support ld80 pseudo-denormal invalid bit patterns; treat them as nan

this is silly, but it makes apps that read binary junk and interpret
it as ld80 "safer", and it gets gnulib to stop replacing printf...

src/math/__fpclassifyl.c

@@ -7,8 +7,11 @@ int __fpclassifyl(long double x)
 {
 	union ldshape u = { x };
 	int e = u.bits.exp;
-	if (!e)
-		return u.bits.m ? FP_SUBNORMAL : FP_ZERO;
+	if (!e) {
+		if (u.bits.m >> 63) return FP_NAN;
+		else if (u.bits.m) return FP_SUBNORMAL;
+		else return FP_ZERO;
+	}
 	if (e == 0x7fff)
 		return u.bits.m & (uint64_t)-1>>1 ? FP_NAN : FP_INFINITE;
 	return u.bits.m & (uint64_t)1<<63 ? FP_NORMAL : FP_NAN;