I have a tiny application in MVC 3.
In this tiny application, I want my URLs very clear and consistent.
There's just one controller with one action with one parameter.
If no value is provided (that is, / is requested by the browser), then a form is displayed to collect that single value. If a value is provided, a page is rendered.
The only route is this one:
routes.MapRoute(
"Default",
"{account}",
new { controller = "Main", action = "Index", account = UrlParameter.Optional }
);
This all works fine, but the account parameter never appears in the address line as a part of the URL. I can manually type test.com/some_account and it will work, but other than that, the account goes as a post parameter and therefore does not appear. And if I use FormMethods.Get in my form, I get ?account=whatever appended to the URL, which is also not what I want and which goes against my understanding. My understanding was that the MVC framework would try to use parameters set in the route, and only if not found, it would append them after the ?.
I've tried various flavours of setting the routes -- one route with a default parameter, or one route with a required parameter, or two routes (one with a required parameter and one without parameters); I've tried mixing HttpGet/HttpPost in all possible ways; I've tried using single action method with optional parameter string account = null and using two action methods (one with parameter, one without), but I simply can't get the thing appear in the URL.
I have also consulted the Steven Sanderson's book on MVC 3, but on the screenshots there are no parameters either (a details page for Kayak is displayed, but the URL in the address bar is htpp://localhost:XXXX/).
The only thing that definitely works and does what I want is
return RedirectToAction("Index", new { account = "whatever" });
But in order to do it, I have to first check the raw incoming URL and do not redirect if it already contains an account in it, otherwise it is an infinite loop. This seems way too strange and unnecessary.
What is the correct way to make account always appear as a part of the URL?
My understanding was that the MVC framework would try to use
parameters set in the route, and only if not found, it would append
them after the ?
Your understanding is not correct. ASP.NET MVC doesn't append anything. It's the client browser sending the form submission as defined in the HTML specification:
The method attribute of the FORM element specifies the HTTP method used
to send the form to the processing agent. This attribute may take two
values:
get: With the HTTP "get" method, the form data set is appended to the URI specified by the action attribute (with a question-mark ("?")
as separator) and this new URI is sent to the processing agent.
post: With the HTTP "post" method, the form data set is included in the body of the form and sent to the processing agent.
ASP.NET MVC routes are used to parse an incoming client HTTP request and redispatch it to the corresponding controller actions. They are also used by HTML helpers such as Html.ActionLink or Html.BeginForm to generate correct routes. It's just that for your specific scenario where you need to submit a user entered value as part of the url path (not query string) the HTML specification has nothing to offer you.
So, if you want to fight against the HTML specification you will have to use other tools: javascript. So you could use GET method and subscribe to the submit handler of the form and inside it manipulate the url so the value that was appended after the ? satisfy your requirements.
Don't think of this as ASP.NET MVC and routes and stuff. Think of it as a simple HTML page (which is what the browser sees of course) and start tackling the problem from that side. How would you in a simple HTML page achieve this?
Related
Let me tel the scenario where I am stuck, I will try to explain simply rather than telling the actual.
1st : I am on page 1. Where few subjects are there with unique names.
2nd : I will click on one subject, which will call an struts2 action (lets call it ActionA)at back end with the unique subject name as a request parameter(Request type = GET).
3rd : Action A will only redirect to "Tutorial page".
4th : While loading it will make an AJAX call to another Struts2 action (lets call it ActionB), which will return JSON containing tutorials for that Subject.
Problem: As I am calling ActionA first and passing the subject name which is just redirecting the page to some other page. On page load I am calling another action to get the JSON. I am not able to get the request parameter value at ActionB that is the one which is returning JSON.
Note: I am using Struts2-JSON plugin thats why not need two actions, one for redirecting the page another for getting the JSON at page load.
Solution tried: I have tried to get the request parameter value that is the Subject name, putting a hidden field in the Tutorial page. But unable to get the value from inside the Angular JS controller.
Here is the example for a shared Service: http://plnkr.co/edit/P2ItVj20RYCJVjdIaXfY
But you are right if you reload the page this doesn't work. I think your scenario needs tweaking if you want to use angular or any other single page framework for that matter. One purpose of single page applications is to minimize reloads, preferable none. If your action A only returns a template where you then want to input the result of action B, I recommend looking at ngRoute or uiRouter, where you define a template (result of action A) and a controller for that view. This template than replaces a section of your page (ng-view) with the new template. If both are new to you I would recommend looking at uiRouter, it is similar but it gives you a lot more possibilities. Both provide a "resolve" function where you can load your action B before the page is rendered.
Code from plunker
angular.module("app", [])
.controller("MainController", ['SharedService', function(SharedService) {
var vm = this;
//bind to service
vm.service = SharedService;
}]);
Service
angular.module("app")
.factory("SharedService", [function() {
var service = {
id: "test"
}
return service;
}]);
I have a GET action for creating records. Because the page is somewhat dynamic, I don't use a model to hold the data. I go off to do some OAuth, only to return to the create screen later on. In order to pass the data back, I am redirecting with a query string. I parse the query string in the GET action, and then show the view. The thing is, the query string is showing up in the browser. This displays pseudo-sensitive data.
Since I am only using the query string for transferring data, I am wondering if I can throw the query string away to prevent it from showing up on the browser.
