MIPS: Remove the obsolete OverwriteMode optimization.

Port 634b0f20f4bbdad305c87fd8e196f4c05efbde1f

BUG=

Review URL: https://codereview.chromium.org/900253002

Cr-Commit-Position: refs/heads/master@{#26467}

Protect against uninitialized lexical variables at top-level.

R=rossberg@chromium.org
BUG=chromium:452510
LOG=Y

Review URL: https://codereview.chromium.org/893993007

Cr-Commit-Position: refs/heads/master@{#26466}

Fix Code::VerifyEmbeddedObjects.

BUG=

Review URL: https://codereview.chromium.org/905473003

Cr-Commit-Position: refs/heads/master@{#26465}

CallIC MISS handler doesn't need receiver.

Looking up and passing this argument is unnecessary.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/898013003

Cr-Commit-Position: refs/heads/master@{#26464}

Externalize deoptimization reasons.

1) The hardcoded strings were converted into DeoptReason enum.

2) Deopt comment were converted into a pair location and deopt reason entries so
the deopt reason tracking mode would less affect the size of the RelocInfo table and heap.

3) DeoptReason entry in RelocInfo reuses kCommentTag value and generates short entry in RelocInfo table.

BUG=452067
LOG=n

Committed: https://crrev.com/c49820e45b57f128a98690940875c049f612dde6
Cr-Commit-Position: refs/heads/master@{#26434}

Committed: https://crrev.com/ec42e002da03adb2db968dd5b7453341ddc59a5c
Cr-Commit-Position: refs/heads/master@{#26448}

Review URL: https://codereview.chromium.org/874323003

Cr-Commit-Position: refs/heads/master@{#26463}

Tiny simplification of FieldIndex::ForInObjectOffset.

Review URL: https://codereview.chromium.org/884163004

Cr-Commit-Position: refs/heads/master@{#26462}

Split --harmony-unicode and --harmony-unicode-regexps.

This way we can ship \u{..} escapes in strings / identifiers before shipping /u
regexps.

BUG=

Review URL: https://codereview.chromium.org/903703002

Cr-Commit-Position: refs/heads/master@{#26461}

Add strong mode.

It doesn't do anything for now, but it implies strict mode. Added tests to
test-parsing.cc to test that.

BUG=

Review URL: https://codereview.chromium.org/898983002

Cr-Commit-Position: refs/heads/master@{#26460}

[turbofan]: Improved source position information

Make sure the initial graph is fully populated with source position information and automatically propagate that information down through newly allocated nodes during reduction passes in the most unobtrusive way that's currently possible.

Review URL: https://codereview.chromium.org/897883002

Cr-Commit-Position: refs/heads/master@{#26459}

Fix try-finally for dead AST-branches in TurboFan.

R=jarin@chromium.org
TEST=mjsunit/regress/regress-crbug-455644
BUG=chromium:455644
LOG=N

Review URL: https://codereview.chromium.org/880443004

Cr-Commit-Position: refs/heads/master@{#26458}

Revert of Externalize deoptimization reasons. (patchset #7 id:120001 of https://codereview.chromium.org/874323003/)

Reason for revert:
This CL breaks RelocInfo::INTERNAL_REFERENCE (and maybe others that come after DEOPT_REASON), which is currently (mostly) unused on tip-of-tree, but will be used for jump tables soon(ish) and should therefore work. It seems to be a problem with implicitly assumptions about the number of reloc info modes. Needs further investigation.

Original issue's description:
> Externalize deoptimization reasons.
>
> 1) The hardcoded strings were converted into DeoptReason enum.
>
> 2) Deopt comment were converted into a pair location and deopt reason entries so
> the deopt reason tracking mode would less affect the size of the RelocInfo table and heap.
>
> 3) DeoptReason entry in RelocInfo reuses kCommentTag value and generates short entry in RelocInfo table.
>
> BUG=452067
> LOG=n
>
> Committed: https://crrev.com/c49820e45b57f128a98690940875c049f612dde6
> Cr-Commit-Position: refs/heads/master@{#26434}
>
> Committed: https://crrev.com/ec42e002da03adb2db968dd5b7453341ddc59a5c
> Cr-Commit-Position: refs/heads/master@{#26448}

TBR=alph@chromium.org,mstarzinger@chromium.org,svenpanne@chromium.org,yurys@chromium.org,loislo@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=452067

Review URL: https://codereview.chromium.org/900223002

Cr-Commit-Position: refs/heads/master@{#26457}

Grokdump workaround for different layout of *_LIST structures written to minidumps on Mac.

