httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@covalent.net>
Subject Re: Release Strategy
Date Mon, 05 Feb 2001 18:48:38 GMT
From: <rbb@covalent.net>
Sent: Monday, February 05, 2001 10:15 AM


> On Mon, 5 Feb 2001, Padwa, Daniel wrote:
> 
> > Why would you build a separate repository instead of using CVS tags to
> > handle this?
> 
> I like the separate repository, because it separates the two trees a bit
> more.  I guess my question for you is how do you move a single change to
> STABLE if there is another change in the same file that you don't want to
> move to STABLE?  This would be possible if we branched the tree, but there
> are a lot of people on this list who really dislike the idea of branching.

So they don't like branching.  Perhaps its time we took the time to learn/use
it and share the pitfalls to avoid them.

Two trees will be a massive amount of effort.  WE ALREADY HAVE THIS PROBLEM!
Just look at the dissynch state between 1.3 and 2.0 bugfixes.  Granted, this
isn't quite apples and apples, but it points out that -someone- will need to
spend the time to identify which bits will move between trees, or each author
will be reponsible for moving between trees.

This is simply not viable.  The entire reason Roy suggests this approach is
to -reduce- the dependence on the RM.  Two trees doesn't accomplish this.

Granted, to have a stable release, a -very small- number of bug fixes may
need to be applied to the candidate before it can go to beta/stable status.
I'd suggest threshholds of 14 days and 28 lines of code.  If we need more
fixes, then we need a new candidate.

The point is very minor changes will be needed to avoid code freezes.  But
we can't permit any aspect to become dependent on a specific individual, or
a massive time sponge, and repatching the stable tree would become.


 



Mime
View raw message