LegHumped.com » ASP.NET » Leghumped

Internal Subnav with BlogEngine.NET

Harry Jennerway — Tue, 03 Jan 2012 15:15:39 +0000

One of the great ASP.NET web apps I’ve started using is BlogEngine.NET, a fairly simple yet powerful CMS. While it handles basic CMS type functionality without problems, there are something which it doesn’t make particularly easy, such as displaying navigation trees.

This can be largely overcome by extending some of the controls which BlogEngine ships with, such as the post list. I changed this to display a list of sub pages for the current page. Once you set a page’s parent page, the subnav is rendered on the masterpage.

We just need one control on page.aspx and some code to bind it in the code-behind:

    private void ServePage(Guid id) {
 	...
  	if(pg.Parent == Guid.Empty)
        {
            // root page
            pageList.RootPageID = id;
        } else
        {
            // sub page of root level page
            pageList.RootPageID = pg.Parent;
        }

Out of the box, the html for the tree is generated once which won’t work if it’s different for every page, so you’ll also need to modify PageList.cs in the ~/App_Code/Controls folder. I kept the caching and replaced it with a dictionary for each page to map between page ID and HTML for the page list.

 static PageList()
        {
            BlogEngine.Core.Page.Saved += (sender, args) => PageList.ClearCachedHtml();
            ClearCachedHtml();
        }
 
        private static void ClearCachedHtml()
        {
            HttpContext.Current.Items.Remove("PageList");
        }
 
	//We also need a few more properties:
	#region Properties
        private static Dictionary<Guid,string> HtmlDictionary
        {
            get
            {
                const string cacheKey = "pageListDict";
                var htmlDict = HttpContext.Current.Cache[cacheKey] as Dictionary<Guid, string>;
                if (htmlDict == null)
                {
                    htmlDict = new Dictionary<Guid, string>();
                    HttpContext.Current.Cache[cacheKey] = htmlDict;
                }
                return htmlDict;
            }
        }
 
        private static string GetPageListForPage(Guid pageId)
        {
            string html;
            if (!HtmlDictionary.TryGetValue(pageId, out html))
            {
                html = BlogEngine.Core.Utils.RenderControl(BindPages(pageId));
                HtmlDictionary.Add(pageId, html);
            }
            return html;
        }
 
        public Guid? RootPageID
        {
            get
            {
                return ViewState["RootPageID"] == null
                           ? (Guid?) null
                           : new Guid(Convert.ToString(ViewState["RootPageID"]));
            }
            set { ViewState["RootPageID"] = value; }
        }
        #endregion

The root page ID specifies the parent pages whose children to render.

The final step is some changes to the private methods which render the lists:

public override void RenderControl(HtmlTextWriter writer)
        {
            writer.Write(GetPageListForPage(RootPageID??Guid.Empty));
           // writer.Write(Environment.NewLine);
        }
 
private static HtmlGenericControl BindPages(Guid? pageId)
        {
            //HtmlGenericControl div = new HtmlGenericControl("div");
            var div = new HtmlGenericControl("ul");
            div.Attributes.Add("class","softwareList");
            div.ID = "pagelist";
            foreach (BlogEngine.Core.Page page in BlogEngine.Core.Page.Pages)
            {
                if (page.ShowInList && page.IsVisibleToPublic && ((pageId.HasValue && pageId.Value == page.Parent) || (!pageId.HasValue && page.Parent.ToString() == "00000000-0000-0000-0000-000000000000")))
                {
                    div.Controls.Add(BuildSubPages(div, page));
                }
            }
            return div;
        }
 
        private static HtmlGenericControl BuildSubPages(Control parentElement, BlogEngine.Core.Page page)
        {
            List<BlogEngine.Core.Page> subPages = BlogEngine.Core.Page.Pages.Where(p => p.Parent == page.Id).ToList();
            HtmlGenericControl li = new HtmlGenericControl("li");
            HtmlAnchor anc = new HtmlAnchor();
            anc.HRef = page.RelativeLink.ToString();
            anc.InnerHtml = page.Title;
            anc.Title = page.Description;
            li.Controls.Add(anc);
            if (subPages.Any())
            {
                HtmlGenericControl ul = new HtmlGenericControl("ul");
                li.Controls.Add(ul);
                foreach (BlogEngine.Core.Page p in subPages)
                {
                    HtmlGenericControl innerLi = new HtmlGenericControl("li");
                    ul.Controls.Add(BuildSubPages(innerLi, p));
                }
            }
            parentElement.Controls.Add(li);
            return li;
        }

Though this should work for unlimited navigation levels, I’ve only tested it with two. Ideally, Umbraco is suited excellently to this due to it’s ability to use XSLT for jobs like this.

