[v8-dev] Re: Fixed arm/mac errors and presubmitting 2324.

LGTM http://codereview.chromium.org/151193 --~--~-~--~~~---~--~~ v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev -~--~~~~--~~--~--~---

[v8-dev] Fixed arm/mac errors and presubmitting 2324.

Reviewers: Mads Ager, Description: Fixed arm/mac errors and presubmitting 2324. Please review this at http://codereview.chromium.org/151193 SVN Base: http://v8.googlecode.com/svn/branches/bleeding_edge/ Affected files: M src/arm/ic-arm.cc M src/bootstrapper.cc M src/code-

[v8-dev] Re: Make global variable initialization consistent with Firefox and Safari...

LGTM. http://codereview.chromium.org/151191 --~--~-~--~~~---~--~~ v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev -~--~~~~--~~--~--~---

[v8-dev] Make global variable initialization consistent with Firefox and Safari...

Reviewers: Kasper Lund, Description: Make global variable initialization consistent with Firefox and Safari behavior. Only force the introduction of a variable directly on the global object if there is an explicit initial value in a variable declaration. BUG=http://crbug.com/12548 Please revie

[v8-dev] Prepare to push version 1.2.11 to trunk.

Reviewers: Mads Ager, Description: Prepare to push version 1.2.11 to trunk. Please review this at http://codereview.chromium.org/150209 SVN Base: http://v8.googlecode.com/svn/branches/bleeding_edge/ Affected files: M ChangeLog M src/version.cc Index: src/version.cc

[v8-dev] Re: question on the performance

2009/7/1 seoji...@gmail.com : > > I guess V8 is very fast compared to other javascript engines. > I'm curious what's the biggest reason for that. > I mean, I know hidden class technique, machine code generation, > efficient gc attribute to that, > but is there number for how each of these contribu

[v8-dev] question on the performance

I guess V8 is very fast compared to other javascript engines. I'm curious what's the biggest reason for that. I mean, I know hidden class technique, machine code generation, efficient gc attribute to that, but is there number for how each of these contributes? For instance, I'd like to know, if w

[v8-dev] [v8 commit] r2323 - Fix the snapshotter when ENABLE_DEBUGGER_SUPPORT is off.

Author: f...@chromium.org Date: Wed Jul 1 08:36:28 2009 New Revision: 2323 Modified: branches/bleeding_edge/src/serialize.cc Log: Fix the snapshotter when ENABLE_DEBUGGER_SUPPORT is off. Review URL: http://codereview.chromium.org/151136 Modified: branches/bleeding_edge/src/serialize.cc ==

[v8-dev] Fix all lines > 80 chars in messages.js so they are <= 80.

LGTM with a few comments. http://codereview.chromium.org/126109/diff/1/2 File src/messages.js (right): http://codereview.chromium.org/126109/diff/1/2#newcode165 Line 165: // arbitrary limit, grab a reasonable slice to report Please start comment with uppercase and end with a dot. http://codere

[v8-dev] Issue 394 in v8: Assessors on global object not handled correctly

Updates: Status: Fixed Comment #1 on issue 394 by sgje...@chromium.org: Assessors on global object not handled correctly http://code.google.com/p/v8/issues/detail?id=394 Fixed in bleeding_e...@2317. -- You received this message because you are listed in the owner or CC fields of this

[v8-dev] Re: Initial implementation of printing the AST as a JSON string. This...

LGTM Should we add something to the debugger protocol to retrieve the AST? http://codereview.chromium.org/131101 --~--~-~--~~~---~--~~ v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev -~--~~~~--~~--

[v8-dev] Re: Experimental: capture 10 frame stack trace by default.

LGTM http://codereview.chromium.org/151145 --~--~-~--~~~---~--~~ v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev -~--~~~~--~~--~--~---

[v8-dev] Revert 2324 until we can fix build and test errors.

Reviewers: bak, Description: Revert 2324 until we can fix build and test errors. tbr=...@chromium.org Please review this at http://codereview.chromium.org/151155 SVN Base: http://v8.googlecode.com/svn/branches/bleeding_edge/ Affected files: M src/bootstrapper.cc M src/code-stubs

[v8-dev] Re: Create a separate space for global property cells

Added Kevin as reviewer. Kevin, could you please review this as well, and especially verify that the use of the constants kSpaceTagSize and kSpaceTagMask are as I think, and that adding the additional space does not limit the upper bound on the old space. Also the max size of the global property

[v8-dev] Re: Use attributes to communicate failed lookup instead of retval.

