reviving Fedora Legacy

[Patrice Dumas]

On Mon, Oct 13, 2008 at 10:35:18AM -0400, Josh Boyer wrote:
> 
> This has impacts on everyone though, not just the people that
> want to keep an old branch active.  You have to worry about:
> 
> 1) Upgrade paths

No more than in fedora releases.

> 2) buildsystem inheritence

That seems like an interesting issue...  Could you be a bit more 
specific?

> 3) Storage (CVS, mirrors, koji, etc)

CVS is not really an issue, we leave old branches already currently.
Others may be problematic. How many updates do we keep in the download
server? Is it only one version or more? If there is only one version,
then there is no problem, in my opinion. For koji, I know that some
builds are already deleted, maybe for LTS we could keep builds around
for some months and delete all except the last one?

> 4) Bug reports (because they certainly will get reported to the
>    overall package maintainers, not just the "LTS" maintainers)

I see 2 phases. In the first phase we consider that EOL product aren't
maintained anymore, that is we respond to those bug reports in the same
way than we respond to bug reports about EOL products today. If there is
enough interest in the project, then we mandate LTS maintainers to be in
watchbugzilla and we find a way to advertise which packages are still
maintained. The most simple would be mass orphaning packages such that a
maintainer has to put himself in the fedora db for a package for that
package to be considered to be still maintained.

> 5a) Depletion of test and development resources to current releases

I don't think we should mandate how fedora contributors allocate their
time.

>  or
> 5b) Lack of participation

Time will tell.

> I'm not saying it's a bad idea, just needs a lot of thought.
> There is a reason the Fedora Legacy project died.

Once again it is very different now. Legacy had different
infrastructure, different rules, and overall was more complicated. And
there has been improvements in the fedora infrastructure since then.

--
Pat