Re: RFC: pentium optimized debian
Joey Hess <joey@kitenet.net> writes:
> [seperate pentium architecture]
> > The dpkg problem I'm not sure. We can likely easially hack it to
> > know that a list of arch's are valid. In truth on some systems
> > target specific optimizations will NOT run on lesser
> > processors. Dpkg should not install arch specific packages in
> > these cases.
> If you want a general solution, you need a table that lists which
> arch's are valid for each other arch.
RPM does this. dpkg should too. This is used this to handle both
"noarch" and machine specific architectures (noarch is the RPM equiv
of our "all" architecture), and it is used in the UltraSparc
distribution to handle installation on 64-bit sparcs, which can handle
both the "sparc" and "usparc" architectures.
If people in Debian are serious about using dpkg on other platforms or
distributions, they are going to have to work on this stuff. RPM is
way beyond dpkg in this respect. Due to active development by people
both inside and outside Red Hat (they seem pretty open to
improvements), RPM is now very portable and is being actively used on
many platforms outside Linux (e.g. HPUX, Solaris, IRIX) and has be
ported to even more platforms (e.g. Cygwin32).
(The OSes that RPM currently knows are: Linux, Irix, Solaris, SunOS,
AmigaOS, AIX, HPUX, OSF1, FreeBSD, SCO, Irix64, NextStep, and BSD/OS)
I definitely think a similar mechanism (for architecture) would be
beneficial to dpkg. For other OSes, I think RPM has lapped us and we
won't catch up, but we would benefit from a multiple OS mechanism if
dpkg ends up being used for HURD.
Steve
dunham@cps.msu.edu
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: