httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marc Slemko <>
Subject Re: The Apache Group's 1st order of business
Date Mon, 12 Jan 1998 01:35:15 GMT
On Sun, 11 Jan 1998, Alexei Kosut wrote:

> On Sun, 11 Jan 1998, Rob Hartill wrote:
> > A slow 2.0 will mean a long shelflife for 1.3, which means people
> > spending time on both trees instead of 2.0.  That's what we have with
> > 1.2 and 1.3 now. My point is that without a plan to get 2.0 out in under
> > 3 months (say), 1.3 will need supporting and that'll slow down 2.0 which
> > means 1.3 needs yet more support and so on.
> Regardless, I don't think this has any bearing on the current discussion.
> Consider: Apache 1.3 will need X amount of time to be released if we do
> not work on 2.0 consecutively. If we also work on 2.0, Apache 1.3 will
> need X' time to be released, with X' < X if we stick with only bug and
> Win32 fixes. Apache 2.0 will take Y amount of time to be completed if we
> start coding it after 1.3.0 ships. It will take Y' time if we start now.
> It is probable that Y' > Y, but I would venture to say that X + Y < Y'
> (note that Y' > X' by definition). And that Y' - X' < Y.
> Now, X, X', Y and Y' are all dependent on many factors, but the above
> analysis suggests that starting a 2.0 branch now, even with no plan
> or schedule whatsoever, will mean *less* shelf time for 1.3, not more, as
> you suggest.
> And, honestly, once we have a 1.3.0 we like, I don't think we'll spend
> much time maintaining it.

Once we have any new tree we won't spend much time maintaining 1.3.  Until
we do, people will push for a lot of things into 1.3.

I am worried about splitting without an idea of where we are going, since
it would really be best to have a clear idea of the major changes before
splitting since they are the onces that would be nice to get started in

When we split, it would be very useful to find someone who can be a
tree-sync-guy.  That means that they would keep a list of things committed
to 1.3, and make sure they get into 2.0 if appropriate.  In the start,
this would be easy.  As they diverge, it takes more effort but hopefully
1.3 changes slow down.

View raw message