Python-2.2.1, Solaris7, make test fails...
Hugh Sasse Staff Elec Eng
hgs at dmu.ac.uk
Thu Apr 18 08:17:50 EDT 2002
On 18 Apr 2002, Martin v. [iso-8859-1] Löwis wrote:
> Hugh Sasse Staff Elec Eng <hgs at dmu.ac.uk> writes:
>
> > So:
> >
> > #
> > # In addition, *static* reverses this effect (negating a previous
> > # *shared* line).
> >
> > is ambiguous, because it does not reverse all the effects of *shared*?
>
> Maybe. This is meant in the context of the Setup file: all modules
> listed after shared are compiled statically. If no modules are listed,
> this has no effect.
>
> If you can propose a rewording, please submit a patch to
> sf.net/projects/python.
I can't yet: I don't know which properties are reversed:
* Modules included in the config.c file
* List of objects to be added to library archive
* linker options added to linker options
* Rules added to Makefile
* Names collected in SHAREDMODS variable.
>
> > I have tried to configure this in different ways:
> > configure
> > configure -- with-gnu-ld
>
> Python's configure does not support --with-gnu-ld, so this option has
> no effect.
>
> > What else does it look for?
>
> Use 'gcc --print-prog-name=ld' to find out which ld binary gcc uses.
/usr/local/sparc-sun-solaris2.7/bin/ld
/usr/local/sparc-sun-solaris2.7/bin/ld --version gives
GNU ld version 2.12
Copyright 2002 Free Software Foundation, Inc.
>
> > If I symling /usr/local/bin/ld to gld will configure find it?
>
> No. If you create an ld link in
Oh, ok, I thought configure might look for gld first.
> /usr/local/lib/gcc-lib/<system>/<version>; that will make gcc use a
> different ld binary.
>
> *However*, there is no need to do that; the system linker should work
> as well. Please invoke the python linker line line manually, adding
> -v:
I think I will delay doing that for now, because most of the tests
worked without using GCC. Another post suggests that this is not common
to Solaris7, (referring to pre-releases of Python-2.2.1). Therefore I
may not really be helping anyone else by exploring that.
>
> Regards,
> Martin
> --
> http://mail.python.org/mailman/listinfo/python-list
>
Thank you again,
Hugh
More information about the Python-list
mailing list