[RESOLVED] How to make my own custom "page" object?

I want to create my own object (dll) that inherits from System.Web.UI.Page so that I can do some custom pre/post processing of a call to a web page that inherits from my class. So for example, lets say you have the following page defined:

[code]
Partial Class TestPage1
Inherits MyCustomDLL.Page

Protected Sub Page_Load(ByVal sender As Object, ByVal e As System.EventArgs) Handles Me.Load

End Sub

Public Function doSomething(param1 as string, param2 as string) as String

End Function

End Class
[code]

So, when someone view this page in a web browser, my DLL should be able to grab the Request object, do something with it, and then pass control to the page. The page will do what it needs to do, and then pass control back to my DLL to issue the final Response object.

Is this possible? Any pointers would be appreciated.

Hey,

I think you are perhaps mixing terminology here.

You would create a Base Page Class, which your Web Form would inherit from. This class could certainly be contained within your own assembly though.

Here is an example:

http://www.4guysfromrolla.com/articles/041305-1.aspx

Hope that helps!

Gary

Not the way you describe it. You might be thinking of an HttpHandler or HttpModule, but a base class doesn't 'call' the child page in the way you're describing. What is the reason you are trying to handle the request, pass the request, then handle it again?

I'm trying to make my own webservice type process. Where the user defines functions in a class, and because it inherits my class, I can handle the output to the http stream. So, I think the better way is using Attributes. So the user can define something like:

<CustomAttribute()> _
Public Function doSomething() as String
Return "something"
End Function

Then, in my base class, somehow I can iterate over all functions the user has decorated with the <CustomAttribute> attribute and take the results, and spit it out via Response.Write or some http writer.

So the user's class, actually doesn't know anything about ASP.NET, it simply inherits from my base class, which handles communication via the HTTP context.

Now I just have to figure out how to do this.

I've seen and worked with frameworks like this before. Using a base class or a class inheriting from Page isn't really relevant here because by doing what you're doing, you're taking over the entire rendering of the page - therefore the page and its properties aren't being used and therefore aren't required. You can do this in a generic handler, a regular aspx (without markup) or an HttpHandler which invokes and reads the attributes on the methods. Since the return value is a string, all you're doing is an HttpContext.Response.Write. So the user's class can be a normal class with a bunch of methods in it that have the [CustomAttribute()] on it.

Quote:

---

Originally Posted by mendhak

I've seen and worked with frameworks like this before. Using a base class or a class inheriting from Page isn't really relevant here because by doing what you're doing, you're taking over the entire rendering of the page - therefore the page and its properties aren't being used and therefore aren't required. You can do this in a generic handler, a regular aspx (without markup) or an HttpHandler which invokes and reads the attributes on the methods. Since the return value is a string, all you're doing is an HttpContext.Response.Write. So the user's class can be a normal class with a bunch of methods in it that have the [CustomAttribute()] on it.

---

Yes, that's exactly what I'm talking about! But I'm still a little unclear here. Let me know if this is correct:

1) I create a base class that defines one or more custom attributes. This base class handles all requests/responses from/to the web browser.

2) The end user creates a regular class and inherits from my class.

3) The end user decorates any methods he wants exposed to my base class with the appropriate attributes

4) When a request comes through, I parse the query string of the URL, to determine what the client is requesting.

5) If I determine it is a function the client wants to call, I iterate through the methods that the end user has decorated.

6) I then get the return value from the function, and write it out via the http context.

Is that basically the idea?

Okay, so I was able to do the following:

Code:

---

Imports Microsoft.VisualBasic
Imports System.Reflection

Public Class CustomPageClass
    Inherits System.Web.UI.Page

    <AttributeUsage(AttributeTargets.Method, Inherited:=True, AllowMultiple:=True)> _
    Public Class CustomAttribute
        Inherits System.Attribute

    End Class

    Private Sub Page_Load(ByVal sender As Object, ByVal e As System.EventArgs) Handles Me.Load
        Dim methods As MethodInfo() = Me.GetType.GetMethods
        For Each method As MethodInfo In methods
            Dim att As WebMethodAttribute = Attribute.GetCustomAttribute(method, GetType(CustomAttribute))
            If att IsNot Nothing Then
                Response.Write(method.Name + "<br>")
            End If
        Next
    End Sub

End Class

---

Which gets all methods from derived class. I do see the method in there that I decorated with the custom attribute, but I'm wondering how to limit the results by GetMethods to just those methods that are decorated with a certain attribute?

