BUG=74627

Review URL: http://codereview.chromium.org/6588116

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7018 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Build all loop graphs without using class HSubgraph.  This also eliminates a
silly goto for the update expression of a for loop.

Support for peeling loops is removed because it's currently untested.

Review URL: http://codereview.chromium.org/6602031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7017 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

It avoids strange performance behavior where hot functions do not
get optimized if there is no GC to promote them, and it does not seem
to hurt us otherwise.

Review URL: http://codereview.chromium.org/6594073

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7016 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6596104

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7015 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TEST=none
BUG=none

Patch by Rodolph Perfetta from ARM Ltd.

Review URL: http://codereview.chromium.org/6594009


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7014 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This fixes issue 1116 and enables the stub in lithium codegen.

Review URL: http://codereview.chromium.org/6588113

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7013 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Patch by Mark Lam from Hewlett-Packard Development Company, LP

Review URL: http://codereview.chromium.org/6357005


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7012 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

LiveObjectList functionality.

Patch by Mark Lam from Hewlett-Packard Development Company, LP

Review URL: http://codereview.chromium.org/6351007


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7011 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

On IA32 we use InstanceofStub::left(), but on x64 we have just used rax and rdx directly. This fixes the assertion crashes on x64 crankshaft.


Review URL: http://codereview.chromium.org/6596101

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7010 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

the Windows WebCore build.

http://build.chromium.org/p/client.v8/builders/Webkit/builds/854/steps/compile/logs/stdio

TBR=cira@chromium.org

Review URL: http://codereview.chromium.org/6596100

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7009 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

If this does not work I will revert the latests changes to the
experimental i8n extension.

TBR=cira@chromium.org

Review URL: http://codereview.chromium.org/6594111

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7008 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6594037/

Revert "Revert "Assignment to read only properties throws in strict mode.""

This reverts commit aefcd82e1d36d458dd071ebf4777340f08aa67b1.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7007 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This is vendor specific, and is prefixed by v8.

WebKit layout tests will be added in a separate CL.

Review URL: http://codereview.chromium.org/6598014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7006 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

We simply need to remove early checks in the resolve eval runtime
functions. CallFunctionStub that follows will handle non-functions in
the right way.

Review URL: http://codereview.chromium.org/6591075

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7005 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

into heap snapshots non-HeapObjects. This is needed as a
preparation for adding DOM subtrees tracking.

BUG=none
TEST=none

Review URL: http://codereview.chromium.org/6596073

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7004 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6594076

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7003 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6591070

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7000 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This fixes an assert when an exception is thrown inside instanceof.

BUG=v8:1207
TEST=mjsunit/regress/regress-1207.js

Review URL: http://codereview.chromium.org/6588083

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6999 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6598066

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6998 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Patch by Mark Lam from Hewlett-Packard Development Company, LP

Review URL: http://codereview.chromium.org/6602034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6997 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6588082

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6996 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6594074

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6995 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6594071

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6994 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Added HandleScope::Escape to HandleScope to allow exiting a value
from a scope.

Review URL: http://codereview.chromium.org/6594075

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6993 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6596070

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6991 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6598063

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6989 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This was wrongly marked as skip since we do not need to patch reloc info on x64 and arm when doing deoptization (issue 1094).

Review URL: http://codereview.chromium.org/6597067

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6984 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fixed Building the simulator with CAN_USE_UNALIGNED_ACCESSES defined.

Fixed a bug in the code generator where it was doing unaliged read and write.
Review URL: http://codereview.chromium.org/6598062

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6983 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This change makes the full code generator always materialize smi constants to match the behavior
 of the environment tracked in the optimizing compiler. The mismatch between environment
and what the non-optimized code causes problems and potentially crashes when doing deoptimization
and OSR.

BUG=67732

Review URL: http://codereview.chromium.org/6592064

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6982 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

When deoptimizing from the key subexpression of a keyed arguments access,
the unoptimized code expects to find the value of the receiver on the
expression stack.  The environment of the optimizing compiler did not
contain this value during evaluation of the key subexpression.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6981 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6592038

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6980 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The counting loop is always run at least once, and checking the loop condition once is sufficient.
Review URL: http://codereview.chromium.org/6592010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6979 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

We do not currently check that a global property is actually
configurable before overwriting it with a new property.


Review URL: http://codereview.chromium.org/6597045

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6978 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This reverts commit 503f2a501e504f330821f247074e731aa649b1f0.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6977 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6594037/

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6976 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6598023/

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6975 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Using const in strict mode yields SyntaxError.

Review URL: http://codereview.chromium.org/6592031/

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6974 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6602007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6973 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The current version includes an extra jump compared to IA32, because
we need to load the jump address into a register and do an indirect
jump, but in the normal case we just jump over this by negating the
deoptimization conditional.

Review URL: http://codereview.chromium.org/6596032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6972 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

I did not use the helper function for adding this builtin function which meant that I missed the removal of the prototype property.

BUG=
TEST=

Review URL: http://codereview.chromium.org/6588050

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6971 ce2b1a6d-e550-0410-aec6-3dcde31c8c00