platform/upstream/v8.git
15 years agoFix Xcode project:
iposva@chromium.org [Sat, 17 Jan 2009 01:08:54 +0000 (01:08 +0000)]
Fix Xcode project:
- Add regexp-stack.[cpp|h]

Review URL: http://codereview.chromium.org/18344

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1100 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoReverting change that made flexo sad.
olehougaard [Fri, 16 Jan 2009 14:53:36 +0000 (14:53 +0000)]
Reverting change that made flexo sad.

TBR=ager
Review URL: http://codereview.chromium.org/18150

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1099 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoWriting snapshot.cc in a form that can be compiled by the crosstool compiler.
olehougaard [Fri, 16 Jan 2009 14:12:56 +0000 (14:12 +0000)]
Writing snapshot.cc in a form that can be compiled by the crosstool compiler.
Changed a cast that caused alignment problems on ARM.
Review URL: http://codereview.chromium.org/18312

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1098 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoPrepare push to trunk. Now working on version 0.4.9.
ager@chromium.org [Fri, 16 Jan 2009 10:09:55 +0000 (10:09 +0000)]
Prepare push to trunk.  Now working on version 0.4.9.
Review URL: http://codereview.chromium.org/18310

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1094 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdded handles to the mirror objects. When a mirror for an object is created
sgjesse@chromium.org [Fri, 16 Jan 2009 09:54:46 +0000 (09:54 +0000)]
Added handles to the mirror objects. When a mirror for an object is created
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

15 years agoChange the file names of a couple of tests for consistency.
ager@chromium.org [Fri, 16 Jan 2009 09:43:42 +0000 (09:43 +0000)]
Change the file names of a couple of tests for consistency.
Review URL: http://codereview.chromium.org/18142

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1092 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoChange the handling of catch blocks to use context extension objects
ager@chromium.org [Fri, 16 Jan 2009 09:42:08 +0000 (09:42 +0000)]
Change the handling of catch blocks to use context extension objects
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

15 years agoAvoid using sprintf_s, which is windows-only.
feng@chromium.org [Thu, 15 Jan 2009 19:42:51 +0000 (19:42 +0000)]
Avoid using sprintf_s, which is windows-only.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1089 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix a bunch of spelling mistakes :\
deanm@chromium.org [Thu, 15 Jan 2009 19:08:34 +0000 (19:08 +0000)]
Fix a bunch of spelling mistakes :\

Review URL: http://codereview.chromium.org/18094

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1088 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix issue 6264 with a test case.
feng@chromium.org [Thu, 15 Jan 2009 17:39:23 +0000 (17:39 +0000)]
Fix issue 6264 with a test case.

The problem is that Disable/EnableAccessCheck on an object may chnage its constructor's behavior
if object's map is the same as constructor's initial map. By copying maps, the constructor's
initial map is not changed.

Review URL: http://codereview.chromium.org/18067

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1087 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoLint error.
erik.corry@gmail.com [Thu, 15 Jan 2009 13:09:23 +0000 (13:09 +0000)]
Lint error.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1081 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoNoone really liked the name "GenerationVariant" so here it gets renamed
erik.corry@gmail.com [Thu, 15 Jan 2009 12:45:48 +0000 (12:45 +0000)]
Noone really liked the name "GenerationVariant" so here it gets renamed
to "Trace".
Review URL: http://codereview.chromium.org/18091

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1080 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoReverted r1078 as it was committed by accident without review.
sgjesse@chromium.org [Thu, 15 Jan 2009 12:29:05 +0000 (12:29 +0000)]
Reverted r1078 as it was committed by accident without review.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1079 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdded handles to the mirror objects. When a mirror for an object is created it is...
sgjesse@chromium.org [Thu, 15 Jan 2009 12:10:34 +0000 (12:10 +0000)]
Added handles to the mirror objects. When a mirror for an object is created 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.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1078 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix issue 191:
ager@chromium.org [Thu, 15 Jan 2009 11:31:08 +0000 (11:31 +0000)]
Fix issue 191:

    http://code.google.com/p/v8/issues/detail?id=191

Accessor setters should not be called for eval-introduced context
extension objects.