[quote]
Yes, that's exactly what I'm talking about! But I'm still a little unclear here. Let me know if this is correct:

1) I create a base class that defines one or more custom attributes. This base class handles all requests/responses from/to the web browser.

Quote:

---

Yes

[qutoe]
2) The end user creates a regular class XXand inherits from my classXX.

---

My point is that they don't have to inherit from anything, the inheritance is a red herring/moot point. It won't make a difference. BBCode doesn't allow for a strikethrough. I tried [s]strike[/s] and [strike]strike[/strike] but can't do it.

Quote:

---

3) The end user decorates any methods he wants exposed to my base class with the appropriate attributes

4) When a request comes through, I parse the query string of the URL, to determine what the client is requesting.

5) If I determine it is a function the client wants to call, I iterate through the methods that the end user has decorated.

---

5) If I determine it is a function the user wants to call, I iterate through the methods that the developer has decorated. <-- You meant that right?

Quote:

---

6) I then get the return value from the function, and write it out via the http context.

Is that basically the idea?

---

Yep

Quote:

---

Originally Posted by dbassettt74

Okay, so I was able to do the following:

Code:

---

Imports Microsoft.VisualBasic
Imports System.Reflection

Public Class CustomPageClass
    Inherits System.Web.UI.Page

    <AttributeUsage(AttributeTargets.Method, Inherited:=True, AllowMultiple:=True)> _
    Public Class CustomAttribute
        Inherits System.Attribute

    End Class

    Private Sub Page_Load(ByVal sender As Object, ByVal e As System.EventArgs) Handles Me.Load
        Dim methods As MethodInfo() = Me.GetType.GetMethods
        For Each method As MethodInfo In methods
            Dim att As WebMethodAttribute = Attribute.GetCustomAttribute(method, GetType(CustomAttribute))
            If att IsNot Nothing Then
                Response.Write(method.Name + "<br>")
            End If
        Next
    End Sub

End Class

---

Which gets all methods from derived class. I do see the method in there that I decorated with the custom attribute, but I'm wondering how to limit the results by GetMethods to just those methods that are decorated with a certain attribute?

---

Not sure what you mean - aren't you doing that already by checking the methodattribute? You're looping through and inspecting each method in the class to see if it has your attribute on it

[QUOTE=mendhak;3666025]

Quote:

---

Yes, that's exactly what I'm talking about! But I'm still a little unclear here. Let me know if this is correct:

1) I create a base class that defines one or more custom attributes. This base class handles all requests/responses from/to the web browser.


My point is that they don't have to inherit from anything, the inheritance is a red herring/moot point. It won't make a difference. BBCode doesn't allow for a strikethrough. I tried [s]strike[/s] and [strike]strike[/strike] but can't do it.



5) If I determine it is a function the user wants to call, I iterate through the methods that the developer has decorated. <-- You meant that right?



Yep

---

But if the developer has created a class that should be handled by my class, and he doesn't inherit from my class, how will my class know this? Also, how would the developer decorate anything if there is no custom attribute defined?

Maybe I should make a little more clear what I'm trying to do. I'm basically trying to re-create my own "webservice", and trying to use the model that ASP.NET uses, unless you don't think that is good? When you create a web service with asp.net, it generates an ASMX page and corresponding VB code behind. The code behind, inherits from System.Web.Services.WebService. I believe that is what allows a request to the ASMX page to be parsed and then uses reflection to iterate through the VB code behind, to find any methods decorated with <WebMethod()>. I'm trying to do something like this. But I want to make it my own and do custom things with it.

Your class code doesn't care because of this line of code.

Dim methods As MethodInfo() = Me.GetType.GetMethods

The GetMethods can occur from anywhere, so you could do it against any other class.

Dim methods As MethodInfo() = SomeOtherClass.GetType().GetMethods()

See what I mean? All you need is [CustomAttribute()] on methods inside SomeOtherClass.

As for the developer knowing whether the attribute exists or not, that's a matter of referencing whatever assembly the CustomAttribute class is in.




Take your webmethod as an example. WebMethodAttribute just happens to be in System.Web.Services but importantly, it inherits from Attribute. Web service codebehind classes inherit from System.Web.Services.WebService. Attributes are a completely different area of the .NET framework as you can see.

