lib/msun: Added fmaximum_mag and fmaximum_num families. Tests and man

[jesuscblazquez_gmail.com]

Summary

Added support for the f{maximum,minimum}_{mag,num} families, the newest standard for maximums in magnitude and number-preferring. This includes modifying fmax.3, on top of D56230, to recommend the use of fmaximum_num and fminimum_num.

Test Plan

Added tests for the new functions on top of fmaximum_fminimum_test.c. Also added a couple new test cases

[fuz]

Nice submission! I'll find some time in the next days to work on this one. Please note that kargl thinks fmaximum and fminimum should never raise FE_INVALID, so some slight rework is needed (we do return (a+b) which triggers FE_INVALID if a or b are signalling nans). He has also asked that the build scripts be in alphabetic order. I have prepared patches for both which I will land in the next days, you can then rebase your feature based on that.

[jesuscblazquez_gmail.com]

In D56236#1286822, @fuz wrote:

Nice submission! I'll find some time in the next days to work on this one. Please note that kargl thinks fmaximum and fminimum should never raise FE_INVALID, so some slight rework is needed (we do return (a+b) which triggers FE_INVALID if a or b are signalling nans). He has also asked that the build scripts be in alphabetic order. I have prepared patches for both which I will land in the next days, you can then rebase your feature based on that.

Great! When I rebase I will also have to fix fmaximum_mag and fminimum_mag to not raise FE_INVALID (I imagine what he said applies to those too). And probably the man pages too. Happy Easter!

[jesuscblazquez_gmail.com]

Rebased on top of the current main (fmaximum and fminimum). Updated _mag to not throw the invalid exception, adapted tests and man pages to the new changes.