Re: [v8-dev] Re: LLVM V8 Code Generator for WebAssembly

Hi, I've created a tracking bug on which we'll apply some more investigation (https://bugs.chromium.org/p/v8/issues/detail?id=10023). We'll have a detailed look in January. Are you interested in upstreaming your work? We're interested in using this for the built-ins and interpreter bytecode han

[v8-dev] Re: Crankshaft: consolidated element loads always deopted on seeing the hole (issue 1329793003 by mvstan...@chromium.org)

Right on, thanks for the help thinking through it too. https://codereview.chromium.org/1329793003/diff/20001/src/hydrogen.cc File src/hydrogen.cc (right): https://codereview.chromium.org/1329793003/diff/20001/src/hydrogen.cc#newcode7447 src/hydrogen.cc:7447: if (*map != isolate()->get_initial_j

[v8-dev] Re: [Interpreter] Add support for property store operations. (issue 1319833004 by rmcil...@chromium.org)

Ah, you can ignore my comment about SLOPPY/STRICT because I see elsewhere that you are only handling SLOPPY for now. Nice tests, again! LGTM. https://codereview.chromium.org/1319833004/diff/1/src/interpreter/interpreter.cc File src/interpreter/interpreter.cc (right): https://codereview.chr

[v8-dev] Crankshaft: consolidated element loads always deopted on seeing the hole (issue 1329793003 by mvstan...@chromium.org)

Reviewers: Jakob, Message: Hi Jakob, Here is the fix we were talking about, thx for the look! --Michael Description: Crankshaft: consolidated element loads always deopted on seeing the hole Update the consolidated load case to carefully chose the load mode based on the consolidated elements kin

[v8-dev] Vector ICs: No more patching for call ics. (issue 1332563003 by mvstan...@chromium.org)

Reviewers: Benedikt Meurer, Jakob, Message: Hi guys, Benedikt - could you review code-stubs*, as you were just in there. Jakob - could you review the changes under src/ic/*? Thanks much! --Michael Description: On a call to Array(), we patched a call ic. This CL makes do with a single dispatche

[v8-dev] Re: Vector ICs: The Oracle needs to report feedback for the count op. (issue 1321993004 by mvstan...@chromium.org)

Hi guys, Addressed Michi's comments, did ports, fixed a bug. Thanks much, --Michael https://codereview.chromium.org/1321993004/diff/40001/src/ast.h File src/ast.h (right): https://codereview.chromium.org/1321993004/diff/40001/src/ast.h#newcode1429 src/ast.h:1429: int ic_slot_count_; On 2015/09/

[v8-dev] Re: [stubs] Unify the various versions of [[Call]] with CallCallableStub. (issue 1311013008 by bmeu...@chromium.org)

LGTM here too. https://codereview.chromium.org/1311013008/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving ema

[v8-dev] Re: Vector ICs: The Oracle needs to report feedback for the count op. (issue 1321993004 by mvstan...@chromium.org)

Reviewers: Michael Starzinger, Jakob, Message: Hi Michi, hi Jakob, PTAL. Jakob - General Michi - Turbofan Thanks, --Michael Description: Vector ICs: The Oracle needs to report feedback for the object literals and the count operation. The refactoring is because it's awkward and error-prone

[v8-dev] Re: [stubs] Unify the various versions of [[Call]] with CallCallableStub. (issue 1311013008 by bmeu...@chromium.org)

Code-stubs-* is looking good. https://codereview.chromium.org/1311013008/diff/120001/src/code-stubs.h File src/code-stubs.h (right): https://codereview.chromium.org/1311013008/diff/120001/src/code-stubs.h#newcode1957 src/code-stubs.h:1957: // and less horrible Invoke family of builtins. oh come

[v8-dev] Re: Adding GetMoreGeneralElementsKind in elements-kind.h (issue 1307743011 by cbr...@chromium.org)

Looks good! LGTM. https://codereview.chromium.org/1307743011/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving

[v8-dev] Reland Vector ICs: platform support for vector-based stores. (issue 1319123004 by mvstan...@chromium.org)

Reviewers: Benedikt Meurer, Jakob, Message: Reland with fix for ARM and ARM64 release build failure. Thx, --Michael Description: Reland Vector ICs: platform support for vector-based stores. The last changes for vector store functionality, they are in 3 areas: 1) The new vector [keyed] store co

