Bug 38754

Summary: REGRESSION(58736?): fast/loader/stateobjects/document-destroyed-navigate-back-with-fragment-scroll.html fails intermittently
Product: WebKit Reporter: Eric Seidel (no email) <eric>
Component: Tools / TestsAssignee: Nobody <webkit-unassigned>
Status: RESOLVED WORKSFORME    
Severity: Normal CC: beidson, jedrzej.nowacki, s3lance
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: PC   
OS: OS X 10.5   
Bug Depends on: 36201    
Bug Blocks:    

Description Eric Seidel (no email) 2010-05-07 09:15:49 PDT
REGRESSION(58736?): fast/loader/stateobjects/document-destroyed-navigate-back-with-fragment-scroll.html fails intermittently

I believe 58736 is the most likely cause.

The test is recording when the hashChanged event happening.  Now that it's async I suspect that the order is no longer deterministic, thus the failures.

We need to either fix/skip the test or rollout the change.
Comment 1 Eric Seidel (no email) 2010-05-07 09:17:12 PDT
http://build.webkit.org/results/GTK%20Linux%2064-bit%20Debug/r58944%20(5684)/fast/loader/stateobjects/pushstate-with-fragment-urls-and-hashchange-diffs.txt

--- /home/slave/webkitgtk/gtk-linux-64-debug/build/layout-test-results/fast/loader/stateobjects/pushstate-with-fragment-urls-and-hashchange-expected.txt	2010-05-07 06:28:08.119099757 -0700
+++ /home/slave/webkitgtk/gtk-linux-64-debug/build/layout-test-results/fast/loader/stateobjects/pushstate-with-fragment-urls-and-hashchange-actual.txt	2010-05-07 06:28:08.119099757 -0700
@@ -19,8 +19,8 @@
 State popped with event null (type object) and last path component some-other.html?withquery#
 Hash change fired and last path component is some-other.html?withquery#
 State popped with event null (type object) and last path component some-other.html?withquery
-Hash change fired and last path component is some-other.html?withquery
 State popped with event null (type object) and last path component some-other.html
+Hash change fired and last path component is some-other.html
 State popped with event null (type object) and last path component pushstate-with-fragment-urls-and-hashchange.html#
 State popped with event null (type object) and last path component pushstate-with-fragment-urls-and-hashchange.html#
 State popped with event null (type object) and last path component pushstate-with-fragment-urls-and-hashchange.html#otherhash
Comment 2 Steven Lai 2010-05-07 09:54:20 PDT
Then we must have this one fixed
http://bugs.webkit.org/show_bug.cgi?id=36335
and modify this test to use the new argument in order to make this work.

Otherwise, r58736 it would make pushState/popState useless.
Comment 3 Steven Lai 2010-05-07 10:03:23 PDT
it's better to rollback 58736 first then and implement the new interface for hashchange/pushstate event :(
Comment 4 Brady Eidson 2010-05-07 11:55:28 PDT
(In reply to comment #3)
> it's better to rollback 58736 first then and implement the new interface for
> hashchange/pushstate event :(

Fixing 36335 is pretty straightforward, we should just do that really quick.  If the intermittent SnowLeopard failure is getting in the way in the mean time (but I don't think we have a SL ews bot quite yet) then we could skip it, and re-enable when landing the fix for 36335
Comment 5 Steven Lai 2010-05-10 19:18:57 PDT
Filed patch to https://bugs.webkit.org/show_bug.cgi?id=36335
Comment 6 Jędrzej Nowacki 2010-05-20 04:47:16 PDT
I think that the bug 39418 could be related to this one.
Comment 7 Jędrzej Nowacki 2010-05-25 04:14:39 PDT
I'm closing the bug, because I think that it is not valid anymore. Feel free to reopen it in case if I'm wrong.

Checked with: http://build.webkit.org/results/GTK Linux 64-bit Debug/r60144 (6316)/fast/