Use Action method selector to differentiate between Ajax and non-ajax request instead of relying on if(Request.isAjaxRequest)? - ajax

Question

I'm following a book called 'Asp.Net MVC4 in Action'. And now at certain point they say, Instead of relying on if statement within our code to check if the request is Ajax or not, we could use an action method selector to differentiate it. And what they have done is create a class AcceptAjaxAttribute with following code
using System;
using System.Reflection;
using System.Web.Mvc;
namespace CustomAjax
{
[AttributeUsage(AttributeTargets.Class | AttributeTargets.Method)]
public class AcceptAjaxAttribute : ActionMethodSelectorAttribute
{
public override bool IsValidForRequest(ControllerContext controllerContext, MethodInfo methodInfo)
{
return controllerContext.HttpContext.Request.IsAjaxRequest();
}
}
}
and the function in controller which looked like this before
var speaker = _repository.FindSpeaker(id);
if(Request.IsAjaxRequest())
{
return Json(speaker, JsonRequestBehaviour.AllowGet);
}
return View();
has changed to something like this
[AcceptAjax]
public ActionResult Details(int id)
{
var speaker = _repository.FindId(id);
return Json(speaker, JsonRequestBehavior.AllowGet);
}
[ActionName("Details")]
public ActionResult Details_NonAjax(int id)
{
var speaker = _repository.FindId(id);
return View();
}
To be honest I have no idea what is being done or why we created new class and used that[AcceptAjax] thingy. Can someone may be explain it to me.

Answer 1

Before you had one action with an if inside, after the refactoring you have 2 actions each returning a different type of result. The ActionMethodSelectorAttribute is used to select the proper action based on some condition. Since the 2 actions have the same name (Details), ASP.NET MVC will use this attribute to select the one or the other based on whether an AJAX request is being used.
But honestly I don't like this approach at all. You now have 2 actions and have repeated the var speaker = _repository.FindId(id); call twice which is not very DRY. Wouldn't it be better if you had this?
[AcceptAjax]
public ActionResult Details(int id)
{
var speaker = _repository.FindId(id);
return View(speaker);
}
If you are like me and think that this is better, then simply replace this AcceptAjaxAttribute you got from the book with an action filter:
[AttributeUsage(AttributeTargets.Class | AttributeTargets.Method)]
public class AcceptAjaxAttribute : ActionFilterAttribute
{
public override void OnActionExecuted(ActionExecutedContext filterContext)
{
if (filterContext.RequestContext.HttpContext.Request.IsAjaxRequest())
{
var result = filterContext.Result as ViewResultBase;
if (result != null && result.Model != null)
{
filterContext.Result = new JsonResult
{
Data = result.Model,
JsonRequestBehavior = JsonRequestBehavior.AllowGet
};
}
}
}
}
The OnActionExecuted method will be invoked after the controller action has finished running and returned some result. Inside this method we verify whether the controller action returned a ViewResultBase (ViewResult or PartialViewResult) and whether a model has been passed. If this is the case we simply replace this result by a JsonResult.
And if you want to avoid decorating all your controller actions with this [AcceptAjax] attribute, you could register it as a global action filter in your ~/App_Start/FilterConfig.cs:
public class FilterConfig
{
public static void RegisterGlobalFilters(GlobalFilterCollection filters)
{
filters.Add(new HandleErrorAttribute());
filters.Add(new AcceptAjaxAttribute());
}
}