Bug 34222 – http/tests/navigation/anchor-basic.html is flaky on Snow Leopard Release Bot

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Eric Seidel (no email)]

http/tests/navigation/anchor-basic.html timed out on Snow Leopard Release Bot

http://build.webkit.org/results/SnowLeopard%20Intel%20Release%20(Tests)/r53938%20(4875)/results.html

Been here for a long time:
http://trac.webkit.org/browser/trunk/LayoutTests/http/tests/navigation/anchor-basic.html

[Ryosuke Niwa]

See new diff:
http://build.webkit.org/results/SnowLeopard%20Intel%20Release%20(Tests)/r95284%20(33129)/http/tests/navigation/anchor-basic-pretty-diff.html

This test fails every now and then on Snow Leopard bots.  Note that this test is skipped on Leopard (see the bug 34222).

[Eric Seidel (no email)]

Do we know what other test result its showing there?  It seems like some other test might be polluting this one?

[Ryosuke Niwa]

(In reply to comment #2)
> Do we know what other test result its showing there?  It seems like some other test might be polluting this one?

http/tests/multipart/stop-crash.html

[Eric Seidel (no email)]

That test is even older.
http://trac.webkit.org/browser/trunk/LayoutTests/http/tests/multipart/stop-crash.html

We obviously could disable one or both of these tests, but best would be to understand the interaction going on and fix the underlying bug in WebCore or DumpRenderTree.

[Tony Chang]

Created attachment 108918 [details]
Patch

[Ryosuke Niwa]

Comment on attachment 108918 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=108918&action=review

> LayoutTests/ChangeLog:10
> +        Add waitUntilDone/notifyDone to eliminate flakiness on mac.  Without
> +        it, the img onload handler doesn't always run.  Also reduce the
> +        multipart timeout (should help to unblock other http requests).

Could you re-format the comment so that it doesn't use double-space? Also it seems awkward for "it" to appear at the beginning of a line or "the" to appear at the end of a line.

[Tony Chang]

Created attachment 108935 [details]
Patch for landing

[WebKit Review Bot]

Comment on attachment 108935 [details]
Patch for landing

Clearing flags on attachment: 108935

Committed r96174: <http://trac.webkit.org/changeset/96174>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Csaba Osztrogonác]

It make http/tests/multipart/stop-crash.html fail on the GTK and the Qt bots:

--- /var/lib/buildbot/build/gtk-linux-32-release/build/layout-test-results/http/tests/multipart/stop-crash-expected.txt 
+++ /var/lib/buildbot/build/gtk-linux-32-release/build/layout-test-results/http/tests/multipart/stop-crash-actual.txt 
@@ -1,6 +1,7 @@
+FAIL: Timed out waiting for notifyDone to be called
 Test for http://bugs.webkit.org/show_bug.cgi?id=13360 REGRESSION: Crash closing live web cam viewer page.
 
 If WebKit does not assert or crash after the test, then it passed.
 
 
-PASS
+

[Csaba Osztrogonác]

It seems it is a bug somewhere in GTK and Qt port.
New bug report filed: https://bugs.webkit.org/show_bug.cgi?id=68979

[Tony Chang]

http/tests/navigation/anchor-basic.html has passed every run after this patch landed.  Looks like that was the problem.

http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=http%2Ftests%2Fnavigation%2Fanchor-basic.html&group=%40ToT%20-%20webkit.org