bryanhunter / cqrs-with-erlang Goto Github PK

Hi,
I have a system with a very similar pattern where a process which's Pid can be looked up using a ID stops after a specific timeout if it doesn't receive any messages. I thought it would be a problem if someone looks up the Pid and then sends a message to it but the process with the Pid decides to terminate in the meantime so I tried to prevent it from happening with some manager process which turned out to be very complicated. What do you think about this? Is it valid to just assume it doesn't happen and let a caller crash from time to time?

Just curious, in your bank_bus.erl, what are you thoughts on making it use gen_event:call instead of notify for the send_command? I am thinking that way the function can return a value (such as command_not_allowed etc.).

https://github.com/bryanhunter/cqrs-with-erlang/blob/ndc-oslo/bank/src/bank_bus.erl