What is difference between Page() and Rediirect() to self in ASP.NET Core Razor pages? - asp.net-core

I have a very simple page that has 2 forms. When I submit one form it resets the other. There is some kind of hidden optimization is going on because when I refresh the page it presents the correct result.
Here is the page:
<div asp-validation-summary="All"></div>
<div class="col-md-3">
<form method="POST">
<fieldset>
<div>Host Name: <input asp-for="ClientConfig.HostName" /></div>
<div>Responses in HTML? <input type="checkbox" asp-for="ClientConfig.Html" /></div>
<input type="submit" asp-page-handler="ClientConfiguration" />
</fieldset>
</form>
<p>Base URL = #Model.ClientConfig.Summary</p>
</div>
<form method="POST">
<fieldset>
<div>Name: <input asp-for="Customer.Name" /></div>
<div>New? <input type="checkbox" asp-for="Customer.New" /></div>
<input type="submit" asp-page-handler="Customer" />
</fieldset>
</form>
<ul>
<li>Customer = #Model.Customer.Summary</li>
</ul>
Here is the model...
public class ClientConfig
{
public static ClientConfig Instance { get; set; } = new ClientConfig();
[Required, StringLength(100)] public string HostName { get; set; } = "LocalHost";
public bool Html { get; set; }
public string Summary => HostName + (Html ? " (Html)" : "");
}
public class Customer
{
public static Customer Instance { get; set; } = new Customer();
[Required, StringLength(100)] public string Name { get; set; } = "Default";
public bool New { get; set; }
public string Summary => Name + (New ? " (New)" : "");
}
public class IndexModel : PageModel
{
public IndexModel()
{
ClientConfig = ClientConfig.Instance;
Customer = Customer.Instance;
}
[BindProperty] public ClientConfig ClientConfig { get; set; }
[BindProperty] public Customer Customer { get; set; }
public async Task<IActionResult> OnPostCustomerAsync()
{
Customer.Instance = Customer;
return Page();
}
public async Task<IActionResult> OnPostClientConfigurationAsync()
{
ClientConfig.Instance = ClientConfig;
return Page();
}
}
So what is "return Page();" doing? According to the documentation it is simply rendering the current page. Not true. To verify this, simply refresh the page. It will be different, accurate with both forms filled in. Also if you replace "return Page()" with "return Redirect("/Index");" the result will also be accurate. So again, what is "return Page()" doing? There is some kind of undocumented optimization that resets all the forms except the one recently submitted.

You have multiple separate forms on your page with separate form values: In one form you are submitting the client configuration object, in the other you are submitting the customer object.
So when you are actually submitting a form, only that form's data is being submitted. For example, if you are submitting the customer form, the client configuration data is not being transferred in the POST request (and the other way around).
As such, when you render the page by returning Page(), only the data that is currently in the page model is being rendered. If you are submitting the customer form, then only the customer data is available (same for the client configuration form).
This happens simply because you only have partial data on a page where you would need more to fill in all forms. If you want to prevent that, you will have to combine the data into a single model and form.
Now, if you refresh the page in the browser, then your browser is typically smart enough not to clear form values immediately. If you do a hard refresh using Ctrl + F5, then the browser should also reset the values.
It's also possible that your browser is performing an auto-fill for the forms here. This will typically only apply for GET requests. So that could be the reason why you are getting this result when you return a Redirect() because that completes the form POST with a GET request.

When I submit one form it resets the other.
That's the expected behavior for the way you coded your page. When the form POSTs to the server, the server does three things:
creates a new IndexModel object using its constructor,
binds the object's properties to the POSTed form values, and
binds the object to its view.
In your code, step (1) resets properties to their default values. Step (2) overwrites those defaults with POSTed form values. Since you're submitting only one form, the other form's values retain their defaults. That's why submitting one resets the other.
So what is "return Page();" doing? According to the documentation it is simply rendering the current page. Not true. To verify this, simply refresh the page. It will be different, accurate with both forms filled in. Also if you replace "return Page()" with "return Redirect("/Index");" the result will also be accurate.
When you submit a form, return Page() renders the page in the context of a POST. On the other hand, when you refresh or redirect, the context is a GET. The difference you see happens because the context is different: the response to a POST is different from the response to a GET.

