Bug 29154 - storage/database-lock-after-reload.html is flakey
Summary: storage/database-lock-after-reload.html is flakey
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: 528+ (Nightly build)
Hardware: PC OS X 10.5
: P2 Normal
Assignee: Eric Seidel (no email)
URL:
Keywords:
Depends on:
Blocks: 33289
  Show dependency treegraph
 
Reported: 2009-09-10 14:53 PDT by Eric Seidel (no email)
Modified: 2012-09-19 13:10 PDT (History)
4 users (show)

See Also:


Attachments
Patch (1.07 KB, patch)
2010-01-28 12:40 PST, Eric Seidel (no email)
hamaji: review+
commit-queue: commit-queue-
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Eric Seidel (no email) 2009-09-10 14:53:36 PDT
storage/database-lock-after-reload.html is flakey

It has now caused 2 commit-queue runs to fail (patches from bug 27632 and bug 29041).

The failing result is:
--- /tmp/layout-test-results/storage/database-lock-after-reload-expected.txt    2009-09-10 14:42:33.000000000 -0700
+++ /tmp/layout-test-results/storage/database-lock-after-reload-actual.txt      2009-09-10 14:42:33.000000000 -0700
@@ -1,3 +1,4 @@
 Inserting some data
+Test failed - database is locked
 Test part 2 Complete

I've not looked through the buildbot logs yet, but I suspect that we'll find a failure of this test on them as well.
Comment 1 Eric Seidel (no email) 2009-09-10 14:55:06 PDT
DumpRenderTree uses its own separate database storage, right?  It wouldn't be using the same storage as Safari?

Also DumpRenderTree makes sure to unlock/re-init/whatever the database before running all the tests, right?
Comment 2 Ben Murdoch 2009-09-11 02:53:40 PDT
(In reply to comment #1)
> DumpRenderTree uses its own separate database storage, right?  It wouldn't be
> using the same storage as Safari?

Yes, it should create it's databases in a separate folder to Safari.

> Also DumpRenderTree makes sure to unlock/re-init/whatever the database before
> running all the tests, right?

There's a method on the layoutTestController to clear the databases that should reset everything but I think each test needs to invoke it during setup. database-lock-after-reload does, however.

Can we tell from what point the test started to become flakey? It was checked in under r45594 about 2 months ago and as far as I know didn't cause any problems then.

Cheers, Ben
Comment 3 Eric Seidel (no email) 2009-09-21 18:54:43 PDT
Failed again just now:
http://build.webkit.org/results/Leopard%20Intel%20Release%20(Tests)/r48615%20(5210)/results.html
Comment 4 Eric Seidel (no email) 2009-09-21 19:09:18 PDT
Unfortunately, we're not tracking flakey tests across time like Chromium does:
http://src.chromium.org/viewvc/chrome/trunk/src/webkit/tools/layout_tests/flakiness_dashboard.html
so I can't tell you when it started failing.
Comment 5 Eric Seidel (no email) 2009-09-25 18:46:59 PDT
It looks like sometimes it causes DumpRenderTree to crash!
http://build.webkit.org/results/Leopard%20Intel%20Release%20(Tests)/r48779%20(5418)/results.html
This is even more bad news bears. :(
Comment 6 Eric Seidel (no email) 2009-10-07 09:40:00 PDT
Bug 30172 may be related.
Comment 7 Eric Seidel (no email) 2010-01-14 15:20:36 PST
This failed on Tiger just now:
http://build.webkit.org/results/Tiger%20Intel%20Release/r53291%20(7775)/storage/database-lock-after-reload-diffs.txt
--- layout-test-results/storage/database-lock-after-reload-expected.txt	2010-01-14 14:58:12.000000000 -0800
+++ layout-test-results/storage/database-lock-after-reload-actual.txt	2010-01-14 14:58:12.000000000 -0800
@@ -1,3 +1 @@
-Inserting some data
-Test part 2 Complete
Comment 8 Eric Seidel (no email) 2010-01-14 16:23:25 PST
This just failed on Leopard Intel Release as well:
http://build.webkit.org/results/Leopard%20Intel%20Release%20(Tests)/r53292%20(9385)/storage/database-lock-after-reload-diffs.txt

This is vying for most-flakey test. :(
Comment 9 Eric Seidel (no email) 2010-01-24 17:16:52 PST
Failed on Tiger again just now:
http://build.webkit.org/results/Tiger%20Intel%20Release/r53784%20(8113)/storage/database-lock-after-reload-diffs.txt

Same results a seen in comment #7
Comment 12 Eric Seidel (no email) 2010-01-28 12:40:59 PST
Created attachment 47642 [details]
Patch
Comment 13 Shinichiro Hamaji 2010-02-04 00:57:44 PST
Comment on attachment 47642 [details]
Patch

I totally agree to disable this test until Bug 34506 is closed! This flakiness is alive long time...
Comment 14 Eric Seidel (no email) 2010-02-05 12:02:47 PST
Attachment 47642 [details] was posted by a committer and has review+, assigning to Eric Seidel for commit.
Comment 15 WebKit Commit Bot 2010-02-08 22:32:19 PST
Comment on attachment 47642 [details]
Patch

Rejecting patch 47642 from commit-queue.

Failed to run "['/Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/svn-apply', '--reviewer', 'Shinichiro Hamaji', '--force']" exit_code: 1
patching file LayoutTests/ChangeLog
Hunk #1 succeeded at 1 with fuzz 3.
patching file LayoutTests/platform/mac/Skipped
Hunk #1 FAILED at 101.
1 out of 1 hunk FAILED -- saving rejects to file LayoutTests/platform/mac/Skipped.rej

Full output: http://webkit-commit-queue.appspot.com/results/251169
Comment 16 Eric Seidel (no email) 2010-02-09 02:02:00 PST
Committed r54540: <http://trac.webkit.org/changeset/54540>
Comment 17 Simon Fraser (smfr) 2012-09-19 13:10:36 PDT
Seems to work now. Going to try unskipping.