Introduce a builddeps make target

In contrast to "dependencies", this only pulls in dependencies required
for building. The dependencies make target will also pull in development
tools and requires depot_tools to be installed and in the path.

BUG=v8:3418
LOG=y
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/363703002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22130 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reland r22082 "Replace HeapNumber as doublebox with an explicit MutableHeapNumber."

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/334323003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22129 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BuildGrowElementsCapacity bounds check used wrong elements kind.

In a refactoring, HGraphBuilder::BuildGrowElementsCapacity() ended up
using the old_elements_kind instead of new_kind when calculating if a
new capacity for an array exceeds Page::kMaxRegularHeapObjectSize.

Repair, such that the check is correct.

R=hpayer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/366583006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22126 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Test mjsunit/allocation-site-info is flaky without explicit gc()

It started failing on arm64 nosnap during a test case where complex nested
literal arrays are created.

R=machenbach@chromium.org
BUG=

Review URL: https://codereview.chromium.org/367703002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22125 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "Add msan build option." (r22109)

Building d8 with msan requires instrumented libc++. Currently the
advised way to do this is from a chromium checkout with d8 as target.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/361963003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22124 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "ARM64: Faster immediate check and fix corner cases"

This reverts r22120 due to build breakage of arm64.debug target.

TBR=m.m.capewell@googlemail.com

Review URL: https://codereview.chromium.org/361973002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22123 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

X87: KeyedLoadIC should have same register spec as LoadIC.

port r22103.

original commit message:
On arm, arm64 and x64 there is a different register specification between LoadIC and KeyedLoadIC.

It would be nicer if these are the same, allowing some key optimizations.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/366583002

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22122 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

ARM64: optimize TruncateDoubleToI

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

BUG=

Review URL: https://codereview.chromium.org/355853003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22121 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

ARM64: Faster immediate check and fix corner cases

Improve the code used to check for encodable logical immediates, fix some corner
cases associated with moving kWMinInt into W registers, and add tests.

BUG=
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/341123003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22120 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "Linux perf tool support update + refactoring." (r22118).

Android ninja build still failing.

TBR=yangguo@chromium.org

Review URL: https://codereview.chromium.org/367633003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22119 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reland "Linux perf tool support update + refactoring."

This relands r22098.

BUG=
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/368433006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22118 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Clean up the global object naming madness.

BUG=
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/352173006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22117 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Disable test-spaces/SizeOfFirstPageIsLargeEnough for snapshot=off build.

R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/368583002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22114 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Make JsNative1JsNative2JsSample cctest work on Win Release

MSVC optimization realizes that CallJsFunction2 is just the same as CallJsFunction, so it eliminates the former making the call stack contain two instances of the same function.

The patch makes two functions distinct.

LOG=N
BUG=v8:3055
R=aandrey@chromium.org, jkummerow@chromium.org, yurys@chromium.org

Review URL: https://codereview.chromium.org/357383003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22113 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Improve error reporting for duplicate object template properties.

BUG=
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/359413007

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22112 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Make Map.set() and Set.add() chainable

From the Harmony draft:

https://people.mozilla.org/~jorendorff/es6-draft.html#sec-map.prototype.set
https://people.mozilla.org/~jorendorff/es6-draft.html#sec-weakmap.prototype.set
https://people.mozilla.org/~jorendorff/es6-draft.html#sec-set.prototype.add
https://people.mozilla.org/~jorendorff/es6-draft.html#sec-weakset.prototype.add

BUG=v8:3410
R=arv@chromium.org, rossberg@chromium.org

Review URL: https://codereview.chromium.org/359173002

Patch from caitp <caitpotter88@gmail.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22111 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Android ARM64: update Makefile for new NDK.

BUG=
R=rmcilroy@chromium.org

Review URL: https://codereview.chromium.org/365593002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22110 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add msan build option.

Currently works only in combination with i18nsupport=off.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/367623002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22109 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Mark long-running tests as TIMEOUT for gc stress.

