Sven Luther wrote: > On Thu, Oct 06, 2005 at 11:06:47AM -0500, Jaldhar H. Vyas wrote: > This may have used to be the case, but should not be a problem anymore, we > have only one kernel per released architecture, and make it easy enough for > them to build modules for the official kernels, the debian kernel team needs > to provide a document on how to build modules probably, but even if it is not > yet fully documented, everything is there to make it happen. They seem to be comparing Debian unstable, with other distros' official releases, which is a bit strange -- presumably they're not claiming to support beta versions of those other distros. Anyway, I seem to remember that they provide the source for the bits that need to go into the guest operating system (I could be wrong, it's been a while since I last played with it). Given that, assuming we can have permission to redistribute binaries, and someone is willing to package them, the bits required to make everything work in the guest could be packaged and distributed (probably in non-free, but distributed nonetheless) by Debian, making it trivially easy for people to install under VMWare. A vmware-guest package could even depend on particular kernel versions if they're that stressed about it (savy admins could always get round that, at their own risk). Alternatively, the postinst could check the environment it's sitting in and put up a warning about it being unsupported, and how to fix that. Either of these would provide more assurance to them than they currently get from an RHEL system with a locally patched kernel. Perhaps this should be pointed out to them, since if that were to happen, we'd be doing their testing for them during the Debian release cycle, and they would just need to confirm the facts at release time. Cheers, Phil.
Description: OpenPGP digital signature