[v8-dev] Re: Vector ICs: platform support for vector-based stores. (issue 1328603003 by mvstan...@chromium.org)

Thanks all! https://codereview.chromium.org/1328603003/diff/40001/src/ast.cc File src/ast.cc (right): https://codereview.chromium.org/1328603003/diff/40001/src/ast.cc#newcode353 src/ast.cc:353: if (saw_computed_name && On 2015/09/03 11:27:39, Jakob wrote: This is surprising: having seen a comp

[v8-dev] Re: MIPS: Vector ICs: platform support for vector-based stores. (issue 1318353003 by paul.l...@imgtec.com)

On 2015/09/03 05:19:53, paul.l... wrote: Hey Michael - here are MIPS ports, l-g-t-m for inclusion in your https://codereview.chromium.org/1328603003/ Thanks tons Paul! Man, that was fast.. https://codereview.chromium.org/1318353003/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://gro

[v8-dev] VectorICs: Cleanup, remove unnecessary arguments from HandleArrayCases() (issue 1314503003 by mvstan...@chromium.org)

Reviewers: Igor Sheludko, Message: Hi Igor, A small cleanup CL to remove unnecessary arguments. PTAL, thanks! --Michael Description: VectorICs: Cleanup, remove unnecessary arguments from HandleArrayCases() BUG= Please review this at https://codereview.chromium.org/1314503003/ Base URL: https:

[v8-dev] Re: Vector ICs: platform support for vector-based stores. (issue 1328603003 by mvstan...@chromium.org)

Mips team, could you guys address mips/mips64 port? Let me know any questions, thanks! https://codereview.chromium.org/1328603003/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Gro

[v8-dev] Vector ICs: platform support for vector-based stores. (issue 1328603003 by mvstan...@chromium.org)

Reviewers: Benedikt Meurer, Jakob, Message: Hi guys, This CL makes vector store support complete in the tree. Benedikt: could you look at code-stubs-.cc, Jakob, everything else. I'll ask for mips porting help in a moment... Thanks! --Michael Description: Vector ICs: platform support for vecto

[v8-dev] Re: Unify ElementsAccessor Array Builtins (issue 1302413006 by cbr...@chromium.org)

lgtm https://codereview.chromium.org/1302413006/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Re: [Interpreter] Add support for property load operations. (issue 1309843007 by rmcil...@chromium.org)

-assembler.cc:239: Node* InterpreterAssembler::CallIC(CallInterfaceDescriptor descriptor, On 2015/09/01 15:53:36, rmcilroy wrote: On 2015/09/01 14:28:29, mvstanton wrote: > If the number of arguments are going to be fixed I'd like to see something in > the name calling it a LoadIC, since StoreI

[v8-dev] Re: [Interpreter] Add support for property load operations. (issue 1309843007 by rmcil...@chromium.org)

Vector lgtm, a couple of comments - love the tests. https://codereview.chromium.org/1309843007/diff/20001/test/cctest/interpreter/test-bytecode-generator.cc File test/cctest/interpreter/test-bytecode-generator.cc (right): https://codereview.chromium.org/1309843007/diff/20001/test/cctest/interpr

[v8-dev] Make type-feedback-vector.h usable without objects-inl.h header (and others). (issue 1309303008 by mvstan...@chromium.org)

Reviewers: Michael Starzinger, Message: Hi Michael, here is the CL we discussed. PTAL, thanks! --Michael Description: Make type-feedback-vector.h usable without objects-inl.h header (and others). Pure refactoring to further reduce the header dependency tangle. BUG= R=mstarzin...@chromium.or

[v8-dev] Re: [runtime] Remove unused TO_NUMBER and TO_STRING functions. (issue 1304363012 by bmeu...@chromium.org)

lgtm https://codereview.chromium.org/1304363012/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Re: Vector ICs: Adapting store ic classes for vectors. (issue 1326483002 by mvstan...@chromium.org)

one kine small ping... https://codereview.chromium.org/1326483002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop recei

[v8-dev] Vector ICs: Adapting store ic classes for vectors. (issue 1326483002 by mvstan...@chromium.org)

Reviewers: Igor Sheludko, Message: Hi Igor, Here are some more adaptations for vector stores. This covers all the changes in IC and the IC compilers. There are additional platform dependent changes to come in a follow-up CL. Thanks! --Michael Description: Vector ICs: Adapting store ic class

