Bug 208046 – [GStreamer][WPE] Add GstGLMemoryEGL support for the video-plane-display

• Status: RESOLVED
• Resolution: FIXED
• Component: Platform

[Philippe Normand]

Also ensure the stride is valid.

[Philippe Normand]

Created attachment 391389 [details]
Patch

[Philippe Normand]

Created attachment 391390 [details]
Patch

[Zan Dobersek]

Comment on attachment 391390 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=391390&action=review

> Source/WebCore/platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:224
> +        if (hasDMABuf() && (m_dmabufStride == -1)) {

Extra parentheses not necessary, but OK.

> Source/WebCore/platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:226
> +            m_isMapped = gst_video_frame_map(&m_videoFrame, &videoInfo, m_buffer.get(), GST_MAP_READ);
> +            RELEASE_ASSERT(m_isMapped);

Is this really a given? or can it fail in some obscure fashion, now or in the future?

[Philippe Normand]

(In reply to Zan Dobersek from comment #3)
> > Source/WebCore/platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:226
> > +            m_isMapped = gst_video_frame_map(&m_videoFrame, &videoInfo, m_buffer.get(), GST_MAP_READ);
> > +            RELEASE_ASSERT(m_isMapped);
> 
> Is this really a given? or can it fail in some obscure fashion, now or in
> the future?

It can fail indeed, especially if no video meta is attached to the buffer. I'll relax this ASSERT.

[Philippe Normand]

Committed r257202: <https://trac.webkit.org/changeset/257202>

[Radar WebKit Bug Importer]

<rdar://problem/59718023>