[PATCH 0/4] math: Fix inaccurate sin/cos/tan for large arguments (BZ 34376)

[Paul Zimmermann]

       Hi Adhemerval,

in addition to Carlos' review, I tested all three functions with all
21892 corner cases I produced, and the largest found error is the same
as before BZ 34376.

If this patch does not go into 2.44, we will have to report huge errors
for trigonometric functions in 2.44 in our new update (which will follow
the 2.44 release).

Paul