Unclear assertion errors due to floating point numbers

[Goju-Ryu]

I recently encountered a an error message that asserted that two identically printed values were not the same. I know this is due to floating point numbers being imperfect representations, but from a user perspective I thought we might want to point it out to the user that this is probably what went wrong.

I would suggest that we check for the two printed values, and if they are the exact same, we point out that floating point imprecision might be the cause. This will especially be helpful for less technical people so they don't assume it's a bug in numbat.

See an example here

[sharkdp]

Great idea, thank you