[v8-dev] Re: Reorder KeyedStoreIC MISS code to avoid unnecessary compilation. (issue 1308073010 by mvstan...@chromium.org)

Ah, no, don't worry that was a mistake after patching this CL into another issue. That crazy REBASE didn't get checked in. I'll delete the patchset, sorry for the confusion! https://codereview.chromium.org/1308073010/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/

[v8-dev] Re: Reorder KeyedStoreIC MISS code to avoid unnecessary compilation. (issue 1308073010 by mvstan...@chromium.org)

Thanks for the comments Igor, addressed 'em... Michael https://codereview.chromium.org/1308073010/diff/20001/src/ic/ic.cc File src/ic/ic.cc (right): https://codereview.chromium.org/1308073010/diff/20001/src/ic/ic.cc#newcode2166 src/ic/ic.cc:2166: if (object->IsJSObject()) { On 2015/08/28 12:25:

[v8-dev] Reorder KeyedStoreIC MISS code to avoid unnecessary compilation. (issue 1308073010 by mvstan...@chromium.org)

Reviewers: Igor Sheludko, Message: Hi Igor, Here is the change we were discussing. I front-load the work that might change after the property set. This will make vector-based stores way easier. Thanks, --Michael Description: Reorder KeyedStoreIC MISS code to avoid unnecessary compilation. W

[v8-dev] Re: [es6] Implement spec compliant ToPrimitive in the runtime. (issue 1306303003 by bmeu...@chromium.org)

LGTM code stubs. https://codereview.chromium.org/1306303003/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving e

[v8-dev] Re: Vector ICs: Make the Oracle gather feedback for vector stores. (issue 1316953003 by mvstan...@chromium.org)

Hi Benedikt, thanks! Okay, I just checked the ic code, and they will MISS if the transitioned map has disappeared from the weak cell. I'll do the same in the vector dispatcher. --Michael https://codereview.chromium.org/1316953003/diff/1/src/type-info.cc File src/type-info.cc (right): https:

[v8-dev] Re: [turbofan] Fix type feedback vector for stub compilation. (issue 1320623004 by mstarzin...@chromium.org)

Smart to do. LGTM. https://codereview.chromium.org/1320623004/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receivin

[v8-dev] Re: [heap] Limit friendship of the Heap class to essentials. (issue 1320843002 by mstarzin...@chromium.org)

lgtm https://codereview.chromium.org/1320843002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Vector ICs: Make the Oracle gather feedback for vector stores. (issue 1316953003 by mvstan...@chromium.org)

Reviewers: Benedikt Meurer, Message: Hi Benedikt, here is the change we discussed. Thanks for the suggestion on regularity of the array, I've addressed that. --Michael Description: Vector ICs: Make the Oracle gather feedback for vector stores. Also, polymorphic element stores have a slightly d

[v8-dev] Re: Vector ICs: Stop iterating the heap to clear keyed store ics. (issue 1314433004 by mvstan...@chromium.org)

Hi Hannes, thanks for the review, here are changes. --Michael https://codereview.chromium.org/1314433004/diff/1/src/heap/heap.cc File src/heap/heap.cc (right): https://codereview.chromium.org/1314433004/diff/1/src/heap/heap.cc#newcode480 src/heap/heap.cc:480: void Heap::ClearAllKeyedStoreICs()

[v8-dev] Re: Vector ICs: Stop iterating the heap to clear keyed store ics. (issue 1314433004 by mvstan...@chromium.org)

https://codereview.chromium.org/1314433004/diff/1/src/heap/heap.h File src/heap/heap.h (right): https://codereview.chromium.org/1314433004/diff/1/src/heap/heap.h#newcode857 src/heap/heap.h:857: void ClearAllKeyedStoreICs(); I did the rename here to narrow the perceived functionality. In general

[v8-dev] Vector ICs: Stop iterating the heap to clear keyed store ics. (issue 1314433004 by mvstan...@chromium.org)

TODO(mvstanton): Do not iterate the heap. - HeapObjectIterator it(code_space()); - - for (Object* object = it.Next(); object != NULL; object = it.Next()) { -Code* code = Code::cast(object); -Code::Kind current_kind = code->kind(); -if (current_kind == Code::FUNCTION || -curr

[v8-dev] Vector ICs: Ensure KeyedAccessStore mode is encoded in all handlers. (issue 1312693004 by mvstan...@chromium.org)

Reviewers: Yang, Message: Hi Yang, Here is the CL we discussed. Validates/protects that the handlers have the same store mode as the IC extra state bits. Description: Vector ICs: Ensure KeyedAccessStore mode is encoded in all handlers. For vector-based keyed store ics, we need to know the cu

[v8-dev] Re: Do not inline array resize operations for outdated prototype maps. (issue 1313303002 by ish...@chromium.org)

Okay, looks like that works! LGTM. https://codereview.chromium.org/1313303002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group a

[v8-dev] Re: Make Simulator respect C stack limits as well. (issue 1314623002 by mstarzin...@chromium.org)

An ingenious solution! LGTM. https://codereview.chromium.org/1314623002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and sto

[v8-dev] Re: Revert of Moving ArraySplice Builtin to ElementsAccessor (issue 1305763002 by cbr...@chromium.org)

lgtm https://codereview.chromium.org/1305763002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Simplify KeyedAccessStoreMode. (issue 1303813004 by mvstan...@chromium.org)

Reviewers: Yang, Message: Hi Yang, Here is a cleanup discovered on my vector store path... Description: Simplify KeyedAccessStoreMode. The map tells us the source elements kind, and holeyness. BUG= Please review this at https://codereview.chromium.org/1303813004/ Base URL: https://chromium.g

[v8-dev] Re: Introduce SharedFunctionInfo::Iterator and Script::Iterator. (issue 1300333003 by yang...@chromium.org)

lgtm https://codereview.chromium.org/1300333003/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Moving ArraySplice Builtin to ElementsAccessor (issue 1293683005 by cbr...@chromium.org)

Nice test coverage, too! LGTM. https://codereview.chromium.org/1293683005/diff/11/src/builtins.cc File src/builtins.cc (right): https://codereview.chromium.org/1293683005/diff/11/src/builtins.cc#newcode180 src/builtins.cc:180: bool ClampedToInteger(Object* object, int* out) { I like t

[v8-dev] Re: Add deserialized scripts to script list. (issue 1296793003 by yang...@chromium.org)

lgtm https://codereview.chromium.org/1296793003/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] VectorICs: New interface descriptor for vector transitioning stores. (issue 1292173003 by mvstan...@chromium.org)

Reviewers: Benedikt Meurer, Message: Hi Benedikt, here is the CL we discussed that allows vector store ICs to successfully carry out transitioning stores. Some platforms (ahem, ia32) need to use stack for stub arguments, and this CL allows that to work. Thanks for the look, --Michael Descri

[v8-dev] Re: Keep track of script objects in a weak fixed array. (issue 1297273005 by yang...@chromium.org)

Hi Michael, can you have a look at this one too? https://codereview.chromium.org/1297273005/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from

[v8-dev] Re: Keep track of script objects in a weak fixed array. (issue 1297273005 by yang...@chromium.org)

Hey, that is very nice :). LGTM. https://codereview.chromium.org/1297273005/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and

[v8-dev] Re: Do not compact weak fixed array when re-allocating new backing store. (issue 1294883004 by yang...@chromium.org)

LGTM with the one nit. https://codereview.chromium.org/1294883004/diff/1/src/objects.cc File src/objects.cc (right): https://codereview.chromium.org/1294883004/diff/1/src/objects.cc#newcode7868 src/objects.cc:7868: } How about calling set_last_used_index(index); here. https://codereview.chromi

[v8-dev] Cleanup: Remove unncessary leave_frame parameter from stub cache. (issue 1299213002 by mvstan...@chromium.org)

Reviewers: Yang, Message: Hi Yang, This parameter is a left-over from the days when vector-based loads were flagged. Now it gets in the way. Time to clean up! --Michael Description: Cleanup: Remove unncessary leave_frame parameter from stub cache. BUG= Please review this at https://codereview.

