Bug 69466 - Added a simpler mechanism for registering one-off finalizers
Summary: Added a simpler mechanism for registering one-off finalizers
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Geoffrey Garen
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-10-05 14:39 PDT by Geoffrey Garen
Modified: 2011-10-05 19:52 PDT (History)
3 users (show)

See Also:

Attachments
Patch (13.47 KB, patch)
2011-10-05 14:42 PDT, Geoffrey Garen 		oliver: review+
Details | Formatted Diff | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Geoffrey Garen 2011-10-05 14:39:33 PDT 
Added a simpler mechanism for registering one-off finalizers
Comment 1 Geoffrey Garen 2011-10-05 14:42:07 PDT 
Created attachment 109864 [details]
Patch
Comment 2 Geoffrey Garen 2011-10-05 15:43:33 PDT 
Committed r96760: <http://trac.webkit.org/changeset/96760>
Comment 3 Adam Roben (:aroben) 2011-10-05 15:49:41 PDT 
Comment on attachment 109864 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=109864&action=review

> Source/JavaScriptCore/heap/Heap.cpp:706
> +    Weak<JSCell> weak(*globalData(), cell, &m_finalizerOwner, reinterpret_cast<void*>(finalizer));

I'm surprised a cast is needed here.
Comment 4 Geoffrey Garen 2011-10-05 18:57:34 PDT 
> I'm surprised a cast is needed here.

C/C++ are based on an abstract machine in which pointers to data (which includes void*) and pointers to functions might be in different address spaces or of different sizes, or where one might be a direct pointer and the other a handle. So, in theory, casting between the two is illegal. That's why reinterpret_cast is required here.
Comment 6 Ryosuke Niwa 2011-10-05 19:52:01 PDT 
Fixed Windows build in http://trac.webkit.org/changeset/96783.