[SciPy-dev] [scikits] openopt SVN instable for the moment

[dmitrey]

Matthieu Brucher wrote:
>
>
> 2008/4/9, dmitrey <dmitrey.kroshko at scipy.org 
> <mailto:dmitrey.kroshko at scipy.org>>:
>
>     So now nlp_ALGENCAN.py and milp_1.py just don't work, mb some
>     others as
>     well.
>
>
> With a fresh SVN checkout, I only get :
> "OO Error:incorrect solver is called, maybe the solver "ALGENCAN" is 
> not installed. Maybe setting p.debug=1 could specify the matter more 
> precisely"
>
> Which is the correct way. If you don't have a fresh SVN, you will have 
> bogus packages that are remaining in your Kernel folder (BTW, this is 
> not my mistake, I didn't add or remove any folder in Kernel, so the 
> issue/bug must be present for some weeks). Restart from scratch and 
> everything will be fine (well, I didn't test with algencan, but I 
> suppose the traceback you got is the one I got).
>
> When it comes to milp_1.py, I just checked on the SVN repository, and 
> it seems that the error is due to revision 900. 
Yes, I have fixed this.

> So you commit code without checking every example as well.
>
> That's what test are made for. I think you should read Test-Driven 
> Development from Kent Beck as well as Refactoring to Patterns from 
> Martin Fowler. Tests are what you are looking forr to automacally 
> check that the solvers are in a clean state.
Unittests are mentioned in my GSoC application.
http://scipy.org/scipy/scikits/wiki/OOTimeLine

D.
>
> Matthieu
> -- 
> French PhD student
> Website : http://matthieu-brucher.developpez.com/
> Blogs : http://matt.eifelle.com and http://blog.developpez.com/?blog=92
> LinkedIn : http://www.linkedin.com/in/matthieubrucher
> ------------------------------------------------------------------------
>
> _______________________________________________
> Scipy-dev mailing list
> Scipy-dev at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-dev
>