Re: [boost] [test] investigating Borland test_fp_comparisons fail ure

Martin Bonner wrote:

The tolerance is a percentage tolerance, rather than an absolute value.

A percentage tolerance makes sense if you regard 1E-10 as far from 1 as 1 is from 1E10. If you do, then 0 is infinitely far from any other finite number. If you don't, then I suggest you are using floating point as an easy substitute for fixed-point arithmetic (which is what I do).

Thanks, that explains my mis-understanding then. Guess I should do a better job reading the friendly manual. Now I understand the issue, I can take another look at why Borland has a regression on this test in 1.33. AlisdairM