<p dir="ltr"></p>
<p dir="ltr">On Sep 19, 2016 9:52 PM, "Mads Kiilerich" <<a href="mailto:mads@kiilerich.com">mads@kiilerich.com</a>> wrote:<br>
><br>
> On 09/19/2016 09:34 PM, Thomas De Schampheleire wrote:<br>
>><br>
>> On Sep 19, 2016 9:30 PM, "Mads Kiilerich" <<a href="mailto:mads@kiilerich.com">mads@kiilerich.com</a>> wrote:<br>
>><br>
>> >> +from kallithea.tests.base import *<br>
>> ><br>
>> ><br>
>> > Would you have any objections to dropping this and just use kallithea.tests.base where it is needed?<br>
>> ><br>
>> > Import * in general is one thing, but having it in a module definition can make it even harder to figure out what is going on.<br>
>><br>
>> Does that work with the override of __all__? That is the real reason to import this way. If there is a better way, yes please.<br>
><br>
><br>
> We only need __all__ in a module when we do 'import *' from it. And as I propose it, there would be no 'import *' from __init__.py but from base.py.<br>
><br>
> (Also __all__ in base.py will unambiguously be everything, and there is thus not need for it. Unless we like explicit.)<br>
></p>
<p dir="ltr">In my reasoning I envisioned not touching the test files, which would require __all__. In the pushed version, you did effectively fix the imports from tests to tests.base, so I agree __all__ is no longer needed. Thanks for taking care of it. I am now rebasing the tg branch on this and other recent changes..</p>
<p dir="ltr">/Thomas</p>