http://codereview.chromium.org/149542/diff/7/1008 File src/objects.cc (right):
http://codereview.chromium.org/149542/diff/7/1008#newcode212 Line 212: return Object::GetPropertyWithDefinedGetter(receiver, On 2009/07/13 20:12:19, Mads Ager wrote: > Since the debugger can cause GCs, you have to wrap 'receiver' in a handle as > well and have a call to GetPropertyWithDefinedGetter inside the if (using the > handles). Moved the code into SetPropertyWithDefinedSetter where the receiver and the setter are already protected. This also covers some more cases of getter/setterr definition, e.g. accessors for indexed properties (tests for those cases are also added). > Would it be possible to add a test that forces a GC in the HandleStepIn code? I wonder if there is a way to enforce GC, say, on each allocation or something like this so that I could run the test with a flag enabling such behavior and check that all objects that need to be wrapped in handles are wrapped in them? http://codereview.chromium.org/149542/diff/7/1008#newcode1627 Line 1627: return SetPropertyWithDefinedSetter(JSFunction::cast(setter), value); On 2009/07/13 20:12:19, Mads Ager wrote: > Same as above. You have to wrap value in a handle as well for the debugger case > and have a call to SetPropertyWIthDefinedSetter inside of the if (using the > handles). This code as well as its counterpart for getters has been moved into SetPropertyWithDefinedSetter where the receiver is already protected by a handle. http://codereview.chromium.org/149542 --~--~---------~--~----~------------~-------~--~----~ v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev -~----------~----~----~----~------~----~------~--~---
