[arch-dev-public] Catalyst naming - looking for suggestions

Roman Kyrylych roman.kyrylych at gmail.com
Sun Feb 10 03:43:29 EST 2008

2008/2/10, Travis Willard <travis at archlinux.org>:
> Ok guys,
> I've been talking to the AMD/ATI guys about getting ArchLinux packages
> bundled with their catalyst releases, and was told something
> interesting - apparently, the driver (kernel/Xorg) is independent of
> the catalyst release, and is still versioned the "old way" - for
> example, we used to have "fglrx 8.45.4" or whatever, and I changed it
> to "catalyst YY.MM" when AMD started releasing their linux drivers as
> 'catalyst'.  The thing with this is, there may be special versions of
> the driver between catalyst releases for certain vendors, for example
> - so catalyst 8.01 may have driver (fglrx) version 8.45.5, but then
> they release a special version for laptop X that's 8.45.6, and isn't a
> catalyst release, as it's geared to the specific laptop's hardware.
> So, basically, I jumped the gun renaming fglrx  to catalyst, as (in
> general) it makes more sense to keep versioning it the old way (which,
> at the time, I thought was going away, and was mistaken).
> So, what I'm wondering is, should I re-rename the packages back to
> fglrx?  Leave them catalyst, and ignore any special versions that may,
> at some point in the future, come out?  Re-rename them to something
> other than 'fglrx' to indicate they're the Radeon Proprietary drivers?
>  I'm not sure how to handle this, and suggestions are more than
> welcome.
> I asked the ATI guys the same thing, and they said they'd prefer to
> have some 'meta-package' called catalyst that depends on the
> appropriate fglrx-* packages, which I don't really like, because then
> the 'catalyst' package is an empty package, used only for the deps it
> pulls in.
> Thoughts?

I don't mind renaming them back to fglrx again. *shrugs*
For extra-special versions - I doubt they will ever be in our repo.
Meta-package is definetely a bad idea, we've got rid of most of them
(except xorg-clients).
Does the decission to include Arch support in the official catalyst
depends on our packaging scheme?

Roman Kyrylych (Роман Кирилич)

More information about the arch-dev-public mailing list