This change fixes the issue, but I think it is time to separate out
the lookup routines for eval-introduced context extension objects.
There is no reason to use the general lookup routines for this.  I'll
do that in a separate changelist.
Review URL: http://codereview.chromium.org/18090

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1076 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFixed some external references in the ARM code generator to use an ExternalReference...
olehougaard [Thu, 15 Jan 2009 08:32:50 +0000 (08:32 +0000)]
Fixed some external references in the ARM code generator to use an ExternalReference wrapper.
Review URL: http://codereview.chromium.org/18266

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1074 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix issue 186:
ager@chromium.org [Wed, 14 Jan 2009 12:13:26 +0000 (12:13 +0000)]
Fix issue 186:

    http://code.google.com/p/v8/issues/detail?id=186

Create a new instance type for context extension objects.  Use it to
not use the __proto__ accessor for context extension objects.
Review URL: http://codereview.chromium.org/18044

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1072 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFixed test failure caused by being more conservative in estimating the
christian.plesner.hansen@gmail.com [Wed, 14 Jan 2009 11:48:52 +0000 (11:48 +0000)]
Fixed test failure caused by being more conservative in estimating the
length of expressions containing backreferences.  Bad plesner for
forgetting to run the tests before submitting!

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1071 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdded clearing of captures before entering the body of a loop. This
christian.plesner.hansen@gmail.com [Wed, 14 Jan 2009 11:32:23 +0000 (11:32 +0000)]
Added clearing of captures before entering the body of a loop.  This
also revealed a bug or two that had to be fixed.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1070 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdd failing test for issue 186:
ager@chromium.org [Wed, 14 Jan 2009 09:20:13 +0000 (09:20 +0000)]
Add failing test for issue 186:

    http://code.google.com/p/v8/issues/detail?id=186

The issue is that the extension object is a normal javascript object
on which assignments to __proto__ changes the prototype chain.
Review URL: http://codereview.chromium.org/18221

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1069 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRefactored the mirror representation of properties. Removed the AssessorMirror and...
sgjesse@chromium.org [Tue, 13 Jan 2009 14:38:44 +0000 (14:38 +0000)]
Refactored the mirror representation of properties. Removed the AssessorMirror and InterceptorPropertyMirror and moved all reflection for properties to PropertyMirror. From a PropertyMirror it can be checked whether a property has getter/setter defined in JavaScript and information on the getter/setter functions are now available. If calling the getter resulted in an exception this is reflected as well.

Properties from interceptors are also reflected through PropertyMirror as the distinction did not make sense seen from a JavaScript debugging perspective. The isNative function on a PropertyMirror can be used to check whether a property is defined natively by the host (or V8).

Simplified the local property lookup in the debug runtime call to just call GetProperty as the property is known to be a local property.
Review URL: http://codereview.chromium.org/17377

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1068 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAvoiding the assumption that the pc pointer of RelocInfo points to the word containin...
olehougaard [Tue, 13 Jan 2009 14:38:12 +0000 (14:38 +0000)]
Avoiding the assumption that the pc pointer of RelocInfo points to the word containing the target address. It wasn't true for ARM. (One step closer to serialization on ARM).
Review URL: http://codereview.chromium.org/17376

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1067 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoThe "enter" opcode was slow and crashes for large arguments on Windows.
lrn@chromium.org [Tue, 13 Jan 2009 10:16:02 +0000 (10:16 +0000)]
The "enter" opcode was slow and crashes for large arguments on Windows.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1066 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix benchmarks to not format scores that are really errors.
ager@chromium.org [Tue, 13 Jan 2009 09:27:47 +0000 (09:27 +0000)]
Fix benchmarks to not format scores that are really errors.

Uploading for Kasper.
Review URL: http://codereview.chromium.org/17641

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1061 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoUnsigned number can't be negated in the VisualStudio compiler.
lrn@chromium.org [Tue, 13 Jan 2009 09:23:26 +0000 (09:23 +0000)]
Unsigned number can't be negated in the VisualStudio compiler.
Fixed a few earlier review comments that had slipped through.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1060 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoIntroducing a kUninitialized constant used for clearing the function cache.
olehougaard [Tue, 13 Jan 2009 07:37:53 +0000 (07:37 +0000)]
Introducing a kUninitialized constant used for clearing the function cache.
Review URL: http://codereview.chromium.org/17639

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1057 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoDo not cache functions until we know they are fully constructed. This is needed in...
olehougaard [Tue, 13 Jan 2009 07:13:33 +0000 (07:13 +0000)]
Do not cache functions until we know they are fully constructed. This is needed in case of a stack overflow during construction.
Review URL: http://codereview.chromium.org/17354

