[SM-Discuss] Architecture-dependent spells

Eric Sandall eric at sandall.us
Fri May 14 14:12:05 EDT 2004


Quoting Arwed von Merkatz <v.merkatz at gmx.net>:
> On Fri, May 14, 2004 at 10:18:35AM -0700, Andrew wrote:
> > > > with ARCHS when unset in DETAILs be set to all ARCHS that we support.
> > > 
> > > Good idea. Or perhaps we should reverse the ARCHS flag and have it
> specify the
> > > architectures this spell does /not/ work on, as that would probably be
> the
> > > smaller list.
> > > 
> > Im going to try and formalize the semantics:
> > 
> > We'll have an ARCHS and a NO_ARCH variable (or whatever you want to
> > call them, i really dont care either way)
> > 
> > If neither of those are set then the spell is assumed to compile anywhere
> > 
> > if ARCHS is set and NO_ARCH is not set then the host architecture must
> >    be in ARCHS, and thus everything is assumed to fail
> > 
> > if NO_ARCH is set and ARCH is not set the host architecture must not be
> >    in the list, and thus everything else is assumed to work
> > 
> > 
> > if BOTH ARCHS and NO_ARCH are set, then we end up in a situation where
> >    we've assumed anything not in either list will both succeed and fail,
> >    so we must choose a default, in the name of consistancy I say we
> >    assume that the spell will compile, which means that we ignore ARCHS
> >    and just make sure that the host architecture is not in NO_ARCH.
> > 
> > -Andrew
>  
> Sounds good. I guess we can assume both ARCHS and NO_ARCH set in the
> same file won't happen except when people doing p4 submit while
> asleep/drunk/otherwise unable to think ;)

Agreed, and this way we don't have to modify all of our spells, just the ones
which don't work on certain architectures for now.

-sandalle

-- 
Eric Sandall                     |  Source Mage GNU/Linux Developer
eric at sandall.us PGP: 0xA8EFDD61  |  http://www.sourcemage.org/
http://eric.sandall.us/          |  SysAdmin @ Inst. Shock Physics @ WSU
http://counter.li.org/  #196285  |  http://www.shock.wsu.edu/

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.



More information about the SM-Discuss mailing list