DevExpress Scheduler: ID of Repeating Appointment

Harry Jennerway — Sat, 01 Oct 2011 17:31:21 +0000

Since I started using the DevExpress Scheduler a few weeks back I’ve found a few limitations in the 2011 version. The killer for me was the lack of appointment IDs for recurring appointments.

After hunting around trying every suggested method from DevExpress, I found a support ticket from another developer having the same problem.

Using the same kind of method I wrote the following core (pretty horrible) to do the job.

	 runat="server"...
		...
		 AppointmentsSelectionChanged="function(s, e) { OnAppointmentsSelectionChanged(s, e.appointmentIds);}" />
        >
 
	 runat="server" ID="appointmentId"/>
         type="text/javascript">
            function OnAppointmentsSelectionChanged(s, appointmentIds) {
                $('#<%= appointmentId.ClientID %>').val(appointmentIds);
            }
        >

This lets you grab the value from the code-behind (in my case by hooking the ASPxScheduler1_PopupMenuShowing event).

protected void ASPxScheduler1_PopupMenuShowing(object sender, PopupMenuShowingEventArgs e) {
        ASPxSchedulerPopupMenu menu = e.Menu;
        // Check whether an empty cell is clicked rather than an appointment.
        if (!menu.Id.Equals(SchedulerMenuItemId.DefaultMenu)) {
		var rawSelectedBookingId = appointmentId.Value; // if it's recurring, you get something like 12_0
                var selectedBookingId = string.IsNullOrWhiteSpace(rawSelectedBookingId) ? -1 : Convert.ToInt32(rawSelectedBookingId.Split(new[] {'_'}, StringSplitOptions.RemoveEmptyEntries).FirstOrDefault());

The selectedBookingId in my case, is the ID of the database object for the user’s right click selection

jQuery: Geocoding for Zip Code Lookups

Harry Jennerway — Sun, 22 May 2011 15:43:50 +0000

Forms which need a user’s address, or even just their country can be tedious to fill out, especially if you need to collect latitude and longitude. Yahoo’s Geolocation API can simplify the process by looking up latitude, longitude and country from a postal/zip code. The function below takes five arguments and uses jQuery to add a click handler. When the button is clicked, an AJAX request is sent to the Yahoo geolocation API via jQuery, the the resulting XML is parsed and the form fields are populated.

This has the benefit of needing just one piece of data (postcode) instead of four. The one problem with this is that the same origin policy prevent making AJAX calls to a separate domain. Unfortunately, this means you need an intermediate web service hosted on the same domain as the page which passes on the request. This just needs to return a URL via cURL (PHP) or WebClient (ASP.NET).

Then in JavaScript, the result can be parsed from the returned XML (unfortunately, Yahoo’s API doesn’t currently support JSON).

	function addGLookupPostCodeListener(searchButtonId, postCodeFieldId, countryFieldId, txtLatitudeId, txtLongitudeId) {
	    var searchButton = $('#' + searchButtonId);
	    searchButton.click(function () {
	        var thePostCode = $('#' + postCodeFieldId).val();
	        if (thePostCode === '')
	            return;
	            $.ajax({type: 'POST', url: '/API.asmx/LookupPostcode',  contentType: "application/json; charset=utf-8", dataType: 'xml', data: JSON.stringify({ postCode: encodeURI(thePostCode)}), success: function (response) {
	                    var xml = response;
	                    var $xml = $(xml);
	                    var haveNodes = false;
	                    var $country = $xml.find("Country");
	                    if ($country != null) { // populat country if it exists
	                        haveNodes = true;
	                        var txtCountry = $('#' + countryFieldId);
	                        switch($country.text()){
	                            case 'GB':// the API returns GB as the country code which is no good 
	                            txtCountry.val('United Kingdom');
	                            break;
	                        default:
	                            txtCountry.val($country.text());
	                            break;
	                    }
	                    var $latitude = $xml.find("Latitude");
	                    var $longitude = $xml.find("Longitude");
	                    if ($latitude != null && $longitude != null) { // populate lat/long if they exist
	                        haveNodes = true;
	                        var txtLatitude = $('#' + txtLatitudeId);
	                        var txtLongitude = $('#' + txtLongitudeId);
	                        txtLatitude.val($latitude.text());
	                        txtLongitude.val($longitude.text());
	                    }
	                    if (!haveNodes) {// if no nodes (due to invalid response or no data), show a message
	                        searchButton.after('Address not found');
	                    }
	                }
	            });
	            return false;
	    });
	}

Draggable Resizable Div With jQuery

Harry Jennerway — Wed, 22 Dec 2010 16:30:01 +0000

I’m a big fan of jQuery, but so far I’ve only used it for effects. I recently did a draggable, resizable football shirt customization widget for an ASP.NET site and ran into some problems on the way. It’s not quite as simple as just calling $(‘#ele’).draggable().resizable(); due to a jQuery bug, so a bit of hackery is needed.

