From: Phil Richards (news_at_[hidden])
Date: 2007-08-03 15:00:34
On 2007-08-03, Stefan Seefeld <seefeld_at_[hidden]> wrote:
> Douglas Gregor wrote:
> > Yes, and we need to fix the trunk, not devise elaborate processes to
> > isolate the "civilized world" of releases from the "wild west" of the
> > development trunk.
> I'd suggest this:
> * establish a 'stable' branch (by copying from RC_1_34_0). [...]
> * establish goals for 1.35, [...]
> * Discouraging developers to check *anything* into trunk any more, but instead,
> encourage the use of branches as a means to backport changes that now are in
> trunk, and should go into stable, whether in time for 1.35 or not.
For what it's worth, I completely agree with the first and third points here.
This development model *does* work. It is the one that I use (and have
used) every day for the last 5 years.[*]
But for it to work it also requires all the tools changes/fixes that
Doug has been saying to get done - the "stable" (or "release branch")
must be being tested (and obviously in a more efficient way than it is
[*] And no, I haven't managed a large open source project release
before; I just manage the release of the 20 or 30 interrelated
libraries/applications used by multiple projects in my day job.
The "process" that we have normally means that I can assemble a release
from the non-broken mainlines pretty quickly. Or highlight the
integration problems, as the case may be.
Of course, I have written quite a lot of tools to help me...
-- change name before "@" to "phil" for email
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk