Thanks to those who replied, both on and off the list.
One respondent shared my concerns about the slippery slope; others seem
to want the facility, though one said it should be low priority.
I'm not so much worried about performance - Exim isn't on the whole CPU
limited - though I still don't like the idea of using floating point
instructions unnecessarily. However, you don't necessarily have to use
floating point instructions to deal with real numbers, though it may be
the easiest way.
Having started my serious computing life writing numerical programs in
Fortran, I'm still very cautious with floating point numbers because of
the well-known problems of precision, rounding, and inequality.
I have put this item on the Wish List. I think it is actually quite a
large item because it impacts on many areas - filters, expansion
comparisons, the expansion "eval" operator are just some of them. It
probably needs a fair bit of thought and planning before anything is
implemented.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