Review URL: https://codereview.chromium.org/902853002

Cr-Commit-Position: refs/heads/master@{#26456}

Remove flush_monomorphic_ics flag.

Since v8:3629 is fixed, ICs and handlers do not retain objects.

Review URL: https://codereview.chromium.org/880253007

Cr-Commit-Position: refs/heads/master@{#26455}

Remove the obsolete OverwriteMode optimization.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/900193002

Cr-Commit-Position: refs/heads/master@{#26454}

Fix HConstant(double, ...) constructor

It must always populate int32_value_, even if that's lossy, because other code (specifically, constant folding for truncating operations) relies on it.

BUG=v8:3865
LOG=y

Review URL: https://codereview.chromium.org/897263002

Cr-Commit-Position: refs/heads/master@{#26453}

Correctly report some internal OOM conditions.

R=hpayer@chromium.org
BUG=chromium:454615
LOG=N

Review URL: https://codereview.chromium.org/900793004

Cr-Commit-Position: refs/heads/master@{#26452}

Add NativeWeakMap to v8.h

A new map wich references its keys weakly is added to v8.h. Internally it uses the same storage as JSWeakMap but doesn't depend on the JavaScript part of WeakMap implementation in weak-collection.js, hence it can be instantiated without entering any context.

BUG=chromium:437416
LOG=Y

Review URL: https://codereview.chromium.org/900123003

Cr-Commit-Position: refs/heads/master@{#26451}

Add a flag to track detached contexts.

When embedder detaches the global objects, its context must be garbage
collected eventually.

BUG=

Review URL: https://codereview.chromium.org/898663005

Cr-Commit-Position: refs/heads/master@{#26450}

Scanner / Unicode decoding: use size_t instead of unsigned.

size_t is the correct data type for this purpose. Our APIs (in particular
ExternalSourceStream::GetMoreData) are already using it, and there were some
static_casts to convert between them.

This CL doesn't intend to fix all of V8, just the minimal sense-making part
around scanner character streams.

BUG=

Review URL: https://codereview.chromium.org/864273005

Cr-Commit-Position: refs/heads/master@{#26449}

Externalize deoptimization reasons.

1) The hardcoded strings were converted into DeoptReason enum.

2) Deopt comment were converted into a pair location and deopt reason entries so
the deopt reason tracking mode would less affect the size of the RelocInfo table and heap.

3) DeoptReason entry in RelocInfo reuses kCommentTag value and generates short entry in RelocInfo table.

BUG=452067
LOG=n

Committed: https://crrev.com/c49820e45b57f128a98690940875c049f612dde6
Cr-Commit-Position: refs/heads/master@{#26434}

Review URL: https://codereview.chromium.org/874323003

Cr-Commit-Position: refs/heads/master@{#26448}

Revert "[turbofan] Make Factory::NewNumber() always return the minus_zero_value."

This reverts commit f578d35ba8d6798afe45cd60f4f06bfa7e2ae6be because
(immutable) heap numbers are surprisingly mutable in V8. Someone else
might want to cleanup the mess, otherwise we'll just keep it this way.

TBR=jkummerow@chromium.org
BUG=chromium:454894
LOG=y

Review URL: https://codereview.chromium.org/898973003

Cr-Commit-Position: refs/heads/master@{#26447}

Update V8 DEPS.

Rolling v8/buildtools to da0df3fdac6036e862addb1155a2d6c11b6c18d5

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/900793003

Cr-Commit-Position: refs/heads/master@{#26446}

x87: Use signaling NaN for holes in fixed double arrays.

