Bug 159068

Summary: [iOS] A WebPageProxy in closed state should not be allowed to hold a process assertion
Product: WebKit Reporter: Chris Dumez <cdumez>
Component: WebKit2Assignee: Chris Dumez <cdumez>
Status: RESOLVED FIXED    
Severity: Normal CC: andersca, barraclough, dbates, kling, rniwa, sam, thorton, webkit-bug-importer
Priority: P1 Keywords: InRadar
Version: WebKit Nightly Build   
Hardware: Unspecified   
OS: iOS 9.3   
Attachments:
Description Flags
Patch none

Chris Dumez
Reported 2016-06-23 12:22:49 PDT
A WebPageProxy in closed state should not be allowed to hold a process assertion. It currently happens that WebPageProxy::close() gets called but that the WebPageProxy is kept alive by Safari (e.g. In case of top-hit preloading). In such cases, we fail to release the foreground process assertion that the WebPageProxy holds. We should clear the foreground process assertion that the WebPageProxy holds as soon as WebPageProxy::close() is called.
Attachments
Patch (1.93 KB, patch)
2016-06-23 12:26 PDT, Chris Dumez 		no flags
DetailsFormatted DiffDiff
View All Add attachment proposed patch, testcase, etc.
Chris Dumez
Comment 1 2016-06-23 12:23:22 PDT
rdar://problem/17665473
Chris Dumez
Comment 2 2016-06-23 12:26:53 PDT
Created attachment 281928 [details] Patch
Tim Horton
Comment 3 2016-06-23 12:52:51 PDT
Comment on attachment 281928 [details] Patch It doesn't make sense for Safari to keep the view alive if it has been closed; you can't come back from being closed, as far as I remember. Are you sure that is intentional?
Chris Dumez
Comment 4 2016-06-23 12:56:16 PDT
(In reply to comment #3) > Comment on attachment 281928 [details] > Patch > > It doesn't make sense for Safari to keep the view alive if it has been > closed; you can't come back from being closed, as far as I remember. Are you > sure that is intentional? It may not be intentional, I'll investigate separately. I definitely see close() getting called but not the destructor. Do you mind if we still land as is though? It cannot hurt to release the assertion as soon as close() is called I believe.
Chris Dumez
Comment 5 2016-06-23 12:59:09 PDT
Comment on attachment 281928 [details] Patch Clearing flags on attachment: 281928 Committed r202393: <http://trac.webkit.org/changeset/202393>
Chris Dumez
Comment 6 2016-06-23 12:59:14 PDT
All reviewed patches have been landed. Closing bug.
Note You need to log in before you can comment on or make changes to this bug.