"Patrick Doyle" <[EMAIL PROTECTED]> wrote:
> Why does Python include the submodules that were explicitly loaded by
> previous imports? Does it go out of it's way to do so? If so, why?
> What purpose does it serve? Or, is it a natural fallout of the manner
> in which imports are processed? If so, could somebody guide my
> intuition as to why this would be such a natural fallout?
>
It is a natural fallout of the manner in which imports are processed.
When you explicitly load a submodule a reference to the submodule is
stored in the parent module. When you do 'from module import *' and the
imported module doesn't define __all__, you create a reference in the
current module to everything referenced by the imported module (except
for variables beginning with '_').
Python makes no distinction at all between the objects named in the
imported module: they all get imported, i.e. assigned into the current
scope no matter how they got there.
If you use your example:
>>> import SDRGen
>>> dir(SDRGen)
['__builtins__', '__doc__', '__file__', '__name__', '__path__']
>>> from SDRGen.TFGenerator import *
>>> dir(SDRGen)
['TFGenerator', '__builtins__', '__doc__', '__file__', '__name__',
'__path__']
>>>
The line 'from SDRGen.TFGenerator import *' always creates the name
TFGenerator in the SDRGen module (and it will create the SDRGen module
if it has to). A subsequent 'from SDRGen import *' is simply doing the
equivalent of:
for name in dir(SDRGen):
if not name.startswith('_'):
setattr(current_module, getattr(SDRGen, name))
except of course the variables name and current_module don't exist.
--
http://mail.python.org/mailman/listinfo/python-list
