[Python-modules-team] python-requests backport
Julien Cristau
julien.cristau at logilab.fr
Thu Jun 6 12:54:53 UTC 2013
On Thu, Jun 6, 2013 at 08:47:04 -0400, Yaroslav Halchenko wrote:
> > Hello Julien and Yaroslav,
>
> > thanks for remindig this: I was wondering about backporting requests but as
> > noted by Yaroslav we have to investigate how many packages will break.
>
> > On Wednesday 05 June 2013 08:30:55 Yaroslav Halchenko wrote:
> > > It would be nice to know if it would break any of the reverse
> > > dependencies if backport gets uploaded (may be add NEWS entry to make it
> > > obvious happen someone pull it from backports)
>
> > I will be happy to work on this, but since I'm not a DD (yet ;) and this is
> > the first time I contront with this kind of situation I was wondering if ftp-
> > masters will allow breckage in our stable release. Is a NEWS entry enough? :)
>
> to say the truth -- I do not know. I do not use/upload to backports.d.o
> much, so this warning comes from the experience with our own
> neuro.debian.net: there we upload lots of backports but we really avoid
> backporting core libraries used by other projects if we know that they
> break compatibility. Some times it possible to provide workarounds at
> some cost of inconvenience -- that is how
> http://neuro.debian.net/pkgs/ipython01x.html
> was born. So we could have stock ipython supported by the release and
> new shiny one co-existing on the same system.
>
> Back to the cows -- according to the backports.d.o:
>
> Backports cannot be tested as extensively as Debian stable, and
> backports are provided on an as-is basis, with risk of incompatibilities
> with other components in Debian stable.
>
> and I do not see otherwise any other recommendation/policy to not upload
> packages known to break "stable" components. So as far as I see it --
> if you are keen on taking the burden of maintaining it there in b.d.o
> and replying to people running into problems -- you could upload. NEWS
> entry though would yet again warn people that this version would break
> things.
>
I think if there's known breakage then the backport can declare Breaks
on appropriate versions of its reverse deps. And actually, those Breaks
should also be in the package in sid/jessie, to handle partial upgrades
properly.
Cheers,
Julien
--
Julien Cristau <julien.cristau at logilab.fr>
Logilab http://www.logilab.fr/
Informatique scientifique & gestion de connaissances
More information about the Python-modules-team
mailing list