[Webkit-unassigned] [Bug 199719] New: [MSE][GStreamer] WebKitMediaSrc rework

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Thu Jul 11 12:55:23 PDT 2019


https://bugs.webkit.org/show_bug.cgi?id=199719

            Bug ID: 199719
           Summary: [MSE][GStreamer] WebKitMediaSrc rework
           Product: WebKit
           Version: WebKit Nightly Build
          Hardware: Unspecified
                OS: Unspecified
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: WebKitGTK
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: aboya at igalia.com
                CC: bugs-noreply at webkitgtk.org

This patch reworks the WebKitMediaSrc element and much of the player
private methods that interacted with it.

In comparison with the old WebKitMediaSrc, in the new one seeks have
been massively simplified.

The new WebKitMediaSrc no longer relies on a bin or appsrc, having
greater control over its operation. This made it comparatively much
easier to implement features such as seek before playback or
single-stream flushing.

Additional tests have been added to check some assumptions, and some
bugs that have surfaced with the changes have been fixed but no new
features (like multi-track support) are implemented in this patch.

One instance of these bugs is `resized` events, which were previously
being emitted when frames with different resolutions where appended.
This is a wrong behavior that has not been preserved in the rework, as
resize events should be emitted when the frames are shown, not
just appended.

There are subtler bugfixes, such as ignoring PTS-less frames in
AppendPipeline::appsinkNewSample(). These frames are problematic for
MSE, yet they were somehow passing through the pipelines. Since
WebKitMediaSrc is stricter with assertions, these have to be filtered.

This test gets rid of !m_mseSeekCompleted assertion failures in tests
and potentially other hard to debug bugs in the previous seek
algorithm.

-- 
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/20190711/e1844585/attachment.html>


More information about the webkit-unassigned mailing list