Bug 64241 - WebKit2 is leaking NSCursors created by leakNamedCursor
Summary: WebKit2 is leaking NSCursors created by leakNamedCursor
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: Anders Carlsson
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-07-10 11:30 PDT by Anders Carlsson
Modified: 2011-07-10 19:08 PDT (History)
1 user (show)

See Also:

Attachments
Patch (12.75 KB, patch)
2011-07-10 11:40 PDT, Anders Carlsson 		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 Anders Carlsson 2011-07-10 11:30:54 PDT 
WebKit2 is leaking NSCursors created by leakNamedCursor
Comment 1 Anders Carlsson 2011-07-10 11:40:04 PDT 
Created attachment 100233 [details]
Patch
Comment 2 Darin Adler 2011-07-10 12:41:42 PDT 
Wouldn’t it be better to cache each of these NSCursor objects exactly once?
Comment 3 Darin Adler 2011-07-10 12:41:53 PDT 
Except for the custom cursor one?
Comment 4 Anders Carlsson 2011-07-10 16:10:41 PDT 
(In reply to comment #2)
> Wouldn’t it be better to cache each of these NSCursor objects exactly once?

We're already caching them in the Cursor singletons in WebKit1, and for WebKit2 that's what the forcing of platformCursor does. I don't like the idea of having three levels of caches for some Cursors (WKSI, CursorMac.mm and Cursor.cpp).
Comment 5 Darin Adler 2011-07-10 17:32:08 PDT 
(In reply to comment #4)
> for WebKit2 that's what the forcing of platformCursor does

So I guess that’s the real fix. Not sure the rest of this patch was needed.
Comment 6 Anders Carlsson 2011-07-10 19:02:15 PDT 
(In reply to comment #5)
> (In reply to comment #4)
> > for WebKit2 that's what the forcing of platformCursor does
> 
> So I guess that’s the real fix. Not sure the rest of this patch was needed.

Right. I figured I'd do that part anyway since I don't like having potential leaks present in the code. In the future I think we can get rid of the Cursor singletons and cache the underlying platform cursors instead.
Comment 7 Anders Carlsson 2011-07-10 19:08:47 PDT 
Committed r90705: <http://trac.webkit.org/changeset/90705>