Change the API of eval to take an array instead

There is really no reason for taking an iterator, just makes things worse for the compiler

ipa-core/src/protocol/ipa_prf/malicious_security/lagrange.rs

@@ -101,14 +101,8 @@ where
     /// that were used to generate this table.
     /// It is assumed that the `y_coordinates` provided to this function correspond the values of the _input_ "x coordinates"
     /// that were used to generate this table.
-    pub fn eval<I>(&self, y_coordinates: I) -> [F; M]
-    where
-        I: IntoIterator + Copy,
-        I::IntoIter: ExactSizeIterator,
-        I::Item: Borrow<F>,
+    pub fn eval(&self, y_coordinates: &[F; N]) -> [F; M]
     {
-        debug_assert_eq!(y_coordinates.into_iter().len(), N);
-
         self.table
             .iter()
             .map(|table_row| {

ipa-core/src/protocol/ipa_prf/malicious_security/verifier.rs

@@ -159,7 +159,7 @@ where
     last_array[1..last_u_or_v_values.len()].copy_from_slice(&last_u_or_v_values[1..]);
 
     // compute and output p_or_q
-    tables.last().unwrap().eval(last_array)[0]
+    tables.last().unwrap().eval(&last_array)[0]
 }
 
 #[cfg(all(test, unit_test))]