Always emit bailout id for inlining property access (even for keyed access).

R=ulan@chromium.org
BUG=chromium:453805
LOG=n

Review URL: https://codereview.chromium.org/887023003

Cr-Commit-Position: refs/heads/master@{#26359}

Load callback data from weak cell instead of embedding it in handler.

BUG=v8:3629
LOG=N

Review URL: https://codereview.chromium.org/877243004

Cr-Commit-Position: refs/heads/master@{#26358}

Do not inline array push for arrays with dictionary mode elements.

BUG=chromium:452878
LOG=n

Review URL: https://codereview.chromium.org/880233002

Cr-Commit-Position: refs/heads/master@{#26357}

Fix the test-code-stubs tests.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/892603002

Cr-Commit-Position: refs/heads/master@{#26356}

[turbofan] Fix usage of ThisFunction parameter in OSR.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/892593002

Cr-Commit-Position: refs/heads/master@{#26355}

Layout descriptor sharing issue fixed.

BUG=chromium:437713, v8:3832
LOG=Y

Review URL: https://codereview.chromium.org/885003002

Cr-Commit-Position: refs/heads/master@{#26354}

MIPS: Add MEGAMORPHIC state support for KeyedLoadIC.

Port 491eb81780873aafd4d7f6638d20f26839383c68

BUG=

Review URL: https://codereview.chromium.org/888843002

Cr-Commit-Position: refs/heads/master@{#26353}

update vector ics to handle megamorphic keyed loads

R=mvstanton@chromium.org

BUG=

Review URL: https://codereview.chromium.org/889863002

Cr-Commit-Position: refs/heads/master@{#26352}

Visit code objects in native contexts only during full GCs.

BUG=

Review URL: https://codereview.chromium.org/887943002

Cr-Commit-Position: refs/heads/master@{#26351}

[turbofan] cleanup MoveOptimizer a little

BUG=

Review URL: https://codereview.chromium.org/888813002

Cr-Commit-Position: refs/heads/master@{#26350}

fix gcc after r26345

BUG=
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/885303002

Cr-Commit-Position: refs/heads/master@{#26349}

Avoid using ProcessArguments for static calls.

R=titzer@chromium.org

Review URL: https://codereview.chromium.org/886853002

Cr-Commit-Position: refs/heads/master@{#26348}

Make GCC happy with test-cpu-profiler.

TBR=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/891663002

Cr-Commit-Position: refs/heads/master@{#26347}

Reland "Initial switch to Chromium-style CHECK_* and DCHECK_* macros.".

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/877753007

Cr-Commit-Position: refs/heads/master@{#26346}

Add MEGAMORPHIC state support for KeyedLoadIC

BUG=

Review URL: https://codereview.chromium.org/878263002

Cr-Commit-Position: refs/heads/master@{#26345}

Collect phantom callback data only once during gc. For scavenges, just consider the young handles referenced by young generation.

BUG=

Review URL: https://codereview.chromium.org/885553005

Cr-Commit-Position: refs/heads/master@{#26344}

Contribution of PowerPC port (continuation of 422063005) - AIX Common1

Contribution of PowerPC port (continuation of 422063005 and 817143002). This patch covers
the key changes needed to the common files needed to support AIX. Subsequent
patches will cover:
- changes to update the ppc directories so they are current with the changes
in the rest of the project.
- remaining AIX changes not resolved by 4.8 compiler
- individual optimizations for PPC

This is based off of the GitHub repository
https://github.com/andrewlow/v8ppc

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

Review URL: https://codereview.chromium.org/866843003

Cr-Commit-Position: refs/heads/master@{#26343}

Revert "Make GCC happy again." and "Initial switch to Chromium-style CHECK_* and DCHECK_* macros.".

This reverts commit 6a4c0a3bae60b1a2239f35e3b759bd215e3886fe and commit
0deaa4b629faeae6832046c294dc7c034a0a6653 for breaking GCC bots.

TBR=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/893533003

Cr-Commit-Position: refs/heads/master@{#26342}

Make GCC happy again.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/889743003

Cr-Commit-Position: refs/heads/master@{#26341}

Initial switch to Chromium-style CHECK_* and DCHECK_* macros.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/888613002

Cr-Commit-Position: refs/heads/master@{#26340}

Update V8 DEPS.

