fix: process cleanup invocations

[BitOfUniverse]

Currently any crashed actor with maxRestarts(0) leads to unrecovered panic.
e.g. in this example -- https://github.com/anthdm/hollywood/blob/master/examples/restarts/main.go

> go run cmd/test.go                                                               
foo got message type actor.Initialized
foo got message type actor.Started
foo started
foo got message type *main.message
foo got message type actor.Stopped
foo got message type actor.Stopped
panic: I failed processing this message [recovered]
        panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x2 addr=0x0 pc=0x1044276ec]

goroutine 36 [running]:
github.com/anthdm/hollywood/actor.(*process).cleanup(0x1400013a2c0, 0x140001440f0?)

the culprit for that is nil argument passed to func param in cleanup function.

After the fix

no crashes, expected log message.

> go run cmd/test.go
foo got message type actor.Initialized
foo got message type actor.Started
foo started
foo got message type *main.message
foo got message type actor.Stopped
foo got message type actor.Stopped
2025/04/10 22:04:12 ERROR Actor crashed too many times pid=foo/6997493373948982811

[tprifti]

looks like same issue that #197 addressing. Also, I think an actor with max restart set to 0 should not execute the recover() block

[BitOfUniverse]

@tprifti if #197 is merged we can close this one. not sure about restart policy for 0 restarts setting. it was used more for issue demonstration. I'd probably keep the same behaviour for 0 restarts.