[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Transition from dpkg to GNU install-info



I went ahead and rewrote a bit the plan and removed stuff that did not
seem to apply to this transition and placed it in a wiki page,
<http://wiki.debian.org/Transitions/DpkgToGnuInstallInfo>. Hope you
don't mind.

On Wed, 2009-03-11 at 08:49:07 +0100, Norbert Preining wrote:
> On Mi, 11 Mär 2009, Guillem Jover wrote:
> > >  + Changes in texinfo
> > >    - a new install-info package replaces dpkg's install-info
> > 
> > Given that we have other info viewers, that seems better than bundling
> > it in the info or texinfo packages.
> 
> Here immediately some questions: I can start preparing a separate
> install-info bin package for experimental, but 
> - should we install GNU install-info into /usr/bin where there is
>   ginstall-info ATM, or into /usr/sbin, where there is dpkg ii?
>   I would suggest /usr/bin.

I think it'd be better to not entangle those two changes, so I'd put it
in the same place as the current install-info from dpkg. A transition
from /usr/sbin to /usr/bin could be done later.

> - can you prepare a dpkg experimental package that does not ship ii but
>   depends on install-info?

As per the transition plan dpkg will still ship a dummy script that
warns the users, this way we don't need dpkg to depend on install-info.
So you only need to add a Replaces against dpkg. Also you should be able
to upload this to unstable whenever you are ready w/o waiting for anyone.

> - I would ship an ii package that conflicts with dpkg < experimental and
>   then we could try the whole thing out.

See above.

> > 
> > >    - This install-info should support the current dpkg's install-info
> > >      options (i.e. no failures, but not necessarily registering the info
> > >      file)
> 
> that would mean a shell/perl script as /usr/bin/install-info that in
> turn calls /usr/bin/ginstall-info, and warns in case we are called with
> a non-supported argument.

Either that, or extend GNU's install-info to accept the other old
arguments as no-ops. Once all Info files have been fixed the wrapper
could be replaced by the real thing.

regards,
guillem


Reply to: