Re: WeakMessageSends/Events for 3.4

Hi Göran,

I don't have a problem with your recommendation.  I am very happy to see the
NuBlu event interface and implementation in 3.4.  If this WeakMessageSend
work is left out, then receivers and arguments of registered actions are
held strongly, so collaborators need to explicitly deregister themselves or
garbage will collect in the corners.  It is probably a good idea to
carefully manage the full lifecycle of objects anyway.   GC is no excuse for
sloppiness!

I can take or leave my tracer changeset as well, although I wish there was a
good tracing facility available in the base.

Finally, I would really like to see ConnectionHandler, SocketStream and
BufferStream in the base, mainly ConnectionHandler.  These have been well
vetted.  SocketStream may be eclipsed by the Network rewrite or flow, or
both...but ConnectionHandler remains useful   (Trying to build support for
streaming mp3s certainly shows the usefulness of the flow model).

So when is 3.4 final scheduled for release?   How about today!?   Just do it
and if there are bugs, then patch it in the update stream.  You are right
that we can always play in 3.5.

cheers,
rob

----- Original Message -----
From: <[email protected]>
To: <[email protected]>
Sent: Monday, December 09, 2002 5:08 AM
Subject: Re: [Squeakfoundation]WeakMessageSends/Events for 3.4


> Hi all!
>
> Without having looked at this, which mean that I could be making a fool
> of myself, I just want to say that we are at *beta* now. Not alpha. In
> my book that means bugfixes, not much else. Is this a bugfix?
>
> Of course, there are always exceptions to the rule, but please - let us
> not swamp 3.4 just before trying to release it. Ok? I mean - when we
> release 3.4 we suddenly have 3.5alpha to happily screw up instead. :-)
>
> regards, Göran
>
> _______________________________________________
> Squeakfoundation mailing list
> [email protected]
> http://lists.squeakfoundation.org/listinfo/squeakfoundation