2b1a6d-e550-0410-aec6-3dcde31c8c00