> I don't have IEEE754 and ANSI floating point issue, so I don't know > whether ANSI defines this behavior or not. If you have IEEE754 or any > related standards, and if such standards say pow(NaN, 0) have to > return NaN, reopen this bug.
C99 draft standard (last freely available version, see http://std.dkuug.dk/jtc1/sc22/open/n2794/n2794.pdf): F.9.4.4(1): pow(x, �0) returns 1 for any x, even a NaN. So it seems glibc is correct.
signature.asc
Description: This is a digitally signed message part