Okay, I see what you mean regarding the attribute. It is available to the developer by simply referencing my assembly, not necessarily inheriting from. But, if the developer does not inherit, I have no knowledge of the developer's class that he created. So how would I use the SomeOtherClass.GetType().GetMethods()? This assumes I know the name of the SomeOtherClass class.

I'm slowly getting this but still a little fuzzy on the overall setup.

This will be an ASP.NET website, that will not necessarily have any actual ASPX pages, because I think this is what the HttpHandler is for. I can have the handler route the request context to my custom class, right? And in that custom class, based on what the user requested, I will then iterate over the developer's classes, looking for particular decorations, do some processing, and then return result via "response" object.

So say the requested URL (could be POST or GET) is:

http://somesite.com/customprocessor/doSomething

The HttpHandler, will capture this, and route to my custom class, and my custom class will see the path "/doSomething" and handle it appropriately.

Is this possible? Can the ASP.NET "engine" be attached to this, even though the user is not explicitly including a file extension like .ASPX?

Looks like someone else is trying to do the same thing, but never got a response:
http://www.vbforums.com/showthread.php?t=379364

I think the developer's class has to inherit from one of my classes, so that I know which class name to use when calling GetType.GetMethods. I just tried it without inheriting, and it doesn't return anything. This is what I would expect because I don't know the developer's class name. So I'm thinking this is why it must inherit, so that you're able to call SomeClassName.GetType.GetMethods, "SomeClassName" being my class that the developer inherits from. If the developer puts some arbitrarily named class such as "SomeCrazyClass", how will I know its name from my compiled dll?

Updated thread at:

http://www.vbforums.com/showthread.php?t=593648

Quote:

---

Originally Posted by dbassettt74

Okay, I see what you mean regarding the attribute. It is available to the developer by simply referencing my assembly, not necessarily inheriting from. But, if the developer does not inherit, I have no knowledge of the developer's class that he created. So how would I use the SomeOtherClass.GetType().GetMethods()? This assumes I know the name of the SomeOtherClass class.

I'm slowly getting this but still a little fuzzy on the overall setup.

This will be an ASP.NET website, that will not necessarily have any actual ASPX pages, because I think this is what the HttpHandler is for. I can have the handler route the request context to my custom class, right? And in that custom class, based on what the user requested, I will then iterate over the developer's classes, looking for particular decorations, do some processing, and then return result via "response" object.

So say the requested URL (could be POST or GET) is:

http://somesite.com/customprocessor/doSomething

The HttpHandler, will capture this, and route to my custom class, and my custom class will see the path "/doSomething" and handle it appropriately.

Is this possible? Can the ASP.NET "engine" be attached to this, even though the user is not explicitly including a file extension like .ASPX?

Looks like someone else is trying to do the same thing, but never got a response:
http://www.vbforums.com/showthread.php?t=379364

---

Yes, that's possible. Because the HttpHandler is an interceptor of sorts, you could theoretically run an entire website from it based on the incoming requests. Look at the process pipeline here:

http://thinkingindotnet.files.wordpr...versusiis7.gif

Request comes in, modules handle it, then it's on to the handler before the page. So you can consider, just for this discussion, that ASPX is a very specific handler for a specific URL. HttpHandlers are generic handlers that are broader. Because it's before ASPX, you have the option of not calling an ASPX at all.

Further, if you're going to do extensionless URLs, then you may want to consider using IIS7, because IIS6 doesn't do extensionless URLs, unless you do a wildcard mapping on the site so that all requests go to the ASP.NET ISAPI DLL.

As for the inheriting bit - do play around with it a bit to see what you can do.

I'm also going to answer a question you asked earlier; I personally don't think it's a good approach because I feel that the benefits are marginal rather than substantial. But hey, you've got your reasons and justifications and that's probably another discussion.

Okay, well that concerns me a bit. I definitely want portability and I want to make my framework assuming the end user/developer does not have access to IIS administration functions. But at the same time, I was trying to make my framework very generic and not "page" based, because no pages will be served from this particular directory of the web site. That is the other key. Only a particular directory will process my framework. So, even if the developer places pages in /processor, they will never show. Only, if they place files in other directories of their site. So I really wanted to employ the use of extensionless URLs, but maybe I can't now?

I suggest you experiment with it first. Create an HttpHandler that returns 'something'. Then in IIS 6 (not the built in web server), do a wildcard mapping. Then try

http://machinename/yourapp/dosomething

See if it gets hit. I believe this will work because it's also the way to get ASP.NET MVC to work in IIS6, but you lose performance.