[v8-dev] Re: Native context: install JS builtins via container object. (issue 1296163003 by yang...@chromium.org)

lgtm https://codereview.chromium.org/1296163003/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Re: Native context: Fix issue when running prologue.js before runtime.js (issue 1287203006 by yang...@chromium.org)

lgtm https://codereview.chromium.org/1287203006/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Re: Native context: run prologue.js before runtime.js (issue 1294803004 by yang...@chromium.org)

lgtm https://codereview.chromium.org/1294803004/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Re: Reenable code recompilation verification. (issue 1267953002 by yang...@chromium.org)

lgtm https://codereview.chromium.org/1267953002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Re: VectorICs: refactoring to eliminate "for queries only" vector ic mode. (issue 1268783004 by mvstan...@chromium.org)

Ugh. It breaks the world because the ROOT_LIST needs the complete definition of TypeFeedbackVector and a circular dependency results. I'm playing with it, but doesn't look awesome... https://codereview.chromium.org/1268783004/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.g

[v8-dev] VectorICs: refactoring to eliminate "for queries only" vector ic mode. (issue 1268783004 by mvstan...@chromium.org)

Reviewers: Yang, Message: Thanks for the look, Yang! Description: VectorICs: refactoring to eliminate "for queries only" vector ic mode. Since we need the notion of a dummy vector ic, we can use that to avoid a special case of the IC constructor. Also, consolidate the two dummy ICs into one. B

[v8-dev] VectorICs: Crankshaft adaptations to deal with vector store ics. (issue 1266983002 by mvstan...@chromium.org)

Reviewers: Benedikt Meurer, Message: Hi Benedikt, thanks for the look! --Michael Description: VectorICs: Crankshaft adaptations to deal with vector store ics. Also, a one line fix in TurboFan to call the correct store ic. BUG= Please review this at https://codereview.chromium.org/1266983002/

[v8-dev] Re: [stubs] Unify (and optimize) implementation of ToObject. (issue 1266013006 by bmeu...@chromium.org)

LGTM one nit. :) https://codereview.chromium.org/1266013006/diff/1/src/ia32/interface-descriptors-ia32.cc File src/ia32/interface-descriptors-ia32.cc (right): https://codereview.chromium.org/1266013006/diff/1/src/ia32/interface-descriptors-ia32.cc#newcode109 src/ia32/interface-descriptors-ia32.

[v8-dev] Re: Create function name const assignment after parsing language mode. (issue 1260053004 by yang...@chromium.org)

Just one comment, and Andreas should generally agree that it's a sane idea :) LGTM. https://codereview.chromium.org/1260053004/diff/1/src/parser.cc File src/parser.cc (right): https://codereview.chromium.org/1260053004/diff/1/src/parser.cc#newcode4446 src/parser.cc:4446: result->Set(0, fact

[v8-dev] Re: Add debug mode flag to force recompiling debug code for verification. (issue 1261973004 by yang...@chromium.org)

Great idea! LGTM. https://codereview.chromium.org/1261973004/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving

[v8-dev] Re: Debugger: correctly redirect code with no stack check. (issue 1269733002 by yang...@chromium.org)

One nit. Looks okay otherwise. LGTM. "Apparently in a test involving eval, strict mode and const, we don't produce the same code on recompile..." Always eval/strict... :p Thanks for finding that, it will be seriously helpful to fix. --Michael https://codereview.chromium.org/1269733002/diff/

[v8-dev] Re: Verify that type feedback vector structure is the same on recompile. (issue 1255383004 by mvstan...@chromium.org)

Hi guys, As per discussion with Jakob, we'll want to check this in next Monday. I'm on vacation but will try and check the commit box that morning before leaving for the airport :p. All the best, --Michael https://codereview.chromium.org/1255383004/ -- -- v8-dev mailing list v8-dev@googlegr

[v8-dev] VectorICs: --print-ast now prints allocated vector slots (issue 1264823003 by mvstan...@chromium.org)

Reviewers: rossberg, Message: Hi Andreas, Here is the ast pretty printing support for vectors we discussed yesterday. Thanks for the look! --Michael Description: VectorICs: --print-ast now prints allocated vector slots Looks like this: --- AST --- FUNC . NAME "foo" . INFERRED NAME "" . RETURN .

