I'd first like to give maximum praise to the gcc-avr project!
I had to use the atan2() func recently and I don't if it's
me or does the atan2(a, b) func supposed to give the
arc tangent of (a/b)?
Because it looks like it's giving the arc tangent of (b/a)
which is contrary to the specs(and IEEE?).
I don't want to patch the assembler source myself,
so could someone with more experience look at
PS: Maybe I have this all backward myself ;)