Inline Array.shift() fast path instead of using a code stub.

TEST=mjsunit/array-shift,mjsunit/array-shift2,mjsunit/array-shift3
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/308793010

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21592 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix HPushArguments instruction.

Use the zone that is passed to New() and fix implementation of
HPushArguments::AddInput() to match HPhi::AddInput().

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/309763003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21591 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove unused debug API.

R=yurys@chromium.org

Review URL: https://codereview.chromium.org/303693010

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21590 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Simplify known successor block lowering.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/308593003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21589 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BuildNumberToString: Check for undefined keys in the cache

BUG=chromium:368114
LOG=y
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/302063004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21586 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Inlined optimized runtime functions: expose Runtime versions for direct testing, skip Hydrogen versions

R=dslomov@chromium.org

Review URL: https://codereview.chromium.org/302703004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21585 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Harden Runtime_FunctionSetPrototype, Isolate::PrintStack

BUG=chromium:377209
LOG=n
R=dslomov@chromium.org

Review URL: https://codereview.chromium.org/306543002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21584 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

HRor and HSar can deoptimize.

BUG=v8:3359
LOG=y
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/309483002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21583 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Don't push eax/rax before calling Runtime_Abort
Additionally delete the unused Throw(BailoutReason)

BUG=
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/308083006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21582 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix typo in test/fuzz-natives/testcfg.py

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/305963002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21581 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

MIPS: Refactor HType to get rid of various hacks.

Port r21578 (dcf13aa)

Original commit message:
- Move HType to it's own file.
- Add HType::HeapObject and some other useful types.
- Get rid of the broken and useless HType::NonPrimitive.
- Introduce HType::FromType() to convert from HeapType to HType.
- Also add unit tests for HType.
- Fix types in Crankshaft.

BUG=
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/301023004

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21580 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

X87: Refactor HType to get rid of various hacks.

port r21578 (dcf13aa)

Original commit message:
- Move HType to it's own file.
- Add HType::HeapObject and some other useful types.
- Get rid of the broken and useless HType::NonPrimitive.
- Introduce HType::FromType() to convert from HeapType to HType.
- Also add unit tests for HType.
- Fix types in Crankshaft.

BUG=

Review URL: https://codereview.chromium.org/307903002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21579 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Refactor HType to get rid of various hacks.

- Move HType to it's own file.
- Add HType::HeapObject and some other useful types.
- Get rid of the broken and useless HType::NonPrimitive.
- Introduce HType::FromType() to convert from HeapType to HType.
- Also add unit tests for HType.
- Fix types in Crankshaft.

R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/300893003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21578 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Simplify, speed-up correct-context ObjectObserve calls

The original patch which ensured that Object.observe did allocations in the correct context regressed performance about 12%. This patch gets back most of that (about 11%) by simply returning the correct function which is then directly callable from JS, rather than by making the call from the runtime function. A side-effect is that their implementation is shorter.

LOG=Y
BUG=NONE
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/307543008

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21575 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Allow microtasks to throw exceptions and handle them gracefully

If the embedder calls V8::TerminateExecution while we're running microtasks, bail out
and clear any pending microtasks.

All other exceptions are simply swallowed. No current Blink or V8 microtasks throw, this
just ensures something sane happens if another embedder decides to pass a throwing
microtask (or if ours unexpectedly throw due to, e.g., stack exhaustion).

BUG=371566
LOG=Y
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/294943009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21574 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

FastCloneShallowArrayStub should not be used it the length of the array is too big as it could eventually exceed the allowed size limit for manually folded allocations.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/300283003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21573 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Split Distributivity test

TBR=verwaest@chromium.org
BUG=

Review URL: https://codereview.chromium.org/298253008

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21572 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Check that all properties are processed in NameDictionary::CopyEnumKeysTo.

BUG=372579
LOG=N
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/306663002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21571 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add new benchmark suite runner.

TEST=./tools/run_benchmarks.py benchmarks/v8.json
TEST=cd tools/unittests; python -m unittest run_benchmarks_test

Does not support custom results processors yet. Will implement that in a future CL.

BUG=374740
LOG=n
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/293023006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21570 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Use non-copying sorting in test harness.

R=machenbach@chromium.org

Review URL: https://codereview.chromium.org/300873006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21569 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix deopt fuzzer after adding a no_sorting flag.

BUG=
TBR=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/305643007

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21568 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix lint complaints in objects.h

R=rossberg@chromium.org
LOG=N

Review URL: https://codereview.chromium.org/304533007

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21567 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add a dummy v8 target to BUILD.gn so we can depend on this from chrome

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

Review URL: https://codereview.chromium.org/300883003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21566 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

CallIC::PrintState() was incorrectly declared as V8_FINAL.

This broke the windows build.

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/304633004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21565 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Refactor CallICStub to use a different stub for each customization.