Port 9eace97bbaab72962c0fda62e5f9011a10604d0d

BUG=
R=chunyang.dai@intel.com

Review URL: https://codereview.chromium.org/895473002

Cr-Commit-Position: refs/heads/master@{#26445}

Revert "MIPS:[turbofan] Improve unordered comparisons for boolean materialization."

Reason for revert:
Fails mjsunit/asm/embenchen/box2d

BUG=

Review URL: https://codereview.chromium.org/893993003

Cr-Commit-Position: refs/heads/master@{#26444}

MIPS: Calls to the CallIC miss handler looked up the receiver incorrectly.

Port 13b97b6214d8448c568d2e547e4a04c5fb468f5d

Original commit message:
On platforms that don't push the return address on the stack at time of call, there was an error.

BUG=

Review URL: https://codereview.chromium.org/864713003

Cr-Commit-Position: refs/heads/master@{#26443}

templates: Don't check IsLineTerminator() if character is negative

BUG=455212
LOG=N
R=arv@chromium.org

Review URL: https://codereview.chromium.org/902703002

Cr-Commit-Position: refs/heads/master@{#26442}

Fix issue with multiple properties and emit store.

We used to have a parse error for conflicting property keys. This check
was removed to match ES6 (SpiderMonkey & Chakra already made this change).
Since this check was removed we ended up with a few new cases when
generating code. For example, accessors always generated code even if
those should have been shadowed by a data property.

BUG=v8:3856
LOG=Y
R=adamk, dslomov@chromium.org

Review URL: https://codereview.chromium.org/895693002

Cr-Commit-Position: refs/heads/master@{#26441}

Add macro for getting a PropertyAttributes for the runtime functions

Motivation: Cleanup

BUG=None
R=adamk
LOG=N

Review URL: https://codereview.chromium.org/897633003

Cr-Commit-Position: refs/heads/master@{#26440}

Calls to the CallIC miss handler looked up the receiver incorrectly.

On platforms that don't push the return address on the stack at time of call, there was an error.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/894753006

Cr-Commit-Position: refs/heads/master@{#26439}

MIPS: Fix assertion in full codegen for holed 'this'.

Port 275e088abec213aa0be3eef748926e4be2c5c79f

BUG=chromium:455141
LOG=N

Review URL: https://codereview.chromium.org/876083006

Cr-Commit-Position: refs/heads/master@{#26438}

Class methods should be non enumerable

Class methods and accessors (both prototype and static) should be
non enumerable.

BUG=v8:3330
LOG=Y
R=dslomov@chromium.org

Review URL: https://codereview.chromium.org/896643003

Cr-Commit-Position: refs/heads/master@{#26437}

Make functions contexts in optimized code weak.

BUG=

Review URL: https://codereview.chromium.org/882913006

Cr-Commit-Position: refs/heads/master@{#26436}

Revert of Externalize deoptimization reasons. (patchset #6 id:100001 of https://codereview.chromium.org/874323003/)

Reason for revert:
it broke the build

Original issue's description:
> Externalize deoptimization reasons.
>
> 1) The hardcoded strings were converted into DeoptReason enum.
>
> 2) Deopt comment were converted into a pair location and deopt reason entries so
> the deopt reason tracking mode would less affect the size of the RelocInfo table and heap.
>
> 3) DeoptReason entry in RelocInfo reuses kCommentTag value and generates short entry in RelocInfo table.
>
> BUG=452067
> LOG=n
>
> Committed: https://crrev.com/c49820e45b57f128a98690940875c049f612dde6
> Cr-Commit-Position: refs/heads/master@{#26434}

TBR=alph@chromium.org,mstarzinger@chromium.org,svenpanne@chromium.org,yurys@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=452067

Review URL: https://codereview.chromium.org/892843007

Cr-Commit-Position: refs/heads/master@{#26435}

Externalize deoptimization reasons.

1) The hardcoded strings were converted into DeoptReason enum.

2) Deopt comment were converted into a pair location and deopt reason entries so
the deopt reason tracking mode would less affect the size of the RelocInfo table and heap.

3) DeoptReason entry in RelocInfo reuses kCommentTag value and generates short entry in RelocInfo table.