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1056 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoSeparately growing stack for irregexp ia32 backtrack stack.
lrn@chromium.org [Mon, 12 Jan 2009 13:05:23 +0000 (13:05 +0000)]
Separately growing stack for irregexp ia32 backtrack stack.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1053 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix comment typo.
ager@chromium.org [Mon, 12 Jan 2009 11:07:02 +0000 (11:07 +0000)]
Fix comment typo.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1052 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAllocate as many object-literal properties as possible inobject.
ager@chromium.org [Mon, 12 Jan 2009 10:59:58 +0000 (10:59 +0000)]
Allocate as many object-literal properties as possible inobject.

This can lead to large objects which wastes a lot of space if we normalize properties.  We therfore clear the inobject properties when normalizing properties.  This is done by adjusting the instance size in the new map and overwriting the inobject properties with a filler.
Review URL: http://codereview.chromium.org/17308

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1051 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdded check that bails out of a repetition when the body is empty.
christian.plesner.hansen@gmail.com [Thu, 8 Jan 2009 12:40:47 +0000 (12:40 +0000)]
Added check that bails out of a repetition when the body is empty.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1047 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFixed lint.
christian.plesner.hansen@gmail.com [Thu, 8 Jan 2009 11:44:06 +0000 (11:44 +0000)]
Fixed lint.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1046 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdded more constness to the api. There are still some methods back
christian.plesner.hansen@gmail.com [Thu, 8 Jan 2009 11:35:34 +0000 (11:35 +0000)]
Added more constness to the api.  There are still some methods back
that could be const but aren't, but now at least all the obvious ones
should be.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1045 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRemove unused code.
sgjesse@chromium.org [Thu, 8 Jan 2009 10:40:09 +0000 (10:40 +0000)]
Remove unused code.
Review URL: http://codereview.chromium.org/17410

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1044 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix for issue 171:
iposva@chromium.org [Wed, 7 Jan 2009 23:26:31 +0000 (23:26 +0000)]
Fix for issue 171:
- Make sure to not destroy the instance type by slightly rearranging the code.
- Added regression test.

Review URL: http://codereview.chromium.org/16594

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1042 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFixed build with no ENABLE_LOGGING_AND_PROFILING
christian.plesner.hansen@gmail.com [Wed, 7 Jan 2009 14:24:08 +0000 (14:24 +0000)]
Fixed build with no ENABLE_LOGGING_AND_PROFILING

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1039 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAllow 2009 as a valid year in the copyright header.
ager@chromium.org [Wed, 7 Jan 2009 11:51:03 +0000 (11:51 +0000)]
Allow 2009 as a valid year in the copyright header.
Review URL: http://codereview.chromium.org/16568

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1038 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix for issue 91 (http://code.google.com/p/v8/issues/detail?id=91)
olehougaard [Wed, 7 Jan 2009 09:58:58 +0000 (09:58 +0000)]
Fix for issue 91 (code.google.com/p/v8/issues/detail?id=91)
Review URL: http://codereview.chromium.org/17232

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1036 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFactored the generation of JSON serialization from beeing part of the mirror objects...
sgjesse@chromium.org [Wed, 7 Jan 2009 09:27:30 +0000 (09:27 +0000)]
Factored the generation of JSON serialization from beeing part of the mirror objects into a seperate class. The purpose of this is to make it easier to change the serialization to contain object identification and references.

Added a test case for the script mirror and modified a number of the other mirror tests.
Review URL: http://codereview.chromium.org/16539

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1035 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoUpdate version number to fix small small whoopsie I committed when
christian.plesner.hansen@gmail.com [Tue, 6 Jan 2009 14:53:10 +0000 (14:53 +0000)]
Update version number to fix small small whoopsie I committed when
pushing to trunk.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1033 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoPreparing push to trunk (0.4.7)
christian.plesner.hansen@gmail.com [Tue, 6 Jan 2009 14:11:19 +0000 (14:11 +0000)]
Preparing push to trunk (0.4.7)

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1030 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdded runtime call to the logging infrastructure. Made some changes
christian.plesner.hansen@gmail.com [Tue, 6 Jan 2009 13:24:52 +0000 (13:24 +0000)]
Added runtime call to the logging infrastructure.  Made some changes
to the way regexps are being logged.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1028 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFixed double printing of exceptions when enabeling the debugger.
sgjesse@chromium.org [Tue, 6 Jan 2009 09:43:18 +0000 (09:43 +0000)]
Fixed double printing of exceptions when enabeling the debugger.