This gives us much more room to customize on different functions, by
using MajorKey to differentiate them.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/295383004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21564 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix processing of partially initialized JSWeakCollection.

R=hpayer@chromium.org
BUG=v8:2070
LOG=N

Review URL: https://codereview.chromium.org/300843009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21563 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Free memory in the mock array buffer allocator.

Turns out ASAN and valgrind notice if you leak 0 allocated bytes.

R=jkummerow@chromium.org, jochen@chromium.org

Review URL: https://codereview.chromium.org/306583006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21562 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add flag to test harness to stop sorting test cases.

R=jkummerow@chromium.org, machenbach@chromium.org

Review URL: https://codereview.chromium.org/307553003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21561 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Refactor after break target computation.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/301563004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21560 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Move LiveEdit-related code.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/300793002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21559 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Changing the attributes of a data property implemented with
ExecutableAccessorInfo turns the property into a field. Better
to keep it as a callback, and correctly deal with the changed
property attributes.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/262053011

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21558 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add Seq[One,Two]ByteString to runtime test generator.

R=dcarney@chromium.org, jkummerow@chromium.org

Review URL: https://codereview.chromium.org/303543006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21557 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Force inline runtime functions into its wrapper.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/307543009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21556 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Cleanup GetPropertyWithCallback / SetPropertyWithCallback API

BUG=
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/305513004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21555 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Drop bogus assert after r21552.

R=hpayer@chromium.org
BUG=v8:2070
LOG=N

Review URL: https://codereview.chromium.org/303493005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21554 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Repairs to FLAG_pretenure_call_new.

R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/307583002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21553 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Make incremental marker post-process JSWeakCollection.

R=hpayer@chromium.org
BUG=v8:2070
LOG=N

Review URL: https://codereview.chromium.org/301553003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21552 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

X87: Fix a typo in below CL

r21515 (6d0ec9a)
Convert ElementsKind into a BitField

BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/301563008

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21551 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

X87: Small changes in preparation for Hydrogen-generated KeyedLoadGeneric

Port r21535(8bb4c2b)

BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/302453011

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21550 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

X87: Convert String array index/length hash to BitField.

Port r21533 (736c779)

BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/298293007

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21549 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

X87: Merge Debugger and Debug.

port r21524 (af88824)

BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/298193004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21548 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

X87: Avoid HeapObject check in HStoreNamedField.

port r21509 (78a54b0)

Original commit message:
This way an HStoreNamedField instruction can never deoptimize
itself, which is another important step towards a working
store elimination.

BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/306483013

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21547 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Extract build configuration into a separate header and move it to the base lib

With this, change, atomicops, once, and lazy instance are no longer dependant
on v8 core. I'll move them in a follow-up change to the libbase as well.

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

Review URL: https://codereview.chromium.org/303463005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21546 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reuse ZoneList memory if possible.

Clear() is rather wasteful for ZoneLists, because it discards
all previously allocated memory, making it impossible to reuse
until the whole Zone is released. Better use Rewind(0) in this
case, which just resets the length to zero.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/302783007

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21545 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix path tracer test.

Path tracer is not compatible with --enable-slow-assert.

R=mstarzinger@chromium.org
BUG=v8:3350
LOG=N

Review URL: https://codereview.chromium.org/297193005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21542 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

X87: Skip write barriers in the fast case when setting up local context.

Port r21481 (5973b48)

Original commit message:
The FastNewContextStub always allocates in new space, so we don't
need to update the write barrier when copying the parameters to
the newly allocated context.

BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/307433004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21541 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

MIPS: Fix improper register use in DoMinMax.

This fixes failure on pidigit benchmark from web-shootout
after 840 digits are calculated.

TEST=
BUG=
R=plind44@gmail.com

Review URL: https://codereview.chromium.org/300853002

Patch from Dusan Milosavljevic <Dusan.Milosavljevic@rt-rk.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21540 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

MIPS: Convert String array index/length hash to BitField.

Port r21533 (736c779)

BUG=
R=plind44@gmail.com

Review URL: https://codereview.chromium.org/297233007

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21539 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix check for empty type

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/307453005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21538 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

MIPS: Use TempDoubleRegister.

Port r21517 (53d6507)

BUG=
R=plind44@gmail.com

Review URL: https://codereview.chromium.org/305563002

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21537 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reland 21529 - "Add a flag to d8 to invoke weak callbacks"

> This will send an idle notification and a low memory notification after
> each test.
>
> For some reason it's not enough to send a low memory notification alone.
>
> BUG=none
> R=yangguo@chromium.org
> LOG=n

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

Review URL: https://codereview.chromium.org/297213006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21536 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Small changes in preparation for Hydrogen-generated KeyedLoadGeneric

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/303583002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21535 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "Make 'name' property on functions configurable."

