Skip to content

Defer cleanup handling, dump core on panic, simplify proxy.go

Benjamin Johannes Lohmar requested to merge benny/defer-cleanup2 into main

#38 Now without changes to logging or free-pools:

  • setup in main to dump core on crashes (see https://go.dev/wiki/CoreDumpDebugging)
  • defer cleanup code preferably right after claiming resources
  • have a single place where server / client connections closed (in handleClient), having failed handshakes just return, after which the deferred Close function is called anyway

The rudimentary replay tests work but that isn't cause for much confidence. What's a good way to check for uncleaned resources? I'd like to include closing the process regularly and irregularly in the replay tests. How would I best avoid the time overhead of having to start multiple ism instances?

Merge request reports