First, the controls to display the image which the user uploads.

     ID="pnlUserImageDragContainer" runat="server" CssClass="userImageContainer">
         ID="pnlUserImageContainer" runat="server" CssClass="userImageResizeContainer">
             ID="imgUserImage" runat="server" CssClass="userImage" />
        >
    >

So there’s a draggable panel, a resizable panel inside it and an image which scales inside that.

The JavaScript is registered when the page loads. When the image is dragged/resized, jQuery’s functions are used to grab the sizes of the panels and store them in textboxes on the form.

    StringBuilder jQueryBuffer = new StringBuilder()
	.AppendFormat(@"$('#{0}')", pnlUserImageContainer.ClientID).Append(@".resizable({  
                stop: function(e, ui) {
               		    //store the position of the resized image for later
                        var imageDiv = ").AppendFormat(@"$('#{0}'); ", pnlUserImageContainer.ClientID).Append(@"
                        ").AppendFormat(@"$('#{0}').val(imageDiv.width()); ", txtCustomImageWidth.ClientID).Append(@"
                        ").AppendFormat(@"$('#{0}').val(imageDiv.height()); ", txtCustomImageHeight.ClientID).Append(@"
 
            		    // turn the background off
                        imageDiv.css({ border: '1px solid transparent', background: 'none' });
                },
            }).parent().draggable({ // the parent is the outer panel (the drag container)
                    stop: function(e, ui) {
                        var containerPos = ").AppendFormat(@"$('#{0}').position(); ", pnlUserImageDragContainer.ClientID).Append(@"
                        ").AppendFormat(@"$('#{0}').val(containerPos.top); ", txtCustomImageTop.ClientID).Append(@"
 
                        // the position of the image inside the container (depends on the size as the inner container is centered)
                        var elementPos = ").AppendFormat(@"$('#{0}').position(); ", pnlUserImageContainer.ClientID).Append(@"
                        ").AppendFormat(@"$('#{0}').val(containerPos.left + elementPos.left); ", txtCustomImageLeft.ClientID).Append(@"
 
                    },
                    ").AppendFormat(" zIndex: '11', ").Append(@"
            }).mouseover(function() { 
                // this is just eye candy for an overlay when the mouse moves over the container
                var imageDiv = ").AppendFormat(@"$('#{0}'); ", pnlUserImageContainer.ClientID).Append(@"
                imageDiv.css({ border: '1px solid #f00', background: 'url(").Append(ResolveUrl("~/images/dragoverlay.png")).Append(@") repeat 0 0' })
            }).mouseout(function() { 
                // remove the eye candy onMouseOut
                var imageDiv = ").AppendFormat(@"$('#{0}'); ", pnlUserImageContainer.ClientID).Append(@"
                imageDiv.css({ border: '1px solid transparent', background: 'none' });
            });

The controls are ASP.NET controls not HTML elements because I access them from code, fundamentally it’s the JavaScript doing the legwork.

This lets a user drag the image around and resize it within the containing div.

Kentico: Global Content Sections

Harry Jennerway — Wed, 17 Nov 2010 16:19:16 +0000

I’ve been working with Kentico recently and I needed a section of text on the MasterPage ediatble by the admin, of which there was only one instance. The problem is, if you use a , you end up with one instance on every page as the property is stored on the page. In Umbraco you’d probably use a either page in the CMS specifically to hold the text or a new property on the root node.

