On Sep 16, 2009, at 11:41 PM, Stefan Behnel wrote:

> Robert Bradshaw wrote:
>> On Sep 16, 2009, at 4:55 PM, Magnus Lie Hetland wrote:
>>> Now, I'd put the code in a package -- except I'd left out the
>>> __init__.py for now (which I'd gotten some warnings about, but I
>>> wasn't planning on doing any importing yet, so I thought it wouldn't
>>> matter; I was using a glob in my setup.py and didn't want to compile
>>> the init file). But as the module name was of the form foo.foo, I
>>> suspected this might be the problem ... and indeed it was.
>>>
>>> So, I guess, if I'd done it "by the book", and included the
>>> __init__.py file, it would have worked from the start. However  
>>> I'd say
>>> the compile failed in a rather non-obvious way... (I.e., it actually
>>> did compile -- it just ignored the .pxd file...)
>>
>> Glad you were able to figure it out. I'm not sure how we should
>> detect this kind of error...
>
> It's impossible to detect the case where an existing .pxd file is  
> not found
> (since there might not actually be one), but we could detect the  
> case where
> the module name does not reflect the package structure, i.e.  
> exactly the
> case where the module code is expected to be inside of a package,  
> but does
> not lie next to an __init__.py file.

We should probably be at least emitting a warning in this case.

- Robert


_______________________________________________
Cython-dev mailing list
[email protected]
http://codespeak.net/mailman/listinfo/cython-dev

Reply via email to