[avr-libc-dev] new atan() and atan2() functions

[Joerg Wunsch]

Hello all,

Reiner Patommel provided completely rewritten implementations for
atan() and atan2() which i've just checked in.  Accuracy is a lot
better in particular for large arguments (where the result becomes
close to Pi/2), but even for smaller arguments, the previous
implementation often experienced errors in the range of 1E-5.

Together with the already rewritten asin() function, the accuracy of
our trigonometrical functions should now be close to what can be
expected from a 32-bit IEEE implementation.
-- 
J"org Wunsch                                           Unix support engineer
address@hidden        http://www.interface-systems.de/~j/