platform/upstream/v8.git
10 years agov8: fix tracking of longest_step in IncrementalMarking.
ernstm@chromium.org [Wed, 16 Jul 2014 13:19:31 +0000 (13:19 +0000)]
v8: fix tracking of longest_step in IncrementalMarking.

longest_step is only output at the end of a mark-compact. We shouldn't
reset it after a Scavenge.

R=hpayer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/395153003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22432 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoChange ScriptCompiler::CompileOptions to allow for two 'cache' modes
vogelheim@chromium.org [Wed, 16 Jul 2014 12:18:33 +0000 (12:18 +0000)]
Change ScriptCompiler::CompileOptions to allow for two 'cache' modes
(parser or code) and to be explicit about cache consumption or production
(rather than making presence of cached_data imply one or the other.)

Also add a --cache flag to d8, to allow testing the functionality.

-----------------------------
API change

Reason: Currently, V8 supports a 'parser cache' for repeatedly executing the same script. We'd like to add a 2nd mode that would cache code, and would like to let the embedder decide which mode they chose (if any).

Note: Previously, the 'use cached data' property was implied by the presence of the cached data itself. (That is, kNoCompileOptions and source->cached_data != NULL.) That is no longer sufficient, since the presence of data is no longer sufficient to determine /which kind/ of data is present.

Changes from old behaviour:

- If you previously didn't use caching, nothing changes.
Example:
  v8::CompileUnbound(isolate, source, kNoCompileOptions);

- If you previously used caching, it worked like this:

  - 1st run:
  v8::CompileUnbound(isolate, source, kProduceToCache);
  Then, source->cached_data would contain the
  data-to-be cached. This remains the same, except you
  need to tell V8 which type of data you want.
  v8::CompileUnbound(isolate, source, kProduceParserCache);

  - 2nd run:
  v8::CompileUnbound(isolate, source, kNoCompileOptions);
  with source->cached_data set to the data you received in
  the first run. This will now ignore the cached data, and
  you need to explicitly tell V8 to use it:
  v8::CompileUnbound(isolate, source, kConsumeParserCache);
-----------------------------

BUG=
R=marja@chromium.org, yangguo@chromium.org

Review URL: https://codereview.chromium.org/389573006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22431 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRename [Load/Delete]ContextSlot for consistency after r22379.
mstarzinger@chromium.org [Wed, 16 Jul 2014 12:03:33 +0000 (12:03 +0000)]
Rename [Load/Delete]ContextSlot for consistency after r22379.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/397973002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22430 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoStore builtin index on the builtin code object.
yangguo@chromium.org [Wed, 16 Jul 2014 09:55:34 +0000 (09:55 +0000)]
Store builtin index on the builtin code object.

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/395823002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22429 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoProperly escape benchmark names when retrieving results.
machenbach@chromium.org [Wed, 16 Jul 2014 09:54:34 +0000 (09:54 +0000)]
Properly escape benchmark names when retrieving results.

This allows benchmarks to contain regular expression characters in their name.

BUG=
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/396203003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22428 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRevert "Remove erroneous uses of CXX / CXX_target / CXX_host from gyp."
rmcilroy@chromium.org [Wed, 16 Jul 2014 09:53:32 +0000 (09:53 +0000)]
Revert "Remove erroneous uses of CXX / CXX_target / CXX_host from gyp."

This reverts commit r22425 for breaking the Arm debug builder.

BUG=384474
TBR=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/393273002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22427 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMake ToPrimitive throw on symbol wrappers
rossberg@chromium.org [Wed, 16 Jul 2014 09:26:11 +0000 (09:26 +0000)]
Make ToPrimitive throw on symbol wrappers

R=mstarzinger@chromium.org
BUG=v8:3442
LOG=Y

Review URL: https://codereview.chromium.org/389263003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22426 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRemove erroneous uses of CXX / CXX_target / CXX_host from gyp.
rmcilroy@chromium.org [Wed, 16 Jul 2014 09:20:37 +0000 (09:20 +0000)]
Remove erroneous uses of CXX / CXX_target / CXX_host from gyp.

The current gyp config tries to test whether the compiler is an arm / mips
compiler and whether it supports -m32 / -m64 based on a test within gyp.
Unfortunately CXX / CXX_target / CXX_host are no no longer set early enough
in gyp for this to work, and so all of these tests were actually being done
using "$(which g++)" rather than the appropriate CXX_target / CXX_host compiler.

This change removes these dynamic tests and instead bases the decisions on
knowledge of the host / target / v8_target architecture combination.  The
the -m32/-m64 flags are set based on host_cxx_is_biarch / target_cxx_is_biarch,
which are set to reasonable defaults of only ia32/x64 or clang compilers
being build for biarch support.  The host_cxx_is_biarch / target_cxx_is_biarch
variables can be overriden if required.

BUG=384474
LOG=NO
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/382343003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22425 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoAllow benchmarks to provide the standard deviation.
machenbach@chromium.org [Wed, 16 Jul 2014 08:53:46 +0000 (08:53 +0000)]
Allow benchmarks to provide the standard deviation.

Some benchmarks include their own runner which provides an
overall average and a standard deviation. This enables
extraction of that value similar to the other measurements.

