[TIP] network testing complexity

[Alfredo Deza]

Hi,

I am trying to test as much as I can in a small project of mine. However,
the project has some network usage: pushes files from one spot to another
with Mercurial via SSH.

For now, I am creating a testing environment where I have some SSH keys
enabled and I am able to test this specific functionality.

However, I also have specific tests for all the methods/functions involved.
I had a user complain about the network tests breaking (which is expected if
he doesn't have a proper setup) but I find it very
useful to have not only tests for methods/functions but also for the final
output: did the file actually made it?

Is this approach OK?

I have had situations where a test for a specific method didn't picked up an
error but the 'test_final_ouput.py' did (which made me correct the tests).

Maybe I should not include them when distributing my project? Or would
documentation explaining how to set everything up so tests can run correctly
would be enough?

Thanks!

-- 
Alfredo Deza
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.idyll.org/pipermail/testing-in-python/attachments/20100406/9fea10af/attachment.html>