Hi, Dmitry ! Sorry for the late answer, but I was in business trip to Moscow and have just retuned.
> Ok, > This was in docs -- there is no special treatment of 0.0 and -0.0 :) > (the same about inf and nan). > > Oleg, if you've got cvs write permission -- please submit the code. If > not, please let me know. I have no CVS writing permission. I will test the "lightweight" implementations of some trigonometric functions, and will think about better __cmpsf implementation (after curing -0.0 problem it uses two words in stack :( ). Will report about my results... All the best ! Oleg