These benchmarks should only be run once. If a benchmarks
specifies multiple runs and provides a standard deviation,
a warning will be issued that makes the build fail on the
buildbot side.

TEST=python -m unittest run_benchmarks_test
BUG=393947
LOG=n
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/395633012

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22424 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years ago[Auto-roll] Bump up version to 3.28.26.0
machenbach@chromium.org [Wed, 16 Jul 2014 08:42:52 +0000 (08:42 +0000)]
[Auto-roll] Bump up version to 3.28.26.0

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/394153003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22423 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoDo not expose all timer events to the API callback.
yangguo@chromium.org [Wed, 16 Jul 2014 08:14:50 +0000 (08:14 +0000)]
Do not expose all timer events to the API callback.

R=fmeawad@chromium.org

Review URL: https://codereview.chromium.org/388783004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22422 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoAdd script to bump up version on bleeding edge.
machenbach@chromium.org [Wed, 16 Jul 2014 07:57:13 +0000 (07:57 +0000)]
Add script to bump up version on bleeding edge.

The version.cc file will from now on be controlled on
bleeding edge and on trunk. The next version to use will
always be max(bleeding_edge_version, trunk_version) + 1.

The trunk and bleeding_edge versions are not semantically
tied together. The trunk branch can be used to make
independent point releases of bleeding_edge revisions.

The version on bleeding_edge will be increased if:
- the lkgr version is smaller than the version of the latest
  revision,
- the lkgr version is not a version change itself,
- the tree is not closed for maintenance.

This will allow chromium to refer directly to bleeding_edge
revisions of v8. The v8 versions will not be unique, but
they will change as often as possible, dependent on the lkgr
cycle time and the frequency this script is called.

BUG=391261
LOG=n
TEST=tools/push-to-trunk/script_test.py
TEST=tools/push-to-trunk/bump_up_version.py --dry_run
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/389353003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22421 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoError.captureStackTrace should define "stack" property as configurable.
yangguo@chromium.org [Wed, 16 Jul 2014 07:55:05 +0000 (07:55 +0000)]
Error.captureStackTrace should define "stack" property as configurable.

R=verwaest@chromium.org
BUG=393988
LOG=N

Review URL: https://codereview.chromium.org/396063008

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22420 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoVerify that source string matches serialized code.
yangguo@chromium.org [Wed, 16 Jul 2014 06:59:14 +0000 (06:59 +0000)]
Verify that source string matches serialized code.

R=mvstanton@chromium.org, vogelheim@chromium.org

Review URL: https://codereview.chromium.org/394793002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22417 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoX87: Drop unnecessary receiver validity checks from {Load,Store}IC_Normal
weiliang.lin@intel.com [Wed, 16 Jul 2014 03:12:29 +0000 (03:12 +0000)]
X87: Drop unnecessary receiver validity checks from {Load,Store}IC_Normal

port r22391

original commit message:
   Drop unnecessary receiver validity checks from {Load,Store}IC_Normal.

   Since these builtins are used as handlers after a map check/dispatch, they don't need to check the receiver again.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/396833004

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22416 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMIPS: Minor fixes to r22396
akos.palfi@imgtec.com [Tue, 15 Jul 2014 19:03:13 +0000 (19:03 +0000)]
MIPS: Minor fixes to r22396

BUG=
R=akos.palfi@imgtec.com

Review URL: https://codereview.chromium.org/394913002

Patch from Balazs Kilvady <kilvadyb@homejinni.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22413 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMIPS: Fix r19500 register usage to match to ASSERT.
akos.palfi@imgtec.com [Tue, 15 Jul 2014 19:01:31 +0000 (19:01 +0000)]
MIPS: Fix r19500 register usage to match to ASSERT.

BUG=
R=akos.palfi@imgtec.com

Review URL: https://codereview.chromium.org/395863002

Patch from Balazs Kilvady <kilvadyb@homejinni.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22412 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMIPS: Drop unnecessary receiver validity checks from {Load,Store}IC_Normal.
akos.palfi@imgtec.com [Tue, 15 Jul 2014 19:00:02 +0000 (19:00 +0000)]
MIPS: Drop unnecessary receiver validity checks from {Load,Store}IC_Normal.

Port r22391 (fe86010)

Original commit message:
Since these builtins are used as handlers after a map check/dispatch, they don't need to check the receiver again.

BUG=
R=akos.palfi@imgtec.com

Review URL: https://codereview.chromium.org/396663004

Patch from Balazs Kilvady <kilvadyb@homejinni.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22411 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMIPS64: Reland r22082 "Replace HeapNumber as doublebox with an explicit MutableHeapN...
dusan.milosavljevic@rt-rk.com [Tue, 15 Jul 2014 16:57:04 +0000 (16:57 +0000)]
MIPS64:  Reland r22082 "Replace HeapNumber as doublebox with an explicit MutableHeapNumber.

Port r22129 (676bb14).

TEST=
BUG=
R=balazs.kilvady@imgtec.com, paul.lind@imgtec.com

Review URL: https://codereview.chromium.org/395803002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22410 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRevert "Remove failure tag.".
hpayer@chromium.org [Tue, 15 Jul 2014 12:22:38 +0000 (12:22 +0000)]
Revert "Remove failure tag.".