R=danno@google.com, danno@chromium.org

Review URL: https://codereview.chromium.org/297163009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21534 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Convert String array index/length hash to BitField

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/300023011

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21533 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove cpp hack to avoid link failures on Windows

TBR=yangguo@chromium.org
BUG=

Review URL: https://codereview.chromium.org/298133005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21532 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "Add a flag to d8 to invoke weak callbacks"

This reverts r21529.

TBR=jochen@chromium.org

Review URL: https://codereview.chromium.org/302543004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21531 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Drop unused includes from worker-thread.cc

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

Review URL: https://codereview.chromium.org/302463007

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21530 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add a flag to d8 to invoke weak callbacks

This will send an idle notification and a low memory notification after
each test.

For some reason it's not enough to send a low memory notification alone.

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

Review URL: https://codereview.chromium.org/306483006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21529 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Establish distributivity for type union & intersection

This requires introducing proper bounds on all leaf types, so that intersection between bitsets and these types can be accurately represented. Extending a union also becomes more involved.

(On the upside, the modified union/intersect algorithm would now allow support for proper variance for function types.)

Not sure if it is worth landing this. Distributivity isn't really a crucial property for our use cases. It seems fine if intersection is slightly lossy.

R=bmeurer@chromium.org, jarin@chromium.org
BUG=

Review URL: https://codereview.chromium.org/251753005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21528 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Replace STATIC_CHECK with STATIC_ASSERT.

It's just an alias, and STATIC_ASSERT is used way more often

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

Review URL: https://codereview.chromium.org/304553002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21527 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Skip PathTracer test for now.

TBR=mstarzinger@chromium.org
BUG=3350
LOG=N

Review URL: https://codereview.chromium.org/305583002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21526 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

MIPS: Avoid HeapObject check in HStoreNamedField.

Port r21509 (78a54b0)

Original commit message:
This way an HStoreNamedField instruction can never deoptimize
itself, which is another important step towards a working
store elimination.

BUG=
R=plind44@gmail.com

Review URL: https://codereview.chromium.org/297143007

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21525 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Merge Debugger and Debug.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/301633005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21524 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix PathTracer.

When tracing, we abuse the map for marking, thereby mutating it.
FixedTypedArrayBase::size() uses the object's map, which causes crash.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/300753002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21523 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Cache optimization status getter in mjsunit.js

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/300003007

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21522 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add defensive assert against a weak callback on an empty persistent.

(This shouldn't happen in the first place, as by definition that callback would never be called. However, the referenced bug sorta looks like this happened. If so, the CHECK should help us pinpoint the culprit.)

R=jochen@chromium.org
BUG=368095
LOG=N

Review URL: https://codereview.chromium.org/265823006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21521 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Do not break in native code (including non-builtin debugger code).

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/300773002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21520 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Replace some calls to MemCopy with small constant sizes with memcpy

MemCopy is only meant for variable size, large (>64bytes) copies,
otherwise, it's probably slower than memcpy due to the call
overhead and the compiler can't optimize it away.

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

Review URL: https://codereview.chromium.org/306453005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21519 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Do not (eagerly) trigger exception in mjsunit.js.

R=jochen@chromium.org

Review URL: https://codereview.chromium.org/301673002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21518 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

ARM: use TempDoubleRegister

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

Review URL: https://codereview.chromium.org/302453008

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21517 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove HType::HeapNumber from mutable heap numbers.

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/293353004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21516 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Convert ElementsKind into a BitField

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/306513002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21515 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix the "PersistentValueMap" memory leak reported here:
http://build.chromium.org/p/client.v8/builders/V8%20Linux64%20ASAN

The bug:
The code assumed that a weak Persistent whose weak callback is being
called would still be weak. That isn't true since the persistent is
un-weakened by the garbage collector before calling the weak callback. [1]

Specifically, PersistentValueMap would funnel all 'remove' actions
through its Release method, which uses PersistentBase::ClearWeak to
obtain the callback data. [2] For 'removes' caused by the weak callback,
ClearWeak always returns a NULL-pointer since by that time the weak
persistent was already un-weakend. The result was a memory leak in
the test, since the code to delete the weak callback data would
delete NULL.

The fix:
I explicity call Traits::DisposeCallbackData from the weak callback
with the data obtained from the v8::WeakCallbackData. To avoid invalid
calls to DisposeCallbackData, I also check whether this instance is
(still) weak before calling it. (That check could easily be elided
if it's expensive, for the price of having two 'remove' code paths.)

Severety:
Probably low. At least in Chromium, noone uses the API in a way to
trigger this; only the test does.

[1] https://code.google.com/p/chromium/codesearch#chromium/src/v8/src/global-handles.cc&q=global-handles.cc&sq=package:chromium&type=cs&l=231
[2] https://code.google.com/p/chromium/codesearch#chromium/src/v8/include/v8-util.h&sq=package:chromium&l=332-345

R=dcarney@chromium.org, dcarney

BUG=

Review URL: https://codereview.chromium.org/297193004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21514 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Escape % in SNPrintF call in pretenuring test.

