Figure out why the StringToBool test is so slow

[edsko]

It turns out, the variant function is called more than 3 million times in this test; that explains why #26 had such a dramatic impact. It's however not entirely clear to me why it's called quite that often. Would be good to investigate, as it might reduce the need to solve #25 .

[edsko]

Actually, I think I might have counted recursive calls here too, so this might be a red herring.

[edsko]

Yeah, re-counted. Only about 38k times, which seems somewhat reasonable, given the long strings, characters being used, and the fact that we keep re-executing the entire tests (which is what #25 is about).