corrected log prob for tophat prior

[ekalosak]

No description provided.

[renesass]

@hvarfner Can you approve this change?

[hvarfner]

Sure. Want me to look or just approve?

[renesass]

It's basically one line only so having a quick look might help. Because in your original implementation, you just returned 0.

[hvarfner]

Now, I actually looked. This should not be related to my work, since it is the standard GP over functions and not the PiBO work with priors over the optimum. I still don't mind having a look at it, but I am not the one who implemented this (assuming that there hasn't been a mixup somewhere).

[renesass]

Oh, I'm sorry, I thought the priors came together with PiBO. But @mfeurer did actually implement the origin implementation long time ago.

[mfeurer]

Hey, just had a look. This should be fine as the result is still a constant.

[renesass]

I figured so, it's already part of v2.0.0a2. Thanks @ekalosak for the PR!