A Smart Form Control for ASP.NET URL Rewriting

• Download SmartFormTag.zip - 53.8 KB

Introduction

One of the websites I work on uses URL rewriting, i.e. when you type in an address www.somesite.com/Products/WidgetA/ you are really browsing to an ASP.NET page somewhere else in the website say www.somesite.com/ProductDisplay.aspx?page=Products/WidgetA/.

For a full explanation of URL rewriting along with a very usable rewriting engine go to 15seconds.com and read Rewrite.NET -- A URL Rewriting Engine for .NET by Robert Chartier.

Background

One problem with this, and believe me there are many problems with URL rewriting, is that it breaks the form tag. The action of the form tag will be the real address of the .aspx page and not the nice friendly address you wanted. To work around this problem MSDN suggests that you create an "Action-less form" meaning you create a custom Form tag that simply doesn't write the action attribute on the form tag.

namespace ActionlessForm {
  public class Form : System.Web.UI.HtmlControls.HtmlForm
  {
     protected override void RenderAttributes(HtmlTextWriter writer)
     {
        writer.WriteAttribute("name", this.Name);
        base.Attributes.Remove("name");

        writer.WriteAttribute("method", this.Method);
        base.Attributes.Remove("method");

        this.Attributes.Render(writer);

        base.Attributes.Remove("action");

        if (base.ID != null)
           writer.WriteAttribute("id", base.ClientID);
     }
  }
}
From "http://msdn2.microsoft.com/en-us/library/ms972974.aspx">URL Rewriting in ASP.NET by Scott Mitchell

The big problem with the above code is that it completely breaks the form tag for any server control that hooks into the postback cycle, validators for instance. If you look at the source code for RenderAttributes method the HtmlForm class in the .net Framework using the excellent Lutz Reflector you will see that there is a whole lot going on in the .net Framework's HtmlForm class that Scott's version does not duplicate. Besides that what if the HtmlForm tag was changed in a future version of the .net Framework?

If we examine exactly how the action attribute gets written we see that, because the GetActionAttribute is private, there is no way to directly change the action attribute value before it gets to the HtmlTextWriter.

Also some of the logic in the RenderAttributes method simply can't be reproduced using a derived class because many of the necessary methods and attributes are private.

writer.WriteAttribute("action", this.GetActionAttribute(), true);

So we need to take a different approach to take control over if / how the Form tag action attribute gets written.

Using the code

My approach was to intercept the attribute before it was written to the HtmlTextWriter. To do this I created a derived class called SelectiveHtmlTextWriter and overrode the WriteAttribute method:

public class SelectiveHtmlTextWriter : HtmlTextWriter
{
    ...
    
    public override void WriteAttribute(string name, string value, bool fEncode) 
    { 
        base.WriteAttribute(name, "/newpagename/goes/here/", fEncode);
    }
}


public class SmartForm : HtmlForm
{
    ...

    protected override void RenderAttributes(HtmlTextWriter writer)
    {
        SelectiveHtmlTextWriter customWriter = new SelectiveHtmlTextWriter(writer);
        base.RenderAttributes(customWriter);
    }
}

A new SelectiveHtmlTextWriter object is created and passed into HtmlForm class RenderAttribute method in place of the normal the HtmlTextWriter.

This works because of polymorphism. All the HtmlForm class is cares about is that it gets a reference to an HtmlTextWriter. However when it calls the RenderAttribute my override gets called instead.

