LG> Python 1.2 (Apr 18 1995) [GCC 2.6.3]
I'm wondering if it could be a gcc library mismatch problem? I built
a quad-fat of python 1.2 on my NS3.3/Sparc with Developer 3.3. I let
configure use the built-in Dev3.3 cc compiler, instead of trying to
use gcc (my general SOP for NEXTSTEP). Anyway, Guido and I tried the
os.listdir test on both Sparc and Intel and it Did The Right Thing. I
didn't try compiling with gcc though.
LG> I think we should eventually focus on correcting Python (and
LG> in my opinion it's not broken) rather than relying on the
LG> supposed POSIX compatibility library available on NS. Maybe NS
LG> Developer 3.3 is better on thin than 3.2, and eventually 4.0
I guess it could also be a Dev3.3/3.2 mismatch problem too. I only
ever had 3.3 for Sparc.
LG> Another point: configure.in uses "uname" to build the name of
LG> the subdirectory of library that contains system specific
LG> code. This is bad, since "uname" is not available
LG> everywhere... This is the result on NeXTSTEP:
Saw the same thing when I built the quad-fat, but I just ignored it
and all seemed to work okay.
-Barry