Right. After quite a long time of pondering this problem, I've finally figured it out. The problem: Razor pages moves in mysterious ways its wonders to perform.
My initial assumption was wrong. The page model constructor is not being bypassed. The page model is being properly constructed from static values. However after construction all bound objects on the page are reset. So this is not an "undocumented optimization"...it is an undocumented impairment.
The fix for this is to reset the page model from static values before returning Page().
public async Task<IActionResult> OnPostCustomerAsync()
{
Customer.Instance = Customer;
ClientConfig = ClientConfig.Instance;
return Page();
}
public async Task<IActionResult> OnPostClientConfigurationAsync()
{
ClientConfig.Instance = ClientConfig;
Customer = Customer.Instance;
return Page();
}
This is obviously a massive kluge, but no elegant solution exists. Anyone?

Related

Identify which MudExpansionPanel is being expanded

I want to present a list of up to 20 panels within a <MudExpansionPanels> component where the expanded child portion of each <MudExpansionPanel> is expensive to render. I tried the following test code but all instances of <LiveAgentSummary> are rendered as the parent is rendered, just to clarify this rendering of <LiveAgentSummary> happens before any panel is manually expanded.
<MudExpansionPanels>
#foreach (var liveAgent in _liveAgents)
{
<MudExpansionPanel Text=#liveAgent.Name>
<LiveAgentSummary AgentId=#liveAgent.Id />
</MudExpansionPanel>
}
</MudExpansionPanels>
I then looked into delaying the render of each <LiveAgentSummary> through use of a RenderFragment that is dynamically built during the <MudExpansionPanel> IsExpandedChanged event. However the event handler does not indicate which panel is being expanded and hence I do not know which liveAgent.Id param value to pass to <LiveAgentSummary> as I build a RenderFragment.
I think <MudExpansionPanels> is missing support for a bind-ActivePanelId property but hopefully I am overlooking an alternative solution to my delayed rendering objective.
This is the official MudBlazor example that prompted me to look into using a RenderFragment.
Update: A long answer briefly appeared yesterday suggesting that I could query the list of panel components on a built-in property that indicates the expanded state. The poster had gone to the trouble of reading the MudBlazor source code but the answer was then deleted.
I am now wondering how from code in an event handler it is possible to iterate over a component hierarchy declared as mark-up. Applying this to my example markup above, how could event handler code obtain a reference to each <MudExpansionPanel> child within <MudExpansionPanels>.
Can't you make use of the bool from the IsExpandedChanged callback? Something like this:
Index.razor
#page "/"
<MudExpansionPanels>
#foreach (var liveAgent in this.liveAgents)
{
<MudExpansionPanel
Text="#($"{liveAgent.Name} ({liveAgent.Data})")"
IsExpandedChanged="#(e => this.Load(e, liveAgent))">
<LiveAgentSummary Agent="#liveAgent" />
</MudExpansionPanel>
}
</MudExpansionPanels>
#code {
private readonly List<Agent> liveAgents = new()
{
new Agent("1", "Agent Smith"),
new Agent("2", "Agent Brown"),
new Agent("3", "Agent Jones")
};
private void Load(bool expanded, Agent agent)
{
if (expanded)
{
agent.Load();
}
}
}
LiveAgentSummary.razor
<MudText>id: #this.Agent.Id, data: #this.Agent.Data</MudText>
#code {
[Parameter]
public Agent Agent { get; set; } = default!;
}
Agent.cs
public record Agent(string Id, string Name)
{
public string Data { get; set; } = "Not loaded";
public void Load()
{
Console.WriteLine($"Loading agent {this.Id}...");
this.Data = "Loaded!";
}
}

