- 17 Oct, 2011 4 commits
-
-
keuchel@chromium.org authored
Review URL: http://codereview.chromium.org/7904008 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9650 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
lrn@chromium.org authored
Just use script being native and FLAG_allow_natives_syntax directly. Review URL: http://codereview.chromium.org/8314018 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9643 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
jkummerow@chromium.org authored
BUG=v8:1767 TEST=mjsunit/numops-fuzz doesn't timeout in debug mode Review URL: http://codereview.chromium.org/8302023 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9642 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
mstarzinger@chromium.org authored
According to the ES5 spec all ">" and "<=" expressions should be be evaluated left-to-right. This obsoletes old hacks for reversing the order to be ES3 compliant. R=lrn@chromium.org BUG=v8:1752 Review URL: http://codereview.chromium.org/8275035 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9641 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
- 14 Oct, 2011 13 commits
-
-
svenpanne@chromium.org authored
Problematic code is e.g.: function typeofViaVariable() { var foo = typeof({}) return foo === "undefined"; } Review URL: http://codereview.chromium.org/8283036 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9640 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
svenpanne@chromium.org authored
Review URL: http://codereview.chromium.org/8291004 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9639 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
fschneider@chromium.org authored
It causes an assertion with deoptimizing from inlined code. Review URL: http://codereview.chromium.org/8277034 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9636 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
fschneider@chromium.org authored
TBR=whesse@chromium.org Review URL: http://codereview.chromium.org/8286023 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9635 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
fschneider@chromium.org authored
TBR=whesse@chromium.org Review URL: http://codereview.chromium.org/8286022 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9633 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
erik.corry@gmail.com authored
Review URL: http://codereview.chromium.org/8276030 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9626 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
ricow@chromium.org authored
We currently report how much memory is available without changing page, i.e., the actual amount of available memory is not correctly stated for new space when doing --trace-gc. Review URL: http://codereview.chromium.org/8275026 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9624 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
erik.corry@gmail.com authored
Review URL: http://codereview.chromium.org/8274032 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9622 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
ricow@chromium.org authored
This has already been fixed, we not actually use the constraints. Review URL: http://codereview.chromium.org/8276029 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9620 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
fschneider@chromium.org authored
1. Record AST id for CallFunctionStub. 2. Correctly extract cached target from CallFunctionStub inline cache. 3. Fix a bug when inling call as a function in effect or value context: Handle abnormal exits correcty. 4. Fix a bug when inlining call as a function in test context: drop function correctly from true and false block. 5. Avoid inlining mutually recursive functions by checking the stack of function states before inlining. This was not a bug, but is just a more general check to avoid recursive inlining. Review URL: http://codereview.chromium.org/8258012 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9619 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
fschneider@chromium.org authored
Omit smi check inside write barriers if the value is known to be a heap object. Refine inferred types of some instructions. Review URL: http://codereview.chromium.org/8256016 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9618 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
ulan@chromium.org authored
Currently, preparser-process crashes when given flags with arguments (e.g. --gc_interval 10). It can be fixed by ignoring everything in the command line except the "throws" command and its immediate arguments. This assumes that no flags appear between "throws" and its arguments. TEST=make ia32.release.check TESTFLAGS="preparser --special-command=\"@ --gc_interval 10\"" Review URL: http://codereview.chromium.org/8268004 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9617 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
danno@chromium.org authored
BUG= TEST= Review URL: http://codereview.chromium.org/8283025 Patch from Paul Lind <plind44@gmail.com>. git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9616 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
- 13 Oct, 2011 14 commits
-
-
yangguo@chromium.org authored
Review URL: http://codereview.chromium.org/8281001 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9615 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
vegorov@chromium.org authored
R=danno@chromium.org Review URL: http://codereview.chromium.org/8278001 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9614 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
rossberg@chromium.org authored
Fix & tweak some proxy-related error messages. R=kmillikin@chromium.org BUG=v8:1543 TEST= Review URL: http://codereview.chromium.org/8229008 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9613 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
svenpanne@chromium.org authored
Previously, comparisons against null/undefined and comparisons of the result of typeof against a constant string were done syntactically. Now we do this via pattern matching on the hydrogen graph, which opens up more opportunities for better code generation, e.g. the following comparisons are now recognized to be special: var bar = typeof foo; var baz = "undefined"; if (bar == baz) ... var blah = undefined; if (hurz == blah) ... If we did this handling of special cases even later at lithium generation time, even more cases could be recognized, but this would involve bigger changes and this CL handles most common cases. Review URL: http://codereview.chromium.org/8242002 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9612 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
fschneider@chromium.org authored
We run with --nolimit-inlining when --stress-opt is on. Review URL: http://codereview.chromium.org/8267007 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9610 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
mstarzinger@chromium.org authored