Rolling v8/build/gyp to 104e21ecf6f28c47d2e729d572aa790b85225ee0

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/891623002

Cr-Commit-Position: refs/heads/master@{#26339}

Implemented new syntax for ImportDeclarations

Also had to split ParseImportsList and ParseExportClause into separate
methods as they have different rules about reserved words and arguments/eval.

Added lots more test cases, including some export cases that were missed before
due to incorrect checking of reserved words.

BUG=v8:1569
LOG=n

Review URL: https://codereview.chromium.org/887843002

Cr-Commit-Position: refs/heads/master@{#26338}

Lexical declarations should not be allowed in Statement

For example let and class should only be allowed inside function/block/script.

We have to continue to support const in statements in sloppy mode for backwards compatibility.

BUG=3831
LOG=Y
R=dslomov@chromium.org, adamk

Review URL: https://codereview.chromium.org/869293002

Cr-Commit-Position: refs/heads/master@{#26337}

Move object literal checking into checker classes

This removes the duplicate property check from object literals.

Instead we repurpose the ObjectLiteralChecker into two cases, implemented
by two subclasses to ObjectLiteralCheckerBase called ObjectLiteralChecker
and ClassLiteralChecker.

The object literal checker now only checks for duplicate __proto__ fields in
object literals.

The class literal checker checks for duplicate constructors, non constructor
fields named constructor as well as static properties named prototype.

BUG=v8:3819
LOG=Y
R=adamk, dslomov@chromium.org

Review URL: https://codereview.chromium.org/873823003

Cr-Commit-Position: refs/heads/master@{#26336}

MIPS: Fix 'Load API accessor from descriptor instead of embedding it in handler.'

Fix a typo in 6ea31f7eee230228c9d34362360b8b9f08e79815

BUG=

Review URL: https://codereview.chromium.org/878023003

Cr-Commit-Position: refs/heads/master@{#26335}

Revert of Use a WeakCell in the CallIC type vector. (patchset #4 id:80001 of https://codereview.chromium.org/881433002/)

Reason for revert:
Breaks a regression test on linux isloates, and linux with gcc 4.8.

Original issue's description:
> Use a WeakCell in the CallIC type vector.
>
> This allows us to clear the IC on a more sedate schedule, just
> like Load and Store ICs.
>
> R=ulan@chromium.org
> BUG=
>
> Committed: https://crrev.com/bcc79d33ca6d97d9ecfcfcf110a6ea84a0225389
> Cr-Commit-Position: refs/heads/master@{#26332}

TBR=ulan@chromium.org,mvstanton@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=

Review URL: https://codereview.chromium.org/886663004

Cr-Commit-Position: refs/heads/master@{#26334}

[turbofan] Fix OSR compilations of for-in.

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

Review URL: https://codereview.chromium.org/890543002

Cr-Commit-Position: refs/heads/master@{#26333}

Use a WeakCell in the CallIC type vector.

This allows us to clear the IC on a more sedate schedule, just
like Load and Store ICs.

R=ulan@chromium.org
BUG=

Review URL: https://codereview.chromium.org/881433002

Cr-Commit-Position: refs/heads/master@{#26332}

Load API holder from prototype chain instead embedding it in handler.

BUG=v8:3629
LOG=N

Review URL: https://codereview.chromium.org/873723005

Cr-Commit-Position: refs/heads/master@{#26331}

Introduce ProcessYoungWeakReferences and process weak list of allocation sites just during full GCs.

BUG=

Review URL: https://codereview.chromium.org/890613002

Cr-Commit-Position: refs/heads/master@{#26330}

store InstructionOperands directly in Instruction

BUG=

Review URL: https://codereview.chromium.org/884503002

Cr-Commit-Position: refs/heads/master@{#26329}

Do not create unresolved variables when parsing arrow functions lazily

Arrow function parameter lists are parsed as expressions. When an identifier
is found a VariableProxy is created and added to the list of unresolved
variables for the scope. When parsing a function lazily, the scope has been
already resolved, so with this patch only the VariableProxy is created,
without adding it as an unresolved variable in the scope.

BUG=v8:3501
LOG=Y

Review URL: https://codereview.chromium.org/880253004

Cr-Commit-Position: refs/heads/master@{#26328}

[turbofan] Make sure there is space for lazy deopt patching before the constant pool.

BUG=chromium:446647
LOG=n
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/874863003

Cr-Commit-Position: refs/heads/master@{#26327}

Load API call data from function instead of embedding it in handler.