Loding pages by posting parameters

The subject might not be clear since I couldn't find a better way to express it.
I am developing a web application using ASP.NET Core 6.0 with Razor Pages. Our previous application was an SPA using Ext JS where any call to server was returning only data and where I was also able to make any kind of call (GET/POST) to get the data.
For example, in the above picture from my old application, I make an ajax call with POST to get the list of periods when I open this page. I make a POST because I am sending the period type in my request payload. Sure I can pass these parameters in a GET request, however my other views have many criteria, so passing these criteria in the query string is not what I want. So, I decided to make it a standard to make my calls with POST method if there are any criteria payload, make GET request only when fething an entity with a simple key parameter (like Id) or GET any list that doesn't have any criteria.
Now, I am quite confused how to do same thing in my new ASP.NET Core Razor Pages web application. Normally, the menu items navigate to the page using link as below, which makes a GET request:
<a asp-area="System" asp-page="/ProfessionList">#AppLocalizer["Profession List"]</a>
<a asp-area="System" asp-page="/PeriodList">#AppLocalizer["Profession List"]</a>
In order to make a POST request, I replaced the menu item for period list as following which makes a POST request with a default periodType payload:
<a asp-area="System" asp-page="/ProfessionList">#AppLocalizer["Profession List"]</a>
<form asp-area="System" asp-page="/PeriodList" method="post">
<input type="hidden" name="periodType" value="1" hidden />
<button type="submit" >#AppLocalizer["Period List"]</button>
</form>
And the corresponding PeriodType.cshtml.cs file is as following:
[Authorize]
public class PeriodListModel: BaseEntityListPageModel<List<JsonPeriodEx>> {
public PeriodListModel(ILogger<BaseEntityListPageModel<List<JsonPeriodEx>>> logger, WebApi webApi) : base(logger, webApi) {
}
public IActionResult OnGet() {
PageData = JsonConvert.DeserializeObject<List<JsonPeriodEx>>(TempData["PageData"].ToString());
return Page();
}
public async Task<IActionResult> OnPostAsync(int periodType) {
var jsonResult = await _WebApi.DoPostAsync<List<JsonPeriodEx>>("/PeriodEx/GetList", new[] { new { Property = "periodType", Value = periodType } });
if (jsonResult.IsLoggedOut)
return RedirectToPage("/Login", new { area = "Account" });
if (jsonResult.Success) {
PageData = jsonResult.Data;
TempData["PageData"] = JsonConvert.SerializeObject(PageData);
return RedirectToPage("/PeriodList");
} else {
return RedirectToPage("/Error");
}
}
}
OnPostAsync successfully binds to the posted periodType parameter and gets the list of periods. Now, at the end of a successful call I want to follow the Post/Redirect/Get pattern and redirect to OnGet with the data from OnPostAsync, which is stored in TempData.
Now, according to the above scenario, is my approach, explained above, correct or should I implement it differently?
Thanks in advance
For these cases I would prefer TempData. Much easier and less code.
public async Task OnGet()
{
TempData["myParamToPass"] = 999;
...
}
public async Task OnPostReadData()
{
if (TempData.ContainsKey("myParamToPass"))
{
var myParamToPassValue = TempData.Peek("myParamToPass") as int?;
...
}
...
}

EditorFor Tag Helper doesn't render validation attributes when using FluentValidator

