When will CVS be replaced by modern version control system?
Christopher Aillon
caillon at redhat.com
Mon Nov 12 14:38:17 UTC 2007
Josh Boyer wrote:
> On Mon, 12 Nov 2007 15:25:32 +0100
> Matej Cepl <mcepl at redhat.com> wrote:
>
>> On Mon, 12 Nov 2007 15:12:43 +0100, Christopher Aillon scripst:
>> > But also keep in mind that much of the current package workflow has
>> > been defined by the limitations of CVS. Our current faux-branching
>> > scheme was in part done because of CVS, for example.
>> > [...]
>> > Returning to your original question, which I'll paraphrase as "what do
>> > we gain by moving away from CVS?" Not much. A small number of users
>> > will take advantage of the features that the new VCS gives them.
>>
>> I think the thing we need to loose is the box where we keep our work-
>> flow. So for example, could somebody explain why in the time or DVCSs we
>> still have tarball-based packaging?
>
> So you can verify that you're using the upstream source as it was
> released, and not some custom fork of the code base.
I think Matej was implying we should point to the upstream repo and
tagname instead of a tarball. Then koji would checkout code from
upstream without relying on us downloading, then subsequently uploading
a tarball via make upload.
More information about the devel
mailing list