TBR=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/398473002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22409 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRemove failure tag.
hpayer@chromium.org [Tue, 15 Jul 2014 11:11:15 +0000 (11:11 +0000)]
Remove failure tag.

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/397613002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22408 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoFix assertion failure. Use Vector instead of List in code deserializer.
yangguo@chromium.org [Tue, 15 Jul 2014 11:01:03 +0000 (11:01 +0000)]
Fix assertion failure. Use Vector instead of List in code deserializer.

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/390353002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22407 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoAllow embedding of ConsString objects into code.
mstarzinger@chromium.org [Tue, 15 Jul 2014 10:51:12 +0000 (10:51 +0000)]
Allow embedding of ConsString objects into code.

R=yangguo@chromium.org
BUG=v8:2803
LOG=N

Review URL: https://codereview.chromium.org/395713002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22406 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoX87: Use the same registers for StoreIC and KeyedStoreIC.
weiliang.lin@intel.com [Tue, 15 Jul 2014 10:41:23 +0000 (10:41 +0000)]
X87: Use the same registers for StoreIC and KeyedStoreIC.

port r22387

original commit message:
  Use the same registers for StoreIC and KeyedStoreIC.

  Based on https://codereview.chromium.org/389283002 ("Use register parameters in ElementsTransitionGenerator") which must land first.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/391933002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22405 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoEnum RegisterInfo renamed to ParameterIndices.
mvstanton@chromium.org [Tue, 15 Jul 2014 10:19:30 +0000 (10:19 +0000)]
Enum RegisterInfo renamed to ParameterIndices.

"Info" is information-poor.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/385953004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22404 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoDo not dump user source code in the code serializer.
yangguo@chromium.org [Tue, 15 Jul 2014 10:17:22 +0000 (10:17 +0000)]
Do not dump user source code in the code serializer.

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/390303002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22403 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoX87: Use register parameters in ElementsTransitionGenerator
weiliang.lin@intel.com [Tue, 15 Jul 2014 09:46:10 +0000 (09:46 +0000)]
X87: Use register parameters in ElementsTransitionGenerator

port r22384

original commit message:
 Use register parameters in ElementsTransitionGenerator
x

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/395633010

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22402 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoUnify the IsShortcutCandidate predicate.
mstarzinger@chromium.org [Tue, 15 Jul 2014 09:19:22 +0000 (09:19 +0000)]
Unify the IsShortcutCandidate predicate.

R=yangguo@chromium.org
BUG=v8:2803
LOG=N

Review URL: https://codereview.chromium.org/390193007

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22399 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoFix up internalized strings after deserializing user code.
yangguo@chromium.org [Tue, 15 Jul 2014 08:46:47 +0000 (08:46 +0000)]
Fix up internalized strings after deserializing user code.

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/387343002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22398 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMIPS: Use the same registers for StoreIC and KeyedStoreIC.
akos.palfi@imgtec.com [Tue, 15 Jul 2014 08:40:26 +0000 (08:40 +0000)]
MIPS: Use the same registers for StoreIC and KeyedStoreIC.

Port r22387 (ff76c666)

BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/390263002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22397 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMIPS: Use register parameters in ElementsTransitionGenerator.
akos.palfi@imgtec.com [Tue, 15 Jul 2014 08:38:16 +0000 (08:38 +0000)]
MIPS: Use register parameters in ElementsTransitionGenerator.

Port r22384 (52caca20)

BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/393693003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22396 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoX87: Clean up and update const / var
weiliang.lin@intel.com [Tue, 15 Jul 2014 08:37:37 +0000 (08:37 +0000)]
X87: Clean up and update const / var

port r22379

original commit message:
This CL simplifies var / const by ensuring the behavior is consistent in itself,
and with regular JS semantics; between regular var/const and eval-ed var/const.

Legacy const is changed so that a declaration declares a configurable, but non-writable, slot,
 and the initializer reconfigures it (when possible) to non-configurable non-writable. This
 avoids the need for "the hole" as marker value in JSContextExtensionObjects and GlobalObjects.
Undefined is used instead.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/397573002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22395 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRemove sequential sweeping mode and perform lazy sweeping when no sweeper threads...
hpayer@chromium.org [Tue, 15 Jul 2014 08:24:44 +0000 (08:24 +0000)]
Remove sequential sweeping mode and perform lazy sweeping when no sweeper threads are active.

BUG=
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/396443004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22394 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agofix the vtune support bug.
danno@chromium.org [Tue, 15 Jul 2014 08:13:42 +0000 (08:13 +0000)]
fix the vtune support bug.

During https://code.google.com/p/v8/source/detail?r=19925 checkin context bound scripts (Script)
and context unbound scripts (UnboundScript) are Distinguished.

And then Sven Panne helped to fix the vtune support compilation
error in https://code.google.com/p/v8/source/detail?r=20955.

The problem is that there is runtime error for vtune
support.
In our original implementation, we encapsulated and passed v8::internal::Script
to V8 API. It will leads to type check error for current V8::Script definition.

So I changed the Handle<Script> definition in JitCodeEvent
to Handle<UnboundScript>
 and add the corresponding change in log.cc.

If you do NOT prefer to change in include/v8.h. I think I can change the definition of
CodeEventLogger::LogRecordedBuffer(...) so that the we can pass the correct
type (JSFunction) as V8::Script to V8 API.