I have a simple form like this which makes use of the #Html.EditorFor extension:
<form method="post">
#Html.EditorFor(x => x.SystemSettings.EmailFromAddress)
<submit-button title="Save"></submit-button>
</form>
I want to take advantage of .NET Core's tag helpers so that my form looks like this instead:
<form method="post">
<editor asp-for="SystemSettings.EmailFromAddress"/>
<submit-button title="Save"></submit-button>
</form>
I also eventually would like to have my own custom tag helpers so I can do something like this instead:
<text-box asp-for="SystemSettings.EmailFromAddress"></text-box>
I have a string template which gets rendered by the #Html.EditorFor extension:
#model string
<div class="form-group">
<label asp-for="#Model" class="m-b-none"></label>
<span asp-description-for="#Model" class="help-block m-b-none small m-t-none"></span>
<div class="input-group">
<input asp-for="#Model" class="form-control" />
<partial name="_ValidationIcon" />
</div>
<span asp-validation-for="#Model" class="validation-message"></span>
</div>
To do that, I saw someone implemented an EditorTagHelper, which looks like this:
[HtmlTargetElement("editor", TagStructure = TagStructure.WithoutEndTag,
Attributes = ForAttributeName)]
public class EditorTagHelper : TagHelper
{
private readonly IHtmlHelper _htmlHelper;
private const string ForAttributeName = "asp-for";
private const string TemplateAttributeName = "asp-template";
[HtmlAttributeName(ForAttributeName)]
public ModelExpression For { get; set; }
[HtmlAttributeName(TemplateAttributeName)]
public string Template { get; set; }
[ViewContext]
[HtmlAttributeNotBound]
public ViewContext ViewContext { get; set; }
public EditorTagHelper(IHtmlHelper htmlHelper)
{
_htmlHelper = htmlHelper;
}
public override async Task ProcessAsync(TagHelperContext context, TagHelperOutput output)
{
if (context == null)
throw new ArgumentNullException(nameof(context));
if (output == null)
throw new ArgumentNullException(nameof(output));
if (!output.Attributes.ContainsName(nameof(Template)))
{
output.Attributes.Add(nameof(Template), Template);
}
output.SuppressOutput();
(_htmlHelper as IViewContextAware).Contextualize(ViewContext);
output.Content.SetHtmlContent(_htmlHelper.Editor(For.Name, Template));
await Task.CompletedTask;
}
}
When I use the EditorTagHelper though, it seems to be missing the unobtrusive Javascript validation attributes:
Using #Html.EditorFor, this gets rendered:
<input class="form-control valid" type="text" data-val="true" data-val-required="Email From Address cannot be empty" id="SystemSettings_EmailFromAddress" name="SystemSettings.EmailFromAddress" value="whatever#test.com" aria-required="true" aria-invalid="false" aria-describedby="SystemSettings_EmailFromAddress-error">
It's got the data-val attributes so client-side validation gets applied.
When I use the EditorTagHelper instead, this gets rendered:
<input class="form-control valid" type="text" id="SystemSettings_EmailFromAddress" name="SystemSettings.EmailFromAddress" value="whatever#test.com" aria-invalid="false">
The unobtrusive validation attributes are not being applied. I am using FluentValidation and I have specified an AbstractValidator like this:
public class SystemSettingsValidator : AbstractValidator<SystemSettings>
{
public SystemSettingsValidator()
{
RuleFor(x => x.EmailFromAddress).NotEmpty()
.WithMessage("Email From Address cannot be empty");
}
}
I found that if I removed the AbstractorValidator and simply added a [Required] attribute to my model property the validation then works properly. This suggests that there is something wrong with FluentValidation. Perhaps there is a configuration issue.
I am using Autofac dependency injection to scan my assemblies and register validator types:
builder.RegisterAssemblyTypes(Assembly.Load(assembly))
.Where(t => t.IsClosedTypeOf(typeof(IValidator<>)))
.AsImplementedInterfaces()
.PropertiesAutowired()
.InstancePerLifetimeScope();
This seems to work fine. In case it wasn't fine, I also tried registering the validators from the fluent validation options like this:
.AddFluentValidation(fv =>
{
fv.RegisterValidatorsFromAssemblies(new List<Assembly>
{Assembly.GetExecutingAssembly(), Assembly.Load(nameof(Entities))});
})
This also seemed to be fine.