R=machenbach@chromium.org

Review URL: https://codereview.chromium.org/360233003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22108 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Move libplatform files into v8::platform namespace

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

Review URL: https://codereview.chromium.org/336313017

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22107 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert 22105 "Remove static initializer from isolate"

Fails to run on mac

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

Review URL: https://codereview.chromium.org/368503003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22106 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove static initializer from isolate

We don't have a default isolate anymore, so we can just create the
thread data and keys on demand

BUG=none
R=dcarney@chromium.org, svenpanne@chromium.org
LOG=y

Review URL: https://codereview.chromium.org/337603010

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22105 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert 22098 "Linux perf tool support update + refactoring."

GetCurrentThreadId doesn't compile on android

Also reverts follow up build fix attempts

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

Review URL: https://codereview.chromium.org/337093005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22104 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

KeyedLoadIC should have same register spec as LoadIC.

On arm, arm64 and x64 there is a different register specification between LoadIC and KeyedLoadIC.

It would be nicer if these are the same, allowing some key optimizations.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/338963003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22103 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix perfjit printf format for filename

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

Review URL: https://codereview.chromium.org/357333002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22102 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add missing cpu.h includes to mips/x87

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

Review URL: https://codereview.chromium.org/358073005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22101 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix build (forgot to update BUILD.gn)

TBR=yangguo@chromium.org

Review URL: https://codereview.chromium.org/341063008

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22100 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

ES6: Add missing Set.prototype.keys function

https://people.mozilla.org/~jorendorff/es6-draft.html#sec-set.prototype.keys

The value of the keys property is just the initial value of the values
function.

BUG=v8:3411
LOG=Y
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/353293003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22099 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Linux perf tool support update + refactoring.

This adds timestamps to allow profiling with code space reuse. Also a couple of updates to reflect the changes in the JIT perf interface + a move of the perf-related stuff into separate files.

Unfortunately, the change only works with the latest patch  to the perf tool from a Linux perf tool contributor (Stephane Eranian).

BUG=
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/255803003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22098 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix external snapshot compilation

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

Review URL: https://codereview.chromium.org/341433005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22097 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Rip out unused inner_scope_contains_with flag

R=marja@chromium.org
BUG=

Review URL: https://codereview.chromium.org/349033006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22096 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Wrap InitializeProperty around SetOwnPropertyIgnoreAttributes and switch over uses

This is a step in the direction of disentangling all uses of SetOwnPropertyIgnoreAttributes so we can provide a more specific implementation for those usecases, and reduce the capabilities of those clients, avoiding subtle bugs.

InitializeProperty only supports adding properties to extensible objects that do not contain the property yet. JSGlobalProxies cannot have properties themselves, so are not supported either.

BUG=
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/352813002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22095 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix compilation on arm after r22092

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

Review URL: https://codereview.chromium.org/360723002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22094 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Parser: add usage counters for "use asm".

R=jochen@chromium.org
BUG=

Review URL: https://codereview.chromium.org/336863007

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22093 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Move platform abstraction to base library

Also split v8-core independent methods from checks.h to base/logging.h and
merge v8checks with the rest of checks.

The CPU::FlushICache method is moved to CpuFeatures::FlushICache

RoundUp and related methods are moved to base/macros.h

Remove all layering violations from src/libplatform

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

Review URL: https://codereview.chromium.org/358363002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22092 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "Fix stack trace accessor behavior."

This reverts r22089.

TBR=verwaest@chromium.org

Review URL: https://codereview.chromium.org/360033002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22091 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Waiting for sweeper threads is last resort in SlowAllocateRaw.

BUG=
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/356403002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22090 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix stack trace accessor behavior.

R=verwaest@chromium.org
BUG=v8:3404
LOG=N

Review URL: https://codereview.chromium.org/343563009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22089 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Introduce debug events for promises.

R=aandrey@chromium.org, rossberg@chromium.org
BUG=v8:3093
LOG=Y

Review URL: https://codereview.chromium.org/357603005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22086 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

[Arm]: Make Assembler::movw only emit a movw instruction.