BUG=
R=danno@chromium.org, svenpanne@chromium.org

Review URL: https://codereview.chromium.org/334263018

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22393 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoFix flags for test cases involving arrow functions
marja@chromium.org [Tue, 15 Jul 2014 08:10:26 +0000 (08:10 +0000)]
Fix flags for test cases involving arrow functions

There are a lot of tests in cctest/test-parsing/ErrorsArrowFunctions,
so it is quite slow. This patch removes some flags to make it faster.
Removing three flags that don't affect the test brings down the run
time one order of magnitude, which is fast enough even for debug/ASAN
bots.

Also, remove the unneeded kArrowFunctions flag from
cctest/test-parsing/NoErrorsYieldSloppyAllModes

BUG=
R=marja@chromium.org

Review URL: https://codereview.chromium.org/387383002

Patch from Adrián Pérez de Castro <aperez@igalia.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22392 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoDrop unnecessary receiver validity checks from {Load,Store}IC_Normal.
jkummerow@chromium.org [Tue, 15 Jul 2014 08:07:25 +0000 (08:07 +0000)]
Drop unnecessary receiver validity checks from {Load,Store}IC_Normal.

Since these builtins are used as handlers after a map check/dispatch, they don't need to check the receiver again.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/390053002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22391 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRemoved some copy-n-paste from StackFrame::Foo API entries.
svenpanne@chromium.org [Tue, 15 Jul 2014 06:59:42 +0000 (06:59 +0000)]
Removed some copy-n-paste from StackFrame::Foo API entries.

BUG=v8:3436
LOG=y
R=dcarney@chromium.org

Review URL: https://codereview.chromium.org/388183002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22390 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoUse the same registers for StoreIC and KeyedStoreIC.
mvstanton@chromium.org [Mon, 14 Jul 2014 20:43:41 +0000 (20:43 +0000)]
Use the same registers for StoreIC and KeyedStoreIC.

Based on https://codereview.chromium.org/389283002 ("Use register parameters in ElementsTransitionGenerator") which must land first.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/390013002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22387 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMIPS: Remove ASSERT since there are tons of different ASSIGN variants
kilvadyb@homejinni.com [Mon, 14 Jul 2014 18:53:49 +0000 (18:53 +0000)]
MIPS: Remove ASSERT since there are tons of different ASSIGN variants

Port r22382 (84a841f)

BUG=
R=akos.palfi@imgtec.com

Review URL: https://codereview.chromium.org/395503002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22386 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMIPS: This CL simplifies var / const by ensuring the behavior is consistent in itself...
kilvadyb@homejinni.com [Mon, 14 Jul 2014 18:22:34 +0000 (18:22 +0000)]
MIPS: This CL simplifies var / const by ensuring the behavior is consistent in itself, and with regular JS semantics; between regular var/const and eval-ed var/const.

Port r22379 (fb62653)

Original commit message:
Legacy const is changed so that a declaration declares a configurable, but non-writable, slot, and the initializer reconfigures it (when possible) to non-configurable non-writable. This avoids the need for "the hole" as marker value in JSContextExtensionObjects and GlobalObjects. Undefined is used instead.

BUG=
R=akos.palfi@imgtec.com

Review URL: https://codereview.chromium.org/389373004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22385 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoUse register parameters in ElementsTransitionGenerator.
mvstanton@chromium.org [Mon, 14 Jul 2014 15:03:38 +0000 (15:03 +0000)]
Use register parameters in ElementsTransitionGenerator.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/389283002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22384 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRemove PropertyAttributes from SetProperty
verwaest@chromium.org [Mon, 14 Jul 2014 14:52:24 +0000 (14:52 +0000)]
Remove PropertyAttributes from SetProperty

BUG=
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/390833003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22383 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRemove ASSERT since there are tons of different ASSIGN variants
verwaest@chromium.org [Mon, 14 Jul 2014 14:42:33 +0000 (14:42 +0000)]
Remove ASSERT since there are tons of different ASSIGN variants

BUG=
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/385173009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22382 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoReplace AddProperty by AddNamedProperty to speed up the common case
verwaest@chromium.org [Mon, 14 Jul 2014 14:05:30 +0000 (14:05 +0000)]
Replace AddProperty by AddNamedProperty to speed up the common case

BUG=
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/384003003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22381 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoFix d8 object printing (symbols, accessors)
rossberg@chromium.org [Mon, 14 Jul 2014 14:01:30 +0000 (14:01 +0000)]
Fix d8 object printing (symbols, accessors)

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/388243002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22380 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoThis CL simplifies var / const by ensuring the behavior is consistent in itself,...
verwaest@chromium.org [Mon, 14 Jul 2014 14:01:04 +0000 (14:01 +0000)]
This CL simplifies var / const by ensuring the behavior is consistent in itself, and with regular JS semantics; between regular var/const and eval-ed var/const.

Legacy const is changed so that a declaration declares a configurable, but non-writable, slot, and the initializer reconfigures it (when possible) to non-configurable non-writable. This avoids the need for "the hole" as marker value in JSContextExtensionObjects and GlobalObjects. Undefined is used instead.

