httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roy T. Fielding" <field...@kiwi.ics.uci.edu>
Subject Re: 1.3.8: One more bug...
Date Wed, 11 Aug 1999 12:54:33 GMT
All these suggestions just make it harder to fix the problems
in a release by introducing new barriers to development and laying
more tasks on the RM.  The RM is not capable of deciding which changes
should be applied to fix problems not found on the RM's machine.
There is never any need to tie everyone's hands while one person
plays catch-up.  The only time the tree needs to be frozen is during
the ten minutes it takes to bump and tag a revision.  And it just
doesn't matter how many "bug fix" release number we go through
between public releases -- what matters is that the bugs are fixed.

This is the last planned 1.3.x release.  Coming up with some new
numbering scheme or convoluted tagging process for 1.3.x is a waste
of time.  If you want to do that for 2.x, fine, but my suggestion for
it is the same one I've been giving for two years now: unbind the
panties and go back to the rational versioning system we were using
way back in the 0.6-0.8 days.  That is, simply bump and tag the minor
version every time that the tree appears to be stable (i.e., nothing
on the showstopper list, nothing easy left to patch-in, and nobody
in mid-commit).  AFTER it has been bumped, decide whether that
particular version should be publically released or not by running
it through tests, doing a sample build, and handing it to the stable
testers.  If it turns out to have a major problem, no big deal -- we
just add the problem to the showstopper list and keep going.

That also means no special alpha/beta numbers.  Describing the tarball
as beta is fine, but is no reason to include that in the revision number.

The last time we did this, we managed a new version every two weeks.

We have three separate numbers in the version already, and it is
ridiculous to sweat so much over releases in the "bug fix" category.
We should be looking for ways to relax the process, not make it more
constipated than it is now.

....Roy

Mime
View raw message