BUG=452067
LOG=n

Review URL: https://codereview.chromium.org/874323003

Cr-Commit-Position: refs/heads/master@{#26434}

test-parsing.cc: Test fix.

The syntax of this test snippet was so wrong ("function function") that it
wasn't testing what it was supposed to test.

R=rossberg@chromium.org
BUG=

Review URL: https://codereview.chromium.org/896043003

Cr-Commit-Position: refs/heads/master@{#26433}

MIPS: Retry "Use a WeakCell in the CallIC type vector."

Port 6fc97a19653e03a891d3cd397069d5a10b9ab325

Original commit message:
The first try failed because I needed to make a better distinction
between clearing ICs according to policy at GC time or unconditional
clearing (say, via %ClearFunctionTypeFeedback).

It was also blocked by an issue in super constructor calls.
This fix (https://codereview.chromium.org/892113002/) needs to land
before checking in this CL.

BUG=

Review URL: https://codereview.chromium.org/896223002

Cr-Commit-Position: refs/heads/master@{#26432}

MIPS: Introduce LanguageMode, drop StrictMode.

Port c7851da4aefb644ab198ead1fa284932fd424797

Original commit message:
This enables adding more language modes in the future.

For maximum flexibility, LanguageMode is a bitmask, so we're not restricted to
use a sequence of language modes which are progressively stricter, but we can
express the language mode as combination of features.

For now, LanguageMode can only be "sloppy" or "strict", and there are
STATIC_ASSERTS in places which need to change when more modes are added.

LanguageMode is a bit like the old LanguageMode when "extended" mode was still
around (see https://codereview.chromium.org/8417035 and
https://codereview.chromium.org/181543002 ) except that it's transmitted through
all the layers (there's no StrictModeFlag).

BUG=

Review URL: https://codereview.chromium.org/890413003

Cr-Commit-Position: refs/heads/master@{#26431}

Revert of Add WeakKeyMap to v8.h (patchset #2 id:20001 of https://codereview.chromium.org/891473005/)

Reason for revert:
Revert this patch due to shared win build compilation failure

http://build.chromium.org/p/client.v8/builders/V8%20Win32%20-%20nosnap%20-%20shared/builds/5030/steps/compile/logs/stdio

Original issue's description:
> Add WeakKeyMap to v8.h
>
> A new map wich references its keys weakly is added to v8.h. Internally it uses the same storage as JSWeakMap but doesn't depend on the JavaScript part of WeakMap implementation in weak-collection.js, hence it can be instantiated without entering any context.
>
> BUG=chromium:437416
> LOG=Y
>
> Committed: https://crrev.com/ee7ed39ac8327124e74dd7ad5f1de0dede988cb7
> Cr-Commit-Position: refs/heads/master@{#26425}

TBR=jochen@chromium.org,mstarzinger@chromium.org,rossberg@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:437416

Review URL: https://codereview.chromium.org/901663002

Cr-Commit-Position: refs/heads/master@{#26430}

Revert of Simplify WeakMap and WeakSet tests slightly. (patchset #1 id:1 of https://codereview.chromium.org/903463002/)

Reason for revert:
Revert this patch as the original CL has to be reverted to due to shared win build failure

http://build.chromium.org/p/client.v8/builders/V8%20Win32%20-%20nosnap%20-%20shared/builds/5030/steps/compile/logs/stdio

Original issue's description:
> Simplify WeakMap and WeakSet tests slightly.
>
> R=rossberg@chromium.org
> TEST=cctest/test-weakmaps,cctest/test-weaksets
>
> Committed: https://crrev.com/a90d7a871da19d33e93c8779186408c8d1615cac
> Cr-Commit-Position: refs/heads/master@{#26428}

TBR=rossberg@chromium.org,mstarzinger@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/899833003

Cr-Commit-Position: refs/heads/master@{#26429}

Simplify WeakMap and WeakSet tests slightly.

