AuroraMiddleware/v8
1
0
Fork 0
Code Issues 2 Pull Requests Releases Wiki Activity
2,341 Commits 1 Branch 0 Tags 839 MiB
35cf7a24ff
Commit Graph

792 Commits

Author SHA1 Message Date
whesse@chromium.org
35cf7a24ff Recommit coderanges putting code objects within a 2 GB range, reserving only a 256 MB range of virtual memory for the code range. 
Review URL: http://codereview.chromium.org/243087

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3018 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-10-05 11:16:25 +00:00
whesse@chromium.org
bd1378eb20 Revert revisions 3013, 3014, and 3016. We need a better solution. 
Review URL: http://codereview.chromium.org/251088

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3017 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-10-05 10:50:55 +00:00
whesse@chromium.org
423de4d776 Add CHECK_INT64_EQ function to avoid operand size ambiguities. 
Review URL: http://codereview.chromium.org/256048

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3016 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-10-03 13:46:22 +00:00
antonm@chromium.org
359ad56264 Fast fix to make tree green. 
Review URL: http://codereview.chromium.org/256041

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3015 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-10-02 18:49:23 +00:00
antonm@chromium.org
8925e8af0a Add a method to convert unsigned C integer into V8 Integer. 
Review URL: http://codereview.chromium.org/260002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3013 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-10-02 15:51:07 +00:00
christian.plesner.hansen@gmail.com
2c51696fc3 Changed Object.keys to return strings for element indices. 
Review URL: http://codereview.chromium.org/246077


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3012 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-10-02 13:43:16 +00:00
erik.corry@gmail.com
4b19c36068 When allocation is forced because we already did two GCs we need to force GCs 
even if we are attempting to allocate in young space.  There were a few cases
where this wasn't done.  Also misc. changes to make diagnosis of errors like
this one easier.
Review URL: http://codereview.chromium.org/251072

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3010 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-10-02 13:35:37 +00:00
ager@chromium.org
e9acdc7a00 Follow the spec in disallowing function declarations without a name. We 
used to allow these for compatibility, but both Safari and Firefox now
disallow them.
Review URL: http://codereview.chromium.org/242124

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3009 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-10-02 12:47:15 +00:00
whesse@chromium.org
1350d0cfb9 Revert change r3004, issue http://codereview.chromium.org/244022, because Linux 64-bit Chrome crashes with more than 10 tabs. Linux may not like 10 processes, each reserving 2 GB of virtual address space. 
Review URL: http://codereview.chromium.org/246064

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3006 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-10-01 15:18:05 +00:00
erik.corry@gmail.com
f2de3fd6d2 Fix the stack limits setting API so it is usable. 
Review URL: http://codereview.chromium.org/242074

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3005 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-10-01 10:33:05 +00:00
whesse@chromium.org
48e6d4706f Allocate all executable code within a 2 GB code range. 
Review URL: http://codereview.chromium.org/244022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3004 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-10-01 09:29:38 +00:00
kmillikin@chromium.org
fb038bf146 Two changes, one a refactoring and one that affects V8's JS semantics. 
1. Change the AST node type CallNew to be a subclass of Expression
   rather than Call.  It's not really a call but it just happens to
   have the same fields.

2. Change our error reporting for invalid left-hand sides in for-in
   statements, pre- and postfix count expressions, and assignments.
   Before we signaled a syntax error at compile time *unless* the LHS
   was a function call or 'new' expression, in which case we signaled
   a reference error at runtime.  Now we signal a reference error at
   runtime in all cases.  This matches the JSC behavior in Safari 4.

Review URL: http://codereview.chromium.org/249039

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2994 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-30 09:49:36 +00:00
sgjesse@chromium.org
70821329a5 Return the original value when assigning to a pixel array. 
After fast-case assignment to a pixel array the original value assigned is now returned. Before the un-tagged smi value was returned causing crashes.

BUG=22913
TEST=cctest/test-api/PixelArray
Review URL: http://codereview.chromium.org/248033

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2985 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-29 14:56:24 +00:00
erik.corry@gmail.com
ae882b44e7 * Fix memory leaks caused by thread local data being lost. 
* Rename some instance variables and accessors to fit code style.
* Don't overwrite existing thread ID.
Review URL: http://codereview.chromium.org/251014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2977 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-28 12:25:21 +00:00
mikhail.naganov@gmail.com
1ef8754597 Heap profiler: correctly determine equivalence of objects having self-refs. 
Review URL: http://codereview.chromium.org/242031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2972 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-28 09:05:06 +00:00
mikhail.naganov@gmail.com
f1a89f3379 Heap profiler: aggregate retainers count of equivalent clusters. 
Also perform some refactoring.

Review URL: http://codereview.chromium.org/247001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2971 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-28 07:12:39 +00:00
christian.plesner.hansen@gmail.com
319097f255 Fixed some cctest flakiness on mac. 
Review URL: http://codereview.chromium.org/242003


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2968 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-25 10:36:00 +00:00
mikhail.naganov@gmail.com
9b4c950963 Heap profiler: count the number of back references for objects. 
Also, perform some refactoring to reuse common code between constructor and retainer profiles.

Review URL: http://codereview.chromium.org/209028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2936 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-18 12:05:18 +00:00
ager@chromium.org
fbc89caf3f Remove failure expectation for mozilla test that now passes. 
TBR=christian.plesner.hansen
Review URL: http://codereview.chromium.org/213005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2914 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-17 11:56:58 +00:00
mikhail.naganov@gmail.com
abc0bd46f6 Add initial version of retainers heap profile. 
The profile is taken together with constructors profile. In theory, it
should represent a complete heap graph. However, this takes a lot of memory,
so it is reduced to a more compact, but still useful form. Namely:

 - objects are aggregated by their constructors, except for Array and Object
   instances, that are too hetereogeneous;

 - for Arrays and Objects, initially every instance is concerned, but then
   they are grouped together based on their retainer graph paths similarity (e.g.
   if two objects has the same retainer, they are considered equal);

Review URL: http://codereview.chromium.org/200132

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2903 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-16 13:41:24 +00:00
sgjesse@chromium.org
40471b41da Fix GC bug and ARM simulator timeout. 
In the Runtime_DebugGetPropertyDetails the raw object pointers from a LookupResult could be used after a GC might have happened. Fixed the bug and restructured the code to make it less likely for changes to the code to re-introduce the bug.

Skipped a long running test from the ARM simulator in debug mode (and renamed the test).
Review URL: http://codereview.chromium.org/204039

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2902 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-16 13:09:26 +00:00
whesse@chromium.org
412a146825 X64: Ensure that unary subtraction returns a zero-extended smi, if it returns a smi. 
Review URL: http://codereview.chromium.org/195101

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2900 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-16 12:40:15 +00:00
sgjesse@chromium.org
677c043fee Handle array construction on native code. 
The construction of arrays when using the the Array function either as a constructor or a normal function is now handled fully in generated code in most cases. Only when Array is called with one argument which is either negative or abowe JSObject::kInitialMaxFastElementArray (which is currently 1000) or if the allocated object cannot fit in the room left in new space is the runtime system entered.

Two new native code built-in functions are added one for normal invocation and one for the construct call. The existing C++ builtin is renamed, but kept. When the normal invocation cannot be handled in generated code the C++ builtin is called. When the construct invocation cannot be handled in native code the generic construct stub is called (which will end up in the C++ builtin through a construct trampoline).

One thing that might be changed is preserving esi (constructor function) during the handling of a construct call. We know precisily what function we where calling anyway and can just reload it. This could remove the parameter construct_call to ArrayNativeCode and remove the handling of this from that function.