Detect the response to the continue command.
Review URL: http://codereview.chromium.org/17071

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1025 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRecognizes character classes like whitespace and non-newline and generates more effic...
lrn@chromium.org [Fri, 2 Jan 2009 12:23:17 +0000 (12:23 +0000)]
Recognizes character classes like whitespace and non-newline and generates more efficient code.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1024 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoUpdate comment and add assert.
ager@chromium.org [Tue, 23 Dec 2008 08:58:56 +0000 (08:58 +0000)]
Update comment and add assert.

Fix typo where underscore is used instead of a dash.
Review URL: http://codereview.chromium.org/16238

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1019 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoInline array loads in loops directly in the code instead of always
ager@chromium.org [Mon, 22 Dec 2008 12:56:32 +0000 (12:56 +0000)]
Inline array loads in loops directly in the code instead of always
calling a stub.  The map to check against is unknown when generating
the code, so we patch the map check in the IC initialization code.

Loop nesting is currently not tracked on ARM.  I'll file feature
request bug reports for implementing this on ARM and add the number to
the TODOs before I commit.
Review URL: http://codereview.chromium.org/16409

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1015 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoSome irregexp optimizations around keeping track of when the current character
erik.corry@gmail.com [Mon, 22 Dec 2008 12:48:14 +0000 (12:48 +0000)]
Some irregexp optimizations around keeping track of when the current character
register contains the next n characters.
Review URL: http://codereview.chromium.org/16410

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1014 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoReverting the BOM changes due to security concerns.
olehougaard [Fri, 19 Dec 2008 13:28:56 +0000 (13:28 +0000)]
Reverting the BOM changes due to security concerns.
Review URL: http://codereview.chromium.org/14890

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1011 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdded support in the profiler for creating 'regions' that cover part
christian.plesner.hansen@gmail.com [Fri, 19 Dec 2008 13:12:43 +0000 (13:12 +0000)]
Added support in the profiler for creating 'regions' that cover part
of the generated code.  These can be used by the profiler to
categorize the ticks that occur within generated code and thereby show
more detailed information about where time is spent in generated code.
For instance, this is what the profiler displayed for a simple regexp
benchmark with irregexp-native before:

[JavaScript]:
   total  nonlib   name
   87.2%   87.2%   RegExp: (?:\w*\W+)*

This is what we can display now:

[JavaScript]:
   total  nonlib   name
   87.2%   87.2%   RegExp: (?:\w*\W+)*
                   -  53.0%  56.7% BranchOrBacktrack
                   -  14.9%  59.8% CheckCharacterLT
                   -  13.7%  20.4% CheckStackLimit
                   -   6.7%   6.7% SafeCall
                   -   2.7%   7.0% CheckCharacterGT
                   -   2.4%   2.4% SafeReturn
                   -   2.1%   2.1% LoadCurrentCharacter
                   -   1.8%   1.8% PushRegister
                   -   0.9%   0.9% PopRegister
                   -   0.9%   0.9% AdvanceRegister
                   -   0.3%   0.3% PopCurrentPosition
                   -   0.3%   0.3% CheckGreedyLoop
                   -   0.0%  20.4% PushBacktrack
                   -   0.0%  22.3% CheckCharacter
                   -   0.0%   2.4% IfRegisterLT

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1010 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix ARM build.
erik.corry@gmail.com [Fri, 19 Dec 2008 12:13:11 +0000 (12:13 +0000)]
Fix ARM build.
Review URL: http://codereview.chromium.org/14887

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1008 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoInstead of jumping over a conditional backtrack, jump to a single instance somewhere...
lrn@chromium.org [Fri, 19 Dec 2008 12:08:29 +0000 (12:08 +0000)]
Instead of jumping over a conditional backtrack, jump to a single instance somewhere else.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1007 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoHandling byte-order marks as specified in Ecmascript-262 and in compliance with Safari.
olehougaard [Fri, 19 Dec 2008 12:06:11 +0000 (12:06 +0000)]
Handling byte-order marks as specified in Ecmascript-262 and in compliance with Safari.
Review URL: http://codereview.chromium.org/15075

