Float generators and documentation

[uint]

A couple related issues.

• The documentation section on generators is outdated. It suggests there's a gen-long-float generator, but that one doesn't seem to exist at all. It also shows defgenerator declarations, although I think those have been dropped in favor of simple function definitions.
• The float generator that does exist (gen-float) doesn't let you specify the upper and lower bounds separately, which is limiting if you want to, say, generate only positive floats.

[sionescu]

Good catch, but first I'll need to try to make the docs generator work again or switch to something else.

[Symbolics]

I was just going to suggest that gen-float should add the ability to generate positive floats. I'll just add my vote here for this, and open another issue for 'x' and 'f' suppression for tests. If you're doing a run of 20K tests for numeric equality, you don't want to see xxxxxxfffxxx... 20K times; a X/Y passed will do.