Ported r8922 (5ea2fb5)

Original commit message:
Remove the try/finally used for with and catch. Instead of using
try/finally to handle break and continue from with or catch,
statically track nesting dept and clean up when compiling break or
continue.
And instead of using try/finally to handle throw to handler in a frame
whose pc is inside a with or catch, store the context that the handler
should run in in the handler itself.

BUG=
TEST=

Review URL: http://codereview.chromium.org/7648026
Patch from Paul Lind <plind44@gmail.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8936 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=v8:1602
TEST=new (new v8Locale().createDateTimeFormat()).getWeekdays should return Error not a Null.
Review URL: http://codereview.chromium.org/7647027

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8935 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Leaving this out can cause crashes when running with --heap-stats (but this is only used when reporting heap statistics, so this is not a stability issue).
Review URL: http://codereview.chromium.org/7647018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8932 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/7645020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8931 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

In the process, make ElementsAccessors work internally more seamlessly with FixedArrayBase.

R=svenpanne@chromium.org
BUG=none
TEST=none

Review URL: http://codereview.chromium.org/7618012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8930 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=kmillikin@chromium.org
BUG=none
TEST=none

Review URL: http://codereview.chromium.org/7617010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8927 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove the try/finally used for with and catch.  Instead of using
try/finally to handle break and continue from with or catch,
statically track nesting dept and clean up when compiling break or
continue.

And instead of using try/finally to handle throw to handler in a frame
whose pc is inside a with or catch, store the context that the handler
should run in in the handler itself.

BUG=
TEST=

Review URL: http://codereview.chromium.org/7618007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8922 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

It is available platforms that have SSE 4.1 and allows us to handle
negative numbers without deoptimization. Before we would deoptimize
on negative inputs to Math.floor. x64 already uses this instruction.

* Change Math.floor unit test to make sure every test case gets
optimized by changing the source code for each test case.

* Fix HIR debug printing for some instructions.
Review URL: http://codereview.chromium.org/7628017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8921 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8920 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

These files already include v8.h so they don't need to define the
namespace alias again.
Signed-off-by: Thiago Farina <tfarina@chromium.org>

Review URL: http://codereview.chromium.org/7640012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8919 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Non-ancient versions of the ARM-ARM explicitly deprecate most uses of the PC
within instructions and older ARM implementations have a non-predictable offset
(8 or 12) for some of these deprecated uses. Avoiding the deprecated instruction
costs us one additional instruction in DirectCEntryStub::GenerateCall, but this
should not cause any significant performance degradation.

The deoptimizer still uses the PC in a stm instruction, but it is a bit unclear
what to do about that, so simply a comment has been added to reconsider this in
the future.
Review URL: http://codereview.chromium.org/7633014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8916 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Dropping removed source file from the gyp build description.

BUG=None
TEST=None
R=iposva@chromium.org



git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8912 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=
TEST=

Review URL: http://codereview.chromium.org/7549008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8911 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/7620012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8909 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=v8:1593

Review URL: http://codereview.chromium.org/7622009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8903 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Also unify Crankshaft code to load array length.

BUG=v8:1493
TEST=external-arrays.js

Review URL: http://codereview.chromium.org/7600025

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8901 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/7623014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8900 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/7620009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8899 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/7617008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8898 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

We output line-by-line now, this avoids hitting StringBuffer ASSERTs when
disassembling huge code objects.
Review URL: http://codereview.chromium.org/7622006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8897 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/7623009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8894 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TEST=cctest/Threaded4

Review URL: http://codereview.chromium.org/7620007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8892 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/7617003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8887 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Changing our builtin JavaScript code slightly, we can make sure that we never
see internal objects as arguments for ToBoolean at runtime. Removing that case
from the stub generator and crankshaft makes things a lot easier.

Heap numbers can never be undetectable (only strings and spec objects can), so
we can leave out a useless test.

Try to re-use a non-null register value when returning 'true' in some cases.

Removed special handling of the 'handle all' case, it will very probably never
happen in real code and only makes things more complicated.

Improved naming of the ToBoolean stubs a bit, reflecting the order in which
cases are handled in the code itself.
Review URL: http://codereview.chromium.org/7497063

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8886 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/7552034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8885 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=kmillikin@chromium.org
BUG=v8:1592
TEST=mjsunit/regress/regress-1592.js

Review URL: http://codereview.chromium.org/7497067

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8884 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TEST=Arm builder on Chromium waterfall successfully compiles

Review URL: http://codereview.chromium.org/7607032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8880 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TBR=whesse@chromium.org
Review URL: http://codereview.chromium.org/7604034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8879 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

* Bug fix for range analysis (contributed by Andy Wingo). Ranges of
double values have to include negative zero. Original code review:
 http://codereview.chromium.org/7514040/

* Fix a bug in optimized Math.round on ARM. When emitting minus-zero checks
we previously return a wrong result because of incorrect register assignment.

* Fix performance problem in IA32 and x64. Refine the checks
for minus zero and avoid unnecessary deoptimizations on Math.floor.

* Improve mjsunit test for Math.round to make sure we also
 get the optimized version of the code for each test case.
Review URL: http://codereview.chromium.org/7604028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8877 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Harmony is intended to make typeof null === "null". This may
break existing programs. Implementing it will allow us to run
some tests on the actual web.

R=kmillikin@chromium.org
BUG=
TEST=

Review URL: http://codereview.chromium.org/7598030

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8876 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=danno@chromium.org
BUG=
TEST=

Review URL: http://codereview.chromium.org/7606026

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8873 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

There is currently no way of specifying that the snapshot build by scons should be created without using the vfp3 instructions.
Review URL: http://codereview.chromium.org/7604031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8872 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TEST="make -j16 check OUTDIR=whatever" works as expected

Review URL: http://codereview.chromium.org/7606025

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8871 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=none
TEST=none

Review URL: http://codereview.chromium.org/7529046

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8870 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=svenpanne@chromium.org,kmillikin@chromium.org
BUG=v8:1584

Review URL: http://codereview.chromium.org/7585006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8868 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=jkummerow@chromium.org

Review URL: http://codereview.chromium.org/7598034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8867 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Replaced the keyword matching state machine with a switch on the first char followed up by inlined char comparisons.

R=lrn@chromium.org
TEST=cctest/test-parsing/ScanKeywords

Review URL: http://codereview.chromium.org/7558017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8866 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Because these are function literals that have an associated name, we were
compiling them as if they were named function expressions.  This is
incorrect, the property name should not be in scope.

R=vegorov@chromium.org
BUG=
TEST=

Review URL: http://codereview.chromium.org/7599024

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8863 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reduced size of first-character table for keyword scanner.

Review URL: http://codereview.chromium.org/7585030

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8862 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/7600019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8860 ce2b1a6d-e550-0410-aec6-3dcde31c8c00