R=rossberg@chromium.org
TEST=cctest/test-weakmaps,cctest/test-weaksets

Review URL: https://codereview.chromium.org/903463002

Cr-Commit-Position: refs/heads/master@{#26428}

try to fix windows after r26424

BUG=

Review URL: https://codereview.chromium.org/901603002

Cr-Commit-Position: refs/heads/master@{#26427}

Move the contents of api-natives.js to c++

R=verwaest@chromium.org

BUG=

Review URL: https://codereview.chromium.org/895053002

Cr-Commit-Position: refs/heads/master@{#26426}

Add WeakKeyMap to v8.h

A new map wich references its keys weakly is added to v8.h. Internally it uses the same storage as JSWeakMap but doesn't depend on the JavaScript part of WeakMap implementation in weak-collection.js, hence it can be instantiated without entering any context.

BUG=chromium:437416
LOG=Y

Review URL: https://codereview.chromium.org/891473005

Cr-Commit-Position: refs/heads/master@{#26425}

[turbofan] Don't allocate UnallocatedOperands in Zone memory during instruction selection

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/889843003

Cr-Commit-Position: refs/heads/master@{#26424}

Fix assertion in full codegen for holed 'this'.

R=rossberg@chromium.org
BUG=chromium:455141
LOG=N

Review URL: https://codereview.chromium.org/902563002

Cr-Commit-Position: refs/heads/master@{#26423}

MIPS:[turbofan] Improve unordered comparisons for boolean materialization.

TEST=
BUG=
R=paul.lind@imgtec.com

Review URL: https://codereview.chromium.org/894883004

Cr-Commit-Position: refs/heads/master@{#26422}

[turbofan] Make ContextScope a proper encapsulation of the current context.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/896983002

Cr-Commit-Position: refs/heads/master@{#26421}

Retry "Use a WeakCell in the CallIC type vector."

The first try failed because I needed to make a better distinction
between clearing ICs according to policy at GC time or unconditional
clearing (say, via %ClearFunctionTypeFeedback).

It was also blocked by an issue in super constructor calls.
This fix (https://codereview.chromium.org/892113002/) needs to land
before checking in this CL.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/866493003

Cr-Commit-Position: refs/heads/master@{#26420}

Introduce LanguageMode, drop StrictMode.

This enables adding more language modes in the future.

For maximum flexibility, LanguageMode is a bitmask, so we're not restricted to
use a sequence of language modes which are progressively stricter, but we can
express the language mode as combination of features.

For now, LanguageMode can only be "sloppy" or "strict", and there are
STATIC_ASSERTS in places which need to change when more modes are added.

LanguageMode is a bit like the old LanguageMode when "extended" mode was still
around (see https://codereview.chromium.org/8417035 and
https://codereview.chromium.org/181543002 ) except that it's transmitted through
all the layers (there's no StrictModeFlag).

BUG=

Review URL: https://codereview.chromium.org/894683003

Cr-Commit-Position: refs/heads/master@{#26419}

MIPS64: Super Constructor Calls need to use a vector slot, not an ic slot.

Port 64f81870ee0bc51a32620be8face35e7cbd0da92

BUG=

Review URL: https://codereview.chromium.org/885283008

Cr-Commit-Position: refs/heads/master@{#26418}

MIPS: new classes: special construct stub for derived classs and TDZ for `this`.

Port 6f97a4948f1495a9b9d60b7f495d0e0efeaa1006

BUG=

Review URL: https://codereview.chromium.org/895243002

Cr-Commit-Position: refs/heads/master@{#26417}

Fix test-api/ObjectProtoToStringES6

Fixes tree closure caused by r26415

BUG=v8:3502
R=arv@chromium.org
LOG=N

Review URL: https://codereview.chromium.org/869773005

Cr-Commit-Position: refs/heads/master@{#26416}

Update harmony ObjectProtoToString() to 2/2/2015 spec

Applies to API implementation of O.p.toString

- Removes special handling of non-string @@toStringTag values (use builtinTag)
- Removes special handling of @@toStringTags which match [[Class]] names (remove ~ prefix)
- Add tests for non-string @@toStringTag values in test-api.cc

BUG=v8:3502
R=arv@chromium.org
LOG=N

Review URL: https://codereview.chromium.org/892393003

Cr-Commit-Position: refs/heads/master@{#26415}

Super Constructor Calls need to use a vector slot, not an ic slot.

