This fixes V8 issue 989.
Before, assignments used the AST ID of the assignment expression to
mark the side effect of the store, which became a target for
deoptimization bailout for code after the assignment. In effect
contexts this environment included the value of the assignment, which
was unexpected by the unoptimized code.
Now we introduce a new assignment ID for AST node types that include
an assignment (Assignment, CountOperation, and ForInStatement) and use
it for the side effect of the store.
Review URL: http://codereview.chromium.org/5682010
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5990 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
SplitBetween (formely known as Split with 3 arguments) should select split position from [start, end] instead of [start, end[. This should also improve allocation quality (remove certain redundant move patterns).
Also some minor renaming and refactoring to make register allocator code more readable.
BUG=v8:962
TEST=test/mjsunit/regress/regress-962.js
Review URL: http://codereview.chromium.org/5720001
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5969 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
see http://code.google.com/p/v8/issues/detail?id=944
This patch makes DateParser::TimeComposer process times that have
millisecond values with only 1 or 2 digits.
Without this patch, Date.parse("2010-11-25T22:02:30.5") returns
1290690150005 and
Date.parse("2010-11-25T22:02:30.5") == Date.parse("2010-11-25T22:02:30.005")
evaluates to true.
With this patch, Date.parse("2010-11-25T22:02:30.5") returns
1290690150500 instead, and
Date.parse("2010-11-25T22:02:30.5") == Date.parse("2010-11-25T22:02:30.005")
evaluates to false.
Review URL: http://codereview.chromium.org/5336005
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5893 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Object.freeze and Object.seal uses GetOwnPropertyNames which
returns an array with local property names. This array will also have
the additional properties defined on Array.prototype or
Object.prototype.
Note that, the implementation of GetOwnPropertyNames (when used as
Object.getOwnPropertyNames) is correct, since the spec says to create
a new Array (which would also have these properties).
Review URL: http://codereview.chromium.org/3137041
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5350 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This adds a check to the fast case string add to ensure that the String object still have the default valueOf function. The default valueOf is sitting on a hidden prototype of String.prototype.
Before using the fast case valueOf the object is checked for a local valueOf property. For slow case objects this check always reports true (the dictionary is not probed, so valueOf might be there) and for fast case objects the descriptor array is checked for the valueOf symbol (just liniar scan). After that the prototype is checked for beeing the initial value of String.prototype. If this all pass (that is the default valueOf is still in place) this result is cached on the map making the check fast the next time.
This is only implemented in the optimizing compiler, as the two usages of %_IsStringWrapperSafeForDefaultValueOf is never hit by the full compiler.
I will port to x64 and ARM when this has been reviewed for ia32.
I will remove the performance counters prior to final commit.
BUG=http://code.google.com/p/v8/issues/detail?id=760
TEST=test/mjsunit/regress/regress-760-1.js
TEST=test/mjsunit/regress/regress-760-2.js
Review URL: http://codereview.chromium.org/3117006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5252 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The existing version will overwrite the existing writable flag with
false even in the case where no new value is given and the existing
writable flag is true.
The reason for the issue is that there is no check to see if the
provided descriptor actually has a writable attribute. This causes us
to use the default value (false) even in the case where nothing was
provided. In addition, the existing tests makes wrong assumptions (that writable is always set to false if not provided) and has been changed to follow the specification.
Review URL: http://codereview.chromium.org/2271001
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4720 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Object.defineProperty with empty property descriptor.
The issue is fixed by implementing step 5 and 6 from DefineOwnProperty in the
specification (ES5 8.12.9).
This also fixes a bug in SameValue when used on boolean values (it
would priorly return a number - not a boolean).
Review URL: http://codereview.chromium.org/2131019
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4708 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This change allows Object.create to take a function as input in the
proto argument. The spec says that Type(O) for the proto argument
should be true but our IS_OBJECT does not check for this. I will make a
IS_SPEC_OBJECT in macros.py and refactor v8natives in another CL.
Review URL: http://codereview.chromium.org/2051001
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4622 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The error was introduced in revision 4557 where support was added for
ES5 date time format strings. Because there was no check for a valid
year a random string starting with a non-digit character would be
parsed.
This change disallows ES5 formatted dates where there is no date
fraction (i.e., with only a timestamp). Since none of the other
browsers support Date.parse on only timestamps I have disabled this
totally instead of just correcting the parser.
Review URL: http://codereview.chromium.org/2017005
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4613 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
We need to be careful to check global property cells for the property
encountered during lookup. Therefore, the ICs have to be specific to
the name of the property if global objects are involved. In
principle, this means that we could get a large number of monomorphic
ICs for the same map if there is a global object in the prototype
chain. However, since this is only done for normal load ICs and not
for keyed load ICs I do not expect this to be a problem. I will
experiment with it once this goes in.
