Closed sameagen-MW closed 1 week ago
However, this should slightly improve performance as now the
M
Page fault may be remote occurrence in this case as I would see this anyways cleaning the the acquired space. Ofr now I feel garbage collection through finally seems reasonable per build.
Not a fix for issue 332. I still need to think through what the consequences for having the temporary scripts self destruct would be. E.g. would we be resilient to page faults at high load, does this severely hinder debuggability, etc...
However, this should slightly improve performance as now the
MatlabCommandRunner
objects can be cleaned up before JVM shutdown. Especially for very long running Jenkins servers / agents this should be an improvement.