Zero-alloc trie lookups (#3998)

* triedb cleanup

* factor out common portion of trie query

* allocate far fewer times in node decoding

* fix bench compilation

* introduce OwnedNode variant to make iter fast again

* generalize recorder trait to Query

* decode trie outputs cost-free in state

* test for passing closure as query

util/benches/bigint.rs

@@ -14,7 +14,7 @@
 // You should have received a copy of the GNU General Public License
 // along with Parity.  If not, see <http://www.gnu.org/licenses/>.
 
-//! benchmarking for rlp
+//! benchmarking for bigint
 //! should be started with:
 //! ```bash
 //! multirust run nightly cargo bench
@@ -24,10 +24,10 @@
 #![feature(asm)]
 
 extern crate test;
-extern crate ethcore_bigint as bigint;
+extern crate ethcore_util;
 
 use test::{Bencher, black_box};
-use bigint::uint::{U256, U512, Uint, U128};
+use ethcore_util::{U256, U512, Uint, U128};
 
 #[bench]
 fn u256_add(b: &mut Bencher) {

util/benches/rlp.rs

@@ -24,12 +24,12 @@
 
 extern crate test;
 extern crate rlp;
-extern crate ethcore_bigint as bigint;
+extern crate ethcore_util as util;
 
 use test::Bencher;
 use std::str::FromStr;
 use rlp::*;
-use bigint::uint::U256;
+use util::U256;
 
 #[bench]
 fn bench_stream_u64_value(b: &mut Bencher) {