Re: [AD] 4.2.1 is up, and a version number proposal

[Evert Glebbeek <eglebbk@xxxxxxxxxx>]

> My reasons are:
>
> - We usually have last minute patches applied without much testing
>   before the release.  If one of those patches is bad then a fix won't
>   appear in a 4.2.x release until a long time later.
>
> - We could try to make sure all patches receive a decent amount of
>   testing before it is included in a release, but that only means
>   releases will get delayed even longer.  I think it's pretty clear that
>   people will only test something once it's released, so it's
>   unrealistic to expect many people to test SVN snapshots.
>
> - Similarly, very few people are willing to test and retest RCs.
>   I think it doesn't fit the volunteer model.
>
> - We don't _need_ a fourth version number, but in the 4.2 series we only
>   have one number left to work with and there's rather a lot of
>   significance attached to it now.  There is a lot of pressure in making
>   a 4.2.x release "perfect" the first time around, which drags the
>   process out.

I pretty much agree with all the above (except maybe the third: at least
some of the RC releases got a decent amount of testing from people).
Having this should really help lower the threshold for making a release,
which is a good thing.

Evert