One thing to note is that an earlier problem I had was that using Autofac assembly scanning was breaking the application when tag helpers were included. I added a filter to ensure that tag helpers are not included when registering these dependencies, e.g.
builder.RegisterAutowiredAssemblyInterfaces(Assembly.Load(Web))
.Where(x => !x.Name.EndsWith("TagHelper"));
I have uploaded a working sample of the code here: https://github.com/ciaran036/coresample2
Navigate to the Settings Page to see the field I am trying to validate.
This issue also appears to affect view components.
Thanks.
I believe the issue is in the tag helper, in that it uses IHtmlHelper.Editor rather than IHtmlHelper<TModel>.EditorFor to generate the HTML content. They are not quite the same.
As you point out FluentValidation injects the validation attributes as you'd expect for #Html.EditorFor(x => x.SystemSettings.EmailFromAddress). However for #Html.Editor("SystemSettings.EmailFromAddress"), which is what your custom tag helper is doing, FluentValidation doesn't inject the validation attributes. So that rules out the tag helper itself and moves the problem to the Editor invocation.
I also noticed that Editor doesn't resolve <label asp-for (or the other <span asp-description-for tag helper you're using) so that suggests it's not a FluentValidation specific issue.
I wasn't able to replicate your success with the Required attribute for the custom tag helper/Editor - the Required attribute only injected the validation attributes when using EditorFor.
The internals for Editor and EditorFor are similar but with a key difference, the way they resolve the ModelExplorer instance used to generate the HTML content differs and I suspect this is the problem. See below for these differences.
Things like PropertyName set to null and Metadata.Property not being set for Editor, but set to EmailFromAddress and SystemSettings.EmailFromAddress for EditorFor are standing out as potential causes for the behaviour we're seeing.
The painful part is the tag helper has a valid ModelExplorer instance via the For property. But there is no built in provision to provide it to the html helper.
As to the resolution, the obvious one seems to be to use EditorFor rather than Editor however it doesn't look easy. It'd likely involve reflection and building an expression.
Another option is, considering the tag helper resolves the ModelExplorer correctly, is to extend HtmlHelper and override the GenerateEditor method - what both Editor and EditorFor end up invoking - so you can pass in the ModelExplorer and work around the problem.
public class CustomHtmlHelper : HtmlHelper, IHtmlHelper
{
public CustomHtmlHelper(IHtmlGenerator htmlGenerator, ICompositeViewEngine viewEngine, IModelMetadataProvider metadataProvider, IViewBufferScope bufferScope, HtmlEncoder htmlEncoder, UrlEncoder urlEncoder) : base(htmlGenerator, viewEngine, metadataProvider, bufferScope, htmlEncoder, urlEncoder) { }
public IHtmlContent CustomGenerateEditor(ModelExplorer modelExplorer, string htmlFieldName, string templateName, object additionalViewData)
{
return GenerateEditor(modelExplorer, htmlFieldName, templateName, additionalViewData);
}
protected override IHtmlContent GenerateEditor(ModelExplorer modelExplorer, string htmlFieldName, string templateName, object additionalViewData)
{
return base.GenerateEditor(modelExplorer, htmlFieldName, templateName, additionalViewData);
}
}
Update your tag helper to use it:
public override async Task ProcessAsync(TagHelperContext context, TagHelperOutput output)
{
if (context == null)
throw new ArgumentNullException(nameof(context));
if (output == null)
throw new ArgumentNullException(nameof(output));
if (!output.Attributes.ContainsName(nameof(Template)))
{
output.Attributes.Add(nameof(Template), Template);
}
output.SuppressOutput();
(_htmlHelper as IViewContextAware).Contextualize(ViewContext);
var customHtmlHelper = _htmlHelper as CustomHtmlHelper;
var content = customHtmlHelper.CustomGenerateEditor(For.ModelExplorer, For.Metadata.DisplayName ?? For.Metadata.PropertyName, Template, null);
output.Content.SetHtmlContent(content);
await Task.CompletedTask;
}
Finally register the new helper, the earlier the better I'd say
services.AddScoped<IHtmlHelper, CustomHtmlHelper>();
Working solution