The Ast Call node is accustomed to using a vector IC slot for the
cases when it uses a CallIC. The super constructor work alters this
somewhat by using a CallConstructStub instead, however the
CallConstructStub expects a vector slot and not a vector ic slot.
This distinction needs to be maintained because slots and ic slots
have different clearing strategies and are handled differently.

R=dslomov@chromium.org
BUG=
LOG=N

Review URL: https://codereview.chromium.org/892113002

Cr-Commit-Position: refs/heads/master@{#26414}

Infer HConstant::NotInNewSpace only if the supplied handle is null.

BUG=chromium:449291
LOG=n
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/898753003

Cr-Commit-Position: refs/heads/master@{#26413}

Verify that code stubs and full code do not have pointers that can retain
context.

BUG=v8:3629
LOG=N

Review URL: https://codereview.chromium.org/879273004

Cr-Commit-Position: refs/heads/master@{#26412}

Update harmony Object.prototype.toString to 2/2/2015 spec

- Removes special handling of non-string @@toStringTag values (use builtinTag)
- Removes special handling of @@toStringTags which match [[Class]] names (remove ~ prefix)

BUG=v8:3502
R=arv@chromium.org
LOG=N

Review URL: https://codereview.chromium.org/895633004

Cr-Commit-Position: refs/heads/master@{#26411}

Check the slots being adding to the SlotsBuffer.
Related to crbug/454297.

Review URL: https://codereview.chromium.org/897523003

Cr-Commit-Position: refs/heads/master@{#26410}

new classes: special construct stub for derived classs and TDZ for `this`.

R=arv@chromium.org,rossberg@chromium.org
BUG=v8:3834
LOG=N

Review URL: https://codereview.chromium.org/867153003

Cr-Commit-Position: refs/heads/master@{#26409}

Whitespace change to trigger bots.

BUG=

Review URL: https://codereview.chromium.org/898783002

Cr-Commit-Position: refs/heads/master@{#26408}

Partial revert of crrev.com/900553003, to unbreak arm64 build.

R=verwaest@chromium.org
BUG=454625
NOTREECHECKS=true

Review URL: https://codereview.chromium.org/892833003

Cr-Commit-Position: refs/heads/master@{#26407}

Turn throws into basic block terminators.

R=bmeurer@chromium.org
TEST=cctest/test-run-jsexceptions/Throw

Review URL: https://codereview.chromium.org/896783002

Cr-Commit-Position: refs/heads/master@{#26406}

Reset inlining limits due to overly long compilation times in Speedometer, Dart2JS

BUG=454625
LOG=y

Review URL: https://codereview.chromium.org/900553003

Cr-Commit-Position: refs/heads/master@{#26405}

[turbofan] Split pure nodes in the scheduler if beneficial.

If a (pure) node has two or more uses, but there exists a path from the
common dominator of these uses to end, which does not contain a use,
then we split the node such that no unnecessary computation takes place.
Note however, that this only applies if the node cannot be hoisted out
of a loop.

BUG=v8:3864
LOG=n
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/899433005

Cr-Commit-Position: refs/heads/master@{#26404}

[turbofan] Parameter nodes are not idempotent.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/896463004

Cr-Commit-Position: refs/heads/master@{#26403}

Revert of Add WeakMap to v8.h (patchset #3 id:40001 of https://codereview.chromium.org/886473005/)

Reason for revert:
Broke compilation on component build http://build.chromium.org/p/client.v8/builders/V8%20Win32%20-%20nosnap%20-%20shared/builds/5007/steps/compile/logs/stdio

Original issue's description:
> Add WeakMap to v8.h
>
> A new map wich references its keys weakly is added to v8.h. Internally it uses the same storage as JSWeakMap but doesn't depend on the JavaScript part of WeakMap implementation in weak-collection.js, hence it can be instantiated without entering any context.
>
> BUG=chromium:437416
> LOG=Y
>
> Committed: https://crrev.com/37d4c57630636f21e3add8d3d1c7c978ff5fc8e0
> Cr-Commit-Position: refs/heads/master@{#26401}

TBR=jochen@chromium.org,mstarzinger@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:437416

Review URL: https://codereview.chromium.org/898763002

Cr-Commit-Position: refs/heads/master@{#26402}

Add WeakMap to v8.h

A new map wich references its keys weakly is added to v8.h. Internally it uses the same storage as JSWeakMap but doesn't depend on the JavaScript part of WeakMap implementation in weak-collection.js, hence it can be instantiated without entering any context.

