start/stop hook guarantees

[Gustavo Niemeyer]

> The confusion stems from the fact thatt I didn't think I was supposed to be
> defining the behaviour of "start" and "stop"; I thought there were accepted
> definitions I was building from, and I assumed that the codebase reflected
> those.

As you know the code base has well known constraints in that area that
you've been fixing. The behavior of start and stop was very well
defined when we take in account these constraints.

(...)
> That is to say, I can comfortably understand either definition of "stop" [0],
> and I don't believe anyone else will have a problem with either definition;
> the confusion stems purely from the fact that we have two competing
> definitions.

There hasn't been a clear definition before of how we wanted these
hooks to behave when facing adveristy, and what the code does has not
mattered so far for the reasons stated. We're now discussing what we
_want_ to happen, which is a natural step forward that we always knew
had to be taken.

> That's great, thank you very much; I look forward to reading about it :).

I'm still researching a bit and thinking about the problem.

> [0] I'm sorry I hastily mischaracterised yours in the previous email; I know
> it doesn't *necessarily* mean "going away forever".

No problem. I knew you actually understood it after our conversation.

-- 
Gustavo Niemeyer
http://niemeyer.net
http://niemeyer.net/plus
http://niemeyer.net/twitter
http://niemeyer.net/blog

-- I'm not absolutely sure of anything.