BUG=
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/379893002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22379 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoReland "Include symbol properties in Object.{create,defineProperties}"
rossberg@chromium.org [Mon, 14 Jul 2014 14:00:33 +0000 (14:00 +0000)]
Reland "Include symbol properties in Object.{create,defineProperties}"

Second try; implementation that doesn't rely on external arrays.

R=mstarzinger@chromium.org
BUG=v8:3440

Review URL: https://codereview.chromium.org/391713002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22378 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoFix error message about read-only symbol properties.
mstarzinger@chromium.org [Mon, 14 Jul 2014 13:57:10 +0000 (13:57 +0000)]
Fix error message about read-only symbol properties.

R=rossberg@chromium.org
BUG=v8:3441
LOG=Y

Review URL: https://codereview.chromium.org/390783003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22377 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRevert "Remove sequential sweeping mode and perform lazy sweeping when no sweeper...
titzer@chromium.org [Mon, 14 Jul 2014 13:02:36 +0000 (13:02 +0000)]
Revert "Remove sequential sweeping mode and perform lazy sweeping when no sweeper threads are active."

Reason: broke win64 build

This reverts commit 221bfdd2da2b6f3c1cbe77c5d197f1ea626b0bd2.

TBR=hpayer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/393523002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22374 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRevert "Include symbol properties in Object.{create,defineProperties}"
rossberg@chromium.org [Mon, 14 Jul 2014 12:27:08 +0000 (12:27 +0000)]
Revert "Include symbol properties in Object.{create,defineProperties}"

TBR=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/394443002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22373 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRemove sequential sweeping mode and perform lazy sweeping when no sweeper threads...
hpayer@chromium.org [Mon, 14 Jul 2014 11:31:22 +0000 (11:31 +0000)]
Remove sequential sweeping mode and perform lazy sweeping when no sweeper threads are active.

BUG=
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/384373002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22372 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoSerialize builtins by referencing canonical ones.
yangguo@chromium.org [Mon, 14 Jul 2014 11:22:03 +0000 (11:22 +0000)]
Serialize builtins by referencing canonical ones.

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/383173002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22371 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoInclude symbol properties in Object.{create,defineProperties}
rossberg@chromium.org [Mon, 14 Jul 2014 10:59:29 +0000 (10:59 +0000)]
Include symbol properties in Object.{create,defineProperties}

R=mstarzinger@chromium.org
BUG=v8:3440
LOG=Y

Review URL: https://codereview.chromium.org/391683002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22370 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoUnify LookupIterator::GetRoot and Objects::GetRootMap
jochen@chromium.org [Mon, 14 Jul 2014 10:54:24 +0000 (10:54 +0000)]
Unify LookupIterator::GetRoot and Objects::GetRootMap

BUG=none
R=verwaest@chromium.org
LOG=n

Review URL: https://codereview.chromium.org/388193002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22369 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoEnsure that each profiling entry ends with a new line.
yangguo@chromium.org [Mon, 14 Jul 2014 10:47:40 +0000 (10:47 +0000)]
Ensure that each profiling entry ends with a new line.

Currently this does not happen if the message builder buffer is full.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/390803002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22368 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoWhitespace change to trigger bots.
machenbach@chromium.org [Mon, 14 Jul 2014 09:27:07 +0000 (09:27 +0000)]
Whitespace change to trigger bots.

TBR=jochen@chromium.org
BUG=

Review URL: https://codereview.chromium.org/386413002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22367 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoImplement handling of arrow functions in the parser
marja@chromium.org [Mon, 14 Jul 2014 07:55:45 +0000 (07:55 +0000)]
Implement handling of arrow functions in the parser

Arrow functions are parsed from ParseAssignmentExpression(). Handling the
parameter list is done by letting ParseConditionalExpression() parse a comma
separated list of identifiers, and it returns a tree of BinaryOperation nodes
with VariableProxy leaves, or a single VariableProxy if there is only one
parameter. When the arrow token "=>" is found, the VariableProxy nodes are
passed to ParseArrowFunctionLiteral(), which will then skip parsing the
paramaeter list. This avoids having to rewind when the arrow is found and
restart parsing the parameter list.

Note that the empty parameter list "()" is handled directly in
ParsePrimaryExpression(): after is has consumed the opening parenthesis,
if a closing parenthesis follows, then the only valid input is an arrow
function. In this case, ParsePrimaryExpression() directly calls
ParseArrowFunctionLiteral(), to avoid needing to return a sentinel value
to signal the empty parameter list. Because it will consume the body of
the arrow function, ParseAssignmentExpression() will not see the arrow
"=>" token as next, and return the already-parser expression.

The implementation is done in ParserBase, so it was needed to do some
additions to ParserBase, ParserTraits and PreParserTraits. Some of the
glue code can be removed later on when more more functionality is moved
to ParserBase.

Additionally, this adds a runtime flag "harmony_arrow_functions"
(disabled by default); enabling "harmony" will enable it as well.

BUG=v8:2700
LOG=N
R=marja@chromium.org

Review URL: https://codereview.chromium.org/383983002

Patch from Adrián Pérez de Castro <aperez@igalia.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22366 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoIntroduce a PrototypeIterator class and use it for prototype access
jochen@chromium.org [Mon, 14 Jul 2014 07:19:49 +0000 (07:19 +0000)]
Introduce a PrototypeIterator class and use it for prototype access

The new pattern is that we first get the map of the root of the
prototype chain using Object::GetMapRoot() and then walk up the
prototype chain using Map::prototype().

