Re: Kernel module packages, the dkms approach ...
"Lee Taylor" <[email protected]>
| Newsgroups | gmane.linux.freshrpms.user |
|---|---|
| Message-ID | <[email protected]> |
Greetings ... Sorry for the delay a reply ... > Matthias Saou wrote : > > > > [...] > > > Feel free to let the list know what you think about this approach. Any > > > bug reports, fixes, enhancements or comments regarding this package > are > > > also very welcome. > > > > > > I'm already thinking of doing a dkms aware package of the lirc kernel > > > module I use on my living room computer, as well as a madwifi one, as > > > tracking Rawhide updates and keeping the remote and the network > working > > > is a waste of time! > > > > No on-list feedback, so I don't really know what to think... :-) > > I'm pleased. I never liked the DKMS approach (in the sense that the > general rule is to not have a compiler installed on end-user systems), but > diversity is always a win. I am providing the dkms packages for other > distributions as well. I'm with Dag on this, but it would make life alot easier fo rsome of my systems. I do like the idea of been able to boot into a new kernel and (hope) that all kernel modules are working without having to wait for all the kernel modulerpm maintainers to catchup ... It's almost a totally win-win without too much to loose (not my data). Would be nice if both DKMS and packaged modules could live together ... or, packages could be make from the DKMS system without needing a compiler on end user systems. I think I did see something like that for mpppe kernel module a little while ago. I guess I need to provide some infrastructure packages. Is there some > documentation on how it would work with RPM ? Something to kickstart me in > providing the same for RHEL ? Always good ... ;-) Mailed Lee