[SciPy-dev] newcore and gcc4/gfortran

[Arnd Baecker]

On Fri, 18 Nov 2005, Travis Oliphant wrote:

> Arnd Baecker wrote:
>
> >On Fri, 18 Nov 2005, Travis Oliphant wrote:
> >
> >[...]
> >
> >
> >
> >>cdfbin_ is a Fortran-compiled subroutine so this looks like a compiler
> >>problem.  It doesn't look like the fortran compiler in gcc is really
> >>ready for primetime.
> >>
> >>At some point it may be possible to test the compiler on these fortran
> >>files in scipy and decide if it will work during configuration.  I'm not
> >>going to start debuggin the gcc fortran compiler at this point, though.
> >>
> >>
> >
> >Which is perfectly understandable!
> >Just for the record - I also tried the gcc4/g77 variant, which gets beyond
> >the above point, but segfaults at another place:
> >
> >check_choose (scipy.special.basic.test_basic.test_choose) ...
> >Program received signal SIGSEGV, Segmentation fault.
> >[Switching to Thread 46912509653888 (LWP 29558)]
> >0x00002aaaabdec67b in PyArray_Choose (ip=<value optimized out>,
> >    op=<value optimized out>) at multiarraymodule.c:1422
> >1422            for(i=0; i<n; i++) Py_XDECREF(mps[i]);
> >(gdb) bt
> >#0  0x00002aaaabdec67b in PyArray_Choose (ip=<value optimized out>,
> >    op=<value optimized out>) at multiarraymodule.c:1422
> >
> >
> Wait...  This is probably a real bug.   I was refactoring code in that
> area last night and may have left the SVN tree in a bad state.   Which
> svn version did you test?

>>> scipy.__core_version__
'0.7.1.1506'
>>> scipy.__scipy_version__
'0.4.2_1443'

Arnd