[v8-dev] Re: Stop overallocating feedback vector slots. (issue 1262803002 by mvstan...@chromium.org)

Hi Benedikt, Here is a fix for the bug we saw the other day, where an IC slot is foolishly allocated for a VariableProxy on the LHS of an assignment operation. (also Property). Thanks for the look, --Michael https://codereview.chromium.org/1262803002/ -- -- v8-dev mailing list v8-dev@googlegr

[v8-dev] Stop overallocating feedback vector slots. (issue 1262803002 by mvstan...@chromium.org)

Reviewers: , Message: Hi Benedikt, Here is a fix for the bug we saw the other day, where an IC slot is foolishly allocated for a VariableProxy on the LHS of an assignment operation. (also Property). Thanks for the look, --Michael Description: Stop overallocating feedback vector slots. When a

[v8-dev] Re: Add test for referring function name for classes. (issue 1264603002 by yang...@chromium.org)

lgtm https://codereview.chromium.org/1264603002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Re: Pretenuring decision of outermost literal is propagated to inner literals. (issue 1263773002 by hpa...@chromium.org)

Right on, LGTM. https://codereview.chromium.org/1263773002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving e

[v8-dev] Re: Pretenuring decision of outermost literal is propagated to inner literals. (issue 1263773002 by hpa...@chromium.org)

Nice. Use site_context->top() and you won't have to pass the new parameter... https://codereview.chromium.org/1263773002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-d

[v8-dev] Bugfix: CCTest test-func-name-inference/InConstructor is broken (issue 1263763002 by mvstan...@chromium.org)

Reviewers: Yang, Message: Hi Yang, here is a fix for the function name inferrer. Recent changes in scoping caused functions that were lazily parsed to be eagerly parsed. It appears that the eager parse case didn't even have the possibility for the function name inferrer to work. I've added

[v8-dev] Bugfix: Incorrect type feedback vector structure on recompile. (issue 1256413005 by mvstan...@chromium.org)

Reviewers: Benedikt Meurer, Message: PTAL, thx! Description: Bugfix: Incorrect type feedback vector structure on recompile. Scoping rules are different on recompile vis-a-vis global loads. BUG=514526 Please review this at https://codereview.chromium.org/1256413005/ Base URL: https://chromium

[v8-dev] Verify that type feedback vector structure is the same on recompile. (issue 1255383004 by mvstan...@chromium.org)

Reviewers: Benedikt Meurer, Message: Hi Benedikt, here is the bug we were looking at... Description: Verify that type feedback vector structure is the same on recompile. Use a CHECK statement to find any violations of this rule. BUG= Please review this at https://codereview.chromium.org/12553

[v8-dev] Re: [stubs] Don't pass name to Load/StoreGlobalViaContext stubs. (issue 1259963002 by bmeu...@chromium.org)

Code stubs: LGTM. https://codereview.chromium.org/1259963002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving

[v8-dev] Re: VectorICs: vector [keyed]store ic MISS handling infrastructure. (issue 1255883002 by mvstan...@chromium.org)

Hi Michael, it's pretty unlikely that those OSR failures are due to this. It only changes code behind a flag that is turned off in the build. https://codereview.chromium.org/1255883002/diff/1/src/ic/ia32/ic-ia32.cc File src/ic/ia32/ic-ia32.cc (right): https://codereview.chromium.org/1255883002/

[v8-dev] Re: Debugger: correctly recompile toplevel eval functions for debugging. (issue 1258503003 by yang...@chromium.org)

one nit, LGTM. https://codereview.chromium.org/1258503003/diff/1/src/compiler.cc File src/compiler.cc (right): https://codereview.chromium.org/1258503003/diff/1/src/compiler.cc#newcode1229 src/compiler.cc:1229: isolate->debug()->OnAfterCompile(script); On 2015/07/27 10:16:49, Yang wrote: Movi

[v8-dev] Re: Remove ExternalArray, derived types, and element kinds (issue 1254623002 by joc...@chromium.org)

ic.cc, the ic and handler compilers, also code-stubs-hydrogen.cc LGTM. https://codereview.chromium.org/1254623002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" grou

[v8-dev] VectorICs: vector [keyed]store ic MISS handling infrastructure. (issue 1255883002 by mvstan...@chromium.org)

Reviewers: Jakob, Message: Hi Jakob, Here is support for vector stores to call MISS handlers appropriately. Also removed an unnecessary assert from ic.cc. Thanks for the look! --Michael Description: VectorICs: vector [keyed]store ic MISS handling infrastructure. BUG= Please review this at htt

[v8-dev] Re: VectorICs: VisitClassLiteral needs adjustment for IC slot usage. (issue 1252303002 by mvstan...@chromium.org)

