64-bit python build

Christos TZOTZIOY Georgiou DLNXPEGFQVEB at spammotel.com
Wed Aug 28 17:37:59 CEST 2002


On Tue, 27 Aug 2002 15:51:19 -0400, rumours say that Tim Peters
<tim.one at comcast.net> might have written:

>[Christos TZOTZIOY]
>I wouldn't bother.  The most important thing to note on SGI boxes is that
>there's a history of (C) compiler optimization bugs nearly as long as
>Python's history.  If you're using an SGI compiler, and don't want to spend
>a lot of time chasing compiler bugs, you're best off disabling optimization
>entirely when building Python.

I agree to that from my own experience with other old projects too.
Just a note for other people building python themselves on SGI boxes:
MIPSPro 7.3 seems to compile python fine with -O3 enabled.  Floating
point accuracy doesn't seem to be influenced, but there is always the
-OPT:roundoff=0:IEEE_arithmetic=0 option.

>> but I got a Bus Error in test_struct ("trying:  xcbhilfd").

>That smells like a compiler optimization bug; the SGI compiler seems prone
>to lifting reads above tests guarding the legality of reads.

Actually, there were many assumptions in structmodule.c about char
pointers being suitably aligned for ints and longs; MIPS 64-bit CPUs are
not very forgiving.  See at end for patch info.

>Sure!  Open source is driven by volunteers scratching their own itches -- if
>you don't, who will?

See SF patch 601369 :)  Hopefully this can also be used for IA64 et al.
-- 
TZOTZIOY, I speak England very best,
Real email address: 'dHpvdEBzaWwtdGVjLmdy\n'.decode('base64')



More information about the Python-list mailing list