[CentOS-devel] Repository structures for SIG and variants in the future

Fri Jan 17 17:37:37 UTC 2014
Mike McLean <mikem at imponderable.org>

On Fri, Jan 17, 2014 at 12:06 PM, Jim Perrin <jperrin at centos.org> wrote:

> On 01/17/2014 10:06 AM, Ljubomir Ljubojevic wrote:
> > So how do you intend to solve this without priorities? Especially if
> > there is a need for some packages to be compiled with different flags?
> Ideally by eliminating package duplication for the SIG/repos who work
> with us, but that's an ivory tower goal that's likely unattainable.

We certainly want to reduce it and box it in. It's clear that we do want to
allow overlap, but we need to try to build some sanity into it and avoid it
when it is not truly necessary

> It will likely be a combination of working with SIGs to coordinate and
> cooperate, changing to SCL builds where possible and appropriate, and
> documentation to better educate the users.
> SCL builds provide a nice name-spaced way to have multiple versions of
> the same packages installed. For example, python27, python33, multiple
> php and httpd versions, etc. All of which are mostly impractical with
> the older rpm packaging style (see php and httpd for example). This is
> by no means a silver bullet, but it does solve a fair chunk of the issue.

Agreed. SCLs are a natural fit here

Without SCLs, the SIG will actually be overriding the base package. There
are a host of potential problems here, repo priority is the least of it.
For major significant changes this could spiral into a dependency
nightmare, or simply break things in the base distro. For minor changes
(small change, bump release), the such issues would be less likely. For
overriding with higher versions, the repo priority plugin would be
unnecessary. For release bumps, the SIG would have to track the base
closely to stay ahead (granted they should probably do that anyway).

At any rate my point is that there are many different reasons to do this
and many different ways it which it might be done. Some approaches will be
sane, and others very, very unwise. I suspect that we'll wind up with a
fairly long policy document delineating these different cases, stating what
is allowed, and listing best practices.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.centos.org/pipermail/centos-devel/attachments/20140117/c6cc346d/attachment-0005.html>