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1006 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years ago* Generate quick checks based on mask and compare for
erik.corry@gmail.com [Fri, 19 Dec 2008 12:02:34 +0000 (12:02 +0000)]
* Generate quick checks based on mask and compare for
  the alternatives in a choice node.  The quick checks
  are conservative in the sense that they only detect
  failure with certainty.  Checks can do 2 or 4 characters
  at a time.
* Inline the quick checks to allow the alternatives to
  be checked without branching in the common case where
  they fail.
Review URL: http://codereview.chromium.org/14194

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1005 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoUnroll + and ? to reduce loop-related work.
erik.corry@gmail.com [Thu, 18 Dec 2008 15:17:24 +0000 (15:17 +0000)]
Unroll + and ? to reduce loop-related work.
Review URL: http://codereview.chromium.org/14836

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1003 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix lint error.
erik.corry@gmail.com [Thu, 18 Dec 2008 15:17:02 +0000 (15:17 +0000)]
Fix lint error.
Review URL: http://codereview.chromium.org/15036

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1002 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRefactored the code for handling debug step in in the runtime system into one functio...
sgjesse@chromium.org [Thu, 18 Dec 2008 14:32:49 +0000 (14:32 +0000)]
Refactored the code for handling debug step in in the runtime system into one function. For constructors this also means that step in will no longer step into the code for the builtins context.
Review URL: http://codereview.chromium.org/15035

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1001 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRemoved propagation of information about preceding nodes by expanding
christian.plesner.hansen@gmail.com [Thu, 18 Dec 2008 14:30:53 +0000 (14:30 +0000)]
Removed propagation of information about preceding nodes by expanding
following nodes.  Found a better solution.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1000 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFixing a subtle bug in receiver resolution when a thrown and caught function is calle...
olehougaard [Thu, 18 Dec 2008 11:28:13 +0000 (11:28 +0000)]
Fixing a subtle bug in receiver resolution when a thrown and caught function is called from a catch-block.
Second attempt - now with better memory efficiency.
Review URL: http://codereview.chromium.org/14834

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@999 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdded option --shell to D8 and updated the usage text.
sgjesse@chromium.org [Thu, 18 Dec 2008 10:39:11 +0000 (10:39 +0000)]
Added option --shell to D8 and updated the usage text.
Review URL: http://codereview.chromium.org/14192

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@998 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRemoved tabs from d8.js.
sgjesse@chromium.org [Thu, 18 Dec 2008 10:38:09 +0000 (10:38 +0000)]
Removed tabs from d8.js.

TBR=ager
Review URL: http://codereview.chromium.org/14835

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@997 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdded command line debugger to D8 shell.
sgjesse@chromium.org [Thu, 18 Dec 2008 10:06:49 +0000 (10:06 +0000)]
Added command line debugger to D8 shell.

  break location [condition]
  clear <breakpoint #>
  backtrace [from frame #] [to frame #]]
  frame <frame #>
  step [in | next | out| min [step count]]
  print <expression>
  source [from line [num lines]]
  scripts
  continue
  help

It is enabled through the option --debugger which is on by default.
Review URL: http://codereview.chromium.org/14509

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@996 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix an issue of a raw pointer being returned after possible allocation.
sgjesse@chromium.org [Thu, 18 Dec 2008 09:39:18 +0000 (09:39 +0000)]
Fix an issue of a raw pointer being returned after possible allocation.
Review URL: http://codereview.chromium.org/14833

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@995 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix issue 142:
iposva@chromium.org [Wed, 17 Dec 2008 17:40:02 +0000 (17:40 +0000)]
Fix issue 142:
- Removed the potential for a NULL pointer access in
  ContextSwitcher::PreemptionReceived.
- Removed a leak of the semaphore in the ContexSwitcher thread, by removing
  the need for this semaphore entirely.
- Added a regression test case which will catch accesses to the ContextSwitcher
  singleton after it has been stopped.