This works great but I have never been fond of hard-coded values so in order to make this more extensible I added an event to the new SelectiveHtmlTextWriter called WritingAttribute which the SmartForm class subscribes to. This allows any attribute, but specifically the action, to be modified before its value is written. In this case I am substituting the Request.RawUrl value which holds the actual URL that the request came from.

public class SmartForm : HtmlForm
{
    public SmartForm()
        : base()
    {
    }

    protected override void RenderAttributes(HtmlTextWriter writer)
    {
        SelectiveHtmlTextWriter customWriter = new SelectiveHtmlTextWriter(writer);
        customWriter.WritingAttribute += new SubstituteValueEventHandler(customWriter_WritingAttribute);
        base.RenderAttributes(customWriter);
    }

    void customWriter_WritingAttribute(object sender, SubstituteValueEventArgs e)
    {
        if (e.Name == "action")
        {
            e.NewValue = Context.Request.RawUrl;
        }
    }
}

In conclusion if I had a choice I would not use URL rewriting at all I would find some other way to make the client happy. But the SmartForm that we outlined here will ease the pain quite a bit. One last caveat, using the SmartForm breaks the Visual Studio designer. I have been unsuccessful at creating a custom designer for this. If anybody comes up with a designer for this email me and I'll add it to the article with the appropriate credits.

You must Sign In to use this message board.

FAQ    Noise Tolerance Very HighHighMediumLowVery Low   Layout Expand Posts & RepliesThread ViewNo JavascriptNo JS + Preview   Per page 102550

Msgs 1 to 8 of 8 (Total in Forum: 8) (Refresh) FirstPrevNext

An easier option is here robert_fairbrother 7:45 27 Aug '07   I think a better, and easier option is to insert just one line in the Page_Load event of the page. Context.RewritePath(System.IO.Path.GetFileName(Request.RawUrl)); See full article here URL Mapping, URL Rewriting Sign In·View Thread·PermaLink Web.Config Xela220 16:41 23 Jul '07   You can use the new section (2.0) in the web.config which will take care of the rewrite. However, the biggest problem that I see is that you cannot use regex. I guess Microsoft plans to add more to it in the future. AvidEnterprise.com - Software and services for businesses Sign In·View Thread·PermaLink Classes derived from HtmlForm break the designer nikkilong 8:08 23 Jul '07   One way I got around this is by creating a ControlAdapter (not a complete example) with my code in and overrided the Render method like so: protected override void Render(System.Web.UI.HtmlTextWriter writer) {       HtmlForm htmlForm = Control as HtmlForm;       writer.AddAttribute(HtmlTextWriterAttribute.Id, htmlForm.ClientID);       writer.AddAttribute(HtmlTextWriterAttribute.Name, htmlForm.Name);       writer.AddAttribute("Method", htmlForm.Method);       // other attributes here       writer.RenderBeginTag(HtmlTextWriterTag.Form);       RenderChildren(writer);       writer.RenderEndTag(); } Then add a .browser file in the App_Browsers file with the following: <!--       You can find existing browser definitions at       <windir>\Microsoft.NET\Framework\<ver>\CONFIG\Browsers --> <browsers>       <browser refID="Default">          <controlAdapters>             <adapter adapterType="MyFormAdapter, MyAssembly"                         controlType="System.Web.UI.HtmlControls.HtmlForm" />          </controlAdapters>       </browser> </browsers> It's not ideal but it was the best way I could find without writing a compatible designer and going totally nuts in the process. Sign In·View Thread·PermaLink Re: Classes derived from HtmlForm break the designer Richard Deeming 11:04 27 Jul '07   A control adapter still won't have access to some of the internals of the HtmlForm control, so the generated HTML will still be missing some of the attributes. The simplest answer is probably to combine the approaches: use a control adapter to modify the output for the form, with a derived HtmlTextWriter class to replace or suppress the action attribute. To minimize the impact, you can use the SetRenderMethodDelegate method to ensure that the original HtmlTextWriter is used to render the child controls. protected override void Render(HtmlTextWriter writer) { if (null != writer && null != this.Control) { HtmlForm form = this.Control as HtmlForm; if (null == form) { Debug.Assert(false, "Only HtmlForm controls are supported."); base.Render(writer); } else { // If an explicit action is specified, use it: string action = form.Attributes["action"]; if (null != action) { form.Attributes.Remove("action"); if (0 != action.Length) action = form.ResolveUrl(action); } SmartFormHtmlTextWriter formWriter = new SmartFormHtmlTextWriter(writer, action); form.SetRenderMethodDelegate(this.FormChildrenRenderMethod); base.Render(formWriter); } } } private void FormChildrenRenderMethod(HtmlTextWriter output, Control container) { Debug.Assert(null != output); Debug.Assert(container == this.Control); if (container.HasControls()) { SmartFormHtmlTextWriter writer = output as SmartFormHtmlTextWriter; if (null != writer) output = writer.InnerHtmlTextWriter; foreach (Control child in container.Controls) { child.RenderControl(output); } } } ... private sealed class SmartFormHtmlTextWriter : HtmlTextWriter { private readonly HtmlTextWriter _innerWriter; private readonly string _formActionValue; private bool _inFormAttributes; public SmartFormHtmlTextWriter(HtmlTextWriter innerWriter, string formActionValue) : base(innerWriter) { Debug.Assert(null != innerWriter); _innerWriter = innerWriter; _formActionValue = formActionValue; } public HtmlTextWriter InnerHtmlTextWriter { get { return _innerWriter; } } public override void Close() { } public override void WriteBeginTag(string tagName) { _inFormAttributes = string.Equals(tagName, "form", StringComparison.OrdinalIgnoreCase); InnerHtmlTextWriter.WriteBeginTag(tagName); } public override void WriteAttribute(string name, string value, bool fEncode) { if (_inFormAttributes && string.Equals(name, "action", StringComparison.OrdinalIgnoreCase)) { // No need to check any further attributes: _inFormAttributes = false; // If we have specified an action, render it: if (!string.IsNullOrEmpty(_formActionValue)) { InnerHtmlTextWriter.WriteAttribute(name, _formActionValue, fEncode); } } else { InnerHtmlTextWriter.WriteAttribute(name, value, fEncode); } } // Override the other public virtual methods of the HtmlTextWriterClass, // and pass the call through to the InnerHtmlTextWriter instance, to // minimize the chances of breaking anything. "These people looked deep within my soul and assigned me a number based on the order in which I joined." - Homer Sign In·View Thread·PermaLink Actually the best workaround would be.... Gevorg 6:05 23 Jul '07   All you need to do is simply rewrite it back to what URL was on the page. Then Form tag would pick it up. Example: protected void Application_BeginRequest(Object sender, EventArgs e) { HttpContext ctx = HttpContext.Current; ctx.Items["OriginalPath"] = ctx.Request.Path; ...blalblalbla } on the page Page_OnInit(..) { HttpContext ctx = HttpContext.Current; ctx.RewritePath((string)ctx.Items["OriginalPath"] ); .... } Inheritance helps. All my pages derived from my StandardPage that does this code. My site - Body Jewelry Not my site - Piercing info Sign In·View Thread·PermaLink 5.00/5 (1 vote) Reverse engineering? Dario Solera 2:42 21 Jul '07   Apart from the interesting article, I think that you cannot decompile the .NET Framework, let alone publishing the source code... If you truly believe you need to pick a mobile phone that "says something" about your personality, don't bother. You don't have a personality. A mental illness, maybe - but not a personality. - Charlie Brooker My Blog - My Photos - ScrewTurn Wiki Sign In·View Thread·PermaLink Re: Reverse engineering? David T Nelson 8:05 23 Jul '07   You may be right about posting the code... I'll take that part out. Anybody willing to decompile it can do so themselves and investigate further. Sign In·View Thread·PermaLink 2.00/5 (1 vote) Re: Reverse engineering? nsimeonov 2:26 24 Jul '07   Actually having the source code of the underlying libraries can be really helpful. Years ago when I was working with delphi the part I loved about it was that I had the source code of all components and I could check how something works instead of guessing or experimenting until I find out. I think developers should be able to browse at least some part of the code if not all of it. Don't get me wrong - I am not a supporter of open source and these ideas, but I would gladly pay an extra fee to have the source code if I have the option. Sign In·View Thread·PermaLink 1.20/5 (2 votes)

Last Visit: 10:11 8 Nov '09     Last Update: 10:11 8 Nov '09 1