GitLab

I've added some tests too.

This is not an optimal solution. I'm working on it.

Moreover, the @note message regarding locking the DUChain will only be written
if the DUChain has to be locked before calling the method.

Note that I've marked the given test as QSKIP. This is because for some reason
I can see this working manually but not in the test :(

We don't have to update the builtins every single time before executing a test
because the DUChain already does this for us. This means that tests will run a
*lot* faster now (as they did before the PST changes :P).

Otherwise, there will be "funny" crashes.

Moreover, I've removed some dangling connect statements.

It makes no sense anymore because we already load modules & classes from other
files.

Moreover, now the tests are passing.

Apparently the standard does not enforce NULL to be defined as nullptr. I
think it's safer to write nullptr and it's supported since GCC 4.6 and Clang
3.0.

Thanks Sergei Ovchinnikov for pointing this out.