When should I send a starting seek?

[amindfv at mailbox.org]

Below is the error I see when I set ```GST_DEBUG=*:WARN```. The same code works without a hitch when I use "uridecodebin" instead of "nleurisource".

Tom


0:00:00.153924403 2533047 0x5613c8dd7ed0 WARN                 basesrc gstbasesrc.c:3688:gst_base_src_start_complete:<source> pad not activated yet
0:00:00.154504971 2533047 0x5613c8dd7ed0 WARN                 basesrc gstbasesrc.c:3688:gst_base_src_start_complete:<source> pad not activated yet
0:00:00.155152264 2533047 0x5613c8dd7ed0 WARN                 basesrc gstbasesrc.c:3688:gst_base_src_start_complete:<source> pad not activated yet
0:00:00.155418976 2533047 0x5613c8dd7ed0 WARN                 basesrc gstbasesrc.c:3688:gst_base_src_start_complete:<source> pad not activated yet
0:00:00.607857950 2533047 0x7f14a0015980 WARN                GST_PADS gstpad.c:4303:gst_pad_peer_query:<decodebin1:src_1> could not send sticky events
0:00:00.608076152 2533047 0x7f14a0015980 WARN                GST_PADS gstpad.c:4303:gst_pad_peer_query:<decodebin1:src_1> could not send sticky events
0:00:00.612837416 2533047 0x7f14a0016360 WARN                GST_PADS gstpad.c:4303:gst_pad_peer_query:<decodebin0:src_1> could not send sticky events
0:00:00.612963924 2533047 0x7f14a0016360 WARN                GST_PADS gstpad.c:4303:gst_pad_peer_query:<decodebin0:src_1> could not send sticky events
0:00:00.613500773 2533047 0x7f14a0015980 ERROR                  libav :0:: env_facs_q 255 is invalid
0:00:00.618308939 2533047 0x7f14a0016360 ERROR                  libav :0:: env_facs_q 255 is invalid
0:00:00.939458436 2533047 0x7f14a80804c0 WARN           matroskademux matroska-demux.c:5869:gst_matroska_demux_loop:<matroskademux1> error: Internal data stream error.
0:00:00.939495600 2533047 0x7f14a80804c0 WARN           matroskademux matroska-demux.c:5869:gst_matroska_demux_loop:<matroskademux1> error: streaming stopped, reason not-linked (-1)
Error received from element matroskademux1: Internal data stream error.
Debugging information: ../gst/matroska/matroska-demux.c(5869): gst_matroska_demux_loop (): /GstPipeline:pipeline0/NleURISource:nleurisource1/GstBin:internal-bin/GstURIDecodeBin:internal-uridecodebin/GstDecodeBin:decodebin0/GstMatroskaDemux:matroskademux1:
streaming stopped, reason not-linked (-1)

(repro:2533047): GStreamer-CRITICAL **: 01:06:51.049: gst_mini_object_unref: assertion 'GST_MINI_OBJECT_REFCOUNT_VALUE (mini_object) > 0' failed
0:00:00.952529891 2533047 0x5613c8dc0de0 WARN           matroskademux matroska-demux.c:5869:gst_matroska_demux_loop:<matroskademux0> error: Internal data stream error.
0:00:00.952568013 2533047 0x5613c8dc0de0 WARN           matroskademux matroska-demux.c:5869:gst_matroska_demux_loop:<matroskademux0> error: streaming stopped, reason not-linked (-1)



On Thu, Apr 21, 2022 at 10:04:49AM +0200, Marianna Smidth Buschle via gstreamer-devel wrote:
> Try increasing the debug level of 'qtdemux', so you can get more information
> on the specific error.
> 
> Moreover since it works with 'gst-launch' but fails with a custom app the
> problem is probably related to dynamic linking.
> One simple option to work around it could be to use 'gst_parse_launch()'
> inside your app and let it handle the linking.
> 
> Otherwise you can go back to 'uridecodebin' and do the SEEK by pre-rolling
> first (I guess that just means waiting for the PAUSED state?).
> 
> The way I have always done my SEEKing is by waiting for the pipeline to be
> into PLAYING and then sending a flushing SEEK to the pipeline object.
> And I mostly use 'gst_parse_launch()' so I don't have to handle the linking
> timings myself.
> 
> Moreover, some elements allow you to set the segment format which might help
> you to get rid of the 'assertion 'segment->format == format'' error.
> Or you can try changing the segment format on the SEEK event.
> 
> Best Regards
> 
> Marianna
> 
> On 20.04.2022 21.28, gstreamer-devel-request at lists.freedesktop.org wrote:
> > I've spent quite a lot of time attempting to debug this with no breakthrough. It may be that I'm missing something very obvious in the documentation, but it's not from lack of trying or RTFM-ing!
> > 
> > Below (and attached) is a smallish reproducer case, the smallest I can make it. It creates a pipeline successfully with "uridecodebin" but fails to with "nleurisource". Any help would be very appreciated!
> > 
> > Any help would be very appreciated!
> > 
> > Thanks,
> > Tom
> 
> -- 
> Best regards / Med venlig hilsen
> “Marianna Smidth Buschle”
>