Gapless events stop replaying

[px03afk]

I've only just moved on to 1.36.7 but I also was trying out 1.37.1 and they both have the same problem. I replay gapless events and many events replay OK but then one of them stalls. Click on the Next button and it goes back to a previous event and then stalls in the same place again. On further digging I find that there is an event which seems to be causing a segmentation fault at line 49 of zm_signal.cpp. There is then a set of backtrace records which has a 'please execute' message on completion but running the command produces lines of ??:? and ??:0.

I can look at the individual frames but whilst looking at them I noticed that the frame count shown on the event is one less than the frame count when you look at the frames themselves so I suspect there is some code which is using running over the frame count from the event list.

[px03afk]

I have been looking deeper into this and I have found three problems with the event data being written into the database:

1. The Frames count may be incorrect when compared with the number of entries in the events filestore - so far I have only found that the Frames count is one less than the filestore
2. The Name for some events remains as New Event and doesn't get the name of the monitor or the event id as other events
3. The EndDateTime is sometimes left NULL

Any one of these causes the playback to crash.

I have written a simple shell script to correct these errors and now I can replay without interruption - unless more issues show up.

[iconnor]

#2 and #3 indicate that zmc crashed during saving the event. Look in your logs.

#1 will require further investigation.

[px03afk]

Thank you. I will check the logs when this happens again.

[px03afk]

I have seen more instances of #2 and #3 and the logs do not show any error, just opening and closing events

[px03afk]

I may have found one of the reasons this was happening. I have a script which was inadvertently issuing a restart periodically by running zmpkg.pl but that seems to have caused some errors. Since correcting the script the problem appears to have gone or at least diminished.