On 2009/07/01 15:44:47, antonm wrote: > http://codereview.chromium.org/151151/diff/1/2#newcode5822 > Line 5822: PropertyAttributes* attrs) { > On 2009/07/01 15:06:59, Mads Ager wrote: > > Why attrs in one method and attributes in another? Make this one attributes > > again? > To fit ln. 5828 int

[v8-dev] [v8 commit] r2322 - Ignore interceptors with no getters when doing any kind of read.

Author: ant...@chromium.org Date: Wed Jul 1 08:06:46 2009 New Revision: 2322 Modified: branches/bleeding_edge/src/ic.cc Log: Ignore interceptors with no getters when doing any kind of read. Review URL: http://codereview.chromium.org/150164 Modified: branches/bleeding_edge/src/ic.cc ==

[v8-dev] [v8 commit] r2321 - Always scavenge when doing GCs.

Author: kasp...@chromium.org Date: Wed Jul 1 06:20:09 2009 New Revision: 2321 Modified: branches/bleeding_edge/src/heap.cc Log: Always scavenge when doing GCs. Review URL: http://codereview.chromium.org/151149 Modified: branches/bleeding_edge/src/heap.cc ===

[v8-dev] Re: Use attributes to communicate failed lookup instead of retval.

Thanks a lot for review, Mads! http://codereview.chromium.org/151151/diff/1/2 File src/objects.cc (right): http://codereview.chromium.org/151151/diff/1/2#newcode5814 Line 5814: return NULL; On 2009/07/01 15:06:59, Mads Ager wrote: > We normally return Heap::undefined_value() when something is a

[v8-dev] Re: Create a separate space for global property cells

First round of comments: http://codereview.chromium.org/151152/diff/1017/13 File src/heap.cc (right): http://codereview.chromium.org/151152/diff/1017/13#newcode89 Line 89: int Heap::old_generation_size_ = 512*MB; Do you need to change this too? http://codereview.chromium.org/151152/diff/1017/1

[v8-dev] Re: Removed virtual behavior from Dictionaries.

LGTM http://codereview.chromium.org/150168/diff/1/3 File src/objects.cc (right): http://codereview.chromium.org/150168/diff/1/3#newcode2312 Line 2312: if (entry != StringDictionary::kNotFound) { NumberDictionary::kNotFound http://codereview.chromium.org/150168/diff/1/8 File src/objects.h (righ

[v8-dev] Create a separate space for global property cells

Reviewers: Kasper Lund, Description: Create a separate space for global property cells. All global property cells are now allocated from a separate space which can only contain this type of objects. The map space and the global property cell space are created from a common super class for creati

[v8-dev] Re: Use attributes to communicate failed lookup instead of retval.

LGTM http://codereview.chromium.org/151151/diff/1/2 File src/objects.cc (right): http://codereview.chromium.org/151151/diff/1/2#newcode5814 Line 5814: return NULL; We normally return Heap::undefined_value() when something is absent. http://codereview.chromium.org/151151/diff/1/2#newcode5822 Li

[v8-dev] Removed virtual behavior from Dictionaries.

Reviewers: Mads Ager, Description: Removed virtual behavior from Dictionaries. Please review this at http://codereview.chromium.org/150168 SVN Base: http://v8.googlecode.com/svn/branches/bleeding_edge/ Affected files: M src/bootstrapper.cc M src/code-stubs.cc M src/factor

[v8-dev] Use attributes to communicate failed lookup instead of retval.

Reviewers: Mads Ager, Description: Use attributes to communicate failed lookup instead of retval. Please review this at http://codereview.chromium.org/151151 SVN Base: http://v8.googlecode.com/svn/branches/bleeding_edge/ Affected files: M src/objects.h M src/objects.cc --~--~

[v8-dev] Re: Ignore interceptors with no getters when doing any kind of read.

LGTM http://codereview.chromium.org/150164 --~--~-~--~~~---~--~~ v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev -~--~~~~--~~--~--~---

[v8-dev] Re: - Avoid allocating an overly large stack frame in the irregexp

I'll fix it. I was wondering what you meant until I looked at the ARM performance. -Ivan On Tue, Jun 30, 2009 at 23:54, Erik Corry wrote: > There must be a way to do this without damaging performance on regexp, > string-unpack-code and string-validate-input. > > 2009/7/1  : >> >> LGTM >> >> htt

[v8-dev] Ignore interceptors with no getters when doing any kind of read.