BUG=chromium:437416
LOG=Y

Review URL: https://codereview.chromium.org/886473005

Cr-Commit-Position: refs/heads/master@{#26401}

Move nested classes of AstGraphBuilder into cc file.

R=titzer@chromium.org

Review URL: https://codereview.chromium.org/896653008

Cr-Commit-Position: refs/heads/master@{#26400}

[turbofan]: Small visualizer cleanup and fix for string handling

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/886313006

Cr-Commit-Position: refs/heads/master@{#26399}

First stab at try-catch and try-finally in TurboFan.

R=titzer@chromium.org,jarin@chromium.org
TEST=cctest/test-run-jsexceptions

Review URL: https://codereview.chromium.org/873423004

Cr-Commit-Position: refs/heads/master@{#26398}

Whitespace change to test goma.

Cr-Commit-Position: refs/heads/master@{#26397}

Use weak cell to embed property cell in StoreGlobal.

BUG=v8:3629
LOG=N

Review URL: https://codereview.chromium.org/898723002

Cr-Commit-Position: refs/heads/master@{#26396}

follow up named interceptor miss with api callback getter

BUG=

Review URL: https://codereview.chromium.org/885763004

Cr-Commit-Position: refs/heads/master@{#26395}

Avoid issuing write barriers for unboxed double fields in Heap::CopyJSObject().

Review URL: https://codereview.chromium.org/880043003

Cr-Commit-Position: refs/heads/master@{#26394}

[V8] Added line, column and script symbols for SyntaxError

For exception in promise we generate v8::Message API object from exception object. And in cases of Syntax or Reference Error we don't have enough information in exception object - we can't restore Error location from top stack frame.
In this patch three aditional private fields introduced for exception object. In case of Syntax Error we store line, column and script on Exception object and receive this information when restoring message.

BUG=443140
LOG=Y
R=yurys@chromium.org

Review URL: https://codereview.chromium.org/885043002

Cr-Commit-Position: refs/heads/master@{#26393}

MIPS: Skip mozilla/js1_5/extensions/regress-355497 in simulator.

BUG=

Review URL: https://codereview.chromium.org/894123002

Cr-Commit-Position: refs/heads/master@{#26392}

Compute the same hash for all NaN values.

Both SameValue and SameValueZero consider different NaNs equal, so we
better assign the same hash value to all NaNs.

BUG=v8:3859
LOG=y
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/897593002

Cr-Commit-Position: refs/heads/master@{#26391}

Clear pending exception on stack overflow in the parser

BUG=450960
LOG=n

Review URL: https://codereview.chromium.org/858213003

Cr-Commit-Position: refs/heads/master@{#26390}

tools: fix postmortem generator

BUG=
R=danno

Review URL: https://codereview.chromium.org/892223002

Cr-Commit-Position: refs/heads/master@{#26389}

MIPS: Use weak cell to embed known map in CompareIC

Port 246a749a6f7e7e3a025bb032bc0c643656c9b08f

BUG=v8:3663
LOG=N

Review URL: https://codereview.chromium.org/881533004

Cr-Commit-Position: refs/heads/master@{#26388}

[turbofan] Put StructuredGraphBuilder out of its misery and merge its remnants back into the AstGraphBuilder.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/894073002

Cr-Commit-Position: refs/heads/master@{#26387}

Use weak cell to embed known map in CompareIC

BUG=v8:3663
LOG=N

Review URL: https://codereview.chromium.org/879333005

Cr-Commit-Position: refs/heads/master@{#26386}

Fetch global object from jsproxy prototype instead of embedding it in StoreGlobal.

BUG=v8:3629
LOG=N

Review URL: https://codereview.chromium.org/892213003

Cr-Commit-Position: refs/heads/master@{#26385}

MIPS64: Megamorphic KeyedLoadIC needs special handling for vector ics.

