systemd can't stop service about mcscripts

[TapeWerm]

fb586c9 seems to have fixed this. The issue only happens if you stop the service after it has been running for at least 75 minutes or so. It has only gone smoothly once, but I anticipate the issue is gone for good. I'll keep this issue open until I test it more. I recommend updating MCscripts as this issue is transient but can break updates and stopping your server in general.

[TapeWerm]

Oh my, it happened again. This definitely seems to be a systemd isssue, and not my fault. Of course, because it's systemd, it does not matter who is at fault, I am the one who needs to change. I'm gonna try changing the socket unit to persist even after the service stops, and see how systemd spazzes out when I do that. systemd--
Edit: When I made the socket persist it stopped, but systemd complained to me. Gonna explore more options.

[TapeWerm]

For now fixes will be pushed to branch systemd_pain. Reason last 2 fixes were pushed is because they featured objective improvements (less required units, fixes to ordering). This is more extraneous unless it works around systemd's ordering bugs.
Edit: 1st test of systemd_pain was successful, I think systemd ignores socket ordering if it isn't listed in Sockets= or maybe I just got lucky. It doesn't derp every time after 75 minutes, just most times.
Edit 2: 2nd test failed, so it still doesn't work. I have given up on the socket being a transient unit, and have embraced it as a persistent one. It will be enabled separately, and tested as such. If it works it will be christened as the trendy new way of playing with systemd 237-3ubuntu10.41. I already have drafted an updated README and DisableServices.sh service migration code.

[TapeWerm]

:galaxybrain: Can't have socket stopping before service if the socket doesn't stop when the service does. I tested the new service and socket and they work, as does the update script, and the README has been updated accordingly for new users.

[TapeWerm]

Well I'm reopening this again, cause guess what, stop still doesn't work. It doesn't kill the socket this time so it can stop cleanly but it is an issue with the IRC bot failing to die properly. Ugh.
Edit: Happy ending, this whole saga only happens if you use the IRC bot. The hourly IRC bot timer I use to revive the bot after a network outage seems to be what started the bot right after it stopped. And when the bot would come back the stop job would cancel, but in doing so systemd would spaz out and stop the socket even though it didn't stop the service. The socket issue has already been worked around by me, now the bot issue is fixed too. I just removed rate limiting from the startup, and it will just keep failing till you stop it. It doesn't eat much CPU so I this seems like a simpler solution.