src/s2/util/math/exactfloat/exactfloat.cc | 16 16 + 0 - 0 !
1 file changed, 16 insertions(+)

 fix bn_ext_get_uint64 for openssl >= 1.1 32bit
 Suggested changes fix compilation error with OpenSSL >= 1.0 for 32bit target.
 The algorithm is similar to what is currently being done in BN_ext_set_uint64,
 but in the opposite direction.

src/Makevars.in | 2 1 + 1 - 0 !
1 file changed, 1 insertion(+), 1 deletion(-)

 manually pass in -latomic on armel to fix ftbfs

R/zzz.R | 2 2 + 0 - 0 !
1 file changed, 2 insertions(+)

 warn that this may be buggy on 32-bit

(onLoad not onAttach as some packages use this via requireNamespace)