BUG=675
Review URL: http://codereview.chromium.org/1559033
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4426 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This re-applies r4220 and r4233, which was reverted in r4254 due to a bug. This bug has now been fixed, with the only change being line 2884 changed from
__ SmiTag(left_side->reg());
to
__ SmiTag(operand->reg());
Added a regression test.
BUG=http://crbug.com/39160
TEST=test/mjsunit/regress/regress-crbug-39160.js
Review URL: http://codereview.chromium.org/1251009
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4261 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
When Array(0) or new Array(0) is handled by the generated code it is handled
by the same code as Array() and new Array(). For this to work the stack is
tweaked to remove the argument of value 0. However the argc was still passed
as 1 if a call to the runtime system was made.
When the stack is tweaked argc is also changed to 0.
BUG=634
TEST=test/mjsunittest/mjsunit/regress/regress-634.js
Review URL: http://codereview.chromium.org/668155
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4038 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Fix for issue 603.
Revision r3484 removed the property name from the call stack for
call ICs. When a non-function was called via a call IC and
Function.prototype.call, an extra value was left on the stack that the
caller could not know to clean up.
Fix is to change the JS builtin used for calling non-functions. It
now gets the callee as receiver, rather than iterating stack frames
and finding it on the expression stack of its JS caller.
Review URL: http://codereview.chromium.org/604064
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3882 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
When functions only have simple assignments of the form this.x = ... the object is created in generated code without actually calling the constructor. In this case the initial map for the function already contains the properties assigned in the constructor. The field descriptors in this initial map now has an enumeration index assigned to make property enumeration order the insertion order. The insertion order here is the order of the this.x assignments in the code.
BUG=http://crbug.com/3867
TEST=test/mjsunit/regress/regress-crbug-3867.js
Review URL: http://codereview.chromium.org/566016
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3768 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
When a function is called with a value type as the receiver this is now boxed as an object.
This is a low-impact solution where the receiver is only boxed when required. For IC calls to the V8 builtins values are not boxed and as most of the functions on String.prototype, Number.prototype and Boolean.prototype are sitting there most IC calls on values will not need any boxing of the receiver.
For calls which are not IC calls but calls through the CallFunctionStub a flag is used to determine whether the receiver might be a value and only when that is the case will the receiver be boxed.
No changtes to Function.call and Function.apply - they already boxed values. According to the ES5 spec the receiver should not be boxed for these functions, but current browsers have not adopted that change yet.
BUG=223
TEST=test/mjsunit/value-wrapper.js
TEST=test/mjsunit/regress/regress-crbug-3184.js
Review URL: http://codereview.chromium.org/542087
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3617 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
On 32-bit the maps are now aligned on a 32-byte boundary in order to encode more maps during compacting GC. The actual size of a map on 32-bit is 28 bytes making this change waste 4 bytes per map.
On 64-bit the encoding for compacting GC is now using more than 32-bits and the maps here are still pointer size aligned. The actual size of a map on 64-bit is 48 bytes and this change does not intruduce any waste.
My choice of 16 bits for kMapPageIndexBits for 64-bit should give the same maximum number of pages (8K) for map space. As maps on 64-bit are larger than on 32-bit the total number of maps on 64-bit will be smaller than on 32-bit. We could consider raising this to 17 or 18.
I moved the kPageSizeBits to globals.h as the calculation of the encoding really depended on this.
There are still an #ifdef/#endif in objects.h and this constant could be moved to globaks.h as well, but I kept it together with the related constants.
All the tests run in debug mode with additional options --gc-global --always-compact as well (except for a few tests on which also fails before this change when run with --gc-global --always-compact).
BUG=http://code.google.com/p/v8/issues/detail?id=524
BUG=http://crbug.com/29428
TEST=test/mjsunit/regress/regress-524.js
Review URL: http://codereview.chromium.org/504026
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3481 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
We did not handle the case where the left-hand-side expression was
fully compiled to control flow. There were also some assertions for
unary and binary expressions that crashed debug builds when the
expression was fully compiled to control flow.
Regression test added.
Review URL: http://codereview.chromium.org/160006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2524 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
With the new representation of the global object adding JavaScript accessors for a property after global inline caches was created for that property did not work property as the inline caches did not take the JavaScript accessor information (fixed array with two elements) that could be present in a global object property cell into account.
This is now fixed by changing the map for a global object when a JavaScript accessor is defined on it.
BUG=394
TEST=test\mjsunit\regress\regress-394.js
Review URL: http://codereview.chromium.org/150162
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2317 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The generic step-in mechanism floods the function called with break points to ensure a break is hit when entering the function. This generic mechanism was also used for function.apply. The code for function.apply contains a keyed load IC which was patched when stepping into function.apply. However function.apply enteres an internal frame not a JavaScript frame. This caused the logic for returning from the break in function.apply to fail as it forced a jump to the IC on the top JavaScript frame. The top JavaScript frame was the frame for the function calling function.apply not the frame for the apply function. Now returning from the break point in the keyed load IC in the apply code caused a jump to the code for the call IC for the function calling function.apply in the first place. Not a pretty sight.
