Re: not later than in testing

[Luk Claes <luk@debian.org>]

Joerg Jaspert wrote:
> On 11014 March 1977, Keith Edmunds wrote:
> 

> So, what do people here on the list think?
> 
> a.) have a second suite for the backports >> etch version and keep a
>     frozen sarge.
> 
>     Advantage: Keeps an upgrade path of sarge+bpo -> clean etch.
> 
>     Disadvantage: Much work on the ftpadmin side. And yet another suite,
>     add to confusion.

I agree that people using backports of lenny to sarge will probably stay using
these backports in etch. People are also advised to only use a couple of
packages from backports as using a lot of packages from backports is not
guaranteed to work. So I don't see many advantages of having an upgrade patch
to clean etch.

> b.) the above, ie. make it a technical requirement that versions have to
>     be etch >> sarge. Where a new hidden setting would be added to the
>     bpo archive that contains the etch version numbers. (Plain etch +
>     security).

This technical requirement would be very welcome IMHO.

>     Advantage: Technically enforced that people who upload package
>     foobar to sarge-bpo have to take the steps to get foobar backported
>     to etch, if they want to upload a version that is not in etch.
> 
>     Disadvantage: Upgrading means to continue to use backports.
>     People have to backport twice if they want it in sarge. Once for
>     etch, once for sarge.

I don't think these should be seen as disadvantages, it are pre-conditions to
have at least some upgrade path available. Plus if you need a backport of a
version of lenny in sarge, you very probably also want that version if you
upgrade.

Cheers

Luk