To reproduce #707 we need to run sequenceserver -d <path/to/dir/with/unformatted/dbs>, which also launches a sequenceserver instance in a subprocess. I had an attempt to reproduce the issue in a test, but the fact we need to shell to a CLI that then shells out to launch a sequenceserver instance makes managing the process state in the test rather tricky. Here's my attempt that successfully captures the issue in #707, but when the issue is resolved it does not kill the sequenceserver process...

https://gist.github.com/tadast/972c653070d4418ee0a161bc4de308a1

Thank you @tadast - can we add your rspec gist to the main rspecs?

can we add your rspec gist to the main rspecs

Sorry @yannickwurm I wasn't explicitly clear - the test in the gist does not quite work fully - it spins up a sequence server web instance as a side effect, and that web app process keeps lingering. I didn't manage to find a nice way to kill that process in the time I'd allocated to this issue. If you think it's worth it, I can have another look