BUG=???
R=verwaest@chromium.org
LOG=n

Review URL: https://codereview.chromium.org/376233002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22365 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoWhiltespace change to trigger bots.
machenbach@chromium.org [Sat, 12 Jul 2014 16:56:32 +0000 (16:56 +0000)]
Whiltespace change to trigger bots.

TBR=jochen@chromium.org
BUG=

Review URL: https://codereview.chromium.org/386163002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22362 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoWhiltespace change to trigger bots.
machenbach@chromium.org [Sat, 12 Jul 2014 15:27:18 +0000 (15:27 +0000)]
Whiltespace change to trigger bots.

TBR=jochen@chromium.org
BUG=

Review URL: https://codereview.chromium.org/384113002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22361 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoFix arm64 deoptimization from double registers (reverts r20613).
jarin@chromium.org [Fri, 11 Jul 2014 19:30:09 +0000 (19:30 +0000)]
Fix arm64 deoptimization from double registers (reverts r20613).

This reverts "ARM64: Use pair memory access in deoptimizer entry", r20613. It does not really make sense to micro-optimize the deoptimizer as it is the ultra-slow path. Moreover, the original code was easier to read (in addition to being correct).

BUG=391313
LOG=N
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/389583003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22360 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRevert "Use the same registers for StoreIC and KeyedStoreIC."
mvstanton@chromium.org [Fri, 11 Jul 2014 14:50:59 +0000 (14:50 +0000)]
Revert "Use the same registers for StoreIC and KeyedStoreIC."

This reverts commit r22358 due to release mode arm64 test failures.

TBR=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/385073007

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22359 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoUse the same registers for StoreIC and KeyedStoreIC.
mvstanton@chromium.org [Fri, 11 Jul 2014 13:53:19 +0000 (13:53 +0000)]
Use the same registers for StoreIC and KeyedStoreIC.

The x64 and ia32 platforms already did this, port to arm and arm64.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/385553004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22358 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMIPS: slightly reduce deopt tables size.
kilvadyb@homejinni.com [Fri, 11 Jul 2014 13:49:33 +0000 (13:49 +0000)]
MIPS: slightly reduce deopt tables size.

Port r22305 (00a3740)

BUG=
R=dusan.milosavljevic@rt-rk.com

Review URL: https://codereview.chromium.org/387473003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22357 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoWhitespace change to trigger bots.
machenbach@chromium.org [Fri, 11 Jul 2014 13:29:12 +0000 (13:29 +0000)]
Whitespace change to trigger bots.

BUG=
TBR=jochen@chromium.org

Review URL: https://codereview.chromium.org/383963005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22356 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoUnbreak build by making old MSVC happy (hopefully).
svenpanne@chromium.org [Fri, 11 Jul 2014 12:55:56 +0000 (12:55 +0000)]
Unbreak build by making old MSVC happy (hopefully).

TBR=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/386023002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22355 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMade printing of special FP values portable.
svenpanne@chromium.org [Fri, 11 Jul 2014 12:42:37 +0000 (12:42 +0000)]
Made printing of special FP values portable.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/386973003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22354 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoWhitespace change to trigger bots.
machenbach@chromium.org [Fri, 11 Jul 2014 12:37:53 +0000 (12:37 +0000)]
Whitespace change to trigger bots.

BUG=
TBR=jochen@chromium.org

Review URL: https://codereview.chromium.org/389613002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22353 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoWhitespace change to trigger bots.
machenbach@chromium.org [Fri, 11 Jul 2014 12:28:22 +0000 (12:28 +0000)]
Whitespace change to trigger bots.

BUG=
TBR=jochen@chromium.org

Review URL: https://codereview.chromium.org/386973002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22352 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMake UBSan happy.
svenpanne@chromium.org [Fri, 11 Jul 2014 12:12:58 +0000 (12:12 +0000)]
Make UBSan happy.

This involves avoiding signed multiplication overflow, shifting too
far and overflow during negation.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/382153003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22351 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoWhitespace change to trigger bots.
machenbach@chromium.org [Fri, 11 Jul 2014 11:54:33 +0000 (11:54 +0000)]
Whitespace change to trigger bots.

BUG=
TBR=jochen@chromium.org

Review URL: https://codereview.chromium.org/385163003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22350 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRevert "Remove sequential sweeping mode and perform lazy sweeping when no sweeper...
dslomov@chromium.org [Fri, 11 Jul 2014 11:33:57 +0000 (11:33 +0000)]
Revert "Remove sequential sweeping mode and perform lazy sweeping when no sweeper threads are active."

This reverts commit r22346 for breaking GC stress tests.

TBR=hpayer@chromium.org

Review URL: https://codereview.chromium.org/386943003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22349 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoDisabling flakes.Filed 3433, 3434, 3435.
dslomov@chromium.org [Fri, 11 Jul 2014 11:20:37 +0000 (11:20 +0000)]
Disabling flakes.Filed 3433, 3434, 3435.

