media/W3C are failing badly on GTK. About 62 of those tests time out. From what I can see the video element tries to load URIs like this: "file:///path/to/WebKit/LayoutTests/media/content/test.wav?Thu Dec 22 2011 13:20:05 GMT+0100 (CET)0.6502569520380348" Which fails because of the parameters.
Created attachment 127597 [details] Patch
Comment on attachment 127597 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=127597&action=review What is the meaning of PASS "maybe" is "maybe" in the new baselines? Please take a look at the suggestion below before landing. > Source/WebCore/platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:288 > + KURL kurl(KURL(), url); > + String cleanUrl(url); I think it would be better here to do: if (kurl.isLocalFile()) { kurl.setQuery(""); kurl.removeFragmentIdentifier(); }
Comment on attachment 127597 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=127597&action=review >> Source/WebCore/platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:288 >> + String cleanUrl(url); > > I think it would be better here to do: > > if (kurl.isLocalFile()) { > kurl.setQuery(""); > kurl.removeFragmentIdentifier(); > } I actually tried this but setQuery() doesn't remove the leading '?'.
(In reply to comment #2) > (From update of attachment 127597 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=127597&action=review > > What is the meaning of PASS "maybe" is "maybe" in the new baselines? > This is how the tests are written. You can see those messages in the other media/W3C baselines too :) I think these new baselines are for tests skipped on Mac, actually. For now I think it's safer to have them in the gtk directory.
Committed r108327: <http://trac.webkit.org/changeset/108327>