R=erik.corry@gmail.com Review URL: http://codereview.chromium.org/8261008 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9609 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
yangguo@chromium.org authored
Review URL: http://codereview.chromium.org/8261007 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9608 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
erik.corry@gmail.com authored
Review URL: http://codereview.chromium.org/8267004 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9607 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
erik.corry@gmail.com authored
Review URL: http://codereview.chromium.org/8256012 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9606 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
yangguo@chromium.org authored
BUG= TEST= Review URL: http://codereview.chromium.org/8241003 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9605 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
rossberg@chromium.org authored
Review URL: http://codereview.chromium.org/8262010 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9602 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
danno@chromium.org authored
This ports the following list of commits to mips. It is provided in this form so that the mips port of the new-gc can be easily tested on your end. This must be downloaded or landed after http://codereview.chromium.org/8106002 This is based on r9585. With these two mips commits, all tests are passing. 9319 Fix a harmless assert and a genuine bug in the GC-safety of stub generation ... 9329 ARM: Pregenerate some stubs that we call from other stubs. 9335 Initialize pre-allocated fields of JSObject with undefined. 9344 Put back the asserts in RememberedSetHelper, but correct this time 9370 Reorganize object type enum, such that proxies are no longer in the middle... 9392 Basic support for tracking smi-only arrays on ia32. 9402 Notify collector about lazily deoptimized code objects. 9411 Porting r9392 to arm (smi-only arrays). 9418 Small refactor to KeyedStoreIC::GenerateGeneric to make it slightly faster. 9447 Tighten up assertions checking GC-safety of stub calls. 9449 Record function call targets, use them for inlining. 9459 Make sure we don't flush the pregenerated stubs, since they need to b 9461 Fix the build on ARM 9466 Move the is_pregenerated flag so it does not overlap other flags.... 9468 Fix the no-VFP3 build on ARM. 9475 Pass correct anchor_slot for EMBEDDED_OBJECT pointers from 9490 Adjust assertions in UpdateSlot to match UpdatePointer in PointersUpdatingVisitor. 9511 Clean list of external references from internal objects like the hole value. 9514 Simplify compares in KeyedStoreIC::GenerateGeneric. 9531 Porting r9456 to arm (Optimize KeyedStoreGeneric for Smi arrays). 9541 Fix load of potentially eval-shadowed let bindings. 9542 Fast allocation of block contexts. 9553 Activate smi-only optimizations for large array literals. 9575 Move declaration of SerializedScopeInfo from variables.h to objects.h 9577 Track elements_kind transitions in KeyedStoreICs. 9583 Fixing a bug in arm as pointed out in issue 1759. 9584 Refactor and fix polymorphic KeyedStoreIC creation. BUG= TEST= Review URL: http://codereview.chromium.org/8112008 Patch from Paul Lind <pling44@gmail.com>. git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9601 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
danno@chromium.org authored
Simplified based on Michael's change Refactor how embedded pointers are visited. (9597) Ported r9328 (bdc13b7) BUG= TEST= Review URL: http://codereview.chromium.org/8106002 Patch from Paul Lind <pling44@gmail.com>. git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9600 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
jkummerow@chromium.org authored
TEST=Compiling Chromium with -Duse_system_v8 no longer prints a GYP warning about "invalid type 'settings'". Review URL: http://codereview.chromium.org/8256008 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9599 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
- 12 Oct, 2011 6 commits
-
-
yangguo@chromium.org authored
Review URL: http://codereview.chromium.org/8240007 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9598 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
mstarzinger@chromium.org authored
This refactoring (almost) gets rid of the requirement to get the target object address for an object pointer embedded in code objects. This is not possible on MIPS as pointers are encoded using two instructions. All usages of RelocInfo::target_object_address() are (almost) obsoleted by this change. The serializer still uses it, so MIPS will not yet work with snapshots turned on. R=danno@chromium.org,vegorov@chromium.org Review URL: http://codereview.chromium.org/8245007 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9597 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
mstarzinger@chromium.org authored
According to the ES5 spec the implementation of DefineOwnProperty() has to special case handling of arrays. This is a preliminary implementation correctly handling definition of array index properties, defining length properties is not completely covered yet. R=rossberg@chromium.org TEST=test262 Review URL: http://codereview.chromium.org/8221002 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9596 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
yangguo@chromium.org authored
Review URL: http://codereview.chromium.org/8229009 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9590 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
rossberg@chromium.org authored
Shorten --harmony-block-scoping to --harmony-scoping. R=keuchel@chromium.org BUG= TEST= Review URL: http://codereview.chromium.org/8226017 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9589 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
vegorov@chromium.org authored
R=fschneider@chromium.org BUG=v8:1763 Review URL: http://codereview.chromium.org/8240004 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9588 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
- 11 Oct, 2011 3 commits
-
-
mstarzinger@chromium.org authored
This is a temporary fix which avoids compaction when incremental marking is restarted during an old-space-step. That could turn the page that holds the chosen free list node into an evacuation candidate. It could also cause several other inconsistencies if it happens during scavenge. R=vegorov@chromium.org Review URL: http://codereview.chromium.org/8228010 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9585 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
jkummerow@chromium.org authored
Review URL: http://codereview.chromium.org/8233011 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9584 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-
yangguo@chromium.org authored
BUG=v8:1759 Review URL: http://codereview.chromium.org/8229012 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9583 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-