[development] Suggestion for people releasing new modules and themes into CVS

Khalid B kb at 2bits.com
Mon Sep 18 14:38:45 UTC 2006


On 9/18/06, Michelle Cox <mcox at charter.net> wrote:
>
> On 9/16/2006 2:35:50 PM, Chris Johnson (chris at tinpixel.com) wrote:
> > Khalid B wrote:
> >
> > > Another problem is that we often
> > don't know if the next release will be
> > > an x+1.0 or x.y+1, so we can't
> > say "ported for 5.0" before Dries makes
> > > that decision. This is another case where code names for future
> > releases
> > > will help.
> >
> > That is probably the single best argument I've seen for code naming
> > releases
> > before they are released -- so we can refer to them easily and accurately
> > in
> > documentation of all kinds (project pages, CVS comments, postings, etc.).
>
> It's been named 5.0 for a while and is far from released. I'm not a dev, so
> maybe I'm off base here, but I don't see a big need to have a definite name
> before the code freeze...

Before it was named, it is known as HEAD (technically, trunk, but that
is the name that is used). It was commonly referred to as 4.8/5.0, and
the naming happens when all the features are in, and a code freeze
happens. At that time Dries decides if it is x.y+1 or x+1.0.

However, during the development cycle (from the time the previous release
is branched, until that final naming, we use HEAD, and that is a moving
target by definition. If someone writes a piece of documentation or refers
to the next release as "HEAD" there is no way to know whick release
this ended up in.

So, what I am advocating is neutral, non-geek code names (flowers, plants,
fish, ...etc.) for the upcoming branch. If we say the next release is red rose,
or anglefish, then we have a point of reference until it is named 5.1 or
whatever.

This should really be a separate thread, but ...


More information about the development mailing list