On Thu, Feb 14, 2008 at 01:32:01PM +0100, Vincent Fourmond wrote:
> On Thu, Feb 14, 2008 at 1:24 PM, Alexander Schmehl <tolimar@debian.org> wrote:
> > > > Just noticed, that we are quite inconsequenz regarding our -data
> > > > packages. Some of them depends on their binary counterpart (e.g.
> > > > warzone2100), some of them don't (e.g. adanaxisgpl).
> > > FTR warzone2100-data actually recommends warzone2100 rather than
> > > depending on it.
> >
> > Yes, that was what I meant but didn't wrote ;)
> >
> > -data packages should IMHO recommend their binary packages, not depends
> > on them.
>
> In case of tight link between data and game (say, version 0.2.3 of
> the game doesn't play with 0.2.3.1 of the data), I think the best is
> to use strict versioned depends on the binary and possibly versioned
> conflicts on the -data package. My guess is that it should take care
> of all version incompatibilities while allowing people to install the
> data independently of the game. (or did I miss something ?).
We all seem to agree that the binary should depend on the data, and
there should be a recommends back. This is also technically correct (in
terms of "can't work without" and "will for most users be useless
without").
Two versions of the same package cannot be installed at the same time.
This means that with a versioned Depends (from the binary on the data),
there is no need for a versioned Recommends (but it doesn't hurt either,
of course). The versioned Depends makes sense as well (for binary and
data packages which are generated from the same source package) IMO: We
only test the package with the same version of the data. If the binary
is available in an archive, the data should be as well (they're built
together after all; the data, being arch-independent, can be installed
before, but not after the binary).
So for packages which build their own data, I think it is good if the
relation is a strict Depends: foo-data (= ${source:~version}). For data
packages from a separate source package, I think it's good to assume
everything works, unless there is reason to know it doesn't.
Thanks,
Bas
--
I encourage people to send encrypted e-mail (see http://www.gnupg.org).
If you have problems reading my e-mail, use a better reader.
Please send the central message of e-mails as plain text
in the message body, not as HTML and definitely not as MS Word.
Please do not use the MS Word format for attachments either.
For more information, see http://pcbcn10.phys.rug.nl/e-mail.html
Attachment:
signature.asc
Description: Digital signature