[SciPy-user] OpenOpt svn broken on Windows

dmitrey dmitrey.kroshko@scipy....
Thu Jul 17 01:24:08 CDT 2008

Matthieu Brucher wrote:
> 2008/7/16 dmitrey <dmitrey.kroshko@scipy.org>:
>> Matthieu Brucher wrote:
>>>> __solverPaths__ is exported by runProbSolver.py, and than converted from None to
>>>> dict. I intended to init it in runProbSolver.py but some issues had
>>>> appeared (Python reported of bugs) so I had moved it to ooMisc.py
>>> It has not been moved to ooMisc. The actual registry is only available
>>> in runProbSolver. Because you are initializing it in runProbSolver and
>>> not in ooMisc.
>> What is not been moved to ooMisc?! I have said I encountered bugs while
>> creating __solverPaths__ in runProbSolver, so now it's created in ooMisc
>> (as None). Of course, then I redefine it as Python dict (in runProbSolver).
> Try and get the type of ooMisc.__solverPaths__. It will always be
> None. Why ? Because you import the variable None as __solverPaths__ in
> rPS. Then you modify the reference, but not the underlying variable.
Do you want to inform me something is wrong with my code?
I have no needs to import ooMisc.__solverPaths__ each time. It is 
imported only once per Python session, no matter how many solvers were 
called, as well as it is redefined in rPS.py to dictionary just a single 
time per session.

>But modifying an external package (algecan) is not the same as enhancing your own package. The first is not your primary job. The latter is.

I just said " it's not my job to get ALGENCAN workaround." I have to do 
my GSoC milestones from OOTimeLine and no more, so it's is up to me to 
decide to do any additional job or not to do, aa well as  choosing which 


More information about the SciPy-user mailing list