Re: [glob2-devel] why the big hurry to make release candidate official? (was: Well, we have a plan then.)

[Martin Voelkle]

On 4/18/07, Joe Wells <address@hidden> wrote:
> "Kieran P" <address@hidden> writes:
>
> >     I still think that delaying with a release candidate defeats the
> >     purpose of having a release candidate. The candidate is nothing
> >     special, just branch and notify us,
> >
> > I agree. Thats why I made the RC 12 hours ago. I'm just about to try the dry
> > run (making the source tarball). If no problems are rpoted, then that tarbal
> > will become official and alpha23 will be out in 24-48 hours.
>
> Why the hurry to make the release candidate official?
>
> Why not some testing first?  A week of at least half a dozen people
> actually using the release candidate would be a good idea, especially
> because the hope is that 0.8.23 will be usable for many months.  This
> testing should be done from the .tar.gz, not from CVS, to help make
> sure that what will go to the world will be okay.
>
> Anyway, I have some crashing bugs that need to be filed in the bug
> tracker and checked out.  Please give me some time to do so.

It's a release candidate for an alpha version. It is perfectly fine if
there are still issues, as long as they are fixed for the final
(alpha) release. If we go ahead with testing alpha RC, we will be soon
making pre-RC alpha releases and testing them...

We are already doing a quite weird release life cycle: RC for alpha
releases. Let's not make it worse.
http://en.wikipedia.org/wiki/Development_stage

Martin