Re: Multilevel vendor branch import

[Mark D. Baushke]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Rodolfo Schulz de Lima <address@hidden> writes:

> On Sun, Oct 26, 2003 at 12:59:48PM -0800, Mark D. Baushke wrote:
> 
> > I have not tried to do it, but you might be able to use multiple vendor
> > branches by using the -b switch to import and then have the vendor be
> > LINUX_0_01, LINUX_0_10
> 
> I've already thought about this solution. It works, but IMHO isn't a
> good way of organizing things. The main problem with "import" stuff is
> that it solves only "one" (but important) revision management problem,
> ie, when we want to hack other's code, and whenever a new vendor version
> comes, we don't have our changes lost. That's all. Every other problem
> involves some odd way to adapt it to the way import works. People keep
> asking for importing into the main branch, etc. I see that one day, if
> CVS want to remain a import concurrent versioning system, those issues
> will have to be addressed

Hmmm... I am not sure I see that myself.

If you have concrete suggestions or a good vision you wish to share on
the future of cvs, please share them here.

> > Do remember that the -b switch needs a list of odd numbers and the final
> > digit must also be odd. I suppose you could make it deeper for sub-releases
> > of each major release
> 
> No, only numbers like 1.1.1, 1.1.3, 1.1.5 are allowed. CVS prints "Only 
> branches with 2 dots are supported", or something like that. Very
> restrictive...

Yes, I see in the source that branching only to a single level is
supported for vendor branches (ie, only digits and two "." characters
are allowing as the argument for the -b switch). I suppose this makes
sense as the deeper you go, the more complicated things become in the
creation of intermediate dead revisions.

        -- Mark

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)

iD8DBQE/nEWn3x41pRYZE/gRAvZnAJ4p6mAzaHlLw/A4EJGJNoJw4ioffQCfT1LG
Yma3RoF7Tox84su8bR7fAA0=
=bk6+
-----END PGP SIGNATURE-----