In the end, I ended up creating a page in Kentico which was only used to hold the text for this section and using a literal on the MasterPage. You can load the page content with:

PageInfo contactInfo = PageInfoProvider.GetPageInfo(CMSContext.CurrentSiteName, "/", "en-GB", "/Global-Sections/Contact-Info", 369, true, CMS.DataEngine.ConnectionHelper.GetConnection());
litContactInfo.Text = contactInfo.EditableRegions["contentpanel"].ToString();

Subscriptions with Paypal Express Checkout

Harry Jennerway — Tue, 09 Nov 2010 11:26:03 +0000

The .NET documentation seems fairly sparese with examples for working PayPal API calls, so here’s my contribution. The CancelSubscription method may not be 100% working, but I believe the problem is with my Paypal account. You also need paypal_base.dll from the SDK. The methods are:

SetExpressCheckout – Sets the token and returns a URL to which to redirect the user.
GetExpressCheckoutDetails – Checks whether a token is valid on returning from PayPal
DoExpressCheckout – Initiates a payment
CreateRecurringPaymentsProfileCode – Creates a subscription with an optional initial amount
CancelSubscription – Cancels a subscription

/// 
/// ShortcutExpressCheckout: The method that calls SetExpressCheckout API
/// 
public bool SetExpressCheckout(double amt, ref string token, ref string retMsg)
{
    string host = "www.paypal.com";
    if (bSandbox) {
        pendpointurl = "https://api-3t.sandbox.paypal.com/nvp";
        host = "www.sandbox.paypal.com";
    }
 
    string baseUrl = "http://" + HttpContext.Current.Request.Url.Authority;
    string[] returnUrlParts = WebConfigurationManager.AppSettings["PaypalReturnUrl"];
             cancelUrlParts = WebConfigurationManager.AppSettings["PaypalCancelUrl"];
    NVPCodec encoder = new NVPCodec();
    encoder["METHOD"] = "SetExpressCheckout";
    encoder["RETURNURL"] = returnURL;
    encoder["CANCELURL"] = cancelURL;
    encoder["AMT"] = amt.ToString();
 
    encoder["LOCALECODE"] = "GB";
    encoder["CURRENCYCODE"] = "GBP"; 
    encoder["VERSION"] = "64.0";
    encoder["NOSHIPPING"] = "1";
    encoder["L_BILLINGTYPE0"] = "RecurringPayments";
    encoder["L_BILLINGAGREEMENTDESCRIPTION0"] = "SubscriptionForGenuineBusiness";
 
    string pStrrequestforNvp = encoder.Encode();
    string pStresponsenvp = HttpCall(pStrrequestforNvp);
 
    NVPCodec decoder = new NVPCodec();
    decoder.Decode(pStresponsenvp);
 
    string strAck = decoder["ACK"].ToLower();
    if (strAck != null && (strAck == "success" || strAck == "successwithwarning")) {
        token = decoder["TOKEN"];
 
        string ECURL = "https://" + host + "/cgi-bin/webscr?cmd=_express-checkout" + "&token=" + token;
 
        retMsg = ECURL;
        return true;
    } else {
        retMsg = "ErrorCode=" + decoder["L_ERRORCODE0"] + "&" +
            "Desc=" + decoder["L_SHORTMESSAGE0"] + "&" +
            "Desc2=" + decoder["L_LONGMESSAGE0"];
 
        return false;
    }
}
 
/// 
///  Checks whether the user has a valid token.
/// 
/// Whether or not the action suceeded.
public bool GetExpressCheckoutDetails(string token, ref string PayerId, ref string retMsg)
{
    if (bSandbox) {
        pendpointurl = "https://api-3t.sandbox.paypal.com/nvp";
    }
 
    NVPCodec encoder = new NVPCodec();
    encoder["METHOD"] = "GetExpressCheckoutDetails";
    encoder["TOKEN"] = token;
    encoder["VERSION"] = "64.0";
 
    string pStrrequestforNvp = encoder.Encode();
    string pStresponsenvp = HttpCall(pStrrequestforNvp);
 
    NVPCodec decoder = new NVPCodec();
    decoder.Decode(pStresponsenvp);
 
    string strAck = decoder["ACK"].ToLower();
    if (strAck != null && (strAck == "success" || strAck == "successwithwarning")) {
        return true;
    } else {
        retMsg = "ErrorCode=" + decoder["L_ERRORCODE0"] + "&" +
            "Desc=" + decoder["L_SHORTMESSAGE0"] + "&" +
            "Desc2=" + decoder["L_LONGMESSAGE0"];
        return false;
    }
}
 
/// 
///Starts payment for a set amount
/// 
/// 
/// 
/// Whether or not the action suceeded.
public bool DoExpressCheckout(string token, string PayerId, string amount, ref string retMsg)
{
    if (bSandbox) {
        pendpointurl = "https://api-3t.sandbox.paypal.com/nvp";
    }
 
