Le Jeu 26 avril 2007 21:30, Axel Thimm a écrit : > As such the punchhole remove/install problem is an embedded issue of > the multilib design (TM). Actually if you want to go to the core issue, that's that no one ever decided what a rpm package was : A is it a container with invariant mandatory content? B is it a container with mixed mandatory and semi-optional content, that may be installed or not depending on system settings, various euristics and the phase of the moon? If the answer is A, puncholes are unacceptable, langpacks must be split... and yum taught smarts to reassemble all the resulting subpackages because raw access will overwhelm users. (dumb packages smart package manager option) If the answer is B loads of smarts must be added to rpm so all the optionnal stuff actually works without side effects. Also that means fat packages and big downloads. (smart packages dumb package manager option) Either way is loads of work because we let out tools bitrot for years, and now we find the workaround pile is not up to the new multi-arch multi-repo world. I personnaly think A is the cleaner design, but that's up to the tool writers to decide (Paul Nasrat, Seith Vidal...). Likewise yum effectively deprecated parallel rpm installs and rpm relocation a few years ago. -- Nicolas Mailhot -- Fedora-maintainers mailing list Fedora-maintainers@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-maintainers -- Fedora-maintainers-readonly mailing list Fedora-maintainers-readonly@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-maintainers-readonly