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/298183002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21513 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reland 21502 - "Move OS::MemCopy and OS::MemMove out of platform to utils"

Verified that arm builds locally.

BUG=none
TBR=jkummerow@chromium.org
LOG=n

Review URL: https://codereview.chromium.org/306473004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21512 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove global pretenuring mode.

BUG=
R=bmeurer@chromium.org, mvstanton@chromium.org

Review URL: https://codereview.chromium.org/304493002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21511 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix d8's interactive shell.

R=danno@chromium.org

Review URL: https://codereview.chromium.org/305553002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21510 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Avoid HeapObject check in HStoreNamedField.

This way an HStoreNamedField instruction can never deoptimize
itself, which is another important step towards a working
store elimination.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/299373005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21509 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix arm64 gc stress issue.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/306483002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21506 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Relax register constraints for LMathSqrt.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/299423005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21505 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert 21502 - "Move OS::MemCopy and OS::MemMove out of platform to utils"

TBR=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/297303004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21504 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Buildfix for arm

TBR=jkummerow@chromium.org
LOG=n
BUG=none

Review URL: https://codereview.chromium.org/299353006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21503 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Move OS::MemCopy and OS::MemMove out of platform to utils

Since both are jitted on some platforms and depend on codegen, they
don't belong to the platform abstraction. At the same time, I can't put
them to codegen.h, as this would introduce cyclic dependencies.

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

Review URL: https://codereview.chromium.org/302563004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21502 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Move NumberOfProcessorsOnline from CPU to OS

It's really more an OS-level information, and this way the default
platform doesn't depend on CPU-level details

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

Review URL: https://codereview.chromium.org/300713002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21501 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

MIPS: Skip write barriers in the fast case when setting up local context.

Port r21481 (5973b48)

Original commit message:
The FastNewContextStub always allocates in new space, so we don't
need to update the write barrier when copying the parameters to
the newly allocated context.

BUG=
R=plind44@gmail.com

Review URL: https://codereview.chromium.org/302633002

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21500 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reland "Customized support for feedback on calls to Array." and follow-up fixes.

Comparing one CallIC::State to another was not done correctly, leading to a failure to patch a CallIC when transitioning from monomorphic Array to megamorphic.

BUG=chromium:377198,chromium:377290
LOG=Y
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/305493003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21499 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Send idle notification on ASAN builder for proper tear down.

BUG=
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/300003005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21498 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

More nuisance

TBR=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/297323004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21497 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

D8 should send idle notification even after the last run.

This, in combination with changes to the bot, should weed out false negatives.

R=marja@chromium.org

Review URL: https://codereview.chromium.org/293393003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21496 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Shut up VS

TBR=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/298073005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21495 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Various extensions to types

R=bmeurer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/296213005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21494 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove high promotion mode.

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/296413004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21493 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Make 'name' property on functions configurable.

R=rossberg@chromium.org
BUG=v8:3333
LOG=N

Review URL: https://codereview.chromium.org/296413003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21492 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Actually fix the memory leak in debugger wrt promises.

In the fuzz test the debugger is not actually loaded...

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/304453003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21491 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reland 21482 - "Merge v8globals.h and globals.h"

> BUG=none
> R=mstarzinger@chromium.org
> LOG=n
>
> Review URL: https://codereview.chromium.org/293363006

BUG=none
TBR=mstarzinger@chromium.org
LOG=n

Review URL: https://codereview.chromium.org/295373003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21490 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix memory leak caused by fuzzing.

%DebugPromiseHandlePrologue expects to be followed up by %..Epilogue,
otherwise we get a leak.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/301553002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21489 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix mem leaks in tests & ScriptCompiler::CompileUnbound.

- Leak in test-parsing/DontRegressPreParserDataSizes
- Leak in test-api/EventLogging
- Leak in ScriptCompiler::CompileUnbound which won't happen during normal
operation, but exposed by test-apöi/CompiledWithInvalidCachedData.

R=yangguo@chromium.org
BUG=

Review URL: https://codereview.chromium.org/302603002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21488 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Always fully qualify the namespace to use in v8.h

Also, remove some unused methods.

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

Review URL: https://codereview.chromium.org/297223006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21487 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "Customized support for feedback on calls to Array." and follow-up fixes.

This reverts r21429, r21434, r21435, r21440, r21445.

BUG=chromium:377198
LOG=y
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/300693002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21484 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "Merge v8globals.h and globals.h"

Because of tree redness.

TBR=jochen@chromium.org

Review URL: https://codereview.chromium.org/306443002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21483 ce2b1a6d-e550-0410-aec6-3dcde31c8c00