Apply effective-affine trick to precomp

src/ecmult_gen_impl.h

@@ -40,7 +40,7 @@ static void secp256k1_ecmult_gen_context_build(secp256k1_ecmult_gen_context_t *c
         VERIFY_CHECK(secp256k1_ge_set_xo_var(&nums_ge, &nums_x, 0));
         secp256k1_gej_set_ge(&nums_gej, &nums_ge);
         /* Add G to make the bits in x uniformly distributed. */
-        secp256k1_gej_add_ge_var(&nums_gej, &nums_gej, &secp256k1_ge_const_g);
+        secp256k1_gej_add_ge_var(&nums_gej, &nums_gej, &secp256k1_ge_const_g, NULL);
     }
 
     /* compute prec. */