Currently Assembler::movw is really the mov macro instruction, leading to raw
emit calls to generate the real movw instruction. Replace all calls of mow
with the mov macro instruction (which will emit a movw if appropriate) and
make movw always emit movw.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/329233002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22085 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

[Arm]: Fix liveedit support if --enable_ool_constant_pool is true.

If the out of line constant pool is enabled, there is an extra value on the stack frame (the
constant pool pointer).  This change fixes GenerateFrameDropperLiveEdit to pull the function
pointer and pop the stack frame correctly whether or not the constant pool pointer is on the
frame.

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/338053003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22084 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "Replace HeapNumber as doublebox with an explicit MutableHeapNumber."

This reverts commit r22082 for breaking arm64 build.

TBR=verwaest@chromium.org

Review URL: https://codereview.chromium.org/360023003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22083 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Replace HeapNumber as doublebox with an explicit MutableHeapNumber.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/355793003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22082 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove kDontInline and simplify compiler hints.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/359733004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22081 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Use simple keyed store again in ArrayPush_JS_Builtin

This reverts the change of array.js in r20839 and recovers 38%
performance regression of DOM-Search case in Browsermark2.0. There is
no performance impact on JavaScript benchmarks (Octane, Kraken and
SunSpider).

BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/353953002

Patch from Weiliang Lin <weiliang.lin@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22080 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Set promotion queue limit before migrating object.

BUG=
R=titzer@chromium.org

Review URL: https://codereview.chromium.org/357263004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22079 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "Add mechanism to postpone interrupts selectively."

This reverts commit r22073.

TBR=yurys@chromium.org

Review URL: https://codereview.chromium.org/362493002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22074 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add mechanism to postpone interrupts selectively.

BUG=v8:3408
LOG=N
R=yurys@chromium.org

Review URL: https://codereview.chromium.org/359723005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22073 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Support non-internalized string key lookups in Hydrogen KeyedLoadIC

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/356213003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22070 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Whitespace change to trigger bots.

TBR=jkummerow@chromium.org
BUG=

Review URL: https://codereview.chromium.org/354203002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22069 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Build "snapshot_blob.bin" only for the host when using separate toolsets.

BUG=389310
LOG=N
R=jochen@chromium.org

Review URL: https://codereview.chromium.org/353143002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22068 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Cannot check the interceptor since it may not be ready yet.

BUG=
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/353143003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22067 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Port the ic part

BUG=
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/354173002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22066 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Don't leak the global object in the Function constructor.

BUG=
R=dcarney@chromium.org

Review URL: https://codereview.chromium.org/359713005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22065 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Split SetProperty(...attributes, strictmode) into DefineProperty(...attributes) and SetProperty(...strictmode)

BUG=
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/351853005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22064 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove script collected debug event.

R=yurys@chromium.org

Review URL: https://codereview.chromium.org/358873005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22063 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove remaining dependencies on allocation.h from a few platform files

I could move parts of allocation to base/ instead, but I see little to
no value in the allocation base classes, soo..

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

Review URL: https://codereview.chromium.org/351683009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22062 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove the duplicated file include.

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/352223007

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22061 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Allow inlining of functions containing %_Arguments.

R=svenpanne@chromium.org
TEST=mjsunit/compiler/inline-arguments

Review URL: https://codereview.chromium.org/356773002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22060 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Unbreak Mac build.

TBR=marja@chromium.org

Review URL: https://codereview.chromium.org/356153006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22059 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Unbreak "os" stuff in shared d8 builds.

Note that "os" still has no properties on Windows, but at least it's there.

BUG=v8:3407
LOG=y
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/356843005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22058 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Whitespace change to trigger bots.

TBR=jkummerow@chromium.org
BUG=

Review URL: https://codereview.chromium.org/352583009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22057 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Unbreak Win64 build (hopefully).

cpplint, our fragile overloading of CheckEqualsHelper and Win64 don't
really play together very well...

TBR=marja@chromium.org

Review URL: https://codereview.chromium.org/354143003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22056 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Let test driver nuke test perf data on errors.

