Merge #1455: doc: improve secp256k1_fe_set_b32_mod doc

3928b7c doc: improve secp256k1_fe_set_b32_mod doc (Coding Enthusiast)

Pull request description:

  As discussed in #1453
  This only changes the `secp256k1_fe_impl_set_b32_mod` comment since I think `secp256k1_fe_set_b32_limit` doc is clear enough.

ACKs for top commit:
  sipa:
    ACK 3928b7c
  theStack:
    ACK 3928b7c

Tree-SHA512: ad62c1b72d6a487473b182e6aadc7765711385add8c6576bf15c2015db82721f19e3d635f7a29316c2ee7e3c73bc55e2cd4f46ec13180be93d6fe8641f47e7d2

src/field.h

@@ -184,7 +184,8 @@ static int secp256k1_fe_equal(const secp256k1_fe *a, const secp256k1_fe *b);
  */
 static int secp256k1_fe_cmp_var(const secp256k1_fe *a, const secp256k1_fe *b);
 
-/** Set a field element equal to a provided 32-byte big endian value, reducing it.
+/** Set a field element equal to the element represented by a provided 32-byte big endian value
+ * interpreted modulo p.
  *
  * On input, r does not need to be initialized. a must be a pointer to an initialized 32-byte array.
  * On output, r = a (mod p). It will have magnitude 1, and not be normalized.