Validation messages from custom model validation attributes are locked to first loaded language

I am working on a multi lingual website using Umbraco 7.2.4 (.NET MVC 4.5). I have pages for each language nested under home nodes with their own culture:
Home (language selection)
nl-BE
some page
some other page
my form page
fr-BE
some page
some other page
my form page
The form model is decorated with validation attributes that I needed to translate for each language. I found a Github project, Umbraco Validation Attributes that extends decoration attributes to retrieve validation messages from Umbraco dictionary items. It works fine for page content but not validation messages.
The issue
land on nl-BE/form
field labels are shown in dutch (nl-BE)
submit invalid form
validation messages are shown in dutch (nl-BE culture)
browse to fr-BE/form
field labels are shown in french (fr-BE)
submit invalid form
Expected behavior is: validation messages are shown in french (fr-BE culture)
Actual behavior is: messages are still shown in dutch (data-val-required attribute is in dutch in the source of the page)
Investigation to date
This is not a browser cache issue, it is reproducible across separate browsers, even separate computers: whoever is generating the form for the first time will lock the validation message culture. The only way to change the language of the validation messages is to recycle the Application Pool.
I doubt that the Umbraco Validation helper class is the issue here but I'm out of ideas, so any insight is appreciated.
Source code
Model
public class MyFormViewModel : RenderModel
{
public class PersonalDetails
{
[UmbracoDisplayName("FORMS_FIRST_NAME")]
[UmbracoRequired("FORMS_FIELD_REQUIRED_ERROR")]
public String FirstName { get; set; }
}
}
View
#inherits Umbraco.Web.Mvc.UmbracoTemplatePage
var model = new MyFormViewModel();
using (Html.BeginUmbracoForm<MyFormController>("SubmitMyForm", null, new {id = "my-form"}))
{
<h3>#LanguageHelper.GetDictionaryItem("FORMS_HEADER_PERSONAL_DETAILS")</h3>
<div class="field-wrapper">
#Html.LabelFor(m => model.PersonalDetails.FirstName)
<div class="input-wrapper">
#Html.TextBoxFor(m => model.PersonalDetails.FirstName)
#Html.ValidationMessageFor(m => model.PersonalDetails.FirstName)
</div>
</div>
note: I have used the native MVC Html.BeginForm method as well, same results.
Controller
public ActionResult SubmitFranchiseApplication(FranchiseFormViewModel viewModel)
{
if (!ModelState.IsValid)
{
TempData["Message"] = LanguageHelper.GetDictionaryItem("FORMS_VALIDATION_FAILED_MESSAGE");
foreach (ModelState modelState in ViewData.ModelState.Values)
{
foreach (ModelError error in modelState.Errors)
{
TempData["Message"] += "<br/>" + error.ErrorMessage;
}
}
return RedirectToCurrentUmbracoPage();
}
}
LanguageHelper
public class LanguageHelper
{
public static string CurrentCulture
{
get
{
return UmbracoContext.Current.PublishedContentRequest.Culture.ToString();
// I also tried using the thread culture
return System.Threading.Thread.CurrentThread.CurrentCulture.ToString();
}
}
public static string GetDictionaryItem(string key)
{
var value = library.GetDictionaryItem(key);
return string.IsNullOrEmpty(value) ? key : value;
}
}
So I finally found a workaround. In attempt to reduce my app to its simplest form and debug it, I ended up recreating the "UmbracoRequired" decoration attribute. The issue appeared when ErrorMessage was set in the Constructor rather than in the GetValidationRules method. It seems that MVC is caching the result of the constructor rather than invoking it again every time the form is loaded. Adding a dynamic property to the UmbracoRequired class for ErrorMessage also works.
Here's how my custom class looks like in the end.
[AttributeUsage(AttributeTargets.Property | AttributeTargets.Field | AttributeTargets.Parameter,
AllowMultiple = false)]
internal class LocalisedRequiredAttribute : RequiredAttribute, IClientValidatable
{
private string _dictionaryKey;
public LocalisedRequiredAttribute(string dictionaryKey)
{
_dictionaryKey = dictionaryKey;
}
public IEnumerable<ModelClientValidationRule> GetClientValidationRules(
ModelMetadata metadata, ControllerContext context)
{
ErrorMessage = LanguageHelper.GetDictionaryItem(_dictionaryKey); // this needs to be set here in order to refresh the translation every time
yield return new ModelClientValidationRule
{
ErrorMessage = this.ErrorMessage, // if you invoke the LanguageHelper here, the result gets cached and you're locked to the current language
ValidationType = "required"
};
}
}

Simple controller which takes POST is not found

I've made some previous question asking for the help with the problems since I updated MVC4 webapi beta to RC. I got most in order now, but here's one I cannot figure out the reason for yet.
For this simple controller I have one which accepts a POST and one that accepts GET. When I try to run those by sending request from a HTML form, only the GET controller is found while the POST one will return me the following error.
{
"Message": "No HTTP resource was found that matches the request URI 'http://localhost/webapi/api/play/test'.",
"MessageDetail": "No action was found on the controller 'Play' that matches the name 'test'."
}
Why is the POST controller not found?
Controllers
public class PlayController : ApiController
{
[HttpPost] // not found
public string Test(string output)
{
return output;
}
[HttpGet] // works
public string Test2(string output)
{
return output;
}
}
HTML form
<form action="http://localhost/webapi/api/play/test" method="post">
<input type="text" name="output" />
<input type="submit" name="submit" />
</form>
<form action="http://localhost/webapi/api/play/test2" method="get">
<input type="text" name="output" />
<input type="submit" name="submit" />
</form>
Web.API is a little bit picky when you want to post "simple" values.
You need to use the [FromBody] attribute to signal that the value is not coming from the URL but from the posted data:
[HttpPost]
public string Test([FromBody] string output)
{
return output;
}
With this change you won't get 404 anymore but output will be always null, because Web.Api requries the posted values in special format (look for the Sending Simple Types section):
Second, the client needs to send the value with the following format:
=value
Specifically, the name portion of the name/value pair must be empty for a simple type. Not >all browsers support this for HTML forms,
but you create this format in script...
So recommend that you should create a model type:
public class MyModel
{
public string Output { get; set; }
}
[HttpPost]
public string Test(MyModel model)
{
return model.Output;
}
Then it will work with your sample froms without modifing your views.