BUG=
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/350913004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22055 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

ARM64: avoid duplicate load of double constants

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

BUG=

Review URL: https://codereview.chromium.org/357973002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22051 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "ARM: Use the shifter operand to merge in previous shift instructions."

This reverts r 22017

Reason: broke Nexus 7 GPU bots (see crbug.com/389198)

BUG=389198
LOG=n
TBR=machenbach@chromium.org, ulan@chromium.org

Review URL: https://codereview.chromium.org/359713004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22050 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Added slim versions of output streams.

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

Review URL: https://codereview.chromium.org/352823003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22049 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove dependency from platform files on v8.h

Pass flag values in through a central initialize call.

We still depend on utils.h for RoundUp and IsAligned.

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

Review URL: https://codereview.chromium.org/353113003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22048 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

X87: Use IC register definitions in platform files.

port r22035

original commit message:
  Use IC register definitions in platform files.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/356133005

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22047 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

MIPS: Use IC register definitions in platform files.

Port r22035 (0eac506)

BUG=
R=palfia@homejinni.com

Review URL: https://codereview.chromium.org/354013003

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22044 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add OnCompileError handler and v8::CompileError debug event.
This event is generated when the parser can not generate code.

R=vsevik@chromium.org, yangguo@chromium.org, yurys@chromium.org

Review URL: https://codereview.chromium.org/264333007

Patch from Alexey Kozyatinskiy <kozyatinskiy@google.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22043 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove unused Debug::AfterCompileFlags flags.

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/353823004

Patch from Alexey Kozyatinskiy <kozyatinskiy@google.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22042 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix regexp=interpreted build.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/350633003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22041 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Added more detailed sweeping timing counters.

BUG=
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/331053003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22040 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Infer whether a variable is assigned in inner functions

R=titzer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/345573002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22039 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reland "Allow inlining of functions containing %_ArgumentsLength."

The layout test failures were unrelated to this change.

R=svenpanne@chromium.org
TBR=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/355993002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22038 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Keep CodeRange::current_allocation_block_index_ in range after r21869.

BUG=305878,388328
LOG=N
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/354003003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22037 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Roll buildtools to fb782d4369d5ae04f17a2fceef7de5a63e50f07b

Also add hooks to download clang-format binaries.

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

Review URL: https://codereview.chromium.org/334403005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22036 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Use IC register definitions in platform files.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/356713003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22035 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Suppress GC stress failure for a test that asserts code being optimized.

The reason is that GC stress causes additional deopts that the test
originally did not foresee.

R=marja@chromium.org

Review URL: https://codereview.chromium.org/355973003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22034 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Avoid adjusting live bytes in JSObject::MigrateFastToFast() if the size delta is zero.

BUG=chromium:388880
LOG=N
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/333903003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22031 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

X87: Cleanup CallInterfaceDescriptor

port r22014

original commit message:
 Make CallInterfaceDescriptor work like CodeStubInterfaceDescriptor, owning it's
 register and representation arrays. This also eliminates a host of TSAN warnings
 xfor static arrays

 This CL depends on https://codereview.chromium.org/352583002 landing first
 (a conceptual dependency at least).

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/343563006

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22030 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Compile optimized code with active debugger but no break points.

R=ulan@chromium.org
BUG=386492
LOG=Y

Review URL: https://codereview.chromium.org/356713004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22029 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

X87: The IC exposes a register definition.

port r22011

original commit message:
  Centralize a register definition in an IC that provides:
  1) symbolic names for the register (like, edx == receiver)
  2) defines ordering when passed on the stack

  Code that implements or uses the IC should use this definition instead of "knowing" what the registers are. Or at least have the definition to validate it's assumptions.

  As a side effect of avoiding runtime static initializers (enforced by tools/check-static-initializers.sh, neat), I gave ownership of the registers array to CodeStubInterfaceDescriptor. This prompted a cleanup of that struct

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/358773002

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22028 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Optimize Map/Set.prototype.forEach

Instead of using an iterator result object and an entries array
(for Map) we introduce a new runtime function that uses an array
as an out param.

On the Map ForEach perf test this leads to a 2.5x performance
improvement. On the overall Map and Set tests this leads to a 18%
and 13% improvement respectively.