BUG=v8:3629
LOG=N

Review URL: https://codereview.chromium.org/883253002

Cr-Commit-Position: refs/heads/master@{#26326}

Retrieve recent v8 release information based on tags.

BUG=chromium:451975
TBR=tandrii@chromium.org
NOTRY=true
LOG=n

TEST=./script_test.py
TEST=tools/release/releases.py -c path/to/src

Review URL: https://codereview.chromium.org/869613007

Cr-Commit-Position: refs/heads/master@{#26325}

[V8] Added Script::is_debugger_script flag for embedders

In DevTools we need one more flag for script origin - is debugger script. We already have "is shared origin" flag. The new flag added by analogy with the old but new has accessor in script object.

R=yurys@chromium.org

Review URL: https://codereview.chromium.org/879553002

Cr-Commit-Position: refs/heads/master@{#26324}

[turbofan] Improve JSON output

Review URL: https://codereview.chromium.org/882973002

Cr-Commit-Position: refs/heads/master@{#26323}

Add try release mode to test runner.

Also refactor the configuration of modes to have the level
of abstraction in one json dict.

In a follow up CL, the new mode could be added to quickcheck
and release trybots.

Review URL: https://codereview.chromium.org/882983002

Cr-Commit-Position: refs/heads/master@{#26322}

Bump max-inlined-nodes-cumulative

BUG=

Review URL: https://codereview.chromium.org/887603002

Cr-Commit-Position: refs/heads/master@{#26321}

Load API accessor from descriptor instead of embedding it in handler.

BUG=v8:3629
LOG=N

Review URL: https://codereview.chromium.org/880333003

Cr-Commit-Position: refs/heads/master@{#26320}

[turbofan] Remove use of generic algorithm from Scheduler.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/882103002

Cr-Commit-Position: refs/heads/master@{#26319}

[turbofan] Gracefully bail out if OSR encounters a loop too deeply nested.

R=jarin@chromium.org
BUG=

Review URL: https://codereview.chromium.org/877553007

Cr-Commit-Position: refs/heads/master@{#26318}

Fix register aliasing after r26306, r26275.

BUG=

Review URL: https://codereview.chromium.org/877343003

Cr-Commit-Position: refs/heads/master@{#26317}

[turbofan] Cleanup the NodeProperties.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/883613006

Cr-Commit-Position: refs/heads/master@{#26316}

Minor refactoring for Zone class and friends.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/885813002

Cr-Commit-Position: refs/heads/master@{#26315}

Don't take iterable path in ArrayFrom if items[@@iterator] is null

BUG=v8:3833
LOG=N
R=arv@chromium.org

Review URL: https://codereview.chromium.org/856303002

Cr-Commit-Position: refs/heads/master@{#26314}

Implement ParseExportDeclaration per latest ES6 spec draft

One missing feature: anonymous function & class declarations
in "export default".

BUG=v8:1569
LOG=n
R=arv@chromium.org

Review URL: https://codereview.chromium.org/882893002

Cr-Commit-Position: refs/heads/master@{#26313}

Add Erik Arvidsson to OWNERS

R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/886463005

Cr-Commit-Position: refs/heads/master@{#26312}

MIPS: Continue learning for calls in crankshaft.

Port 7d363783e14433032a4e8dd426fe0e27d1637fe5

Original commit message:
The type feedback vector makes this easy to do.
This is a re-land of https://codereview.chromium.org/868453005/
with a fix for the DCHECK failure.

BUG=

Review URL: https://codereview.chromium.org/877253003

Cr-Commit-Position: refs/heads/master@{#26311}

Extract LoadAccessor into a masm function.

BUG=

Review URL: https://codereview.chromium.org/886503002

Cr-Commit-Position: refs/heads/master@{#26310}

[turbofan] Add missing deopt for the assignment in the for-in statement.

BUG=chromium:416359
LOG=n
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/881303002

Cr-Commit-Position: refs/heads/master@{#26309}

Only add API-triggered compilation to compile-script histogram.

R=vogelheim@chromium.org

Review URL: https://codereview.chromium.org/879273003

Cr-Commit-Position: refs/heads/master@{#26308}

[turbofan] Use unboxed doubles in range types.

BUG=
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/882063002

Cr-Commit-Position: refs/heads/master@{#26307}

Load setter from map descriptor instead of embedding it in handler.

BUG=v8:3629
LOG=N

Review URL: https://codereview.chromium.org/879213003

Cr-Commit-Position: refs/heads/master@{#26306}

Add missing FrameState to JSToName nodes.

