From: David Abrahams (dave_at_[hidden])
Date: 2002-10-26 07:35:13
"Jeff Garland" <jeff_at_[hidden]> writes:
> > > So I think we are in agreement that the current release procedure
> > > could be improved....
> > I hope we also agree on a specific way to improve it.
> > Do we?
> Let's see:
> 1) You have convinced me that we should be merging out to
> the release candidate instead of into the main, so that's
> one change.
> 2) We will need to make sure that regression tests are
> frequently run on the main during the release period as
> well as on the release candidate branch.
> 3) Any fixes made on the release candidate branch to
> fix regression problems not detectible on the main
> branch -- say because the main is broken for some reason --
> should be merged immediately to the main. This should
> be very rare.
> How's that?
I'm happy with all of that. I think #2 can be slightly de-emphasized
(but only slightly): the main trunk naturally gets a lot more testing
than the RC branch because developers are using it.
-- David Abrahams dave_at_[hidden] * http://www.boost-consulting.com
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk