XFAIL a test following the +nan fix.

This test is expecting "nan", while the fix enh made to libc makes it
now be "+nan". I have a feeling this is really a bug with either the
test or Darwin/the BSDs, since we found this was actually a bug in the
BSD libc, but just XFAIL it for us for now until I diagnose properly.

Change-Id: Icf5c7f7a0fdd9a8696842201bee8464a11e7c6f1

test/localization/locale.categories/category.numeric/locale.nm.put/facet.num.put.members/put_long_double.pass.cpp

@@ -15,6 +15,7 @@
 
 // TODO GLIBC uses a different string for positive and negative NAN numbers.
 // XFAIL: linux-gnu
+// XFAIL: android
 
 #include <locale>
 #include <ios>