[SciPy-user] (no subject)

Travis Oliphant oliphant.travis at ieee.org
Fri Feb 24 05:50:04 CST 2006


Adrián López wrote:

> Very strange ... If I change just that single line as you said, it 
> runs correctly !!! THANKS!!! How did you know the problem was on that 
> line?

I'm not sure what the issue is exactly but it might be a bug in Python 
itself (though I don't tend to jump to that conclusion quickly).  I 
figured that line might be a problem, because I kept getting errors (but 
not segfaults) from the code at that line.

The errors made no sense to me, but I wondered if on your system, 
whatever was causing the problem I was seeing was showing up as a segfault.

It's an interesting test-case, though.   I'll look into it some more.  I 
don't like to leave things like this floating because it may be pointing 
to a problem I'm overlooking...

It's hard to say why the numpy version prints an extra line.  It 
probably has something to do with rounding modes and the fact that in 
numpy x[3] is not a Python float but an array-scalar float that (for 
now) uses array-math instead of scalar math.  The scalar math is coming.

-Travis



More information about the SciPy-user mailing list