The X64 and ARM implementations are not part of this changelist.
Review URL: http://codereview.chromium.org/193125

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2899 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-16 11:17:57 +00:00
christian.plesner.hansen@gmail.com
5c3fd26620 [ES5] Made properties of the arguments array enumerable. 
Review URL: http://codereview.chromium.org/200141


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2897 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-16 08:25:54 +00:00
christian.plesner.hansen@gmail.com
0c5b0dbed0 Fixed typo in mozilla expectations file. 
Review URL: http://codereview.chromium.org/206014


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2894 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-15 14:36:29 +00:00
christian.plesner.hansen@gmail.com
c209d82788 Fixed mozilla test expectations. 
Review URL: http://codereview.chromium.org/196121


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2893 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-15 14:02:31 +00:00
christian.plesner.hansen@gmail.com
6519951e06 Added test suite adapter for es5conform. 
Review URL: http://codereview.chromium.org/193112


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2892 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-15 13:30:46 +00:00
christian.plesner.hansen@gmail.com
77ff957f58 Implemented Object.keys. 
Review URL: http://codereview.chromium.org/201114


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2890 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-15 11:51:40 +00:00
christian.plesner.hansen@gmail.com
681e67b775 Make 'hidden' the default visibility for gcc. Add build option, 
visibility=[hidden|default], that controls visibility and make
'hidden' the default.  Export a few variables that had been forgotten.

Review URL: http://codereview.chromium.org/206011


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2884 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-15 11:11:09 +00:00
sgjesse@chromium.org
34d6ff7606 Implemented missing pieces of the debugger for ARM. 
The main piece of this change was to add support for break on return for ARM. On ARM the normal js function return consist of the following code sequence.

  mov sp, fp
  ldmia sp!, {fp, lr}
  add sp, sp, #4
  bx lr

to a call to the debug break return entry code using the following code sequence

  mov lr, pc
  ldr pc, [pc, #-4]
  <debug break return entry code entry point address>
  bktp 0

The values of Assembler::kPatchReturnSequenceLength and Assembler::kPatchReturnSequenceLength are somewhat misleading, but they fit the current use in the debugger. Also Assembler::kPatchReturnSequenceLength is used in the IC code as well (for something which is not related to return sequences at all). I will change that in a separate changelist.

For the debugger to work also added recording of the return sequence in the relocation info and handling of source position recording when a function ends with a return statement.

Used the constant kInstrSize instead of sizeof(Instr).

Passes all debugger tests on both simulator and hardware (only release mode tested on hardware).
Review URL: http://codereview.chromium.org/199075

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2879 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-14 06:57:24 +00:00
lrn@chromium.org
158dcbc39d X64: Extract all smi operations into MacroAssembler macros. 
First step in changing Smi representation.

Review URL: http://codereview.chromium.org/196077


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2869 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-10 12:55:27 +00:00
erik.corry@gmail.com
adaab82197 Compile fixes for ARM and miscellaneous spolling. 
Review URL: http://codereview.chromium.org/199056

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2853 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-09 09:35:34 +00:00
christian.plesner.hansen@gmail.com
d391c5491e Made function prototype property DontEnum for compatibility with 
Safari.  Fixed const error in api.

Review URL: http://codereview.chromium.org/200056


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2852 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-09 09:30:33 +00:00
yurys@chromium.org
bbc3c97c2c Support stepping out for recursive functions.Simply flooding JS function from the calling stack frame with one shot breakpoints is not enough to support step out action in all cases since the function on top of the stack may be turn recursive and we may end up flooding itself. To overcome this a pointer to the stack frame where the debugger should be invoked after stepping out is strored in the debugger.Chromium bug:http://code.google.com/p/chromium/issues/detail?id=17967 
Review URL: http://codereview.chromium.org/200041

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2850 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-09 08:40:59 +00:00
yurys@chromium.org
ae062d5df1 Added new scope type ScopeIterator::ScopeTypeCatch for with generated from catch blocks. 
Removed a false assertion in ScopeIterator that assumed context extension to never be a JSContextExtensionObject.

The context extension object in a 'with' context is JSContextExtensionObject iff the 'with' statement is generated from a catch block in order to extend its local scope with a variable holding exception object. This is how we differentiate 'catch' scope from 'with' scope.

