git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24181 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=bmeurer@chromium.org
BUG=v8:3589
LOG=n

Review URL: https://codereview.chromium.org/598963002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24180 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

LOG=n
BUG=v8:3589
TEST=compiler-unittests,cctest
R=titzer@chromium.org

Review URL: https://codereview.chromium.org/596703004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24179 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Boring semi-mechanical stuff...

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/598953002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24178 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=bmeurer@chromium.org
TEST=compiler-unittests/JSBuiltinReducerTest.MathSqrt

Review URL: https://codereview.chromium.org/595963002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24177 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Make heap numbers detection more consistent on arm64.
All the tested benchmarks (octane2, kraken, sunspider, v8-v4 and lua) are unchanged (a57 and a53).

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

Review URL: https://codereview.chromium.org/577273002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24176 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Port r24161 (008f335)

BUG=
R=dusan.milosavljevic@imgtec.com

Review URL: https://codereview.chromium.org/593313004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24175 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=416416
LOG=N
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/587363002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24174 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TBR=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/598943002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24173 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This is preliminary work, required to properly support bounds checking for typed array loads/stores.

TEST=compiler-unittests,cctest
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/602563002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24172 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/597083002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24171 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=dslomov@chromium.org

Review URL: https://codereview.chromium.org/595253002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24170 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r24146.

original commit message:
   Made the detailed reason for deopts mandatory on ia32. Unified and improved things.

   The number of different detail strings has been reduced a bit without
   any loss of information. Furthermore, DoubleToI is a bit more
   informative now. Tiny reason fixes.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/597003003

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24169 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/598603002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24168 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Besides addressing a fundamental flaw, this significantly simplifies
several aspects of the system.  The downside is a loss of precision
and a loss of algebraic properties.

Range types are now fully implemented.

R=rossberg@chromium.org
BUG=

Review URL: https://codereview.chromium.org/558193003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24163 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/596783002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24161 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r24133.   On X87 this check should be unnecessary too.

original commit message:
    Removed superfluous deopt for ROR on ia32.

    We only synthesize a ROR when we have an expression with a toplevel
    '|', which returns a signed value by definition, so deopting on
    negative values is not needed. ia32 was the only platform where this
    was done, all other didn't have the check.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/589383003

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24160 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24156 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TBR=tandrii@chromium.org
TEST=script_test.py

Review URL: https://codereview.chromium.org/594773003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24154 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24153 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

A new message was always generated if there is a the top-most verbose
TryCatch, even when rethrowing an exception from a TryCatch that is going
out of scope, and we already have a message.

BUG=v8:3583
LOG=Y
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/587703002

Patch from Jens Widell <jl@opera.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24152 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Forgot a few kNear occurences, only happens with --deopt-every-n-times.
Death to manual jump distance calculation! :-P

TBR=jarin@chromium.org

Review URL: https://codereview.chromium.org/598573002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24151 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/594823002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24148 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=bmeurer@chromium.org
TEST=compiler-unittests/JSBuiltinReducerTest.MathMax

Review URL: https://codereview.chromium.org/598523002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24147 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The number of different detail strings has been reduced a bit without
any loss of information. Furthermore, DoubleToI is a bit more
informative now. Tiny reason fixes.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/591393002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24146 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

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

Review URL: https://codereview.chromium.org/598513002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24145 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=bmeurer@chromium.org
TEST=compiler-unittests/JSBuiltinReducerTest.MathImul

Review URL: https://codereview.chromium.org/591373003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24144 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

- Remove an unused feature that allowed to check for required
  data before each script step.
- Use a relative path to the version file. In the production
  environment, this will point to the cwd/version_file, while
  in the test environment it is fake_cwd/version_file.

TBR=tandrii@chromium.org

Review URL: https://codereview.chromium.org/591783003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24140 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=ishell@chromium.org

Review URL: https://codereview.chromium.org/592113002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24139 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The code for EnsureSpaceForLazyDeopt is taken from lithium-codegen-*.

BUG=
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/562033003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24138 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=svenpanne@chromium.org
TEST=cctest/test-js-typed-lowering/BuiltinMathMax

Review URL: https://codereview.chromium.org/590993003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24135 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This is based on a regular expression matching anything in a namespace
ending in test, or a method containing the words ForTest, ForTesting,
for_test, or for_testing.

It is possible to blacklist entire directories or individual files.

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

Review URL: https://codereview.chromium.org/589123002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24134 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

We only synthesize a ROR when we have an expression with a toplevel
'|', which returns a signed value by definition, so deopting on
negative values is not needed. ia32 was the only platform where this
was done, all other didn't have the check.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/597503003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24133 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/595773002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24130 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r24123.  R24123 does not cover all required code change in X87.

original commit message:
   Further improve deopt reason output.
   * Make the detailed deopt reason mandatory on x64, other platforms will
     follow in separate CLs.

   * Extracted and improved jump table entry sharing logic: When --trace-deopt
     is on, we get separate entries for different deopt reasons. This enables
     us to distinguish the several reasons single instructions can have.

   * Don't emit superfluous jump table comments: The bailout ID is still visible,
     and the jump table entry number is not interesting (but easy to determine
     if really needed).

   * Unify the internal name of the jump table member across platforms.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/589313002

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24129 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r24120.

original commit message:
   Make KeyedLoads from a sloppy arguments array use a handler.
   Before, a custom stub was installed

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/594773002

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

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24128 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=yangguo@chromium.org
BUG=

Review URL: https://codereview.chromium.org/578263004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24127 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Port r24120 (eed5963)

Original commit message:
Before, a custom stub was installed.

BUG=
R=paul.lind@imgtec.com

Review URL: https://codereview.chromium.org/594563003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24126 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=416449
LOG=y
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/592813002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24125 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The synthetic roots exist in every snapshot and there are no actual objects corresponding to them (there was attempt to create some virtual objects to reuse existing snapshot builder but it just complicated the things).

Previously synthetic root wouldn't have been added if it was empty. Current implementation always adds all synthetic roots in the snapshot no matter if they are empty or not.

BUG=None
LOG=N
R=alph@chromium.org, mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/590833002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24124 ce2b1a6d-e550-0410-aec6-3dcde31c8c00