EventTimer.Stop panics if called more than once

[agukrapo]

We use quickfix extensively in our systems and we are getting this panic several times per week across our services

panic: close of closed channel
goroutine 39832 [running]:
github.com/quickfixgo/quickfix/internal.(*EventTimer).Stop(...)
	/go/pkg/mod/github.com/quickfixgo/[email protected]/internal/event_timer.go:48
github.com/quickfixgo/quickfix.(*session).run.func3()
	/go/pkg/mod/github.com/quickfixgo/[email protected]/session.go:795 +0x4c
github.com/quickfixgo/quickfix.(*session).run(0xc00060cc00)
	/go/pkg/mod/github.com/quickfixgo/[email protected]/session.go:823 +0x446
github.com/quickfixgo/quickfix.(*Initiator).handleConnection.func1()
	/go/pkg/mod/github.com/quickfixgo/[email protected]/initiator.go:142 +0x25
created by github.com/quickfixgo/quickfix.(*Initiator).handleConnection in goroutine 39831
	/go/pkg/mod/github.com/quickfixgo/[email protected]/initiator.go:141 +0xd3

A solution could be to make EventTimer.Stop idempotent:

func (t *EventTimer) Stop() {
	if t == nil {
		return
	}

	t.once.Do(func() {
		close(t.done)
	})

	t.wg.Wait()
}

This proposal is included in PR #580

[melkanas]

I had the same panic, managed to avoid this in my application by having a state of initiator and testing the state before initiator.Start() or initiator.Stop() to prevent calling them multiple times