Reviewers: Mads Ager, Kasper Lund, Description: Ignore interceptors with no getters when doing any kind of read. Please review this at http://codereview.chromium.org/150164 SVN Base: http://v8.googlecode.com/svn/branches/bleeding_edge/ Affected files: M src/ic.cc --~--~-~--~

[v8-dev] Re: X64: Move remembered set to a safe location on x64 platform.

LGTM http://codereview.chromium.org/151148/diff/1/2 File src/spaces-inl.h (right): http://codereview.chromium.org/151148/diff/1/2#newcode107 Line 107: // The reset address is computed as: reset -> rset (or RSet?) http://codereview.chromium.org/151148/diff/1/2#newcode108 Line 108: //page_ad

[v8-dev] [v8 commit] r2320 - X64: Add Math.random, fix UnarySubStub

Author: whe...@chromium.org Date: Wed Jul 1 04:59:10 2009 New Revision: 2320 Modified: branches/bleeding_edge/src/x64/codegen-x64.cc Log: X64: Add Math.random, fix UnarySubStub Review URL: http://codereview.chromium.org/151077 Modified: branches/bleeding_edge/src/x64/codegen-x64.cc ===

[v8-dev] Re: X64: Implement InstanceofStub. Prohibit zero-size code objects.

All comments adressed. http://codereview.chromium.org/151142 --~--~-~--~~~---~--~~ v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev -~--~~~~--~~--~--~---

[v8-dev] Re: X64: Add Math.random, fix UnarySubStub

http://codereview.chromium.org/151077/diff/1/2 File src/x64/codegen-x64.cc (right): http://codereview.chromium.org/151077/diff/1/2#newcode5518 Line 5518: // Also enter it if the value of the smi is MIN_INT. On 2009/07/01 10:59:44, Lasse Reichstein wrote: > MIN_INT -> Smi::kMinValue Done. http:

[v8-dev] [v8 commit] r2319 - Treat the builtins object like other global objects (with

Author: kasp...@chromium.org Date: Wed Jul 1 04:44:37 2009 New Revision: 2319 Modified: branches/bleeding_edge/src/arm/stub-cache-arm.cc branches/bleeding_edge/src/bootstrapper.cc branches/bleeding_edge/src/factory.cc branches/bleeding_edge/src/factory.h branches/bleeding_edg

[v8-dev] X64: Move remembered set to a safe location on x64 platform.

Reviewers: Mads Ager, Lasse Reichstein, Message: The remembered set on x64 is 1/2 the size of the set on ia32, so we needed to move it to an offset in the available space. Description: X64: Move remembered set to a safe location on x64 platform. Please review this at http://codereview.chromium.

[v8-dev] [v8 commit] r2317 - Handle JavaScript accessors on the global object.

Author: sgje...@chromium.org Date: Wed Jul 1 04:20:33 2009 New Revision: 2317 Added: branches/bleeding_edge/test/mjsunit/regress/regress-394.js (contents, props changed) Modified: branches/bleeding_edge/src/objects.cc Log: Handle JavaScript accessors on the global object. With the

[v8-dev] [v8 commit] r2318 - ARM: Don't always go into the runtime system just because a

Author: erik.co...@gmail.com Date: Wed Jul 1 04:32:08 2009 New Revision: 2318 Modified: branches/bleeding_edge/src/arm/codegen-arm.cc Log: ARM: Don't always go into the runtime system just because a mul gave zero. Review URL: http://codereview.chromium.org/151144 Modified: branches/bleedi

[v8-dev] Re: Treat the builtins object like other global objects (with...

STV! http://codereview.chromium.org/151146 --~--~-~--~~~---~--~~ v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev -~--~~~~--~~--~--~---

[v8-dev] Re: ARM: Don't always go into the runtime system just because a

http://codereview.chromium.org/151144/diff/1/2 File src/arm/codegen-arm.cc (right): http://codereview.chromium.org/151144/diff/1/2#newcode5340 Line 5340: __ mov(r0, Operand(Smi::FromInt(0)), LeaveCC, pl); On 2009/07/01 11:11:14, Lasse Reichstein wrote: > Is it more efficient to use r3 instead of

[v8-dev] Re: Changed RegExp parser to use a recursive data structure instead of stack-based recursion.

LGTM with comments if it lints. http://codereview.chromium.org/149069/diff/3003/2006 File src/parser.cc (right): http://codereview.chromium.org/149069/diff/3003/2006#newcode608 Line 608: struct RegExpParserState : public ZoneObject { Please make this a real object. http://codereview.chromium.o

[v8-dev] Treat the builtins object like other global objects (with...