Step-in now handles function.apply as a separate case where the actual JavaScript function called through apply is flodded with breakpoints instead of the function.apply function.
BUG=269
BUG=8210@chromium.org
Review URL: http://codereview.chromium.org/63055
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1683 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
an object that holds a setter. If there are no store ics then no
flushing is done. The implementation has been tweaked so that no ICs
are cleared during normal context creation.
This may cost us some performance but I'm submitting it as it is and
if there are problems we can either decide to be smarter about when,
what and/or how we clear, or back this change out altogether.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1509 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Moved the registrered debug event listener from the context to a global handle in the Debugger class. Storing it in the context did not make much sense.
Changed a lot of tests to handle the API change.
BUG=1242707
Review URL: http://codereview.chromium.org/19753
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1212 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Allows backtracking to clear registers instead of pushing and popping
them to restore state.
Redo of 1135 with bug fixed.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1156 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Reduces number of pushes when flushing a trace. Some are converted to clears
in the undo-code instead, and some just ignored if they have no value worth restoring.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1136 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
it is assigned a numeric handle. Handles are used to make a 1:1
correspondence between objects and mirrors. Currently the mirrors are
cached in a JavaScript array and when creating a mirror this cache is
checked to see if a mirror already exists for the object. This cache is
cleared when leaving the debugger.
Changed the serialization format to take advantage of these handles. When
an object is serialized referenced objects are represented just by their
handle id serialized as '{ref:<handle>}'. During serialization the
referenced handles are collected and the serializer can provide a
serialization of all the referenced objects.
Removed the special handling of array properties. Indexed properties and
the length property are now rendered as named properties in the
serialization.
Removed the special serialization handling of RegExp properties. The
properties 'source', 'global', 'ignoreCase' and 'multiline' are serialized
with the rest of the properties.
Changed a lot of tests to handle the changed format.
Review URL: http://codereview.chromium.org/18092
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1093 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
instead of normal JSObjects.
This ensures that __proto__ and accessors on the Object prototype do
not interfere with catch scopes. Also, it fixes the bug that catch
variables were not DontDelete (issue 74).
Next step is to create special lookup routines for context extension
objects and remove the special handling of context extension objects
from the general javascript object lookup routines.
Review URL: http://codereview.chromium.org/18143
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1091 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Changed the serialization format to take advantage of these handles. When an object is serialized referenced objects are represented just by their handle id serialized as '{ref:<handle>}'. During serialization the referenced handles are collected and the serializer can provide a serialization of all the referenced objects.
Removed the special handling of array properties. Indexed properties and the length property are now rendered as named properties in the serialization.
Removed the special serialization handling of RegExp properties. The properties 'source', 'global', 'ignoreCase' and 'multiline' are serialized with the rest of the properties.
Changed a lot of tests to handle the changed format.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1078 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
are always odd if the bit-field bits are included.
Modified a couple of debugger tests that relied on the ordering of
elements in descriptor arrays. Descriptor arrays are sorted by
hash-code values, so the order changes if we change the hash code.
Review URL: http://codereview.chromium.org/6489
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@440 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
All uses of %AddProperty are replaced by %SetProperty. A few uses of
IgnoreAttributesAndSetLocalProperty are added, and the runtime version
of it adds a PropertyAttributes argument. Only the JSObject class
in objects.cc now uses AddProperty, and it can become private.
Review URL: http://codereview.chromium.org/6445
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@426 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
are evaluated with an extra element on the stack, which needs
to be taken into account when breaking and continuing.
I'll clean up the code and add an abstraction for manipulating
the break stack height in a future CL -- I want to try to get
rid of the separate local variable we keep around for the "state"
when running in a finally block.
Review URL: http://codereview.chromium.org/5625
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@400 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
- Added option to test runner that allows us to run the tests under
valgrind.
- Added test status for the failing arm simulator tests.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@116 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Added better test support.
Added load, quit and version functions to the shell sample so it's easier to run benchmarks and tests.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
In the shell sample don't print the result of executing a script, only
evaluating expressions.
Fixed issue when building samples on Windows using a shared V8
library. Added visibility option on Linux build which makes the
generated library 18% smaller.
Changed build system to accept multiple build modes in one build and
generate seperate objects, libraries and executables for each mode.
Removed deferred negation optimization (a * -b => -(a * b)) since this
visibly changes operand conversion order.
Improved parsing performance by introducing stack guard in preparsing.
Without a stack guard preparsing always bails out with stack overflow.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
