RepoFunnel and the Software Component Pipeline

[Nick Coghlan]

Hi folks,

Based on discussions with various people at Flock last week, I've
started hacking on a proof of concept for a "repo funnel" service that
allows people to define their own custom repositories in Pulp as a
filtered view of the smorgasboard offered by COPR. That way, they can
say "I want to use these repos" once in an online service, and
subscribe their client machines to that single custom defined repo,
rather than having a proliferation of COPR repos directly on client
devices, which are then hard to replicate across machines (e.g. a
separate laptop and workstation, or multiple remote servers).

However, I want to design it in a way that means it also ends up being
useful for improvements to the Fedora Review Process, which means
putting some thought into the overall workflow this might enable
*before* I get too deep into building it.

Accordingly, I put together a page on the wiki at
https://fedoraproject.org/wiki/Env_and_Stacks/Projects/SoftwareComponentPipeline
to describe what *I'd* be hoping to get out of this.

That page also captures my own input on "What should the role of E&S
be in further defining the Fedora Rings concept?" - at the very least,
I think we should be the ones responsible for the overall developer
experience of the distro (encompassing not just client systems, but
relevant online services), that we should be setting expectations for
the contributions individual language SIGs make to that process, and
defining the handover to the Base WG and the Edition WGs for the
assembly of the available software components into the Fedora
Editions, EPEL, and the main Fedora repositories.

Regards,
Nick.

-- 
Nick Coghlan   |   ncoghlan at gmail.com   |   Brisbane, Australia