> Yes, actually:
>
> cc -mips4 -n32 -L/server/people/staff/easmith/lib32 -L/usr/lib32/mips4
> -L/usr/lib32 -L/lib32
>-Wl,-rpath,/server/people/staff/easmith/lib32:/usr/lib32/mips4:/usr/lib32:/lib32
>-Wl,-multigot -woff 1110
>
> Hmm... interesting. Unsetting this results in Makefile.PL succeeding
> in producing a Makefile, which successfully does a make and make test:
A-ha!
> I'm afraid it didn't produce any further info. It looks to be the
> checking for $ENV{CC} that's the problem. I'm also wondering about the
> 'basename' in there if $cc isn't executable - it won't be if $cc is
> more than just "cc" or "gcc". Ah. Yes. I stuck in lines to print out
> $cc before and after that step, and got:
>
> cc is cc -mips4 -n32 -L/server/people/staff/easmith/lib32 -L/usr/lib32/mips4
>-L/usr/lib32 -L/lib32
>-Wl,-rpath,/server/people/staff/easmith/lib32:/usr/lib32/mips4:/usr/lib32:/lib32
>-Wl,-multigot -woff 1110
> cc is lib32 -Wl,-multigot -woff 1110
Okay, the basename() logic obviously needs a rethink.
> > I don't think the BSD::Resource currently supports the *64() versions.
>
> Well, no, but I thought you might like to know about them... I would
I knew of them but haven't bothered so far, there be icky portability
issues there that I'd rather let sleep in peace...
> be curious if this is how other mixed 32/64-bit systems (if there are
> any other than IRIX) are doing things.
Of all systems I know of IRIX makes the most effort for a mixed mode,
others take more either-or approach.
> > --
> > $jhi++; # http://www.iki.fi/jhi/
> > # There is this special biologist word we use for 'stable'.
> > # It is 'dead'. -- Jack Cohen
>
> As I previously remarked, as a biologist I do appreciate this
> quote... :-}
:-)
--
$jhi++; # http://www.iki.fi/jhi/
# There is this special biologist word we use for 'stable'.
# It is 'dead'. -- Jack Cohen
