[Mageia-sysadm] list of things to do in sysadmin
rdalverny at gmail.com
Fri Apr 8 12:06:36 CEST 2011
On Fri, Apr 8, 2011 at 11:03, nicolas vigier <boklm at mars-attacks.org> wrote:
> First we need to check if it's doing what we want. And check that we
> all agree on what it should do. I didn't look in details, but :
> - I think gitorious allows anybody to create his own repositories. But
> I think what we want is shared repositories, and I don't know if we
> want to host personal ones too.
The whole interest of gitorious is precisely that (easily forking a
repository for pushing patches afterwise). If we cut this off, we
better not use this solution at all.
> - We need to see how we manage commit permissions. We probably want to
> use ldap groups for this, I don't know if it's possible with
About gitorious, I don't know if we can interface with its own ACLs,
but there's a web-based UI for managing that there, per repository.
Better use it.
> - gitorious or redmine include wikis, forums, bug tracker. Maybe we
> don't want that, so we should be able to disable them.
Yep. Or consider that is a separate forge.
> - if we only use this tool to create new repositories and have a web
> interface to show commits but use none of the advanced features of
> gitorious/redmine, maybe a few scripts in puppet and gitweb is more
> simple than customising gitorious/redmine to do exactly what we want.
The point is: do we use a tool that already builds a multi-projects
forge, or do we build our own?
For the distribution and Mageia project itself, everything revolves
around a few monolithic bits at this time: Bugzilla, SVN, buildsystem.
If we want to give separate "upstream" identities to software in /soft
(or to some tools still buried down here and there), each one should
have its own:
- fully customizable web page (with an obvious path to it) - with a
default template provided for those who don't need/want to customize
- bug/feature tracker
- code repository + web output of it
- perhaps a mini blog, but a forum seems less likely to happen there
If we want to give some flexibility into the project to incumbate
small projects to embbed into or derive from Mageia, creating such a
new "upstream" project should be, at least for some people, as easy as
using a single web form - so yes, it would be somehow like what
tuxfamily, savannah, indefero, google code, github, etc. - shall we
use these instead then? I believe we are not all in sync about this.
So the question is: do we build this with something existing that
duplicates somehow what we have already, and integrate it as we can?
or do we build this on top of what we already have?
> So someone needs to see exactly what is needed, with people that will use
> the repositories, and look at all available tools to be able to decide
> what we should use.
We've got to decide at some points too. Or we will still be at this
point in 6 months.
More information about the Mageia-sysadm