Otherwise, is there a way to go to another action without redirecting? I've found, if I call the "other" action method directly, it tries to find the view of the original action. I can explicitly change the return View(viewModel) line to return View("create", viewModel) but that seems really dirty.
You should consider changing the action to accept POST requests. At least this will prevent the sensitive information from appearing in the browser. For extra security, your site should be served via SSL.
The other thing you can try is encrypting the sensitive values or the entire query string. The only problem is that this, too, will be preserved in the browser's history unless you require users to log in.
It looks like your action method is trying to do too much. Authentication/authorization is a separate concern which should not be part of the action method. It is better to move the authentication work in to an action filter.
Create an class that extends authorization attribute and override its OnAuthorization method to do your authorization work.
This frees your controller action method to accept POST requests.
I have footer view that's included on all my pages which contains a form. I would like to be able to make use of CI's form validation library to validate the form. Is that possible?
Currently the form posts back to the current page using the PHP_SELF environment variable. I don't want to get it to post to a controller because when validation fails it loads the controller name in the address bar, which is not the desired behaviour.
Any suggestions gratefully received.
Thanks,
Gaz
One way, whilst far from ideal, would be to create a "contact" function in every controller. This could be in the form of a library/helper.
CI doesn't natively let you call one controller from another, although I believe there are extensions that enable this.
Another option would be an AJAX call instead, which would allow you to post to a generic controller, validate etc whilst remaining on the current page.
In this use case, I would definitely go for an AJAX call to a generic controller. This allows you to show errors even before submitting in the origin page.
Another way (slightly more complex), involves posting your form data to a generic controller method, passing it a hidden input containing the current URL.
The generic controller method handling your form can then redirect to the page on which the user submitted the form, passing it the validation errors or a success message using flash session variables: $this->session->set_flashdata('errors',validation_errors()) might do the trick (untested)
The good thing about this is that you can use the generic form-handling method for both the ajax case (suppressing the redirect) and the non-ajax case
AJAX would be best, just like everyone else says.
I would redirect the form to one function in one controller, you could make a controller just for the form itself. Then have a hidden value with the return URL. As far as errors go you could send them back with flashdata.
Just remember to never copy paste code, it a bad practice and guarantees bugs.
//make sure you load the proper model
if ($this->form_validation->run() == FALSE){
// invalid
$redirect = $this->input->post('url');
$this->session->set_flashdata('errors',validation_errors());
redirect($redirect);
} else {
/*
success, do what you want here
*/
redirect('send them where ever');
}
I know that CodeIgniter already elegantly handles URL's. What I have is a form with multiple elements (date, keyword, location = optional). Is it possible to set up CI to create a URL that looks like:
mysite.com/class/function/date/keyword?
If you are using CodeIgniter's form helper, it sends your form data via POST, so you can't easily have your field's value displayed in the url.
What you could have, though, is a controller method that collects your form data and redirects to the url you want.
If you need further clarification, please let me know.
Yes, you can use the anchor function by passing the relevant info to it.
I switched from Intelligencia's UrlRewriter to the new web forms routing in ASP.NET 4.0. I have it working great for basic pages, however, in my e-commerce site, when browsing category pages, I previously used querystrings that were built into my pager control to control paging and now am not sure how to handle this using routing.
I defined a MapPageRoute as:
routes.MapPageRoute("cat-browse", "Category/{name}_{id}", ~/CategoryPage.aspx");
This works great. Now, somebody clicks to go to page 2. Previously I would have just tacked on ?page=2 to the url. How do I handle this using web forms routing? I know I can do something like:
http://www.mysite.com/Category/Arts-and-Crafts_17/page/2
But in addition to page, I can have filters, age ranges, gender, etc.
Should I just keep defining routes
that handle these variables like
above?
Should I continue using querystrings
and if so, how do you define a route
to handle that?
The main reason to use url routing is to expose clean, user-and-SEO-friendly, URLs. If this is your goal, then try to stick to it and not use querystring parameters. Note: I don't believe we need to completely ban the use of querystrings and, depending on your situation, you may decide it best to use querystring parameters for parameters that are not used frequently, or where no real value is added by making the information more semantically meaningful.
So here's what I would do:
Define a catch-all for all your other parameters:
routes.MapPageRoute("cat-browse", "Category/{name}_{id}/{*queryvalues}", "~/CategoryPage.aspx");
In /CategoryPage.aspx, access the router parameter and then parse as appropriate:
Page.RouteData.Values["queryvalues"]
Instead of using the syntax of Arts-and-Crafts_17/**page/2/age/34** for these parameters, I perfer to use the following syntax: Arts-and-Crafts_17/pg-2/age-34/
If you do this, the catch-all parameter 'querystring', will equal pg-2/age-34. You can now easily parse this data and add each name/value to the page context. Note that you will need to do something along these lines since each of these parameters are optional on your site.
You can take advantage of C# 4.0 named and optional parameters. Please have a look at this example from haacked
If you are using a lower version of the framework, you can also use code from the link above. But instead of declaring the method as
public ActionResult Search(int? page=0)
{}
you can declare it as
public ActionResult Search(int? page)
{
if(page == null)
{
page=0;
}
}
HTH