Re: Testing/Working on PyUNO?

On Wed, Feb 19, 2014 at 5:55 PM, Bjoern Michaelsen <
bjoern.michaelsen@canonical.com> wrote:

> On Wed, Feb 19, 2014 at 11:37:00AM -0500, Kohei Yoshida wrote:
> > On Wed, 2014-02-19 at 17:27 +0100, Bjoern Michaelsen wrote:
> > > On Wed, Feb 19, 2014 at 11:19:53AM -0500, Kohei Yoshida wrote:
> > > > Yet, it's the core developers who will likely be called in to answer
> > > > "hey my build failed with this python test, tell me what's wrong!!!"
> > >
> > > In subsequentcheck? If that should indeed be the case even then, we
> can make a
> > > separate "make pythoncheck" target.
> > >
> > > > Now you are the one being academic here.  Stick to the topic please,
> > > > which is whether or not Python tests should be used to test core
> > > > functionality.  Nobody is talking about boost here.
> > >
> > > I am talking about testing the _product_ here.
> >
> > And I'm talking about what tools should be used to test what part of the
> > product.  I think you are at one level above us.  No wonder we are
> > talking past each other.
>
> Yes, there is a difference between:
>
>  https://en.wikipedia.org/wiki/Integration_testing
>  https://en.wikipedia.org/wiki/Unit_testing
>
> and both have their use, but are very different beasts. I think, if we
> dont mix
> them up, we can have both, handle them differently and all be happy.
All our higher level tests are already integration tests. Only some low
level classes have real unit tests. There is a third option that is more in
the line of python and java tests and that are API tests. So making sure
that the functionality that we provide through the API is not broken. That
is fundamentally different from testing the core.