I think this actually turns the code in question into dead code, because the current definition of UseSparseVariant only allows arrays in DictionaryMode, and these never return encoded ranges from %GetArrayKeys. This is a subtle link between the definition of %GetArrayKeys and UseSparseVariant, so I would prefer to keep the code rather than having an invisible depeendency.

BUG=73940

Review URL: http://codereview.chromium.org/6596004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6946 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Resubmit of patch for issue 1145 with a few additions:
- Now also clears exceptions when calling Runtime_LazyRecompile.
- Sets function where parsing fails to not be optimizable.

BUG=v8:1145
TEST=test/mjsunit/regress/regress-1145.js

Review URL: http://codereview.chromium.org/6469050

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6945 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6581036

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6940 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Check result before and bail out if exception has been thrown.

BUG=v8:1172
TEST=test/mjsunit/regress/regress-1172-bis.js

Review URL: http://codereview.chromium.org/6580030

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6939 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Direct call accessor getter callbacks (arm implementation).

Original review: http://codereview.chromium.org/6462029/

Review URL: http://codereview.chromium.org/6576035

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6938 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The loading the register used for the null object was to early and it could be clobbered.

TBR=karlklose@chromium.org
BUG=none
TEST=mozilla/ecma_3/Statements/regress-131348
Review URL: http://codereview.chromium.org/6578020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6937 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6576030

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6936 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6576033

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6935 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Also fix issue 1175 and 1177.

BUG=v8:1175

Review URL: http://codereview.chromium.org/6568007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6934 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

r5445: Implement for-in cache validity checking in the full codegen on IA-32.

BUG=v8:867
TEST=none
Review URL: http://codereview.chromium.org/6576028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6933 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6579036

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6932 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The existing version does not have a correct implementation of
Generate_OnStackReplacement since we do not correctly find the loop
depth. This is mainly because EmitStackCheck in full-codegen-x64.cc
did not write it and partly due to us reading at the wrong offset
(which had no effect when it was not written in the first place, we
simply got a random number from the next instruction).

The DoComputeOsrOoutputFrame is very similar to the ia32 version.



Review URL: http://codereview.chromium.org/6581028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6931 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

r5445: Implement for-in cache validity checking in the full codegen on IA-32.
Review URL: http://codereview.chromium.org/6581031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6930 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

r6635: Remove the redundant load on every context lookup.
r6659: Do not compile the unreachable body of functions with illegal redeclarations.

The crankshaft part of these changes are already ported.
Review URL: http://codereview.chromium.org/6581029

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6929 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Since writting a code comment can include a pc-jump the assertion
should just check that we are greater than or equal to
kMinRelocCommentSize (renamed from kRelocCommentSize), not that we are
actually equal to.


Review URL: http://codereview.chromium.org/6580028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6928 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=1811
TEST=test/mjsunit/regress/regress-1181.js

Review URL: http://codereview.chromium.org/6573004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6925 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6580027

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6924 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6922 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TBR=ricow@chromium.org
Review URL: http://codereview.chromium.org/6580026

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6921 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

r6635: Remove the redundant load on every context lookup.
r6659: Do not compile the unreachable body of functions with illegal redeclarations.
Review URL: http://codereview.chromium.org/6572003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6920 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Original cl:
http://codereview.chromium.org/6542066/

Landing to ensure sync with buildbot update.


Review URL: http://codereview.chromium.org/6579032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6918 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

When printing the sub-parts of a live range, only print the use intervals
belonging to each part.

Add a flag to print all use positions (incl. environment uses)


Review URL: http://codereview.chromium.org/6566005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6917 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Temporarily disable DoStoreContextSlot in x64 lithium codegen to diagnose non-reproducible buildbot error.  Fix presubmit error.
Review URL: http://codereview.chromium.org/6572002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6915 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6570008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6914 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

In a lot of cases where we were passing a subgraph (a pair of basic blocks),
we actually only needed a basic block.

Review URL: http://codereview.chromium.org/6570006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6913 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6568004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6912 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TBR=whesse

Review URL: http://codereview.chromium.org/6568005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6911 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6538080

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6910 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6532020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6909 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6567003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6908 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6570005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6907 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

entry in lithium backend.

Review URL: http://codereview.chromium.org/6566002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6906 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

We can wrongly assume that exception which is not intended to be caught
by external try/catch should be caught if this exception inherits
external catcher from some previous exception.  To prevent that,
clear external catcher when processing exceptions which cannot be
externally caught.

BUG=v8:1184
TEST=test/mjsunit/regress/regress-1184.js

Review URL: http://codereview.chromium.org/6538081

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6905 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6469083/

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6904 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Allocate the context in the prologue. Two issues had to be solved:
(1) deoptimization needs to handle functions with a local context,
(2) we need a safepoint in the prologue.
(Thanks to Kevin.)

Review URL: http://codereview.chromium.org/6534022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6903 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The main issue was due to multiple recompilations of functions.  Now
code objects are grouped by function using SFI object address.
JSFunction objects are no longer tracked, instead we track SFI object
moves. To pick a correct code version, we now sample return addresses
instead of JSFunction addresses.

tools/{linux|mac|windows}-tickprocessor scripts differentiate
between code optimization states for the same function
(using * and ~ prefixes introduced earlier).

DevTools CPU profiler treats all variants of function code as
a single function.

ll_prof treats each optimized variant as a separate entry, because
it can disassemble each one of them.

tickprocessor.py not updated -- it is deprecated and will be removed.

BUG=v8/1087,b/3178160
TEST=all existing tests pass, including Chromium layout tests

Review URL: http://codereview.chromium.org/6551011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6902 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6541069

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6897 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Analyses full minidump (.dmp) files.

Shows the processor state at the point of exception including the
stack of the active thread and the referenced objects in the V8
heap. Code objects are disassembled and the addresses linked from the
stack (pushed return addresses) are marked with "=>".

Review URL: http://codereview.chromium.org/6312058

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6896 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/6549016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6895 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add more generic version of reloc info padding to ensure enough space for reloc patching during deoptimization (fixes issue 1174).

The old version only added extra space when we did indirect calls, but
the problem remains the same with normal calls that can be represented
as a single byte. When doing patching each call will always be at
least 2 bytes long because we use RUNTIME_ENTY as the reloc mode.


Review URL: http://codereview.chromium.org/6541053

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6894 ce2b1a6d-e550-0410-aec6-3dcde31c8c00