R=machenbach@chromium.org

Review URL: https://codereview.chromium.org/382083003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22348 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMIPS: update owners.
akos.palfi@imgtec.com [Fri, 11 Jul 2014 10:02:48 +0000 (10:02 +0000)]
MIPS: update owners.

BUG=
R=danno@chromium.org, plind44@gmail.com

Review URL: https://codereview.chromium.org/389463002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22347 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRemove sequential sweeping mode and perform lazy sweeping when no sweeper threads...
hpayer@chromium.org [Fri, 11 Jul 2014 09:51:34 +0000 (09:51 +0000)]
Remove sequential sweeping mode and perform lazy sweeping when no sweeper threads are active.

BUG=
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/382793002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22346 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoUnflake and speedup JsNative*Sample cpu profile tests
alph@chromium.org [Fri, 11 Jul 2014 09:06:12 +0000 (09:06 +0000)]
Unflake and speedup JsNative*Sample cpu profile tests

Instead of running cpu profiler for a hundred milliseconds,
collecting samples distributed in a non-deterministic way all along
the code, make the tests rely on a single sample we collect on
the profiler start.

R=bmeurer@chromium.org, yurys@chromium.org

Review URL: https://codereview.chromium.org/301603005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22345 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRemove AddCode and RemoveCode GDB JIT entry hooks.
mstarzinger@chromium.org [Fri, 11 Jul 2014 07:51:25 +0000 (07:51 +0000)]
Remove AddCode and RemoveCode GDB JIT entry hooks.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/387533003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22344 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoX87: Use a register spec for StoreIC and KeyedStoreIC.
weiliang.lin@intel.com [Fri, 11 Jul 2014 07:47:58 +0000 (07:47 +0000)]
X87: Use a register spec for StoreIC and KeyedStoreIC.

port r22328

original commit message:

  Use a register spec for StoreIC and KeyedStoreIC.

  This continues refactoring already applied for LoadIC in r22035 (https://code.google.com/p/v8/source/detail?r=22035).

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/382123003

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22343 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoChange ASSERTs in serializer macro to STATIC_ASSERTs.
yangguo@chromium.org [Fri, 11 Jul 2014 07:04:51 +0000 (07:04 +0000)]
Change ASSERTs in serializer macro to STATIC_ASSERTs.

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/387523002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22340 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoFix assertion in ScriptData.
yangguo@chromium.org [Fri, 11 Jul 2014 06:58:25 +0000 (06:58 +0000)]
Fix assertion in ScriptData.

R=marja@chromium.org

Review URL: https://codereview.chromium.org/383963002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22339 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoTemporarily remove newly introduced assertion.
yangguo@chromium.org [Fri, 11 Jul 2014 06:46:24 +0000 (06:46 +0000)]
Temporarily remove newly introduced assertion.

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/388753003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22338 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRevert "Implement handling of arrow functions in the parser"
marja@chromium.org [Fri, 11 Jul 2014 06:39:31 +0000 (06:39 +0000)]
Revert "Implement handling of arrow functions in the parser"

This reverts revision 22320.

Reason: ASAN still detects leaks!

Conflicts:
src/preparser.h

TBR=aperez@igalia.com,marja@chromium.org
BUG=

Review URL: https://codereview.chromium.org/389503002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22337 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMIPS: Use a register spec for StoreIC and KeyedStoreIC.
akos.palfi@imgtec.com [Fri, 11 Jul 2014 01:50:12 +0000 (01:50 +0000)]
MIPS: Use a register spec for StoreIC and KeyedStoreIC.

Port r22328 (92275b7a)

Original commit message:
This continues refactoring already applied for LoadIC in r22035 (https://code.google.com/p/v8/source/detail?r=22035).

BUG=
R=gergely@homejinni.com

Review URL: https://codereview.chromium.org/383913002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22336 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRevert "Enable out-of-line constant pool on Arm"
rmcilroy@chromium.org [Thu, 10 Jul 2014 22:39:10 +0000 (22:39 +0000)]
Revert "Enable out-of-line constant pool on Arm"

This reverts r22332 due to Arm breakages.

TBR=dslomov

Review URL: https://codereview.chromium.org/384953002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22335 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoAdd compile-only win64 trybot.
machenbach@chromium.org [Thu, 10 Jul 2014 19:24:33 +0000 (19:24 +0000)]
Add compile-only win64 trybot.

BUG=353487
LOG=n
R=jochen@chromium.org
TBR=jochen@chromium.org

Review URL: https://codereview.chromium.org/385583004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22334 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoEnsure we update attributes in CopyGeneralizeFieldRepresentation
verwaest@chromium.org [Thu, 10 Jul 2014 18:30:19 +0000 (18:30 +0000)]
Ensure we update attributes in CopyGeneralizeFieldRepresentation

BUG=
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/381113002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22333 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoEnable out-of-line constant pool on Arm
rmcilroy@chromium.org [Thu, 10 Jul 2014 18:06:39 +0000 (18:06 +0000)]
Enable out-of-line constant pool on Arm

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/228633002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22332 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoFix windows compile, again.
hpayer@chromium.org [Thu, 10 Jul 2014 17:13:27 +0000 (17:13 +0000)]
Fix windows compile, again.

