Skip / fast forward not working - random intervals and direction

[sub]

Sorry I thought Martin was asking about something unrelated to your posts. I've just reviewed and saw it was the same issue.

Martin, what exactly is wrong with it? It shouldnt cause mvpmc to disconnect should it? In the logs it looked like it had stopped, and is sending a complete screen refresh. The sequencing of this stuff is pretty difficult to change, so we need to be pretty sure of what needs to change. I can see changes in this area easily breaking things for NMT playback etc, leading to even more work to fix it.

[mvallevand]

Can you post a log showing this?

WKleeven's log.rar above is a good example, NPVR starts sending the frame, but also continue to stream data after the stop for a couple of blocks

Code:

2011-10-04 08:11:28.462    [DEBUG][3]    Sending 'Stop' Request
2011-10-04 08:11:28.462    [DEBUG][3]    Recordings.RELOAD_RECORDINGS
2011-10-04 08:11:28.462    [DEBUG][3]    Recordings.RELOAD_RECORDINGS@1
2011-10-04 08:11:28.462    [DEBUG][3]    LoadList@0
2011-10-04 08:11:28.462    [DEBUG][3]    LoadList@1
2011-10-04 08:11:28.478    [DEBUG][3]    LoadList@2
2011-10-04 08:11:28.478    [DEBUG][3]    LoadList@3
2011-10-04 08:11:28.494    [DEBUG][3]    LoadList@setobjects
2011-10-04 08:11:28.494    [DEBUG][3]    LoadList@exit
2011-10-04 08:11:28.494    [DEBUG][3]    Recordings.RELOAD_RECORDINGS@2
2011-10-04 08:11:28.494    [DEBUG][3]    PlaybackPositionHelper.GetLastPlaybackDate(E:\Recordings\Tommy Teleshopping\Tommy Teleshopping_20111003_08000900-1.ts)
2011-10-04 08:11:28.494    [DEBUG][3]    PlaybackPositionHelper.GetPlaybackPosition(E:\Recordings\Tommy Teleshopping\Tommy Teleshopping_20111003_08000900-1.ts)
2011-10-04 08:11:28.494    [DEBUG][3]    Recordings.RELOAD_RECORDINGS@exit
2011-10-04 08:11:28.494    [DEBUG][3]    UiStatic.Dispose()
2011-10-04 08:11:28.494    [DEBUG][3]    OnKeyDown() complete
[b]2011-10-04 08:11:28.494    [DEBUG][3]    MVP sending frame[/b]
2011-10-04 08:11:28.556    [DEBUG][7]    Received: STREAM_GETDATA
2011-10-04 08:11:28.556    [DEBUG][7]    RollingFile.Position returning 3400000
2011-10-04 08:11:28.556    [DEBUG][7]    Request 200000 bytes (file size:111465200   file location: 3400000)
2011-10-04 08:11:28.556    [DEBUG][7]    about to read 200000 from location 3400000  (current length = 111465200)
2011-10-04 08:11:28.556    [DEBUG][7]    RollingFile.Position returning 3600000
2011-10-04 08:11:28.556    [DEBUG][7]    Request completed. 200000 bytes sent
2011-10-04 08:11:28.665    [DEBUG][7]    Received: STREAM_GETDATA
2011-10-04 08:11:28.665    [DEBUG][7]    RollingFile.Position returning 3600000
2011-10-04 08:11:28.665    [DEBUG][7]    Request 200000 bytes (file size:111465200   file location: 3600000)
2011-10-04 08:11:28.665    [DEBUG][7]    about to read 200000 from location 3600000  (current length = 111465200)
2011-10-04 08:11:28.665    [DEBUG][7]    RollingFile.Position returning 3800000
2011-10-04 08:11:28.665    [DEBUG][7]    Request completed. 200000 bytes sent
2011-10-04 08:11:28.665    [DEBUG][7]    Received: STREAM_STOP
2011-10-04 08:11:28.743    [DEBUG][7]    RollingFile.Close()
2011-10-04 08:11:28.743    [DEBUG][7]    Posting PLAYBACK_ENDED
2011-10-04 08:11:28.977    [DEBUG][3]    MVP frame sent
2011-10-04 08:11:28.977    [DEBUG][3]    MVP request processed
2011-10-04 08:11:28.977    [DEBUG][3]    MVP waiting for data
2011-10-04 08:11:28.977    [DEBUG][3]    Handling PLAYBACK_ENDED
2011-10-04 08:11:28.977    [DEBUG][3]    PLAYBACK_ENDED@2.0
2011-10-04 08:11:29.118    [DEBUG][3]    trying to force complete refresh2

Martin

[sub]

WKleeven's log.rar above is a good example, NPVR starts sending the frame, but also continue to stream data after the stop for a couple of blocks

Why should that cause a problem though?

[sub]

You could see if this helps. Its adds a half second sleep in there.

[mvallevand]

In GBPVR the GUI and streaming socket ended together, in NPVR they don't. I believe that screen send is blocking thread [3] so the PTS updates don't get sent in a timely manner. I do have a one screen protection against this but if NPVR sends a second screen during period I am seeing a crash.

Martin

[sub]

Give the patch a go and let me know if it helps.

[WKleeven]

I'm sorry, the mvp still crashed. This time I was playing a video, then pushed back/exit.
This test (see logs) it took 3 attempts before it crashed. Sometimes it crashed the first time though.
I was thinking: I already reported that it doesn't crash when pushing "stop", only when pushing "back/exit", so maybe the problem is how NPVR responds to the "back/exit"key, not something else?

[sub]

Martin, what do those most recent logs look like to you? In those it looks like delay I put in did do UI update after the media stream had stopped.

[mvallevand]

I won't have time to try the patch until much later but I see it with both Stop and Back/Exit. If it was just Back/Exit I would have told you to learn to use Stop, I've never liked Back/Exit as stop anyway.

Looking at the logs the delay works, I still think [3] is being blocked, I don't see the PTS values being read after the Escape is hit.

Martin

[sub]

Looking at the logs the delay works, I still think [3] is being blocked, I don't see the PTS values being read after the Escape is hit.

The delay I put in is a sleep will definitely block other things on the same thread, but there is no easy way around that. Surely that shouldnt matter from the mvpmc side though? ie, it can still send any pts values, even if NextPVR hasnt read it from the socket. Can you tell from the mvpmc side what it is unhappy with?