| Summary: | [GTK][GStreamer] Flaky ASSERTION FAILED: m_lock.isHeld() in TextureMapperPlatformLayerProxy | ||||||
|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Alicia Boya García <aboya> | ||||
| Component: | WebKitGTK | Assignee: | Philippe Normand <pnormand> | ||||
| Status: | RESOLVED FIXED | ||||||
| Severity: | Normal | CC: | bugs-noreply, calvaris, pnormand | ||||
| Priority: | P2 | ||||||
| Version: | WebKit Nightly Build | ||||||
| Hardware: | Unspecified | ||||||
| OS: | Unspecified | ||||||
| Attachments: |
|
||||||
|
Description
Alicia Boya García
2019-04-09 11:35:02 PDT
What's your platform video decoder? The trace indicates the flush is done from the player destructor, that means a v4l2 is used. Can you confirm? (In reply to Philippe Normand from comment #1) > What's your platform video decoder? The trace indicates the flush is done > from the player destructor, that means a v4l2 is used. Can you confirm? I don't think so. According to pipeline dumps a decoder has not even been added to the pipeline at all in this test. 🤨 Oh I know, m_videoDecoderPlatform likely is initialized to the first enum value be default, that's no good. (In reply to Philippe Normand from comment #4) > Oh I know, m_videoDecoderPlatform likely is initialized to the first enum > value be default, that's no good. It could become a WTF::Optional<T>. I'll write a patch tomorrow if no one else does it first. Created attachment 368118 [details]
Patch
Comment on attachment 368118 [details]
Patch
I tend to think that the proper solution would be:
enum class WebKitGstVideoDecoderPlatform { Other, ImxVPU, Video4Linux };
WebKitGstVideoDecoderPlatform m_videoDecoderPlatform { Other }
Seems to me like a clear case for Optional. Comment on attachment 368118 [details] Patch (In reply to Philippe Normand from comment #8) > Seems to me like a clear case for Optional. I don't have a strong opinion so r+ Committed r244587: <https://trac.webkit.org/changeset/244587> |