|
Boost : |
From: Beman Dawes (bdawes_at_[hidden])
Date: 2002-08-16 12:25:03
At 06:48 AM 8/16/2002, Gennaro Prota wrote:
>On Thu, 15 Aug 2002 16:22:23 -0400, Beman Dawes <bdawes_at_[hidden]>
>wrote:
>
>>The 1.29.0 "branch-for-release" date has been pushed back a week, until
>>next Thursday, August 22.
>>
>>We need to be seeing much more consistent test results before
>>branching.
>
>
>In fact, I have a (stupid) question: what is the criterion for
>deciding when to do a release?
Hardly a stupid question. It's the answer that's stupid:-)
Namely, we don't have any real criteria.
Until this year, we did a lot of releases, and if people complained about
problems with a release, we just did a ++.n release.
This year we are trying to do releases on a less frantic schedule, and
working harder to test a release before shipping it. This will be the
third release we branched first, and then hold the release until obvious
glitches are fixed.
What we've been doing essentially is to query the list, look at the tests
for regressions from prior results, and then just going ahead when it felt
OK.
--Beman
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk