[SciPy-Dev] request review of some test cleaning

[Ralf Gommers]

2010/5/28 Stéfan van der Walt <stefan at sun.ac.za>

> On 27 May 2010 09:52, Warren Weckesser <warren.weckesser at enthought.com>
> wrote:
> > On my "to do" list for scipy.linalg is the issue of the 'disp' argument
> > and return values of funm, logm, signm and sqrtm.  It would be nice to
> > just get rid of the 'disp' argument.  But that issue will have to wait
> > until 0.9.
>
> I also picked that up while reviewing.  If we leave 'disp' in place
> and raise a warning on error (instead of printing), we're not
> technically breaking the API.  We can also raise a pending deprecation
> warning whenever `disp` is used.
>
> You can also do nothing instead of printing. disp is like full_output, if
you care about the accuracy you should just grab the error as a return value
- printing or warning with an arbitrary criterion is just annoying.

Warren, if you have no time for this now I'll just merge the tests like they
are now on Sunday. Otherwise I'll adapt it.

Cheers,
Ralf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scipy-dev/attachments/20100528/133885bf/attachment.html>