Review URL: http://codereview.chromium.org/14483

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@994 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFixed bug in interest propagation caused by following the loop edge
christian.plesner.hansen@gmail.com [Wed, 17 Dec 2008 13:16:38 +0000 (13:16 +0000)]
Fixed bug in interest propagation caused by following the loop edge
out of a loop choice node before the continuation edge.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@990 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoChanged (1<<31)-1 into INT_MAX.
lrn@chromium.org [Wed, 17 Dec 2008 12:08:27 +0000 (12:08 +0000)]
Changed (1<<31)-1 into INT_MAX.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@989 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoEach RegExtTree node can now report the min and max size of strings it can match.
lrn@chromium.org [Wed, 17 Dec 2008 10:59:14 +0000 (10:59 +0000)]
Each RegExtTree node can now report the min and max size of strings it can match.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@988 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoMinor changes, mostly cosmetic, in string search.
lrn@chromium.org [Wed, 17 Dec 2008 09:53:06 +0000 (09:53 +0000)]
Minor changes, mostly cosmetic, in string search.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@987 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRemoved some unused code to make Linux builders happy.
sgjesse@chromium.org [Wed, 17 Dec 2008 09:02:31 +0000 (09:02 +0000)]
Removed some unused code to make Linux builders happy.

TBR=ager
Review URL: http://codereview.chromium.org/14189

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@986 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRefactored the recording of source position in the generated code. The code generator...
sgjesse@chromium.org [Wed, 17 Dec 2008 08:45:42 +0000 (08:45 +0000)]
Refactored the recording of source position in the generated code. The code generator now has two methods

  void CodeForStatement(Node* node)
  void CodeForSourcePosition(int pos)

The first is used to indicate that code is about to be generated for the given statement and the second is used to indicate that code is about to be generated for the given source position.

Added position information for some statements which was missing whem.

Updated the code generator for ARM to emit source position the same way as for IA-32.

Added an assert to ensure that deferred code stubs will always have a source source position as if it has not it will take whatever source position before which makes no sense.

The passing test on ARM has only been tested using the simulator.
Review URL: http://codereview.chromium.org/14170

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@985 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRemoved rep-cmps{w,b} from CheckCharacters to improve performance.
lrn@chromium.org [Tue, 16 Dec 2008 14:21:00 +0000 (14:21 +0000)]
Removed rep-cmps{w,b} from CheckCharacters to improve performance.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@984 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdd missing files (debug-<architecture>.cc) to the Xcode project.
iposva@chromium.org [Mon, 15 Dec 2008 20:40:12 +0000 (20:40 +0000)]
Add missing files (debug-<architecture>.cc) to the Xcode project.
Review URL: http://codereview.chromium.org/14120

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@983 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoGive an error when setting break points in functions either defined through the API...
sgjesse@chromium.org [Mon, 15 Dec 2008 09:15:05 +0000 (09:15 +0000)]
Give an error when setting break points in functions either defined through the API or in functions which are part of the V8 builtins.

BUG=178
Review URL: http://codereview.chromium.org/13785

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@977 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoIntroduced a TempAssign utility because I just couldn't watch this anymore. YMMV.
olehougaard [Mon, 15 Dec 2008 08:55:35 +0000 (08:55 +0000)]
Introduced a TempAssign utility because I just couldn't watch this anymore. YMMV.
Review URL: http://codereview.chromium.org/14422

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@976 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRegexp-IA32 tests doesn't work with snapshots.
lrn@chromium.org [Fri, 12 Dec 2008 11:40:03 +0000 (11:40 +0000)]
Regexp-IA32 tests doesn't work with snapshots.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@975 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoPreemption code for irregexp-native-ia32. Regexps can not only succeede or
lrn@chromium.org [Fri, 12 Dec 2008 10:49:00 +0000 (10:49 +0000)]
Preemption code for irregexp-native-ia32. Regexps can not only succeede or
fail, but also report a thrown exception.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@974 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoQuantified look-aheads are sometimes removed entirely, leaving only a
lrn@chromium.org [Fri, 12 Dec 2008 10:22:56 +0000 (10:22 +0000)]
Quantified look-aheads are sometimes removed entirely, leaving only a
single atom node. A flag was not set in this case, leading the wrapper
code to think the pattern was equal to the atom and use the pattern
in the indexOf operation.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@971 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix lint issue.
ager@chromium.org [Thu, 11 Dec 2008 13:04:22 +0000 (13:04 +0000)]
Fix lint issue.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@969 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoPrepare pushing new v8 version to trunk (now working on version 0.4.7).
ager@chromium.org [Thu, 11 Dec 2008 11:38:02 +0000 (11:38 +0000)]
Prepare pushing new v8 version to trunk (now working on version 0.4.7).
Review URL: http://codereview.chromium.org/13385

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@967 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoContinue running if failing to make a debug event into a JSON event for sending to...
sgjesse@chromium.org [Thu, 11 Dec 2008 11:25:06 +0000 (11:25 +0000)]
Continue running if failing to make a debug event into a JSON event for sending to the debugger.

