From: hipster <hipster@...4all.nl>
Date: 2000-10-27T22:19:23+09:00
Subject: [ruby-talk:5911] Re: [RFC] Towards a new synchronisation primitive

On Fri, 27 Oct 2000  21:26:37 +0900, Dave Thomas wrote:
> > I'd say the point of definition of a method is at the end of the def
> > expression, before any method body expressions. Therefore only the
> > method body is subject to sync-behaviour.
> 
> I can see cases where that wouldn't work, such as:
> 
>    class RingBuffer
> 
>       def sync initialize
>          @inPtr = @outPtr = 0
>       end
> 
>       def sync printBuff(from = @inPtr, to = @outPtr)
>         # ...
>       end
>    end
> 
> If synchronization only applies to the method body, its possible that
> from gets assigned @inPtr in one state, then something else runs, and
> to gets @outPtr in another state. Not critical in this case, but I
> guess it could be.

You're absolutely right. My point-of-definition argument has nothing
to do with this. Parameter initializers should be considered a kind of
'method constructors', and should be part of the same method synch
scope.

	thanks,
	Michel