Alan Cox wrote:
>
> > Quite a while back, we agreed that applications should be built for a 386
> > for portability. If an application used a feature of a more recent
> > processor,
> > then it was responsible for checking and doing something appropriate if the
> > test failed.
>
> In which case I don't think we need to worry about i386/i586/athlon/transmeta
> etc.
I think we might have to for the "requires", since by the time the app
runs it's really too late. It's rather necessary to be able to tell
ahead of time.
> Perhaps we should acquire an "optimised for:" and "cpu required:" in RPM or
> the tool the lsb chooses ?
Sounds like a good idea to me...
-hpa
--
<[EMAIL PROTECTED]> at work, <[EMAIL PROTECTED]> in private!
