I've been debugging a some problems i have with the windows filters. The problem is that some reads requests seems to get stuck.
Here is an example: ------------------ Lets say i start playback of livetv. Backend starts recording. Now i try to connect a spliter filter (windows filer) that tries to read at a non existan file offset (filters sometimes need to read different parts of the file while connecting). I know this filter will not be able to connect, but it shoud atleast not cause lockups. First a SEEK command is sent to the backend, it returns with the same file position i sent. A few read requests is sent that eventualy times out and backend says: "Couldn't read from the capture card in 15 seconds. Stopping" The following reads fail and properly return zero from REQUEST_BLOCK command. If, at this point a new SEEK command is sent to another position, all the following REQUEST_BLOCK commands blocks indefinitely waiting for a reply from the backend. Any ideas on how to solve this? I coud send a GET_FILE_POSITION for every read, but that woud spam the backend with lots of extra commands. Or if the SEEK command returned something different from what i sent when seeking to a position that don't exist. Or maybe the proper fix woud be to figure out why REQUEST_BLOCK gets stuck after a read have timed out. Maybe i'm using the protocol in a way that i'm not supposed to? _______________________________________________ mythtv-dev mailing list [email protected] http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
