On Tue, Apr 17, 2001 at 02:17:00PM -0400, Chuck Murcko wrote:
> David's got a valid point though: determinism and risk management for
> 2.0 release. And also about the schizoid nature of httpd going for
> release, stuff going in and stuff testing all at once.
Right. We can talk about "safe" patches all we want, but every one of them
could destabilize. Bugs exist because we miss things. There is nothing that
says we won't miss things in the patches.
> Maybe it's time to start actually putting things on the list for 2.1,
> like rethinking the scoreboard implementation or fully instrumenting the
> server. I mean, not everything's got time to make it into 2.0, so IMO we
> should start consciously slating big stuff for 2.1-5 timeframe.
Ah! YES. This is a great way to help release some pressure for feature
development.
Hmm. But we can only go hog-wild on a 2.1 with new features if it is
separate from the 2.0 development (where we want to bugfix). That means two
trees (whether that is two CVS repositories, or one rep with two branches).
Bugger on maintenance since we will probably tweak 1.3 for a while, too.
I'm not sure what the right approach is, but it would be good to start the
list in the STATUS file, methinks.
>...
> And I know that multiprotocol support has been on the list for 2.0 for a
> little while now. We still have to talk about removing the proxy: hack.
A "little while" ?? Heh. mod_echo was first checked in on August 23, 1999.
Cheers,
-g
--
Greg Stein, http://www.lyra.org/