Port 16843e239d95f5e8eaaa06d2a0a6e371298fe11f

Original commit message:
When --vector-ics is true, we still tail-call to the hand-written
megamorphic KeyedLoadIC (formerly "generic"). Now that this code uses
the megamorphic cache, it needs to deal properly with the vector and
slot registers. Achieve this with a sentinel vectors/slot combo.

BUG=

Review URL: https://codereview.chromium.org/894053002

Cr-Commit-Position: refs/heads/master@{#26384}

Whitespace change to test goma activation.

Cr-Commit-Position: refs/heads/master@{#26383}

Initialize CompilationInfo::osr_expr_stack_height correctly.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/889383002

Cr-Commit-Position: refs/heads/master@{#26382}

Megamorphic KeyedLoadIC needs special handling for vector ics.

When --vector-ics is true, we still tail-call to the hand-written
megamorphic KeyedLoadIC (formerly "generic"). Now that this code uses
the megamorphic cache, it needs to deal properly with the vector and
slot registers. Achieve this with a sentinel vectors/slot combo.

R=dcarney@chromium.org
BUG=

Review URL: https://codereview.chromium.org/892173002

Cr-Commit-Position: refs/heads/master@{#26381}

Check global object behind global proxy for extensibility

BUG=454091
LOG=Y

Review URL: https://codereview.chromium.org/895573002

Cr-Commit-Position: refs/heads/master@{#26380}

Use weak cells in CheckPropertyCell.

BUG=v8:3629
LOG=N

Review URL: https://codereview.chromium.org/815143002

Cr-Commit-Position: refs/heads/master@{#26379}

Speed up tests for OSR of for-in and for-of loops.

R=marja@chromium.org
BUG=

Review URL: https://codereview.chromium.org/889293003

Cr-Commit-Position: refs/heads/master@{#26378}

[turbofan] push virtual register field down to InstructionOperand

BUG=

Review URL: https://codereview.chromium.org/893913004

Cr-Commit-Position: refs/heads/master@{#26377}

[turbofan] Fix OSR into functions where the expression stack is not empty.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/890903002

Cr-Commit-Position: refs/heads/master@{#26376}

shell.cc race condition

BUG=v8:3850
LOG=N

Review URL: https://codereview.chromium.org/892083002

Cr-Commit-Position: refs/heads/master@{#26375}

MIPS64: Fix bugs after hole-nan value change.

TEST=mjsunit/regress/regress-undefined-nan{2,3}, mjsunit/elements-transition, mjsunit/elide-double-hole-check-*
BUG=

Review URL: https://codereview.chromium.org/892853002

Cr-Commit-Position: refs/heads/master@{#26374}

Whitespace change to trigger bots.

Cr-Commit-Position: refs/heads/master@{#26373}

Revert of Fix --max_old_space_size=4096 integer overflow. (patchset #1 id:1 of https://codereview.chromium.org/890563003/)

Reason for revert:
Win64 compile failure.

Original issue's description:
> Fix --max_old_space_size=4096 integer overflow.
>
> R=yangguo@chromium.org
>
> Committed: https://crrev.com/6253aa8908c55535473ddbe1db8e5a6f5e559b73
> Cr-Commit-Position: refs/heads/master@{#26371}

TBR=svenpanne@chromium.org,ben@strongloop.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/875053005

Cr-Commit-Position: refs/heads/master@{#26372}

Fix --max_old_space_size=4096 integer overflow.

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/890563003

Cr-Commit-Position: refs/heads/master@{#26371}

add maps and sets to zone containers

R=bmeurer@chromium.org

BUG=

Review URL: https://codereview.chromium.org/889963002

Cr-Commit-Position: refs/heads/master@{#26370}

Update V8 DEPS.

Rolling v8/build/gyp to 4d7c139b1820c5fcb993868c61f170a02cda8a40

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/894493002

Cr-Commit-Position: refs/heads/master@{#26369}

MIPS: Remove unnecessary printf lines from assembler tests.

BUG=

Review URL: https://codereview.chromium.org/888093002

Cr-Commit-Position: refs/heads/master@{#26368}