[SciPy-user] Re: [SciPy-dev] Scheduled release

[Pearu Peterson]

On Thu, 25 Sep 2003, Travis Oliphant wrote:

> I would like feedback on a change I made recently to CVS.   I noticed 
> that LAPACK functions were locking up the interpreter when nans (and/or 
> infs) where in the array sent to the routine.
> 
> Has this behavior been observed before or is it due to use of ATLAS, 
> recently compilation flags etc...
> 
> I've "solved" the problem by elminating all nans before calling the 
> ATLAS routines.
> 
> If anyone has other suggestions for this problem, let me know.

As I said in my previous mail, I could not reproduce the problem.
However, now that running scipy.special tests I noticed a "hang". During 
this "hang" python's CPU was up during test_chebyc indicating that some 
calculations where in progress and the "hang" might be caused by some 
"infinite" loop or whatever. Rebuilding special magically fixed this, now 
I cannot reproduce hangs anymore (I am still not using nan_to_num).

So, I would still like to see examples of hangs from LAPACK functions as,
IMHO, nan_to_num is not a proper solution. The whole point of supporting 
NaN/Infs in scipy is lost by this as well as the acctual cause for the 
hangs might still be there to show up in some other weird situation. IMHO.

Pearu