This partly fixes Chromium issue 5349 (http://code.google.com/p/chromium/issues/detail?id=5349).
Review URL: http://codereview.chromium.org/13384

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@966 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoImprove mark-compact object grouping interface.
deanm@chromium.org [Thu, 11 Dec 2008 11:20:04 +0000 (11:20 +0000)]
Improve mark-compact object grouping interface.

The main goal was to improve O(n^2) behavior when there are many object groups.  The old API required the grouping to be done on the v8 side, along with a linear search.  The new interface requires the caller to do the grouping, passing V8 entire groups at a time.  This removes the group id concept on the v8 side.

  - Changed AddObjectToGroup to AddObjectGroup.
  - Removed the group id concept from the V8 side.
  - Remove a static constructor while I'm here, lazily initialize
    the object groups list.
  - Cleaned up return by non-const references to return pointers.

Review URL: http://codereview.chromium.org/13341

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@965 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years ago- Added lookbehind propagation for the initial node; now, if the
christian.plesner.hansen@gmail.com [Thu, 11 Dec 2008 11:13:13 +0000 (11:13 +0000)]
- Added lookbehind propagation for the initial node; now, if the
  initial node is interested in what precedes it the automaton is
  given an initial all-consuming character class that determines it.
- Added verification of some node information invariants.  We now
  check that if a node expresses interest in what precedes it that
  information is available to it after assertion expansion.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@964 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoMake sure that the API does not change templates pass in from the
ager@chromium.org [Thu, 11 Dec 2008 11:09:51 +0000 (11:09 +0000)]
Make sure that the API does not change templates pass in from the
user.  The API should never change templates behind the back of the
user.
Review URL: http://codereview.chromium.org/13741

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@963 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdd handling of empty stack in the backtrace debug request.
sgjesse@chromium.org [Thu, 11 Dec 2008 09:35:44 +0000 (09:35 +0000)]
Add handling of empty stack in the backtrace debug request.

Add accessor function for the exception in an ExceptionEvent.
Review URL: http://codereview.chromium.org/13382

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@961 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoAdded test for bug 176 (zero length matches should fail in quantifiers).
erik.corry@gmail.com [Thu, 11 Dec 2008 09:01:55 +0000 (09:01 +0000)]
Added test for bug 176 (zero length matches should fail in quantifiers).
Review URL: http://codereview.chromium.org/13381

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@960 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoChanged the debugger break handling to support situations where there are no stack...
sgjesse@chromium.org [Thu, 11 Dec 2008 08:03:24 +0000 (08:03 +0000)]
Changed the debugger break handling to support situations where there are no stack frames. This can happen when an exception is thrown when compiling code.

This is related to Chromium issue 5349 (http://code.google.com/p/chromium/issues/detail?id=5349).
Review URL: http://codereview.chromium.org/13720

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@959 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoA recursive call to the code generator could (conceivably) clobber any
kmillikin@chromium.org [Wed, 10 Dec 2008 13:49:19 +0000 (13:49 +0000)]
A recursive call to the code generator could (conceivably) clobber any
register.  Generate code to evaluate both arguments of FastCharCodeAt
before popping either from the stack.
Review URL: http://codereview.chromium.org/13706

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@958 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoArm codegen could emit const pool in the middle of jump table.
lrn@chromium.org [Wed, 10 Dec 2008 08:37:58 +0000 (08:37 +0000)]
Arm codegen could emit const pool in the middle of jump table.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@956 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoMake sure that the generic stubs for keyed load and store and for
ager@chromium.org [Wed, 10 Dec 2008 08:05:10 +0000 (08:05 +0000)]
Make sure that the generic stubs for keyed load and store and for
dictionary probing respects access check bit.
Review URL: http://codereview.chromium.org/13663

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@955 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRemoved a few calls to Top::ReportPendingMessages() that caused the messages to be...
olehougaard [Tue, 9 Dec 2008 13:18:33 +0000 (13:18 +0000)]
Removed a few calls to Top::ReportPendingMessages() that caused the messages to be reported prematurely and not at the C++/JS boundary.
Review URL: http://codereview.chromium.org/13287

