Catch :noproc errors in Xandra.Connection

[harunzengin]

Closes #383

Since we're catching the :noproc errors on Xandra.Connection level, we don't need to catch them in Xandra.Cluster.

[whatyouhide]

This is wrapping the whole case into a try block. I think that's not right, as it would hide exits that could happen anywhere.

Instead, we should wrap just the :gen_statem.call/2 into try. That would also mean significantly more readable code, as this starts to be a really convoluted nested piece of code.

Basically, do this:

case gen_statem_call_trapping_noproc(conn_pid, {:checkout_state_for_next_request, req_alias}) do
  # ...

  # New clause:
  {:error, :noproc} ->
     # ...
end

# Then:
defp gen_statem_call_trapping_noproc(pid, call) do
  :gen_statem.call(pid, call)
catch
  :exit, {:noproc, _} ->
    {:error, :noproc}
end

Does that make sense? It will also, incidentally, make the diff a lot slimmer 🙃

[harunzengin]

Makes a lot of sense

[whatyouhide]

Fantastic work 💟