Add benchmarks for dyn-traits

They are roughly half as fast due to the dynamic call overhead.

Author: Byron

benches/usage.rs

@@ -2,6 +2,7 @@ use criterion::*;
 use prodash::{
     messages::MessageLevel,
     tree::{root::Options as TreeOptions, Root as Tree},
+    BoxedDynProgress,
 };
 
 fn usage(c: &mut Criterion) {
@@ -13,15 +14,20 @@ fn usage(c: &mut Criterion) {
         .create()
         .into()
     }
-    c.benchmark_group("Tree::add_child")
-        .throughput(Throughput::Elements(4))
-        .bench_function("add children to build a tree of tasks and clear them (in drop)", |b| {
+    c.benchmark_group("BoxedDynProgress::set")
+        .throughput(Throughput::Elements(5))
+        .bench_function("set tree 5 times", |b| {
             let root = small_tree();
+            let mut progress = root.add_child("the one");
+            progress.init(Some(20), Some("element".into()));
+            let mut progress = BoxedDynProgress::new(progress);
+            use prodash::Progress;
             b.iter(|| {
-                let mut c = root.add_child("1");
-                let _one = c.add_child("1");
-                let _two = c.add_child("2");
-                let _three = c.add_child("3");
+                progress.set(1);
+                progress.set(2);
+                progress.set(3);
+                progress.set(4);
+                progress.set(5);
             });
         });
     c.benchmark_group("tree::Item::set")
@@ -38,6 +44,17 @@ fn usage(c: &mut Criterion) {
                 progress.set(5);
             });
         });
+    c.benchmark_group("Tree::add_child")
+        .throughput(Throughput::Elements(4))
+        .bench_function("add children to build a tree of tasks and clear them (in drop)", |b| {
+            let root = small_tree();
+            b.iter(|| {
+                let mut c = root.add_child("1");
+                let _one = c.add_child("1");
+                let _two = c.add_child("2");
+                let _three = c.add_child("3");
+            });
+        });
     c.benchmark_group("tree::Item::message")
         .throughput(Throughput::Elements(1))
         .bench_function(