BUG=
R=dslomov@chromium.org

Review URL: https://codereview.chromium.org/380263002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22331 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoOnly install JitCodeEventHandler for GDB JIT if requested.
mstarzinger@chromium.org [Thu, 10 Jul 2014 16:31:57 +0000 (16:31 +0000)]
Only install JitCodeEventHandler for GDB JIT if requested.

TBR=dslomov@chromium.org

Review URL: https://codereview.chromium.org/381883005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22330 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoFix compilation with shared library.
mstarzinger@chromium.org [Thu, 10 Jul 2014 15:19:29 +0000 (15:19 +0000)]
Fix compilation with shared library.

TBR=dslomov@chromium.org

Review URL: https://codereview.chromium.org/381103002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22329 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoUse a register spec for StoreIC and KeyedStoreIC.
mvstanton@chromium.org [Thu, 10 Jul 2014 14:59:47 +0000 (14:59 +0000)]
Use a register spec for StoreIC and KeyedStoreIC.

This continues refactoring already applied for LoadIC in r22035 (https://code.google.com/p/v8/source/detail?r=22035).

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/381633002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22328 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoStray assertion
rossberg@chromium.org [Thu, 10 Jul 2014 14:54:35 +0000 (14:54 +0000)]
Stray assertion

TBR=marja@chromium.org
BUG=

Review URL: https://codereview.chromium.org/381093002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22327 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoFix Windows compile error.
hpayer@chromium.org [Thu, 10 Jul 2014 14:50:32 +0000 (14:50 +0000)]
Fix Windows compile error.

BUG=
R=dslomov@chromium.org

Review URL: https://codereview.chromium.org/386603003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22326 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRemove GDB JIT hooks from the position recorder.
mstarzinger@chromium.org [Thu, 10 Jul 2014 14:41:06 +0000 (14:41 +0000)]
Remove GDB JIT hooks from the position recorder.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/385583003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22325 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoParser sync tests for `let` identifiers
rossberg@chromium.org [Thu, 10 Jul 2014 14:39:35 +0000 (14:39 +0000)]
Parser sync tests for `let` identifiers

R=marja@chromium.org
BUG=

Review URL: https://codereview.chromium.org/385613002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22324 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoMake `let` usable as an identifier in ES6 sloppy mode.
rossberg@chromium.org [Thu, 10 Jul 2014 14:06:37 +0000 (14:06 +0000)]
Make `let` usable as an identifier in ES6 sloppy mode.

All of our mjsunit suite now runs through with --harmony-scoping enabled, up to expected failures (tests checking syntax errors for const/function in strict mode).

R=marja@chromium.org, ulan@chromium.org
BUG=v8:2198
LOG=Y

Review URL: https://codereview.chromium.org/378303003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22323 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoFix Windows compile error.
hpayer@chromium.org [Thu, 10 Jul 2014 13:57:07 +0000 (13:57 +0000)]
Fix Windows compile error.

BUG=
R=dslomov@chromium.org

Review URL: https://codereview.chromium.org/384743002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22322 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoSwitch type of the-hole to be internal.
mstarzinger@chromium.org [Thu, 10 Jul 2014 12:37:08 +0000 (12:37 +0000)]
Switch type of the-hole to be internal.

R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/381063002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22321 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoImplement handling of arrow functions in the parser
marja@chromium.org [Thu, 10 Jul 2014 12:27:07 +0000 (12:27 +0000)]
Implement handling of arrow functions in the parser

Arrow functions are parsed from ParseAssignmentExpression(). Handling the
parameter list is done by letting ParseConditionalExpression() parse a comma
separated list of identifiers, and it returns a tree of BinaryOperation nodes
with VariableProxy leaves, or a single VariableProxy if there is only one
parameter. When the arrow token "=>" is found, the VariableProxy nodes are
passed to ParseArrowFunctionLiteral(), which will then skip parsing the
paramaeter list. This avoids having to rewind when the arrow is found and
restart parsing the parameter list.

Note that the empty parameter list "()" is handled directly in
ParsePrimaryExpression(): after is has consumed the opening parenthesis,
if a closing parenthesis follows, then the only valid input is an arrow
function. In this case, ParsePrimaryExpression() directly calls
ParseArrowFunctionLiteral(), to avoid needing to return a sentinel value
to signal the empty parameter list. Because it will consume the body of
the arrow function, ParseAssignmentExpression() will not see the arrow
"=>" token as next, and return the already-parser expression.

The implementation is done in ParserBase, so it was needed to do some
additions to ParserBase, ParserTraits and PreParserTraits. Some of the
glue code can be removed later on when more more functionality is moved
to ParserBase.

Additionally, this adds a runtime flag "harmony_arrow_functions"
(disabled by default); enabling "harmony" will enable it as well.

BUG=v8:2700
LOG=N
R=marja@chromium.org

Review URL: https://codereview.chromium.org/385553003

Patch from Adrián Pérez de Castro <aperez@igalia.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22320 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

10 years agoRevert "Precisely sweeping of scan-on-scavenge pages."
hpayer@chromium.org [Thu, 10 Jul 2014 12:22:01 +0000 (12:22 +0000)]
Revert "Precisely sweeping of scan-on-scavenge pages."

BUG=
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/387483002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22319 ce2b1a6d-e550-0410-aec6-3dcde31c8c00