Add test for forward-sexp-function

[cruegge]

Added a basic test for #1148. Turned out simpler than I thought ;)

[gracjan]

Nice! Can I have a wish or two:

1. Use :expected-result :failed like in other failing test cases.
2. Add two more, successful cases.

When you are done force-push it to the same branch, github will pick up the change.

Thanks.

[cruegge]

While thinking of useful successful test cases, I stumbled upon another weird behaviour, which occurs when calling haskell-forward-sexp with a prefix arg larger than 1.

• When point is at an opening paren, the function delegates to scan-sexp, which does not know about Haskell tokens. So C-u 4 C-M-f may lead to different results than 4 times C-M-f.
• When point is not on an opening paren, arg is ignored completely.

I wrote an additional successful test case for the original problem, and for each of the new ones a failing and a successful one (I'm going to push them in a moment). I also have a possible fix for all cases; if you like, I can open a new PR, or push it here as well.

[gracjan]

You can add the fix here.

[gracjan]

@cruegge, we need the backward case implemented, too. It is tracked in #759 where there is also a description of an idea how to get it done. Can you get a go at it?

[cruegge]

I see what I can do.