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

Re: A plan to get rid of unnecessary package dependencies

On Mon, Sep 25, 2006 at 02:40:49PM -0700, Kevin B. McCarty wrote:
> (Sorry if you get this twice, I don't know whether you subscribe to
> debian-devel.)

I am subscribed here, also sorry about the broken MFT in the
-announce mail :-/

> Thank you for this very cool effort!  Might we see checklib packaged for
> Debian soon?

Thanks :)

Aurélien GÉRÔME started an alioth project, I'm also member there:

Nothing there yet, as I wanted to get the announcement out before
starting to work on the codebase again to clean it up (which is
badly needed :-/).

The current version of the source is linked on the bottom of the
pages (updated 10 minutes ago, so please reload ;)

> One thing I noticed is that there are a lot of "problems" (in your
> terminology) caused by unneeded dependencies on libgcc1
> (/lib/libgcc_s.so.1).  From my quick investigation, it appears that the
> C++ and Fortran compilers (g++, g77, gfortran) introduce this dependency
> automatically to programs linked with them.  However, if gcc is instead
> used in the linking step, no such dependency is created (at least on
> amd64 where I'm testing).

I'm not sure if libgcc_s.so.1 doesn't do some magic that checklib
can't detect. Maybe one of the gcc maintainers could clarify if/when
it's needed.

If it really is needed it could be ignored for the results, that
wouldn't be much of a problem.

> Even though you are technically correct that libgcc1 is an unneeded
> dependency, it may be hard for anyone (except maybe the gcc maintainers)
> to do anything to fix this; could you have an option for libcheck to
> ignore this library?  I suspect if this was done, the proportion of
> "problems" in your pie chart would shrink a fair bit.

Could be done, however if the dependency is really unneeded, it
might be interesting to fix it in gcc if possible and desirable
(i.e. not unpractical for reasons unknown to me).

One general comment: Checklib problems concern binary packages. So
if there's a problem on one package, that doesn't necessarily mean
that the corresponding source package is the correct place to fix

As in this case the problem might originate in gcc, another common
reason are broken .la/.pc files in -dev packages. Checklib just
can't know where the problems come from, it can only see where they
manifest themselves.

Richard Atterer wrote:
> Hmm, maybe the functionality could be included in lintian?

I'd have to talk with the lintian maintainers, but it's not easy.
Checklib needs a local Debian Mirror to work, and that sounds a bit
problematic for inclusion in lintian :)

There might be ways around this, I'll try to explore that.

Christian Aichinger

Attachment: signature.asc
Description: Digital signature

Reply to: