R=rossberg@chromium.org
TEST=mjsunit/apply,test262

Review URL: http://codereview.chromium.org/8342034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9709 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=kmillikin@chromium.org

BUG=v8:1782
TEST=mjsunit/debug-step-3.js

Review URL: http://codereview.chromium.org//8356001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9708 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

In built-in code we use arrays for internal computations. This makes it
possible to affect the built-in code by putting getters or setters on
the array prototype chain. Using internal arrays prevents those issues.

Related to: http://code.google.com/p/v8/source/detail?r=7040

R=svenpanne@chromium.org
TEST=test262/15.4.4.20-9-b-6

Review URL: http://codereview.chromium.org/8353006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9707 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Note that this is not an observable behavior, but following the principle of
least surprise, we should follow the spec. Additional (extremely tiny) bonus:
Some ICs see fewer values => better code.
Review URL: http://codereview.chromium.org/8352004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9706 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TEST=mjsunit/elements-kind passes both with and without --smi-only-arrays flag

Review URL: http://codereview.chromium.org/8356002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9705 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/8353002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9704 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This reverts commits
r9673: "Scope tree serialization and ScopeIterator cleanup."
r9674: "Use OS::SNPrintF instead of snprintf."
r9675: "Use int instead of size_t, StrLength instead of strlen."

Review URL: http://codereview.chromium.org/8353003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9703 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TEST=mjsunit/elements-kind

Review URL: http://codereview.chromium.org/8305001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9702 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=ulan@chromium.org
BUG=
TEST=

Review URL: http://codereview.chromium.org/8345038

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9701 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Port r9641 (2876c37)

Note that this port does not include crankshaft portions of this
commit. They have been included in today's update to
http://codereview.chromium.org/7934002/

Also, this fixes a changed function prototype for:
void Deoptimizer::RevertStackCheckCodeAt()

BUG=
TEST=

Review URL: http://codereview.chromium.org/8348028
Patch from Paul Lind <plind44@gmail.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9700 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Named function expression have an implicit local variable that
refers to the current function (ThisFunction). Before we only could inline
anonymous function expressions like:

A.prototype.foo = function() {}

as opposed to

A.prototype.foo = function foo() {}

This change enables inlining function of expressions like this.
Review URL: http://codereview.chromium.org/8346032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9699 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=none
TEST=test/mjsunit/array-literal.js

Review URL: http://codereview.chromium.org/8258015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9698 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

RegExpMacroAssembler::CheckStackGuardState should update input string pointer when it is moved or changed by GC.

If input string was cons-string it might undergo short-circuiting during GC. This does not change input start if underlying seq-string (first element of cons-string) does not move but this makes input-string pointer on the native regexp's frame invalid.

R=lrn@chromium.org

Review URL: http://codereview.chromium.org/8343001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9697 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=arm debug test of mjsunit/elements-transition segfaults

Review URL: http://codereview.chromium.org/8342032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9696 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/8345040

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9695 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This is not covered by test262 yet, but it really makes sense and matches Firefox's behaviour.

TEST=mjsunit/define-properties.js
Review URL: http://codereview.chromium.org/8349031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9694 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=svenpanne@chromium.org
BUG=chromium:100702
TEST=mjsunit/regress/regress-100702

Review URL: http://codereview.chromium.org/8347034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9693 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=
TEST=

Review URL: http://codereview.chromium.org/8341008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9692 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This fixes updating of property attributes for data elements when
attributes are already present on a dictionary element but get removed
by a subsequent redefinition of that element.

R=rossberg@chromium.org
BUG=v8:1772
TEST=test262/15.2.3.6-4-82-18

Review URL: http://codereview.chromium.org/8337017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9691 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/8329022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9690 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Android releases.  See http://code.google.com/p/v8/issues/detail?id=1780
Review URL: http://codereview.chromium.org/8347032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9689 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This time, we initially leave the HTypeof instruction in the Hydrogen graph,
even for the special cases. We later try to remove this instruction (and any
HConstant) in the canonicalization pass, if possible. Always removing the
HTypeof during the initial graph construction is wrong if e.g. it is used in an
HSimulate.