Quote:

---

Originally Posted by mendhak

I suggest you experiment with it first. Create an HttpHandler that returns 'something'. Then in IIS 6 (not the built in web server), do a wildcard mapping. Then try

http://machinename/yourapp/dosomething

See if it gets hit. I believe this will work because it's also the way to get ASP.NET MVC to work in IIS6, but you lose performance.

---

When you say "do a wildcard mapping", does that mean on the IIS server as an administrator, or within the web.config of the web app? As I mentioned before, I need to assume the developer does not have any access privileges to the IIS server configuration/administration.

Quote:

---

Originally Posted by mendhak

I suggest you experiment with it first. Create an HttpHandler that returns 'something'. Then in IIS 6 (not the built in web server), do a wildcard mapping. Then try

http://machinename/yourapp/dosomething

See if it gets hit. I believe this will work because it's also the way to get ASP.NET MVC to work in IIS6, but you lose performance.

---

Hmmm... That URL doesn't seem to work in II6. In the local ASP.NET development server, it works, correctly calling my custom HttpHandler to handle the request. But in IIS6, it simply tells me "404 not found". Probably because it thinks I'm looking for a directory named /dosomething, when in fact it doesn't actually exist.

So maybe this is a problem? Again, I need to assume developer doesn't have access to IIS administration/configuration, beyond the web.config file. Is there maybe something else I can put in the web.config to make IIS6 handle this? This is what my HttpHandlers currently looks like:

Code:

---

<httpHandlers>
        <remove verb="*" path="*.asmx"/>
        <add verb="*" path="*" type="MyHttpHandler.Handler, MyHttpHandler"/>
        <add verb="*" path="*.asmx" validate="false" type="System.Web.Script.Services.ScriptHandlerFactory, System.Web.Extensions, Version=3.5.0.0, Culture=neutral, PublicKeyToken=31BF3856AD364E35"/>
        <add verb="*" path="*_AppService.axd" validate="false" type="System.Web.Script.Services.ScriptHandlerFactory, System.Web.Extensions, Version=3.5.0.0, Culture=neutral, PublicKeyToken=31BF3856AD364E35"/>
        <add verb="GET,HEAD" path="ScriptResource.axd" type="System.Web.Handlers.ScriptResourceHandler, System.Web.Extensions, Version=3.5.0.0, Culture=neutral, PublicKeyToken=31BF3856AD364E35" validate="false"/>
</httpHandlers>

---

Quote:

---

Originally Posted by dbassettt74

When you say "do a wildcard mapping", does that mean on the IIS server as an administrator, or within the web.config of the web app? As I mentioned before, I need to assume the developer does not have any access privileges to the IIS server configuration/administration.

---

Yes, on the IIS server as administrator. If the developer can't do it... tell them to get it done :P

Quote:

---

Originally Posted by dbassettt74

Hmmm... That URL doesn't seem to work in II6. In the local ASP.NET development server, it works, correctly calling my custom HttpHandler to handle the request. But in IIS6, it simply tells me "404 not found". Probably because it thinks I'm looking for a directory named /dosomething, when in fact it doesn't actually exist.

So maybe this is a problem? Again, I need to assume developer doesn't have access to IIS administration/configuration, beyond the web.config file. Is there maybe something else I can put in the web.config to make IIS6 handle this? This is what my HttpHandlers currently looks like:

Code:

---

<httpHandlers>
        <remove verb="*" path="*.asmx"/>
        <add verb="*" path="*" type="MyHttpHandler.Handler, MyHttpHandler"/>
        <add verb="*" path="*.asmx" validate="false" type="System.Web.Script.Services.ScriptHandlerFactory, System.Web.Extensions, Version=3.5.0.0, Culture=neutral, PublicKeyToken=31BF3856AD364E35"/>
        <add verb="*" path="*_AppService.axd" validate="false" type="System.Web.Script.Services.ScriptHandlerFactory, System.Web.Extensions, Version=3.5.0.0, Culture=neutral, PublicKeyToken=31BF3856AD364E35"/>
        <add verb="GET,HEAD" path="ScriptResource.axd" type="System.Web.Handlers.ScriptResourceHandler, System.Web.Extensions, Version=3.5.0.0, Culture=neutral, PublicKeyToken=31BF3856AD364E35" validate="false"/>
</httpHandlers>

---

---

When adding the wildcard mapping did you uncheck "Verify that file exists"?