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@954 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoRefactor the convertion of a target address into a code object from the debugger...
sgjesse@chromium.org [Tue, 9 Dec 2008 12:53:59 +0000 (12:53 +0000)]
Refactor the convertion of a target address into a code object from the debugger to the Code class.
Review URL: http://codereview.chromium.org/13285

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@952 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFixed lint error.
sgjesse@chromium.org [Tue, 9 Dec 2008 12:30:05 +0000 (12:30 +0000)]
Fixed lint error.
Review URL: http://codereview.chromium.org/13661

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@950 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFixed Linux build by removing unused code.
sgjesse@chromium.org [Tue, 9 Dec 2008 11:24:53 +0000 (11:24 +0000)]
Fixed Linux build by removing unused code.
Review URL: http://codereview.chromium.org/13284

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@948 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoMoved the code generation for debug break stubs from builtins* to debug*. From builti...
sgjesse@chromium.org [Tue, 9 Dec 2008 11:12:14 +0000 (11:12 +0000)]
Moved the code generation for debug break stubs from builtins* to debug*. From builtins* all code generation delegate to debug*. Added files debug-ia32.cc and debug-arm.cc for platfoem specific code generation of debugger stubs.

Removed a dead file reference (frames-ia32-inl.h) from the Visual Studio project.

Updated the ARM Visual Studio project to build again by adding all the RegExp files.
Review URL: http://codereview.chromium.org/13657

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@947 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoSimplify the logic determining whether to report an exception.
olehougaard [Tue, 9 Dec 2008 10:36:31 +0000 (10:36 +0000)]
Simplify the logic determining whether to report an exception.
Review URL: http://codereview.chromium.org/13283

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@946 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFixing a bug where a try-finally block obscured a try-catch block when registering...
olehougaard [Tue, 9 Dec 2008 10:16:38 +0000 (10:16 +0000)]
Fixing a bug where a try-finally block obscured a try-catch block when registering an external try-catch handler.
Review URL: http://codereview.chromium.org/13658

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@945 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFixed compiler warning C4244 when compiling with the ARM simulator in Windows.
sgjesse@chromium.org [Tue, 9 Dec 2008 09:43:16 +0000 (09:43 +0000)]
Fixed compiler warning C4244 when compiling with the ARM simulator in Windows.
Review URL: http://codereview.chromium.org/13282

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@944 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoMake Logger::SuspectReadEvent extract the class name from the object. This makes...
deanm@chromium.org [Tue, 9 Dec 2008 09:19:02 +0000 (09:19 +0000)]
Make Logger::SuspectReadEvent extract the class name from the object.  This makes the lone caller cleaner, and also avoids a tiny bit of work in the default case of not running with --log-suspect.

Review URL: http://codereview.chromium.org/13655

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@943 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoFix build (someone tell gcc you can't take the address of a static
erik.corry@gmail.com [Tue, 9 Dec 2008 09:17:41 +0000 (09:17 +0000)]
Fix build (someone tell gcc you can't take the address of a static
const int and someone tell MSVC it's OK to define a static const int
in a .cc file).
Review URL: http://codereview.chromium.org/13656

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@942 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years ago* Have an ASCII and a UC16 interpreter for Irregexp bytecodes -
erik.corry@gmail.com [Tue, 9 Dec 2008 08:30:49 +0000 (08:30 +0000)]
* Have an ASCII and a UC16 interpreter for Irregexp bytecodes -
  never have to convert an ASCII string to UC16 for Irregexp.
* Generate slightly different code when we know the subject string
  is ASCII.
Review URL: http://codereview.chromium.org/13247

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@941 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

15 years agoSkip some debugger related tests on ARM for nowas they have different behaviour based...
sgjesse@chromium.org [Mon, 8 Dec 2008 15:11:02 +0000 (15:11 +0000)]
Skip some debugger related tests on ARM for nowas they have different behaviour based on platform and mode. The debugger infrastructure is currently not working on the ARM platform. This also fix the "syntax" error of the previous change to mjsunit.status.
Review URL: http://codereview.chromium.org/13622

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@940 ce2b1a6d-e550-0410-aec6-3dcde31c8c00