[Webkit-unassigned] [Bug 177355] [GTK] Initial webm support in MSE

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Tue Sep 26 01:05:07 PDT 2017


Xabier Rodríguez Calvar <calvaris at igalia.com> changed:

           What    |Removed                     |Added
 Attachment #321540|review?                     |review-
              Flags|                            |

--- Comment #4 from Xabier Rodríguez Calvar <calvaris at igalia.com> ---
Comment on attachment 321540
  --> https://bugs.webkit.org/attachment.cgi?id=321540

View in context: https://bugs.webkit.org/attachment.cgi?id=321540&action=review

>> Source/WebCore/platform/graphics/gstreamer/mse/AppendPipeline.cpp:143
>> +        ASSERT_NOT_REACHED();
> This is the only concern I have. I suppose AppendPipeline wouldn't get created if the type doesn't match one of the two conditions?

A test here would be interesting if there is none. Feel free to submit it to the spec tests if there is none about this yet. If it is not accepted because of whatever, I think it still pays off to keep it as WebKit test. I guess the test would consist of creating a SourceBuffer with crap as type, try appending some allowed format and try to play it, expecting an error.

If the test fails already without calling this constructor, then we are good to go with the ASSERT_NOT_REACHED. If not, I agree with Žan here. I think the proper thing to do here would be to setting the AppendPipeline to Invalid state.

Then I see another problem coming and it is that apparently all state transitions from Invalid are "ok" and I think it shouldn't be like that. Only Invalid -> Invalid should be allowed. Everything else should be ok = false. I guess then we would have some ASSERTs in Debug mode that we'd have to fix in cases where somebody appends something to a pipeline in Invalid state. Am I missing anything?

You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20170926/c5d70e03/attachment-0001.html>

More information about the webkit-unassigned mailing list