Re: Test whether or not destructors appear in any expression

[Chris Warburton]

I've decided to only sample names which appear as theorem dependencies.

This clears up the destructor/destructor-function issue (i.e. whether
they count as "real" TIP expressions, or whether destructors aren't
valid and we were just misinterpreting some sort of strict/sanity check)
and it also clears up whether we should include custom-*, etc.