[SciPy-dev] scipy_distutils/*_info.py files

eric eric at scipy.org
Mon Feb 18 11:04:51 CST 2002


> > It is not a very important matter but may be somebody (Eric ?) had a
> > vision on this (unified specification,etc) that we should try to follow.
> 
> Why not have a generic SystemInfo class which all instantiate?
> Something like (bare  pseudocode):
> 
> class SystemInfo:
>       def __init__(self,what_I_need_to_find):
>       ...
> 
>       def get_info(self):
>           ....
>           return dict_with_info
> 
> 
> Then you could do:
> 
> x11_info = SystemInfo(lib = 'X11',...)
> atlas_info = SystemInfo(lib = 'atlas',...)

The idea of a class is good.  I think the specific cases are special enough that they should be sub-classes instead of variables handed in so that:

class system_info:
    ...
    def get_info(self):
        ...
    def search_local(self):
       """ Search system for directories. """
    def download_from_remote(self):
       """ Some provision for grabbing libraries from remote locations
           like scipy.org if the person wants this to happen. 
       """

class atlas_info(system_info):
   """ Special cased to handle atlas specific stuff
   """

> 
> and so on. I know this is the barest, roughest of sketches. But hopefully my
> point is clear. The advantage of this would be to have a unified interface
> both for specifying what one is looking for (files, directories, libraries,
> versions, etc.) and for the format of the output.
> 
> I think it would be worth doing it that way as knowing what's installed in a
> system is a fairly common and generic problem.

Agreed.  As long as we don't get in the business of trying to write autoconf. :-)

eric

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://projects.scipy.org/pipermail/scipy-dev/attachments/20020218/fc7e87c8/attachment.html 


More information about the Scipy-dev mailing list