I reworked the patches for the Bessel et.al. math functions into an extension.
Direct link:  http://users.rcn.com/m3ha11/ruby/mathext-0.1.tar.gz
Tested on 1.8.0preview6 on OS X last night.

I don't know why this bugs me so much.  I guess it's just a matter of:
if ruby has access to _some_ functions in 'libm', then it should have
access to all of them.   That doesn't really make sense, I know.

I agree with Josef that if we were to rewrite our own Bessels (etc.),
we'd want someone who really knows what they are doing.
I re-read 'What Every Computer Scientist Should Know About Floating Point"
last night -- very scarey stuff!  :-)