After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 379826 - [dvddemux] Do NOT expose pads that have multiple structures
[dvddemux] Do NOT expose pads that have multiple structures
Status: RESOLVED INCOMPLETE
Product: GStreamer
Classification: Platform
Component: gst-plugins-ugly
git master
Other Linux
: Normal normal
: NONE
Assigned To: GStreamer Maintainers
GStreamer Maintainers
Depends on:
Blocks:
 
 
Reported: 2006-11-27 16:17 UTC by Edward Hervey
Modified: 2007-05-07 17:34 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Edward Hervey 2006-11-27 16:17:37 UTC
When dvddemux adds pads, it exposes pads that have multiple structures. This causes issues when deciding what data is actually going to flow.
Decoders/demuxers should, when adding a pad from a pad template with multiple structures, know what type of data is going to flow through that pad. It should therefore add pads with only one structure.

Ex : This is a pad template:
  SRC template: 'current_audio'
    Availability: Always
    Capabilities:
      audio/mpeg
            mpegversion: 1
      audio/x-lpcm
                  width: { 16, 20, 24 }
                   rate: { 48000, 96000 }
               channels: [ 1, 8 ]
          dynamic_range: [ 0, 255 ]
               emphasis: { false, true }
                   mute: { false, true }
      audio/x-ac3
      audio/x-dts

When the pad is created it's one of those structures, not ALL ! :)

Please fix accordingly.
Comment 1 Tim-Philipp Müller 2006-12-11 17:48:37 UTC
You mean it adds pads without setting caps on them first, resulting in _get_caps() returning the template caps? (you can't "add a pad with multiple structures", pad caps are always fixed) Could you be a bit more precise? Does this happen with all pads or only specific ones?

I can't see the culprit in the code right now (could be one of the private pads though, not sure if we set caps on those if they're exposed).

Comment 2 Tim-Philipp Müller 2007-01-18 15:58:17 UTC
Edward: ping?