Future changes in the new package and new branch processes

Pierre-Yves Chibon pingou at pingoured.fr
Mon Sep 8 17:57:55 UTC 2014


On Mon, Sep 08, 2014 at 06:46:10PM +0200, Ralf Corsepius wrote:
> On 09/08/2014 06:12 PM, Till Maas wrote:
> >On Sun, Sep 07, 2014 at 05:23:15AM +0200, Ralf Corsepius wrote:
> >
> >>I think the only safe way is to create an empty branch and not to populate
> >>it, because there are many constraints to be considered before a package can
> >
> >My proposal is to point new branches to the first commit in the master
> >branch in each repo.
> 
> I fail to understand the rationale for this. It doesn't make any sense to
> me. The pkgdb simply doesn't have any knowledge about what is right or
> wrong.
> 
> Imagine you were going to branch a package for EPEL5. If you are going to
> populate it from master, the likelihood, this package will be incompatible
> to EPEL5 is almost 100%.

I think you mis-read or mis-understood Till's proposal. Basically Till is
proposing that we point every new branch to the commit that created the git
repo, for example for perl-MooX-Types-MooseLike-Numeric that you maintain it
would be:
http://pkgs.fedoraproject.org/cgit/perl-MooX-Types-MooseLike-Numeric.git/commit/?id=101a14a33

But looking through other repo, Till, it seems that older repo do not have this
commit.

Pierre


More information about the devel mailing list