Bug 115090

Summary: REGRESSION (r143102): Cannot correctly display http://atomsforpeace.info
Product: WebKit Reporter: aattyy
Component: Layout and RenderingAssignee: Nobody <webkit-unassigned>
Status: RESOLVED WONTFIX    
Severity: Normal CC: ap, jchaffraix, robert, simon.fraser, zalan
Priority: P1    
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   
URL: http://atomsforpeace.info
Attachments:
Description Flags
Screen shots
none
test reduction
none
test reduction none

aattyy
Reported 2013-04-23 23:28:50 PDT
The positions and sizes of the youtube videos are not right.
Attachments
Screen shots (3.24 MB, application/zip)
2013-04-24 09:11 PDT, aattyy 		no flags
Details
test reduction (214 bytes, text/html)
2013-04-24 10:56 PDT, alan 		no flags
Details
test reduction (200 bytes, text/html)
2013-04-24 11:16 PDT, alan 		no flags
Details
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.
Alexey Proskuryakov
Comment 1 2013-04-24 08:58:45 PDT
Could you please describe the issue in more detail? This site looks the same to me in Safari as it does in Firefox.
aattyy
Comment 2 2013-04-24 09:11:43 PDT
Created attachment 199493 [details] Screen shots Attached is how my computer display the first video (you have to scroll to the right to see it). The second and the third have the same issue.
Alexey Proskuryakov
Comment 3 2013-04-24 09:32:07 PDT
Regressed in <http://trac.webkit.org/r143102>. That said, the page is broken in Firefox too.
Robert Hogan
Comment 4 2013-04-24 10:38:34 PDT
(In reply to comment #3) > Regressed in <http://trac.webkit.org/r143102>. > > That said, the page is broken in Firefox too. Opera too - Opera/9.80 (X11; Linux i686) Presto/2.12.388 Version/12.15
Robert Hogan
Comment 5 2013-04-24 10:48:22 PDT
I don't think this is a regression - the video is in an iframe whose nearest ancestor with a specified height is its grandparent - which has a height of 600px and a margin of -300px. The position and height of the video look right to me given the CSS.
alan
Comment 6 2013-04-24 10:56:36 PDT
Created attachment 199501 [details] test reduction
alan
Comment 7 2013-04-24 11:16:38 PDT
Created attachment 199503 [details] test reduction This is about being in strict mode vs. sizing the replaced element based on the ancestor's auto height. Before r143102, we traversed from the RenderReplaced up to the RenderView and if _any_ of the ancestor's had non-auto height, we considered the RenderReplaced non-auto as well. r143102 changed it to restrict the check to the containing block only. This is not a regression.
Alexey Proskuryakov
Comment 8 2013-04-24 12:19:57 PDT
I would most definitely call this a regression, because a real web site got broken. However, the issue will need to be fixed server side, not in WebKit per the above analysis.
Note You need to log in before you can comment on or make changes to this bug.