tree 58b2fbcd87f2c1e581315f83a1bae982689dc267
parent 552a4e5170cbef752a3cfc4a101c298399adbf38
author Sergiu Deitsch <sergiu.deitsch@gmail.com> 1637278599 +0100
committer Sameer Agarwal <sameeragarwal@google.com> 1637291636 +0000

fix fmin and fmax NaN handling

fmin and fmax do not handle NaNs correctly. Also, the comparison
operator for floating-point numbers may raise FE_INVALID if one of the
arguments is NaN. Both functions, however, are not subject to any of the
error conditions specified by the error handling for related
floating-point operators and functions.

Change-Id: Ic6bb65f18568066dba3c739a2df06f5fc3131a80