On Sun, 9 Jan 2005 17:53:25 +0000
Ciaran McCreesh <[EMAIL PROTECTED]> wrote:

> On Sun, 09 Jan 2005 18:40:42 +0100 Benjamin Schindler
> <[EMAIL PROTECTED]> wrote:
> | > Do all amd64 boxes support sse? If they do, you can do:
> | Yes, both sse and sse2
> 
> Right. Then you should not be offering USE flag settings for sse/sse2
> on amd64. The only purpose of those flags is to allow us to build,
> say, i686 stages on a pentium4 without having to create a whole new
> arch to do so.
> 
> Even if they didn't, according to remarks on irc supporting sse and
> friends on amd64 is a very different task from supporting them on x86.
> In which case, have different use flags, and plain old use.mask will
> work just fine.
> 
> | It would be much cleaner not needing to do this. But lets see
> | whether portage people crucify us *G*
> 
> Given that you'll be breaking SRC_URI, IUSE, emerge -pv output,
> LICENSE and the cache, and that you'll be tinkering with things that
> reaaaallly aren't supposed to be touched, I don't fancy your chances.

Yep, if anyone starts to play with portage internals inside an eclass
you better prepare for some nasty treatment, I think we had enough
issues with the debug eclass in that regard.
IMHO the proper way would be to use USE_EXPAND for these flags, but that
would require changes to all affected ebuilds as the generic
sse/mmx/... flags would be renamed to arch-specific ones.

Marius

-- 
Public Key at http://www.genone.de/info/gpg-key.pub

In the beginning, there was nothing. And God said, 'Let there be
Light.' And there was still nothing, but you could see a bit better.

Attachment: pgp3Nx7u2qmiQ.pgp
Description: PGP signature

Reply via email to