Subject: None
To: Jason Thorpe <>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: tech-toolchain
Date: 07/26/1998 14:30:38
>Todd is already fixing the Makefiles.  It's not that hard, 0xd00d.

``Breaking''` them would be more accurate. Of course going through and
doing the change is easy; i've neveer said otherwise. It's just not
very maintainable and not the right engineering decisison.  and we
have to go ``fix'' them again when we introduce a 64-bit mips.

We should be doing this Really Right(tm), and IMNSO, just forking off
${MACHINE_ARCH} at each point is not Really Right(tm).

Better to just define a variable to "mips" somewhere and _use_ it,
everywhere where the semantic intent is to test for ``any mips'',
surely?  That's the key point I've been arguing.  (I still think it'd
make more sense for that to be ${MACHINE_ARCH, but so it goes).

What do you think about the idea that an ONLY_FOR_ARCHes of "mipseb"
should also imply "mipseb64" (assuming a 32-bit emul, as discussed for
sparc)?  Should we maintain those dependencies by hand forever?
Will the pkg peolpe who do that understand the issues?