Eval should never return an error

Only awaiting eval should error.
Eval should always be available, even if using its methods returns errors.

packages/desktop/headless_tests/events.rs

@@ -45,7 +45,7 @@ fn mock_event(id: &'static str, value: &'static str) {
                 value, id
             );
 
-            eval(&js).unwrap().await.unwrap();
+            eval(&js).await.unwrap();
         });
     })
 }

packages/desktop/headless_tests/rendering.rs

@@ -38,7 +38,6 @@ fn use_inner_html(id: &'static str) -> Option<String> {
                     return element.innerHTML"#,
                 id
             ))
-            .unwrap()
             .await
             .unwrap();
 

packages/desktop/src/eval.rs

@@ -15,8 +15,8 @@ impl DesktopEvalProvider {
 }
 
 impl EvalProvider for DesktopEvalProvider {
-    fn new_evaluator(&self, js: String) -> Result<GenerationalBox<Box<dyn Evaluator>>, EvalError> {
-        Ok(DesktopEvaluator::create(self.desktop_ctx.clone(), js))
+    fn new_evaluator(&self, js: String) -> GenerationalBox<Box<dyn Evaluator>> {
+        DesktopEvaluator::create(self.desktop_ctx.clone(), js)
     }
 }