Bug 44038

Summary: HTML5 fragment parsing doesn't load iframes
Product: WebKit Reporter: Adam Barth <abarth>
Component: Page LoadingAssignee: Nobody <webkit-unassigned>
Status: RESOLVED FIXED    
Severity: Normal CC: commit-queue, eric
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: All   
OS: All   
Attachments:
Description Flags
reduced test case
none
Patch
none
Patch for landing none

Description Adam Barth 2010-08-15 14:27:19 PDT
Created attachment 64454 [details]
reduced test case

This is causing some layout test failures with --html5-treebuilder turned on.  I've attached a reduction.
Comment 1 Adam Barth 2010-08-15 14:37:47 PDT
The problem is that Document::adoptNode is calling setRemainsAliveOnRemovalFromTree and confusing the frame.  I think we're supposed to construct the fragment completely detached but somehow it's getting attached?
Comment 2 Eric Seidel (no email) 2010-08-15 20:49:03 PDT
Fragments shouldn't get attached at all. I suspect we have a bug in HTMLConstructionSite.
Comment 3 Adam Barth 2010-08-15 22:14:27 PDT
Created attachment 64465 [details]
Patch
Comment 4 Alexey Proskuryakov 2010-08-16 00:21:31 PDT
Comment on attachment 64465 [details]
Patch

r=me

The test output doesn't really make it obvious that it's a pass. Did you intent to keep the numbered alerts in?
Comment 5 Adam Barth 2010-08-16 00:28:53 PDT
Created attachment 64473 [details]
Patch for landing
Comment 6 WebKit Commit Bot 2010-08-16 01:04:42 PDT
Comment on attachment 64473 [details]
Patch for landing

Clearing flags on attachment: 64473

Committed r65396: <http://trac.webkit.org/changeset/65396>
Comment 7 WebKit Commit Bot 2010-08-16 01:04:47 PDT
All reviewed patches have been landed.  Closing bug.
Comment 8 Eric Seidel (no email) 2010-08-16 07:20:43 PDT
Comment on attachment 64473 [details]
Patch for landing

Thanks.  I agree the test output is a bit confusing. :)  But I've very glad for the bug fix!