R=jarin@chromium.org
TEST=mjsunit/regress/regress-crbug-451770
BUG=chromium:451770
LOG=N

Review URL: https://codereview.chromium.org/880963002

Cr-Commit-Position: refs/heads/master@{#26305}

Removed TRYSERVER_ROOT since this is not used anymore

Instead PROJECT property is used by Rietveld to set patch_project, which is then
mapped to correct path in the recipe.

R=machenbach@chromium.org

Review URL: https://codereview.chromium.org/878573003

Cr-Commit-Position: refs/heads/master@{#26304}

Let release scripts determine version based on tags.

Also create new v8 versions based on the roll ref, which
has some more verification steps than the candidate ref.

Improve bailout regarding existing versions. Only create a
new version if the revisions is in the future.

BUG=chromium:451975
TBR=tandrii@chromium.org
NOTRY=true
LOG=n

Review URL: https://codereview.chromium.org/878913002

Cr-Commit-Position: refs/heads/master@{#26303}

Continue learning for calls in crankshaft.

The type feedback vector makes this easy to do.
This is a re-land of https://codereview.chromium.org/868453005/
with a fix for the DCHECK failure.

Review URL: https://codereview.chromium.org/885593002

Cr-Commit-Position: refs/heads/master@{#26302}

Reland of "Steps towards unification of number bitset and range types."

This reverts commit 76193749791eac9032f47db2f04efd78c0e120a9.

BUG=

Review URL: https://codereview.chromium.org/877643002

Cr-Commit-Position: refs/heads/master@{#26301}

Update V8 DEPS.

Rolling v8/tools/clang to 0825693c3460f23b1d78ad272999f850447ad333

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/873313008

Cr-Commit-Position: refs/heads/master@{#26300}

Begin modernization of --harmony-modules

The approach taken in this CL is to incrementally move toward the
currently-specced version of modules in ES6. The biggest change in this
patch is separating the parsing of modules from the parsing of scripts,
getting rid of the 'module' keyword and thus disallowing modules-in-scripts
as well as modules-in-modules.

The syntax supported by import/export declarations has not yet been significantly
changed, with the major exception being that import declarations require a string
as the 'from' part.

Most of the existing tests have been disabled, with a first new test added
in cctest/test-parsing.

BUG=v8:1569
LOG=n

Review URL: https://codereview.chromium.org/881623002

Cr-Commit-Position: refs/heads/master@{#26299}

MIPS: Use a trampoline stub to load the type feedback vector for CallICs.

Port 2a567faa3e7242359e29c63f8dc8c18104af0610

BUG=

Review URL: https://codereview.chromium.org/882683002

Cr-Commit-Position: refs/heads/master@{#26298}

Land test case for RegExp.source.

BUG=chromium:447561
LOG=N
TBR=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/878033003

Cr-Commit-Position: refs/heads/master@{#26297}

Reland "Only use FreeSpace objects in the free list"

Review URL: https://codereview.chromium.org/882633002

Cr-Commit-Position: refs/heads/master@{#26296}

Rescale histogram timers.

R=vogelheim@chromium.org

Review URL: https://codereview.chromium.org/875873002

Cr-Commit-Position: refs/heads/master@{#26295}

[turbofan] Ensure that NTLs are always properly connected to the end.

Up until now we used a special Terminate node to artifically connect non
terminating loops to the End node, but this was kind of adhoc and didn't
work for the CFG. So without all kinds of weird hacks, the end block in
the CFG will not be connected to NTLs, which makes it impossible to
compute post dominance / control dependence in the current setting.

So instead of Terminate, we add a special Branch to NTLs, whose
condition is the special Always node, which corresponds to True, except
that it cannot be folded away. This way we don't need any special
machinery in the scheduler, since it's just a regular Branch.

R=titzer@chromium.org

Review URL: https://codereview.chromium.org/875263004

Cr-Commit-Position: refs/heads/master@{#26294}

Revert of Continue learning for calls in optimized code when we have no type feedback. (patchset #4 id:60001 of https://codereview.chromium.org/868453005/)

Reason for revert:
Serializer tests broke. Need to debug and fix.

Original issue's description:
> Continue learning for calls in optimized code when we have no type feedback.
>
> Based on CL https://codereview.chromium.org/871063002/ which needs to land first.
>
> BUG=
>
> Committed: https://crrev.com/f5f2692b5ff70ac3cd06a903b7846174b97a2e55
> Cr-Commit-Position: refs/heads/master@{#26292}

TBR=verwaest@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=

Review URL: https://codereview.chromium.org/878103002

Cr-Commit-Position: refs/heads/master@{#26293}

Continue learning for calls in optimized code when we have no type feedback.

Based on CL https://codereview.chromium.org/871063002/ which needs to land first.

BUG=

Review URL: https://codereview.chromium.org/868453005

Cr-Commit-Position: refs/heads/master@{#26292}

fix performance regression on intel call api stubs

additionally, make the interface match the JSFunction interface

BUG=451405
LOG=N

Review URL: https://codereview.chromium.org/883613003

Cr-Commit-Position: refs/heads/master@{#26291}

Use a trampoline stub to load the type feedback vector for CallICs.

BUG=

Review URL: https://codereview.chromium.org/871063002

Cr-Commit-Position: refs/heads/master@{#26290}

Do not generalize field representations when making elements kind or observed transition.

BUG=chromium:448711
LOG=y

Review URL: https://codereview.chromium.org/861173004

Cr-Commit-Position: refs/heads/master@{#26289}

[x86] Disable AVX unless the operating system explicitly claims to support it.

BUG=chromium:452033, v8:3846
LOG=y
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/878063002

Cr-Commit-Position: refs/heads/master@{#26288}

Revert of Only use FreeSpace objects in the free list. (patchset #3 id:40001 of https://codereview.chromium.org/876613002/)

Reason for revert:
Test failures

Original issue's description:
> Only use FreeSpace objects in the free list.
>
> This solves an issue with the custom startup snapshot, in cases where
> deserializing the isolate requires more than one page per space.
>
> R=hpayer@chromium.org
>
> Committed: https://crrev.com/66964395108f03220cb6f45ddc73c5965e2c76a9
> Cr-Commit-Position: refs/heads/master@{#26285}

TBR=hpayer@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/882443004

Cr-Commit-Position: refs/heads/master@{#26287}

[turbofan] Only replace nodes eagerly during simplified lowering if the types stay the same.

BUG=chromium:452427
LOG=n
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/871373010

Cr-Commit-Position: refs/heads/master@{#26286}

Only use FreeSpace objects in the free list.

This solves an issue with the custom startup snapshot, in cases where
deserializing the isolate requires more than one page per space.

R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/876613002

Cr-Commit-Position: refs/heads/master@{#26285}

Refactor version increment in release scripts.

BUG=chromium:451975
TBR=tandrii@chromium.org
NOTRY=true
LOG=n

Review URL: https://codereview.chromium.org/870903003

Cr-Commit-Position: refs/heads/master@{#26284}

[turbofan] Some cleanup for scheduler unittests.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/877033002

Cr-Commit-Position: refs/heads/master@{#26283}

Don't use internal gmock helper classes.

BUG=v8:3835
LOG=n

Review URL: https://codereview.chromium.org/881653003

Cr-Commit-Position: refs/heads/master@{#26282}

[turbofan] Handle cyclic dependencies in context typing.

BUG=chromium:451012
LOG=n
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/874983002

Cr-Commit-Position: refs/heads/master@{#26281}

[turbofan] Remove GenericAlgorithm from verifier and graph replay.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/879583002

Cr-Commit-Position: refs/heads/master@{#26280}

MIPS: Refactoring after 'MIPS: split api call stubs into accessor and function call stubs'.

Change stack handling to clarify the usage of reserved MIPS argument slots for mips32.

BUG=

Review URL: https://codereview.chromium.org/867183003

Cr-Commit-Position: refs/heads/master@{#26279}

MIPS64: fix wrong register use in LoadIndexedStringStub.

TEST=mjsunit/string-externalize, others...
BUG=

Review URL: https://codereview.chromium.org/861193008

Cr-Commit-Position: refs/heads/master@{#26278}

MIPS: Fixed Hydrogen environment handling for mul-i ARM and ARM64.

Port a7d67a64f11100434b196143e2ba516f8c13697a

Original commit message:
The whole logic in DoMul makes me cry, so I made only the minimal
change to fix the issue...

BUG=v8:451322
LOG=y

Review URL: https://codereview.chromium.org/879473002

Cr-Commit-Position: refs/heads/master@{#26277}

[turbofan] Simplify reduction if IfTrue and IfFalse and fix bugs.

