Latest post Wed, Oct 2 2019 9:08 AM by andym. 2 replies.
Page 1 of 1 (3 items)
Sort Posts: Previous Next
  • Tue, Sep 10 2019 7:22 PM

    AMT3 - ChainManager not responding

    Dear Avid AMT3 Dev Team,

    I have a problem that you probably can resolve. When ChainManager is running with a ISampleBasedPushInputAsset as source, I create a second task to call the IMPSink::write(MPBufferHolder buffer) function on each frame.


    On error exception the ChainManager does not respond to the IMPSink::write(MPBufferHolder buffer) function and blocks the second task. Is it possible to respond to the write function task before raising the exception ? 

    Last point, it should be nice if the return value of the write function could be a clear error status ?

    Thank you,
    GC.

  • Wed, Oct 2 2019 9:08 AM In reply to

    • andym
    • Not Ranked
    • Joined on Thu, Jan 13 2011
    • Posts 116
    • Points 1,390
    • Avid Developer Moderator

    Re: AMT3 - ChainManager not responding

    Hi Grégoire,

    sorry for taking a while to come up with a reply. It looks like you have run into the following issue

    [MPI-29483] Updated MediaProcessor to close input sample-based sinks when error occurs on performing job

    which is already fixed but not yet released as AMT3. We'll have that version coming soon.

    What you can do on your side as a temporary mitigation technique is to close the sinks that you get when you detect an error so you will avoid the problem until a fixed AMT3 version becomes available. Once you signal 'close' after receving the ChainManager exception, the write operation will unblock.

    Regarding your 'last point', we don't do this intentionally due to the asynchronous nature of the operation.

     

    Hope this helps

    Andy

  • Wed, Oct 2 2019 9:08 AM In reply to

    • andym
    • Not Ranked
    • Joined on Thu, Jan 13 2011
    • Posts 116
    • Points 1,390
    • Avid Developer Moderator

    Re: AMT3 - ChainManager not responding

    Hi Grégoire,

    sorry for taking a while to come up with a reply. It looks like you have run into the following issue

    [MPI-29483] Updated MediaProcessor to close input sample-based sinks when error occurs on performing job

    which is already fixed but not yet released as AMT3. We'll have that version coming soon.

    What you can do on your side as a temporary mitigation technique is to close the sinks that you get when you detect an error so you will avoid the problem until a fixed AMT3 version becomes available. Once you signal 'close' after receving the ChainManager exception, the write operation will unblock.

    Regarding your 'last point', we don't do this intentionally due to the asynchronous nature of the operation.

     

    Hope this helps

    Andy

Page 1 of 1 (3 items)

© Copyright 2011 Avid Technology, Inc.  Terms of Use |  Privacy Policy |  Site Map |  Find a Reseller