The removals can be generalized a bit, but this will happen in a separate CL.

TEST=mjsunit/optimized-typeof.js
Review URL: http://codereview.chromium.org/8334021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9688 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The new helper avoids expensive FindCodeForInnerPointer invocation when we have
the host code object available. It is used when patching stack checks.

Also some comments on the ARM platform are corrected.
Review URL: http://codereview.chromium.org/8330021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9687 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=erik.corry@gmail.com
BUG=v8:1773

Review URL: http://codereview.chromium.org/8337014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9686 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This fixes updating of property attributes for getters and setters on
dictionary elements while redefining. This just updates the property
details on the existing element.

R=rossberg@chromium.org
BUG=v8:1772
TEST=test262

Review URL: http://codereview.chromium.org/8337013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9685 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This is reapplying r9501 with this single change which seemed to be causing most (all) of the failures for r9501.

--- a/src/debug.cc
+++ b/src/debug.cc
@@ -2230,6 +2230,7 @@ Debugger::Debugger(Isolate* isolate)
       compiling_natives_(false),
       is_loading_debugger_(false),
       never_unload_debugger_(false),
+      force_debugger_active_(true),
       message_handler_(NULL),
       debugger_unload_pending_(false),
       host_dispatch_handler_(NULL),

R=kmillikin@chromium.org

BUG=
TEST=

Review URL: http://codereview.chromium.org//8337009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9684 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Also makes func.caller return null if the caller is a bound function,
matching JSC.
Fix bug preventing poisoned setters from triggering.

TEST=mjsunit/function-bind, mjsunit/strict-mode

Review URL: http://codereview.chromium.org/8333019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9681 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=
TEST=

Review URL: http://codereview.chromium.org/8337008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9680 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TEST=GCC 4.5.3 successfully compiles d8.cc even with -Werror enabled.

Review URL: http://codereview.chromium.org/8296018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9679 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

If a GC happened during initialization (when allocating the elements array)
of the non_strict_arguments_boilerplate, heap verification would fail with the following assert:

ASSERT_EQ((map()->has_fast_elements() || map()->has_fast_smi_only_elements()),
            (elements()->map() == GetHeap()->fixed_array_map() ||
             elements()->map() == GetHeap()->fixed_cow_array_map()));

This was not harmful since the boilerplate was setup 
correctly immediatly afterwards.


Simplified the setup code by removing a call to GetElementsTransitionMap. It always return the same map as 
the input object in this case and is therefore unnecessary.


Added more assertions to verify well-formed non-strict
arguments backing store.

BUG=v8:1520
TEST=no more flaky tests with failing this assert.
 
Review URL: http://codereview.chromium.org/8336021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9678 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Implement same on Mips.

BUG=v8:1774
TEST=mjsunit/function-bind

Review URL: http://codereview.chromium.org/8337012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9677 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The LookupResult utility class is used in handlified code, but it can
contain a raw pointer to the lookup's holder object.  Create a per-thread
stack of live LookupResults and iterate all the live ones on GC.

R=vegorov@chromium.org,erik.corry@gmail.com
BUG=
TEST=

Review URL: http://codereview.chromium.org/8341009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9676 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/8339013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9675 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/8339011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9674 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The intention is to store enough scope information for the debugger to
handle stack allocation of block scoped variables introduced by
http://codereview.chromium.org/7860045/ .

This CL is based on
http://codereview.chromium.org/7904008/ .

Review URL: http://codereview.chromium.org/7979001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9673 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/8321002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9666 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/8321001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9665 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/8318003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9664 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=lrn@chromium.org
BUG=v8:893
TEST=test262

Review URL: http://codereview.chromium.org/8320001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9663 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/8312015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9662 ce2b1a6d-e550-0410-aec6-3dcde31c8c00