Fix first reccard test

[cydanil]

This PR aims to close #32, where we fix the tests for the record card.

These are more of integration tests, ranging from GUI action, all the way through the widgets, and checking that the data to be stored matches the inputs.

I apologise if the first commit seems so huge: it's the culmination of all PRs I've done on this project thus far.

It's still a work in progress, but I'm proud to say that the first test passes!

Since it's been a long way, I would recommend not looking at the diff, but rather look at the new content from scratch.

Thanks!

[cydanil]

I'm torn between merging the one test that passes, and adding the others as they get fixed, and merging once all are fixed.

It's certainly going to take a while, as a lot of the interface has changed since the tests were written.
For instance, there's no more Redo action.

[maweki]

Since broken tests are no use to anyone, I am fine with merging early. Might also give other people an opening to fix smaller stuff.

[takluyver]

Should this be in an if __name__ == '__main__': block or something? Putting it like this means the module will run its own tests on import, and then the test runner (nose) will run them again.

[takluyver]

Oh, I get it - the test functions need a reccard instance. I think this block should still be in a test function or method, itself, so that it's not trying to run the tests when it's imported.

[cydanil]

Certainly! But first I'd like the whole shebang to work as intended. However, the structure of the application has changed quite a bit since these tests were implemented, as you noted earlier.

And so I'm currently attempting to make it work, and then I will restructure the tests :)

I think they needn't be interdependent, and each could have their own reccard instance.

Thanks for checking though, I'd have missed that point!

[takluyver]

This appears to be passing on CI now :-)

[cydanil]

After over two months of work leading to it, I'm not sure I can let it go ever again!... :p

I wanted to refactor it more: a good thing to do would be to decouple the tests from one another. But it's becoming too time consuming, when I could focus on making the application more usable.

Therefore, I'll rebase and update once #132 is in, and then leave it so. It isn't perfect, but at least it will provide insights in case of regressions.

[cydanil]

That's it.

[cydanil]

One test has been disabled, as it's becoming too time consuming to keep working on these tests. That test, which tests the sensitivity of the save push button upon undoing, is better suited to be reworked with dogtail.

I still left the code, such that in the future, when someone asks why it's disabled, we can task them with fixing it.

Other than that, the three other tests work and have been verified to be correct!
If there are no further comments, I will merge over the weekend :)