math: support invalid ld80 representations in fpclassify
authorSzabolcs Nagy <nsz@port70.net>
Thu, 5 Sep 2013 18:05:07 +0000 (18:05 +0000)
committerSzabolcs Nagy <nsz@port70.net>
Thu, 5 Sep 2013 18:05:07 +0000 (18:05 +0000)
commitf657fe4b9f734d7fdea515af8dffbf7c28ce4fbc
tree901aaa9611b662fc526ae766796dc2d6b04aa636
parentf4d9bfb3a986f2c25ec023426b4fb222130af75f
math: support invalid ld80 representations in fpclassify

apparently gnulib requires invalid long double representations
to be handled correctly in printf so we classify them according
to how the fpu treats them: bad inf is nan, bad nan is nan,
bad normal is nan and bad subnormal/zero is minimal normal
src/math/__fpclassifyl.c