=sci-physics/geant-data-4.10.00:4 depends on the exact matching package-version of sci-physics/geant, so =sci-physics/geant-4.10.00_p01:4 is blocked by it. My suggestion would be to loosen the dependency of =sci-physics/geant-data-4.10.00:4 - the dependency of geant upon the datafiles (when USE=data) should anyways be sufficient, is the dependency the other way around really needed? Output from portage (shortened): [ebuild N ] sci-physics/geant-4.10.00:4 [ebuild N ] sci-physics/geant-data-4.10.00:4 [ebuild N ] sci-physics/geant-4.10.00_p01:4 !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: sci-physics/geant:4 (sci-physics/geant-4.10.00_p01:4/4::gentoo_prefix, ebuild scheduled for merge) pulled in by =sci-physics/geant-4.10.00_p01 (Argument) (sci-physics/geant-4.10.00:4/4::gentoo_prefix, ebuild scheduled for merge) pulled in by ~sci-physics/geant-4.10.00:4 required by (sci-physics/geant-data-4.10.00:4/4::gentoo_prefix, ebuild scheduled for merge) Reproducible: Always
The upstream release geant and geant-data together. Any mismatch in the versions of the two is not supported. Our packaging practice is just a reflect of this reality. Why do you want the two versions of geant and geant-data in the first place?
You are right, but I do not want to mix versions: - Upstream did not release new data files for 4.10.00_p01 over 4.10.00 (unless I missed that) - there is no =geant-data-4.10.00_p01 in the tree, which geant-data should I use with =geant-4.10.00_p01? This basically means USE=data emerge =geant-4.10.00_p01 has to fail unless either the dependencies of =geant-data-4.10.00 are loosened or =geant-data-4.10.00_p01 is provided.
Ah, I see. Sorry I didn't understand the problem on the first sight. Yes the patch level is a pain in the version. I've thought on this issue months ago, but without a conclusion. @sebastian, do you have any insight on this?
+ 29 May 2014; Sébastien Fabbro <bicatali@gentoo.org> geant-data-4.10.00.ebuild: + Relaxed hard dependency on exactly same version of sci-physics/geant + i don't use geant, but feel free to do something else (such as slotting, or different versioning r1 vs p01) if you think geant data version has to be strictly compatible with geant