BUG=None
LOG=Y
R=danno@chromium.org

Review URL: https://codereview.chromium.org/355663002

Patch from Erik Arvidsson <arv@chromium.org>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22027 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

MIPS: Fix r22021 “The IC exposes a register definition.” after r22018.

BUG=
R=palfia@homejinni.com

Review URL: https://codereview.chromium.org/344543004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22024 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

MIPS: Cleanup CallInterfaceDescriptor.

Port r22014 (234866c)

Original commit message:
Make CallInterfaceDescriptor work like CodeStubInterfaceDescriptor, owning it's
register and representation arrays. This also eliminates a host of TSAN warnings
for static arrays.

This CL depends on https://codereview.chromium.org/352583002 landing first
(a conceptual dependency at least).

BUG=
R=palfia@homejinni.com

Review URL: https://codereview.chromium.org/356773003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22023 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Revert "Allow inlining of functions containing %_ArgumentsLength."

This cuased layout test failures.

TBR=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/349923009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22022 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

MIPS: The IC exposes a register definition.

Port r22011 (be22370)

Original commit message:
Centralize a register definition in an IC that provides:
1) symbolic names for the register (like, edx == receiver)
2) defines ordering when passed on the stack

Code that implements or uses the IC should use this definition instead of "knowing" what the registers are. Or at least have the definition to validate it's assumptions.

As a side effect of avoiding runtime static initializers (enforced by tools/check-static-initializers.sh, neat), I gave ownership of the registers array to CodeStubInterfaceDescriptor. This prompted a cleanup of that struct.

BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/354863002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22021 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Allow inlining of functions containing %_ArgumentsLength.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/355833002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22020 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Collect garbage with kReduceMemoryFootprintMask in IdleNotification.

BUG=350720
LOG=Y
TBR=hpayer@chromium.org

Review URL: https://codereview.chromium.org/345263004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22019 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove distinction between hidden and normal runtime functions

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

Review URL: https://codereview.chromium.org/346413004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22018 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

ARM: Use the shifter operand to merge in previous shift instructions.

When possible, we transform sequences of code of the form
    lsl r1, r2, #imm
    add r0, r5, r1
into
    add r0, r5, r2 LSL #imm

This is an adaptation of r21161.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/312173002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22017 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Whitespace change to trigger bots.

TBR=jkummerow@chromium.org

BUG=

Review URL: https://codereview.chromium.org/348843010

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22016 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix tools/whitespace.txt. Huge perf improvements.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/330913006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22015 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Cleanup CallInterfaceDescriptor

Make CallInterfaceDescriptor work like CodeStubInterfaceDescriptor, owning it's
register and representation arrays. This also eliminates a host of TSAN warnings
for static arrays.

This CL depends on https://codereview.chromium.org/352583002 landing first
(a conceptual dependency at least).

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/350293003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22014 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reorder full code for while loops to better reflect statement positions.

R=jkummerow@chromium.org
BUG=v8:2047
LOG=N

Review URL: https://codereview.chromium.org/353823002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22013 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Better fix for --trace-ic source positions: Use HOptimizedGraphBuilderWithPositions

This reverts the changes made in r22008.

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/355823003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22012 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The IC exposes a register definition.

Centralize a register definition in an IC that provides:
1) symbolic names for the register (like, edx == receiver)
2) defines ordering when passed on the stack

Code that implements or uses the IC should use this definition instead of "knowing" what the registers are. Or at least have the definition to validate it's assumptions.

As a side effect of avoiding runtime static initializers (enforced by tools/check-static-initializers.sh, neat), I gave ownership of the registers array to CodeStubInterfaceDescriptor. This prompted a cleanup of that struct.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/352583002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22011 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Blacklist failing test from GC stress runs.

The test expects a function not to deopt. However GC stress causes
the code to be marked for deopt for a reason we are not testing against.

R=marja@chromium.org

Review URL: https://codereview.chromium.org/354853002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22010 ce2b1a6d-e550-0410-aec6-3dcde31c8c00