Reviewers: Søren Gjesse, Description: Treat the builtins object like other global objects (with cells in old space for properties) and avoid allocating tons of unused properties in new space when creating new contexts. Please review this at http://codereview.chromium.org/151146 SVN Base: http://

[v8-dev] Re: Handle JavaScript accessors on the global object...

LGTM. http://codereview.chromium.org/150162 --~--~-~--~~~---~--~~ v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev -~--~~~~--~~--~--~---

[v8-dev] Experimental: capture 10 frame stack trace by default.

Reviewers: Søren Gjesse, Description: Enable capture of the top of the stack on error instantiation. Performance impact on v8 benchmarks seems limited. Will be backed out if chrome performance regresses. Please review this at http://codereview.chromium.org/151145 Affected files: M src/bootst

[v8-dev] Re: ARM: Don't always go into the runtime system just because a

LGTM http://codereview.chromium.org/151144/diff/1/2 File src/arm/codegen-arm.cc (right): http://codereview.chromium.org/151144/diff/1/2#newcode5340 Line 5340: __ mov(r0, Operand(Smi::FromInt(0)), LeaveCC, pl); Is it more efficient to use r3 instead of Operand(Smi::FromInt(0))? (We should know t

[v8-dev] Re: X64: Add Math.random, fix UnarySubStub

LGTM http://codereview.chromium.org/151077/diff/1/2 File src/x64/codegen-x64.cc (right): http://codereview.chromium.org/151077/diff/1/2#newcode5518 Line 5518: // Also enter it if the value of the smi is MIN_INT. MIN_INT -> Smi::kMinValue http://codereview.chromium.org/151077 --~--~-~-

[v8-dev] Re: X64: Implement InstanceofStub. Prohibit zero-size code objects.

LGTM http://codereview.chromium.org/151142/diff/1/5 File src/x64/codegen-x64.cc (right): http://codereview.chromium.org/151142/diff/1/5#newcode5765 Line 5765: __ movq(rax, Operand(rsp, 2 * kPointerSize)); // 2 ~ return address, function Is there no constant that can be used here? The comment i

[v8-dev] Handle JavaScript accessors on the global object...

Reviewers: Kasper Lund, Description: Handle JavaScript accessors on the global object. 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

[v8-dev] ARM: Don't always go into the runtime system just because a

Reviewers: Lasse Reichstein, Description: ARM: Don't always go into the runtime system just because a mul gave zero. Please review this at http://codereview.chromium.org/151144 SVN Base: http://v8.googlecode.com/svn/branches/bleeding_edge/ Affected files: M src/arm/codegen-arm.cc Ind

[v8-dev] X64: Implement InstanceofStub. Prohibit zero-size code objects.

Reviewers: Lasse Reichstein, Description: X64: Implement InstanceofStub. Prohibit zero-size code objects. Please review this at http://codereview.chromium.org/151142 SVN Base: http://v8.googlecode.com/svn/branches/bleeding_edge/ Affected files: M src/x64/assembler-x64.cc M src/x

[v8-dev] [v8 commit] r2315 - Reimplement profiler sampler on Mac OS X to get it working under Chromium.

Author: mikhail.naga...@gmail.com Date: Wed Jul 1 01:46:59 2009 New Revision: 2315 Modified: branches/bleeding_edge/src/log.cc branches/bleeding_edge/src/platform-freebsd.cc branches/bleeding_edge/src/platform-linux.cc branches/bleeding_edge/src/platform-macos.cc branches/ble

[v8-dev] [v8 commit] r2316 - Fix presubmit error.

Author: mikhail.naga...@gmail.com Date: Wed Jul 1 02:11:33 2009 New Revision: 2316 Modified: branches/bleeding_edge/src/platform-macos.cc Log: Fix presubmit error. tbr=ipo...@chromium.org Review URL: http://codereview.chromium.org/150158 Modified: branches/bleeding_edge/src/platform-maco

[v8-dev] Re: - Avoid allocating an overly large stack frame in the irregexp

You can use a RegExpStack to avoid allocating a new memory area every time. It keeps a 1K buffer that can be reused, and it is only freed if you expand it beyond the default size.It would even be appropriately named :) /L On Wed, Jul 1, 2009 at 8:54 AM, Erik Corry wrote: > > There must be a way

[v8-dev] Issue 394 in v8: Assessors on global object not handled correctly

Status: Accepted Owner: sgje...@chromium.org Labels: Type-Bug Priority-Medium New issue 394 by sgje...@chromium.org: Assessors on global object not handled correctly http://code.google.com/p/v8/issues/detail?id=394 After changing the representation of the global object the handling of accessor