[Webkit-unassigned] [Bug 89122] [GStreamer] Audio device not closed after playing sound

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=89122





--- Comment #22 from Martin Robinson <mrobinson at webkit.org>  2012-08-23 08:35:20 PST ---
(From update of attachment 160122)
View in context: https://bugs.webkit.org/attachment.cgi?id=160122&action=review

This patch looks good to me, though I think that someone like Eric should look at the platform-independent changes.

> Source/WebCore/platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:343
> +        // Position queries don't work on a null pipeline and when we're at the end of the stream the pipeline is null.
> +        if (m_seeking)
> +            return m_seekTime;
> +        if (m_mediaDuration)
> +            return m_mediaDuration;
> +    }

Hrm. I guess if the sync can handle position queries gracefully, then the comment is inaccurate?

> Source/WebCore/platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:497
> +    bool prerollRequired = m_isEndReached && !m_player->mediaPlayerClient()->mediaPlayerLoop();
> +    if (prerollRequired) {
> +        LOG_MEDIA_MESSAGE("Seek on EOS, pre-rolling pipeline again.");
> +        gst_element_set_state(m_playBin, GST_STATE_PAUSED);
> +    }
> +
> +    // Avoid useless seeking.
> +    if (time == playbackPosition()) {
> +        if (prerollRequired)
> +            gst_element_set_state(m_playBin, GST_STATE_NULL);
> +        return;
> +    }
> +

Is it important to go from PAUSED back to NULL for no-op seeks here? If not, why can't you just do:

// Avoid useless seeking.
if (time == playbackPosition())
    return;


if (m_isEndReached && !m_player->mediaPlayerClient()->mediaPlayerLoop()) {
    LOG_MEDIA_MESSAGE("Seek on EOS, pre-rolling pipeline again.");
    gst_element_set_state(m_playBin, GST_STATE_PAUSED);
}

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.