[Webkit-unassigned] [Bug 132923] [GTK] Layout Test http/tests/security/cross-origin-plugin-private-browsing-toggled.html is failing

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed May 14 18:25:20 PDT 2014


Carlos Alberto Lopez Perez <clopez at igalia.com> changed:

           What    |Removed                     |Added
                 CC|                            |jhoneycutt at apple.com,
                   |                            |jpfau at apple.com,
                   |                            |mario at webkit.org

--- Comment #1 from Carlos Alberto Lopez Perez <clopez at igalia.com>  2014-05-14 18:25:40 PST ---
(In reply to comment #0)
> Investigating the specific revision that made it start failing again is pending.

This is the summary of my investigation for this test on the GTK port (release build).

The following timeline was obtained with the help of https://github.com/clopez/webkit-testhunter

http://trac.webkit.org/changeset/156532 <-- Test stop failing on GTK (msanchez)
http://trac.webkit.org/changeset/159744 <-- Test get removed from GTK TestExpectations (msanchez)
http://trac.webkit.org/changeset/168668 <-- Test starts failing again on GTK (jhoneycutt)

Before r168668 the test only works if you run many tests in parallel (Fails otherwise)

For example:

Try to build r168667 and test to run all the http/* tests in parallel with:

Tools/Scripts/run-webkit-tests --no-show-results --no-new-test-results --no-sample-on-timeout --results-directory layout-test-results --debug-rwt-logging --release --webkit-test-runner --gtk --retry-failures http

You will see that http/tests/security/cross-origin-plugin-private-browsing-toggled.html passes.

Now, on r168667 run only the test http/tests/security/cross-origin-plugin-private-browsing-toggled.html alone:

Tools/Scripts/run-webkit-tests --no-show-results --no-new-test-results --no-sample-on-timeout --results-directory layout-test-results --debug-rwt-logging --release --webkit-test-runner --gtk --retry-failures http/tests/security/cross-origin-plugin-private-browsing-toggled.html

It will fail!.

Now, if you build r168668, and try the tests again, you will see that this test fails always. No matter if you run the test alone or in parallel with the other tests.

This drove me crazy because I was not able to reproduce the behaviour reported by the bot, until I tried to run all the test suite (like the bot does) instead of only running this test alone.


-> Before r168668: The test fails if you run it alone, but works if you run it in parallel with the other tests.

-> On r168668: The test fails always.

So maybe there is some race condition here?

Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

More information about the webkit-unassigned mailing list