On Tue, Dec 27, 2005 at 01:39:14PM -0500, seth vidal wrote: > > Is this intentional? It seems like it could be beneficial to have the > > same package syntax that is supported in other places (i.e. anaconda, > > yum itself) work in the groups format. > > > > It would be beneficial, especially when testing an application that's > > being ported, to be able to groupadd and groupremove groups that are > > defined by arch. > > > > it makes it a pain to share this groups file across architectures of > the same release, though if you have to constantly account for all these > arch-specific packages. > > -sv I can see that argument. I guess I would still find it useful, though. The 32-64 bit transition makes for a unique, and potentially long-lived PITA where these kinds of workarounds need to be employed. Perhaps a tag, instead of the name.arch syntax would be better? I.e. (from memory, since I'm about 4,000 miles from work right now) <group>abi_transition</group> <required abi_bits=32>package1</required> would be less intrusive, since no matter what platformn the name is the same, for those where there is no 64-bit abi the abi_bits would be ignored, and for those that are 64-bit this could be honored by passing a flag to yum to parse on the client side, i.e. yum groupinstall abi_transition --abi_bits=32 and this should become purely a parsing trick on the client side. Any thoughts on whether this would be acceptable? Perhaps it could be a modular add-on to yum rather than a part of the core? All it needs is for the groups file have this syntax and have it agreed on. Thanks, -Peter -- The 5 year plan: In five years we'll make up another plan. Or just re-use this one.
