problem building Python in MachTen Unix

William C. Brennan brennanw at
Tue May 2 20:34:29 EDT 2000

[[ This message was both posted and mailed: see
   the 'To' and 'Newsgroups' headers for details. ]]

On Sun, 30 Apr 2000 17:48:36 -0400, sue wrote
(in message <390CAA4D.6C554E23 at>):

Sue wrote:
> Does ../Include/mach-o/dyld.h or ../mach-o/dyld.h exist?  If
> not, the include path is wrong.

No, they don't exist anywhere on my machine.  I presume they're a
NeXt/Rhapsody kind of thing, at least that's what the comments
in the configure file led me to believe.

> I have found it's futile to argue with configure.  

Yes, I think I've just learned this lesson. ;-)  Thanks for the

> Edit the makefile directly and save a copy elsewhere.  Get the binary
> built, and then backtrack.  You may need to check config.h.

After examining the makefile and config.h, I took a stab and removed
the definition of "WITH_DYLD" from config.h.  After that, the making
of Python proceeded perfectly.

Well, *almost* perfectly.  I still fail a test in the math module.
My newly build Python reports

    >>> math.modf(-7.1)
    (-0.9, -8.0)

which is wrong.  The correct answer is (-0.1, -7.0).

If you (or anyone else) happens to have any inspiration as to why
my math.modf fuction may be broken, I'm eagerly accepting suggestions.
But python seems basically usable now, and for that I'm happy.

Thanks for your help!

-- bill
Bill Brennan
brennanw at

More information about the Python-list mailing list