Patch to sort out libgstreamer0.9 versioning

[Loïc Minier]

        Hi,

On Tue, Sep 13, 2005, Rob Taylor wrote:
> no, this is always true without loss of generality where A and B are not
> identical.

 The formula might be always true, but I maintain that the
 simplification of using one dep only (the current result of the
 formula) is only ok for the current version, in the current use case.

> any situation where A!=B will have either A>B, or B<A (as versioning is
> totally ordered), if B<A, simply rename the variables, and the argument
> holds.

 Yes, the formula is going to be correct, but only with respect to
 inter-dependencies, and other packages will get a too high dep.

> The above argument only makes sense to me if packaging system and katie
> transitions consider dependencies generated by shlibdeps differently
> somehow. I don't believe this is the case, i could be wrong of course...

 Shlibs are bundled dependencies, in /var/lib/dpkg/info/*.shlibs.  If
 you build other packages with ELF "DEPENDS" on "SONAME"s in these
 files, then dh_shlibdeps will add a dep as specified in the shlibs
 file.

 That means that aside from inter-dependencies between binary packages
 from the same upstream source, there are other dependencies on these
 binary packages from other packages build-depending on this source.

 You only look at the current deps in the GStreamer source package you
 consider, please take into account what the other packages will see as
 shlibs.....

   Bye,

-- 
Loïc Minier <lool at dooz.org>