Thx for the comments, note my comment over in ast-graph-builder.cc, VisitArrayLiteral(). https://codereview.chromium.org/1252303002/diff/20001/src/compiler/js-generic-lowering.cc File src/compiler/js-generic-lowering.cc (right): https://codereview.chromium.org/1252303002/diff/20001/src/compiler

[v8-dev] Re: Debugger: correctly redirect eval code. (issue 1258583002 by yang...@chromium.org)

lgtm https://codereview.chromium.org/1258583002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] VectorICs: VisitClassLiteral needs adjustment for IC slot usage. (issue 1252303002 by mvstan...@chromium.org)

Reviewers: Michael Starzinger, Message: Hi Michael, here is the thing I mentioned at lunch, thx! --Michael Description: VectorICs: VisitClassLiteral needs adjustment for IC slot usage. Also, generic lowering for keyed stores needs to handle the case when there is no IC slot available (it can us

[v8-dev] Re: Remove d8's interactive Javascript debugger. (issue 1243213004 by yang...@chromium.org)

lgtm https://codereview.chromium.org/1243213004/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Re: Reduce duplicate code in full-codegen across platforms. (issue 1255613002 by yang...@chromium.org)

lgtm https://codereview.chromium.org/1255613002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Re: Allow HydrogenCodeStubs to tolerate and use stack arguments. (issue 1250563004 by mvstan...@chromium.org)

ping. https://codereview.chromium.org/1250563004/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Eliminate redundant descriptor ElementTransitionAndStoreDescriptor. (issue 1248973002 by mvstan...@chromium.org)

Reviewers: Jakob, Message: Hi Jakob, Here is the change we discussed, thanks for the look! --Michael Description: Eliminate redundant descriptor ElementTransitionAndStoreDescriptor. It's just the same as StoreTransitionDescriptor. BUG= Please review this at https://codereview.chromium.org/124

[v8-dev] Re: Fix pushing of register in CallConstructStub outside frame. (issue 1245043003 by mstarzin...@chromium.org)

lgtm. https://codereview.chromium.org/1245043003/diff/20001/src/x64/code-stubs-x64.cc File src/x64/code-stubs-x64.cc (right): https://codereview.chromium.org/1245043003/diff/20001/src/x64/code-stubs-x64.cc#newcode1795 src/x64/code-stubs-x64.cc:1795: // rdi : the function to call edi, edx, etc.?

[v8-dev] Re: Allow for optimized code map to have zero entries. (issue 1249543005 by mstarzin...@chromium.org)

Totally awesome! LGTM. Do you have a test that creates turbofan context-independent code that gets found with this new approach? Thanks, --Michael https://codereview.chromium.org/1249543005/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received

[v8-dev] Allow HydrogenCodeStubs to tolerate and use stack arguments. (issue 1250563004 by mvstan...@chromium.org)

Reviewers: danno, Message: Hi Danno, Here is the enhancement we discussed. The other platforms will look like x64, continuing to use a register to pass a map to the transitioning store. I'll upload them shortly... --Michael Description: Allow HydrogenCodeStubs to tolerate and use stack argume

[v8-dev] Special printing for type feedback vectors. (issue 1225403005 by mvstan...@chromium.org)

Reviewers: Benedikt Meurer, Message: Hi Benedikt, Here is the printing function we discussed a little while ago. Description: Special printing for type feedback vectors. Gdb macro jfv on an object will print it as a feedback vector. Printouts look like this: DebugPrint: 0x5dc0d2ad: [TypeFeedba

[v8-dev] Scoping error caused crash in CallICNexus::StateFromFeedback (issue 1231343003 by mvstan...@chromium.org)

Reviewers: rossberg, Message: Hi Andreas, here are the changes. I noticed that in test-parser.cc there are now arrow function tests of the kind you proposed, like (eval) => { "use strict"; }, so I didn't need to add them. Thanks, --Michael Description: Scoping error caused crash in CallICNe

[v8-dev] Re: Switch CallConstructStub to take new.target in register. (issue 1237813002 by mstarzin...@chromium.org)

LGTM on the rest of the CL. https://codereview.chromium.org/1237813002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop

[v8-dev] Re: Remove unused bailout reasons. (issue 1237623003 by mstarzin...@chromium.org)

Cool! LGTM. https://codereview.chromium.org/1237623003/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving email

[v8-dev] Re: Debugger: do not compile IC for accessors when debugging. (issue 1220283009 by yang...@chromium.org)

Cool! Okay, LGTM. https://codereview.chromium.org/1220283009/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving

[v8-dev] Re: Add unoptimized/optimized variants of MathFloor TF code stub (issue 1225943002 by da...@chromium.org)

LGTM modulo a couple comments. Nice to see this coming together! https://codereview.chromium.org/1225943002/diff/20001/src/code-stubs.h File src/code-stubs.h (right): https://codereview.chromium.org/1225943002/diff/20001/src/code-stubs.h#newcode102 src/code-stubs.h:102: V(MathFloor)

[v8-dev] Some extra checks to debug a crash failure. (issue 1219533003 by mvstan...@chromium.org)

Reviewers: Jakob, Message: Hi Jakob, Could you have a look? I'd be keen to get this in a chrome build. Thanks for the help, --Michael Description: Some extra checks to debug a crash failure related to type feedback vector. It appears that we have some path that recompiles code that looks quite

[v8-dev] Re: Debugger: clear ICs on activating step-in to correctly flood accessor pairs. (issue 1212253009 by yang...@chromium.org)

lgtm https://codereview.chromium.org/1212253009/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Re: Reland concurrent sweeping of code space. (issue 1225733002 by hpa...@chromium.org)

lgtm https://codereview.chromium.org/1225733002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from

[v8-dev] Re: Make context register implicit for CallInterfaceDescriptors (issue 1211333003 by da...@chromium.org)

Hi Danno, LGTM, just one question. If you can keep the const it would be nice. right on, --Michael https://codereview.chromium.org/1211333003/diff/80001/src/code-stubs.h File src/code-stubs.h (right): https://codereview.chromium.org/1211333003/diff/80001/src/code-stubs.h#newcode217 src/code-stu

[v8-dev] Re: Debugger: consistently ignore native and extension scripts for debugging. (issue 1216193002 by yang...@chromium.org)

Cool, I like the wrapping up of two predicates into IsSubjectToDebugging(). LGTM. https://codereview.chromium.org/1216193002/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v

[v8-dev] Re: Vector ICs: Introduce an InstanceType for the type feedback vector. (issue 1217943004 by mvstan...@chromium.org)

Okay, thanks! Landing... https://codereview.chromium.org/1217943004/diff/20001/src/heap/heap.cc File src/heap/heap.cc (right): https://codereview.chromium.org/1217943004/diff/20001/src/heap/heap.cc#newcode4726 src/heap/heap.cc:4726: // object = EnsureDoubleAligned(this, object, size); On 2015/0

[v8-dev] Re: Debugger: allow recompiling toplevel code for debugging. (issue 1214953003 by yang...@chromium.org)

Looks good. LGTM. https://codereview.chromium.org/1214953003/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving

[v8-dev] Re: Debugger: allow recompiling toplevel code for debugging. (issue 1214953003 by yang...@chromium.org)

Looks good. LGTM. https://codereview.chromium.org/1214953003/ -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving

[v8-dev] Vector ICs: Introduce an InstanceType for the type feedback vector. (issue 1217943004 by mvstan...@chromium.org)

Reviewers: Benedikt Meurer, Hannes Payer, Message: Hi Benedikt, hi Hannes, Benedikt, could you review the CL generally, and Hannes, the heap changes specifically? Thanks yall, --Michael Description: Vector ICs: Introduce an InstanceType for the type feedback vector. For better validation, prin

[v8-dev] Vector ICs: Changes to the IC system to support vector-based stores. (issue 1213773002 by mvstan...@chromium.org)

Reviewers: Yang, Message: Hi Yang, Here is one more in the series, thanks for the help with lithium before. This CL has the necessary changes to IC classes to support [keyed]store ICs that accept a vector and slot. Thanks for the look! --Michael Description: Vector ICs: Changes to the IC sy

[v8-dev] Re: VectorICs: Lithium support for vector-based stores. (issue 1209903003 by mvstan...@chromium.org)

Thanks for the review :) --Michael https://codereview.chromium.org/1209903003/diff/20001/src/arm64/lithium-codegen-arm64.cc File src/arm64/lithium-codegen-arm64.cc (right): https://codereview.chromium.org/1209903003/diff/20001/src/arm64/lithium-codegen-arm64.cc#newcode5501 src/arm64/lithium-cod

  1   2   3   4   5   6   7   8   9   10   >