Bug 88727

Summary: [GTK] http/tests/xmlhttprequest/origin-exact-matching.html is failing
Product: WebKit Reporter: Chris Dumez <cdumez>
Component: WebCore Misc.Assignee: Nobody <webkit-unassigned>
Status: RESOLVED FIXED    
Severity: Normal CC: zan
Priority: P2 Keywords: Gtk, LayoutTestFailure
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   
Bug Depends on: 88896    
Bug Blocks:    
Attachments:
Description Flags
Diff none

Chris Dumez
Reported 2012-06-10 04:10:31 PDT
The http/tests/xmlhttprequest/origin-exact-matching.html test introduced in r119911 (Bug 88139) is failing for EFL and GTK port. It looks like our behavior is different than expected for a few origins. We are also missing the "CONSOLE MESSAGE: XMLHttpRequest cannot load..." messages. Those are not printed because our error domain is "soup_http_error_quark" instead of errorDomainWebKitInternal.
Attachments
Diff (10.10 KB, text/plain)
2012-06-10 04:12 PDT, Chris Dumez 		no flags
Details
View All Add attachment proposed patch, testcase, etc.
Chris Dumez
Comment 1 2012-06-10 04:12:45 PDT
Created attachment 146729 [details] Diff
Chris Dumez
Comment 2 2012-06-10 04:19:20 PDT
Skipping test case for both EFL and GTK ports in Bug 88728.
Chris Dumez
Comment 3 2012-06-12 22:42:41 PDT
This does not fail for EFL anymore, unskipping test in Bug 88960. I don't see the tests in the unexpected passes on the GTK bot though so I'm keeping this bug open for GTK.
Zan Dobersek
Comment 4 2012-06-13 01:55:17 PDT
I went on and bisected libsoup changes, narrowing down the offending commit to this: http://git.gnome.org/browse/libsoup/commit/?id=9effb5ca942412ecde9242c745f2df6da80853a3 Brave new world it is.
Zan Dobersek
Comment 5 2012-06-13 01:57:37 PDT
Oops, wrong bug! Consider comment #4 irrelevant.
Zan Dobersek
Comment 6 2012-10-17 03:40:18 PDT
The test has been passing consistently lately so the expectation was removed. Closing the bug.
Note You need to log in before you can comment on or make changes to this bug.