% ./configure --without-threads --with-pymalloc
% gmake
...

./python setup.py
...
Then it dumps core. Removing the  --with-pymalloc and
it runs just fine.
[281] % uname -a
HP-UX wssgped B.10.20 A 9000/782 2001125167 two-user
license
[282] % gcc --version
2.95.1

Logged In: YES 
user_id=6380

Alas, I can't reproduce this on Linux.

I'm afraid building Python on HP-UX is braindead, and I
don't know wht the problem is.  No two HP-UX users ever seem
to run in the same problem, nor do I ever seem to get help
from folks with HP-UX machines in debugging those
problems...

So, while I am sympathetic, I can't help.  Somebody else
please help!

Logged In: NO 

[213] % gcc --version
2.95.1
[214] % uname -a
HP-UX wvenus B.10.20 A 9000/782 2009382447 two-user license


If I build Objects/obmalloc.w WITHOUT the -O2 option then
./python works.
If I build Objects/obmalloc.w WITH the -O1 option then
./python works.

If I build Objects/obmalloc.w WITH the -O2 (the default)
option then ./python
fails as seen below. Using -O3 and -O4 also causes ./python
to crash.


[212] % gdb ./python ./setup.py build
Excess command line arguments ignored. (build)
GNU gdb 5.0
Copyright 2000 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public
License, and you are
welcome to change it and/or distribute copies of it under
certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show
warranty" for details.
This GDB was configured as "hppa2.0-hp-hpux10.20"...
"/tmp_mnt/net/roti/wssg/hpux10x/usrlocal/packages/python2.1/Python-2.1b1/./setup.py"
is not a core dump: File format not recognized
(gdb) run
Starting program:
/tmp_mnt/net/roti/wssg/hpux10x/usrlocal/packages/python2.1/Python-2.1b1/./python 
warning: Unable to find __d_pid symbol in object file.
warning: Suggest linking with /opt/langtools/lib/end.o.
warning: GDB will be unable to track shl_load/shl_unload
calls

Program received signal SIGBUS, Bus error.
_PyCore_ObjectMalloc (nbytes=28) at Objects/obmalloc.c:417
417                             if ((pool->freeblock =
*(block **)bp) != NULL) {

Logged In: NO 

With an updated version of gcc it (python2.1b1) still fails
with -O2 but in a different place.

% gcc --version
2.95.3
% uname -a
HP-UX wvenus B.10.20 A 9000/782 2009382447 two-user license
% gdb
GNU gdb 5.0
(gdb) run
Starting program:
/tmp_mnt/net/roti/wssg/hpux10x/usrlocal/packages/python2.1/Python-2.1b1/./python 
warning: Unable to find __d_pid symbol in object file.
warning: Suggest linking with /opt/langtools/lib/end.o.
warning: GDB will be unable to track shl_load/shl_unload
calls

Program received signal SIGSEGV, Segmentation fault.
0x4d984 in PyNumber_CoerceEx (pv=0x7b03c4b4, pw=0x7b03c4b0)
    at Objects/object.c:1175
1175            if (v->ob_type->tp_as_number &&
v->ob_type->tp_as_number->nb_coerce) {


If I remove the -O2 when compiling Objects/object.o then
python builds without a crash.
% gcc -c -g -Wall -Wstrict-prototypes -I. -I./Include
-DHAVE_CONFIG_H  -o Objects/object.o Objects/object.c

Logged In: NO 

Downloaded python2.1beta2. Used gcc 2.95.3 on HPUX 10.2

This time it blew up in different manner:
PYTHONPATH= ./python ./setup.py build
Fatal Python error: unknown scope for name in __init__(0) in
./Lib/site.py
symbols: {'data': 12, 'self': 12, 'files': 12, 'name': 12,
'None': 264, 'dirs': 12}
locals: {'self': 0}
globals: {'None': 1}

/bin/sh: 22414 Abort
gmake: *** [sharedmods] Error 134

Removed Objects/object.o and rebuilt it without the -O2
option and then the system fully built with no crashes.

Logged In: YES 
user_id=6380

I'm closing this -- it seems to be a compiler problem.

The moral of the story: turn off -O2 when you're
experiencing weird crashes.