Re: [Gnu-arch-users] Re: Making microbranches popular

[David Allouche]

On Tue, Jan 27, 2004 at 08:36:49PM +1100, Robert Collins wrote:
> On Tue, 2004-01-27 at 16:20, Jeffrey Yasskin wrote:
> 
> > I seem to remember Tom saying that he was willing to make the version
> > numbers more flexible if someone could provide a sort function for them.
> > Am I delusional, and if not, is that offer still open? Would that fix
> > this problem?
> 
> No. This problem is trying to shoehorn a square block into a round hole.
> 
> Tla's versioning infrastructure is not targeted at complex version
> strings - see how tom id's his preX and actual releases for an example.
> I think I sent an email saying 'use configs'. And I KNOW that Tom sent
> one too.
> 
> configs are the Right Way to do this. (For more than tla - for CVS, svn
> and more).


For all that it's worth, some people also seem to express a desire to
have some "alias" scheme as a user convenience. For example, have DEVO
be the current main development branch (aka. HEAD in CVS), which may
resolve to a different version over time. Or have "frob-4.2-epsilon-7"
resolve to "frob--4.2.5.7".

Of course, such a convenience feature needs _no_ support inside tla. You
can just implement a prototype using "aba" for the front-end and the
"archive-metada" branch I have been discussing (w/o raising much
interest it seems) to store the bindings.

Bah... Anyway the metadata issue is clear to me now. You will see me
waving my "use archive-metada category" from time to time.

Other than that, I agree that making the Arch namespace any more messy
would be a bad thing. I'm pretty happy with both the version part and
the anonymous branch feature, but I see no need to add up on it. Though
I may change my mind if someone gives me a better argument than "my
stupid lazy users who actually do not have a damn clue about arch think
it would be better".

-- 
                                                            -- ddaa