Please find attached ebuild for madwifi cvs version.
Created attachment 20458 [details]
Any objections to changing the header to the standard?
# Copyright 1999-2003 Gentoo Technologies, Inc.
# Distributed under the terms of the GNU General Public License v2
# $Header: $
I can copy the old header to the changelog so you both will be credited as
No problem with header, do what you want :)
Great, it is committed.
I ran into a problem where re-mergeing this would fail on cvs update. Deleting
the distfile/cvs-src/madwifi-cvs directory took care of the problem. Could
you see if this happens to you as well? This is the first time I am using
the cvs eclass so I don't know how it behaves.
CVS ebuilds are considered a Bad Thing (tm) unless *absolutely* necessary.
PLEASE don't commit them unless they've been more widely discussed and deemed
warranted. I know the madwifi people have taken ages to release a new tarball.
Until they do, can we please remove this and instead take a CVS snapshot,
get it tested, and add it to the normal "madwifi-driver" package. Thank you.
Stephen: I have no problems at all with cvs update when re-merging.
Peter: It makes no sense to use the madwifi-driver package as it is right
now, the code is very old and cvs version is a lot more stable and reliable.
I agree that cvs ebuilds are generally not a good idea, but in the madwifi
particular case it is very useful to have a cvs ebuild as the driver is being
developed and it is far from a stable version. If cvs ebuilds are not suitable
for gentoo policy, then a regularly updated cvs snapshot in portage would
be very helpful too.
I was planning to switch over to a snapshot eventually but since it is that
much against policy I'll take a snapshot this afternoon/tonight and rework
the ebuild. I am at work now so cannot remove the package but feel free
to wipe it out if tonight is too late. I'll try to keep up at least a monthly
Why do we have the cvs eclass if it is against policy?
Yeah, all i was suggesting is that CVS snapshots are better suited as they
let us verify a specific ebuild will actually merge, instead of relying on
the current madwifi cvs not to be borked. I have a new atheros miniPCI card
i'll be able to do this testing on soon as i get my stupid T30 with it's
annoying BIOS check to accept the card.
the cvs eclass exists in those extreme cases where a CVS ebuild is required..
there's even been some discussion of removing it altogether (not that i see
that happening any time soon).
Not trying to make a fuss, just want us to have a stable and up-to-date driver
for atheros cards. (:
Added todays CVS snapshot as madwifi-driver-0.1_pre20031110. Please test.