R=mstarzinger@chromium.org
BUG=chromium:451958
LOG=Y

Review URL: https://codereview.chromium.org/880533002

Cr-Commit-Position: refs/heads/master@{#26276}

Load getter from map descriptor instead of embedding it in handler.

BUG=v8:3629
LOG=N

Review URL: https://codereview.chromium.org/872723003

Cr-Commit-Position: refs/heads/master@{#26275}

Cache WeakCell for optimized code.

BUG=

Review URL: https://codereview.chromium.org/873233002

Cr-Commit-Position: refs/heads/master@{#26274}

Add HeapNumber fast path to v8::Value::{Uint,Int}32Value()

This has the added benefit that these functions are now guaranteed not to throw when v8::Value::Is{Uint,Int}32() returned true, even when calling into JavaScript would throw a stack limit error.

BUG=chromium:446097
LOG=y

Review URL: https://codereview.chromium.org/873983003

Cr-Commit-Position: refs/heads/master@{#26273}

Perform full GC if incremental marking is completed or marking deque is empty.

BUG=

Review URL: https://codereview.chromium.org/874993003

Cr-Commit-Position: refs/heads/master@{#26272}

External name changes of release scripts.

Moved and renamed the scripts to fit with git.

These changes require documentation updates, updates of
the release spreadsheet and the buildbot side.

BUG=chromium:451975
TBR=tandrii@chromium.org
NOTRY=true
LOG=n

Review URL: https://codereview.chromium.org/868473004

Cr-Commit-Position: refs/heads/master@{#26271}

[turbofan] Better narrow the derived type for the right shift operation.

Currently the derived type of a right shift does not narrow the input
type based on the actual shift amount - well it does some narrowing
but more can be down. For patterns such as u32[i>>2], which is very
common is asm.js code, this limits the ability to later prove that an
index bounds check is unnecessary which can have a significant
performance impact.

Review URL: https://codereview.chromium.org/873143002

Cr-Commit-Position: refs/heads/master@{#26270}

Clean up release scripts.

This makes some internal renaming, e.g. trunk -> candidates,
bleeding edge -> master, without changing the api.

Also remove some unused bailout steps from the push script.

Remove unused bump version script.

BUG=chromium:451975
TBR=tandrii@chromium.org
NOTRY=true
LOG=n

Review URL: https://codereview.chromium.org/873213002

Cr-Commit-Position: refs/heads/master@{#26269}

Add missing FrameState for Runtime_CreateArrayLiteral.

R=jarin@chromium.org
TEST=mjsunit/regress/regress-crbug-451013
BUG=chromium:451013
LOG=N

Review URL: https://codereview.chromium.org/873973003

Cr-Commit-Position: refs/heads/master@{#26268}

Convert compiler cctest to unittests: SchedulerTest

R=mstarzinger@chromium.org
LOG=N

Review URL: https://codereview.chromium.org/863213003

Cr-Commit-Position: refs/heads/master@{#26267}

Probe same-isolate compilation cache before deserializing.

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/875863002

Cr-Commit-Position: refs/heads/master@{#26266}

[turbofan] make register allocator verifier independent of phi assignment

Review URL: https://codereview.chromium.org/875853002

Cr-Commit-Position: refs/heads/master@{#26265}

[turbofan] Don't embed pointer to builtins object in generic lowering.

Accessing the builtins object this way prevents sharing of code across
native contexts; instead we now load the builtin from the context of the
JS operation instead.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/879433002

Cr-Commit-Position: refs/heads/master@{#26264}

[turbofan] Add new JSIntrinsicsLowering reducer.

The lowering of intrinsics is therefore now decoupled from the general
inlining logic.

TEST=cctest,unittests
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/872363002

Cr-Commit-Position: refs/heads/master@{#26263}

Fix regexp=interpreted build wrt zone related changes.

Broke after c7b09aac317dd810d29b82061973199b80363b65

TBR=danno@chromium.org

Review URL: https://codereview.chromium.org/875833002

Cr-Commit-Position: refs/heads/master@{#26262}

Fixed Hydrogen environment handling for mul-i on ARM and ARM64.

The whole logic in DoMul makes me cry, so I made only the minimal
change to fix the issue...

BUG=v8:451322
LOG=y

Review URL: https://codereview.chromium.org/873703002

Cr-Commit-Position: refs/heads/master@{#26261}

Update V8 DEPS.

Rolling v8/tools/clang to 853a3150b369e5dcd62d7429998bfafe8158e16e

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/868373002

Cr-Commit-Position: refs/heads/master@{#26260}