Chrome bug:
http://code.google.com/p/chromium/issues/detail?id=17229
Review URL: http://codereview.chromium.org/202005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2843 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-08 10:20:28 +00:00
kasperl@chromium.org
6621a43833 Add regression test case for http://crbug.com/18639 which 
was fixed in r2642.
Review URL: http://codereview.chromium.org/192037

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2840 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-08 07:22:35 +00:00
sgjesse@chromium.org
e26c1c8d52 Use more socket ports in debugger agent test. 
This is an attempth to address the flakiness of the test cctest/test-debug/DebuggerAgent on the Mac build-bot.
Review URL: http://codereview.chromium.org/200037

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2839 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-07 13:24:43 +00:00
yurys@chromium.org
ab54e95377 Support stepping in functions called using CallFunction stub. When Debug::PrepareStep is called to prepare 'step in' and current code target is CallFunction stub, the debugger will find function being called on the expression stack and flood it with one shot breakpoints. 
BreakLocationIterator changed to treat 'debugger;' statements as a possible break location. Since 'debugger;' statement should always invoke debugger it is hanled in a special way.

Related Chromium issue:
http://code.google.com/p/chromium/issues/detail?id=17978
Review URL: http://codereview.chromium.org/195015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2830 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-07 07:20:05 +00:00
mikhail.naganov@gmail.com
65dcf75c2c Linux profiler: check whether signal handler is called in the VM thread. 
I have several Chromium's core files having SIGPROF signal handler called in the context of an arbitrary thread, causing a crash. This change introduces checking of current thread in the signal handler.

Review URL: http://codereview.chromium.org/171115

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2825 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-04 11:16:51 +00:00
lrn@chromium.org
643d5cde71 Marked Mozilla Date test as flaky. 
Seems to err in a few cases, depending on the current time.

Review URL: http://codereview.chromium.org/192022


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2822 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-04 08:09:58 +00:00
ager@chromium.org
d7dcf8d633 Add verification flags to debug mjsunit test runs. 
Review URL: http://codereview.chromium.org/186006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2816 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-02 12:40:00 +00:00
lrn@chromium.org
11ef516137 ARM RegExp fix bug 432. 
Review URL: http://codereview.chromium.org/186005


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2814 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-02 09:10:49 +00:00
ager@chromium.org
74241d5116 Change ASSERT to CHECK in test-heap. Use int instead of intptr_t to 
make MacOS compiler happier.

We need to clean this stuff up.

TBR=whesse@chromium.org
Review URL: http://codereview.chromium.org/182044

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2802 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-01 15:51:51 +00:00
ager@chromium.org
41722bc6bc Change cache test to only test for correct behavior on collisions and 
not the exact value.

On MacOS, there is a difference in the last decimal digit from the
other platforms.

TBR=erik.corry@gmail.com
Review URL: http://codereview.chromium.org/180062

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2801 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-01 15:40:19 +00:00
ager@chromium.org
876049ba16 Change a cast to make MacOS compiler happy. 
We need to find a nicer way to handle these.

TBR=whesse@chromium.org
Review URL: http://codereview.chromium.org/182043

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2800 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-01 15:28:31 +00:00
ager@chromium.org
d8af3528dd Fix the handling of termination exceptions thrown when creating error 
objects in the runtime system.
Review URL: http://codereview.chromium.org/179062

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2796 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-01 13:55:45 +00:00
whesse@chromium.org
13d59a069d Add explicit integer type-casts to make WIN64 build without errors. 
Review URL: http://codereview.chromium.org/178054

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2795 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-01 13:08:16 +00:00
erik.corry@gmail.com
09f65846d0 Use more sane inputs to get more consistent answers from 
transcendental functions.
Review URL: http://codereview.chromium.org/173644

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2794 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-01 11:47:08 +00:00
erik.corry@gmail.com
be70d6fc4d Cache the results of slow math operations on machines that don't 
have hardware fpu instructions to execute them.
Review URL: http://codereview.chromium.org/179059

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2790 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-09-01 09:03:58 +00:00
ager@chromium.org
90abdd285a Mark native regexp macro assembler tests as flaky on ARM until Lasse 
can look at them.

The code seems to work on all of our tests.

TBR=lrn@chromium.org

Review URL: http://codereview.chromium.org/179040

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2788 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2009-08-31 16:24:44 +00:00