Closed cclaessens closed 7 years ago
I'm not able to replicate this problem. Are you sure you're using the latest commit? I had something like this happening yesterday, but I fixed the bug in that case. I'm writing this as of commit 3fce23f.
Nevermind my last comment. I hadn't been performing a run. I do replicate the problem.
This was fixed in commit 8da44ce643886435d550c2a13db9468bf80b96c0
Currently only packets with matching f and t ID are processed and missing packets result are skipped and a new event is started.
Running psyllid with the fpa receiver resulted in a lot of warnings from the roach_receiver about Improper packet sizes and a small egg file. After adding a warning which is printed when the packet size is as expected it seems that in a 100ms run 1 out of 2 packets have the wrong size. To produce the following output I also added a debug line ("acquisition break") which is written when the streaming writer doesn't get the packet it expects. Acquisitions are always only 1 record long.
fpa_run_output.txt
In a run with the socket receiver the packet sizes seem to always be as expected and acquisitions have more records. Breaks occur regularly too.
socket_run_output.txt