Re: How about a db 5.1.25 backport
On 2011-04-18 Gerfried Fuchs <rhonda@deb.at> wrote:
> * Andreas Metzler <ametzler@downhill.at.eu.org> [2011-04-17 13:46:58 CEST]:
> > I am considering uploading a backport of berkeley db 5.1.25. Packages
> > in sid are (supposed to be) switching from 4.8 and later to 5.1, which
> > is not available in squeeze, making backports hard.
> It makes backports hard in the way that to work around the issue you
> raised the backport will require a versioned dependency on libdb-dev. If
> that's clear for everyone involved then it's fine with me, but I fear
> the issue will be moved to a different level.
Hello,
Afaiui the db using packages (should) fall in two camps:
1. Packages with an unversioned (or >= squeeze) build-dependency on
libdb-dev. (Probably >= 3.2) They can be switched to new bdb by
binNMUing and therefore must automatically handle file format upgrades
in their maintainerscripts, without relying checks of the package
version.
2. Packages build-depending on the the versioned -dev packages (e.g.
libdb4.8-dev). Database upgrades only happen when the maintainer wants
them to happen, since they require a sourceful upload. The maintainer
can add custom handholding for each b-d update in its maintainer
scripts.
I currently fail to see how adding db 5.1 to backports.org breaks either
of these: The packages in 1. will usually not have b-d on the backported
db package, and will therefore be built against squeeze's db. The
packages in 2. will suddenly be backportable in straightforward way,
without big source changes.
Your point seem to be rather new broken packages in sid. - They have
been uploaded recently with an unversioned libdb-dev b-d without
handling db upgrades. This bug has not yet been triggered in sid,
because not db upgrade has hit them.
Did I understand the point correctly?
thanks, cu andreas
--
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
Reply to: