[SciPy-Dev] SciPy 0.12.0/Numpy 1.7.1 sgerqf segfault

Pauli Virtanen pav@iki...
Mon Apr 15 16:32:38 CDT 2013


16.04.2013 00:13, Orion Poplawski kirjoitti:
[clip]
>> (scipy.linalg doesn't have it's own python_xerbla.c, so it ends up
>> looking for the symbol in numpy.linalg.lapack_lite --- rather unexpected!)
> 
> So, any suggestions for this?  Should flapack.pyf.src not mark them as thread 
> safe?  Should scipy provide its own xerbla?

Yes, to the latter.

The actual fix to this bug is with high probability a replacement from

integer optional,intent(in),depend(n),check(lwork>=m||lwork==-1) ::
lwork=3*m

to

integer optional,intent(in),depend(m),check(lwork>=m||lwork==-1) ::
lwork=3*m

on line 653 of scipy/linalg/flapack.pyf.src. This bug doesn't manifest
always as it depends on ordering of dict entries inside f2py...

[clip]
> Not sure where to look next.  Time to move to numpy list?

It's a Scipy bug, so this list should be fine.

-- 
Pauli Virtanen



More information about the SciPy-Dev mailing list