Use media-file.js to select a valid file type. Fixing this exposed a few other issues (also fixed): 1) update video-load-and-stall.cgi to take a time to stall for. Previously the cgi script would just exit on stall, which can (should?) be interpreted as a network error rather than a stall (which requires 3 seconds before firing). 2) video-play-before-metadata.html was waiting for a suspend event which isn't expected (in the *-expected.txt file). The suspend event is 'likely' to come after the onload, but might come before, making this test flakey.
Created attachment 88169 [details] Patch
Comment on attachment 88169 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=88169&action=review Great change, but I would prefer to have it split into three bugs so r- for now. > LayoutTests/ChangeLog:16 > + Fix the following which became exposed after unblocking the core format > + problem: > + > + - update video-load-and-stall.cgi to take a time to stall for. Previously the > + cgi script would just exit on stall, which can (should?) be interpreted as a > + network error rather than a stall (which requires 3 seconds before firing). > + > + - video-play-before-metadata.html was waiting for a suspend event which isn't > + expected (in the *-expected.txt file). The suspend event is 'likely' to come > + after the onload, but might come before, making this test flakey. I think it is worth fixing each of these in separate bugs because you are doing more than changing the test file used.
Created attachment 88506 [details] Patch
Good idea :-) Reduced down to just dealing with the file format.
Comment on attachment 88506 [details] Patch Forgot to reset the cq? flag with the new patch... Could you cq+? Thanks...
Comment on attachment 88506 [details] Patch Clearing flags on attachment: 88506 Committed r83210: <http://trac.webkit.org/changeset/83210>
All reviewed patches have been landed. Closing bug.