    NVPCodec encoder = new NVPCodec();
    encoder["METHOD"] = "DoExpressCheckoutPayment";
    encoder["TOKEN"] = token;
     encoder["VERSION"]="64.0";
     encoder["PAYERID"]=PayerId;
     encoder["PAYMENTACTION"] = "Sale";
     encoder["AMT"]=amount;
     encoder["CURRENCYCODE"]="GBP";
 
    string pStrrequestforNvp = encoder.Encode();
    string pStresponsenvp = HttpCall(pStrrequestforNvp);
 
    NVPCodec decoder = new NVPCodec();
    decoder.Decode(pStresponsenvp);
 
    string strAck = decoder["ACK"].ToLower();
    if (strAck != null && (strAck == "success" || strAck == "successwithwarning")) {
        return true;
    } else {
        retMsg = "ErrorCode=" + decoder["L_ERRORCODE0"] + "&" +
            "Desc=" + decoder["L_SHORTMESSAGE0"] + "&" +
            "Desc2=" + decoder["L_LONGMESSAGE0"];
        return false;
    }
}
 
/// 
/// Create a PayPal subscription.
/// 
/// Will contain either an error message or the Paypal ProfileID
/// Whether or not the action suceeded.  will contain either an error message or the Paypal ProfileID
public bool CreateRecurringPaymentsProfileCode(string token, string amount, DateTime profileDate, string billingPeriod, string billingFrequency, ref string retMsg)
{
    NVPCallerServices caller = new NVPCallerServices();
    IAPIProfile profile = ProfileFactory.createSignatureAPIProfile();
 
    profile.APIUsername = this.APIUsername;
    profile.APIPassword = this.APIPassword;
    profile.APISignature = this.APISignature;
    profile.Environment = "sandbox";
    caller.APIProfile = profile;
    string host = "www.paypal.com";
    if (bSandbox) {
        pendpointurl = "https://api-3t.sandbox.paypal.com/nvp";
        host = "www.sandbox.paypal.com";
    }
 
    NVPCodec encoder = new NVPCodec();
    encoder["VERSION"] = "64.0";
    encoder["METHOD"] = "CreateRecurringPaymentsProfile";
    encoder["TOKEN"] = token;
    encoder["AMT"] = amount;
    //encoder["CURRENCYCODE"] = "GBP";
    encoder["PROFILESTARTDATE"] = profileDate.ToString("yyyy-MM-ddT00:00:00");   //Date format from server expects Ex: 2006-9-6T0:0:0
    encoder["BILLINGPERIOD"] = billingPeriod;
    encoder["BILLINGFREQUENCY"] = billingFrequency;
    encoder["DESC"] = "SubscriptionForGenuineBusiness";
    encoder["INITAMT"] = amount;
    string pStrrequestforNvp = encoder.Encode();
 
    caller = new NVPCallerServices();
    profile = ProfileFactory.createSignatureAPIProfile();
    profile.APIUsername = this.APIUsername;
    profile.APIPassword = this.APIPassword;
    profile.APISignature = this.APISignature;
    profile.Environment = "sandbox";
    caller.APIProfile = profile;
    string pStresponsenvp = caller.Call(pStrrequestforNvp);
 
    NVPCodec decoder = new NVPCodec();
    decoder.Decode(pStresponsenvp);
    string strAck = decoder["ACK"];
    bool success = false;
    if (strAck != null && (strAck == "Success" || strAck == "SuccessWithWarning")) {
        success = true; // check decoder["result"]
        retMsg = decoder["ProfileID"];
    } else {
        success = false;
        retMsg = "ErrorCode=" + decoder["L_ERRORCODE0"] + "&" +
            "Desc=" + decoder["L_SHORTMESSAGE0"] + "&" +
            "Desc2=" + decoder["L_LONGMESSAGE0"];
    }
 
    return success;
}
 
/// 
/// Cancels a subscription
/// 
/// Whether or not the action suceeded.
public bool CancelSubscription(string profileId, ref string retMsg)
{
    if (bSandbox) {
        pendpointurl = "https://api-3t.sandbox.paypal.com/nvp";
    }
 
    NVPCodec encoder = new NVPCodec();
    encoder["METHOD"] = "ManageRecurringPaymentsProfileStatus";
    encoder["PROFILEID"] = profileId;
    encoder["NOTE"] = "Subscription cancelled";
    encoder["ACTION"] = "Cancel";
 
    string pStrrequestforNvp = encoder.Encode();
    string pStresponsenvp = HttpCall(pStrrequestforNvp);
 
    NVPCodec decoder = new NVPCodec();
    decoder.Decode(pStresponsenvp);
    string strAck = decoder["ACK"];
    if (strAck != null && (strAck == "success" || strAck == "successwithwarning")) {
        return true;
    } else {
        retMsg = decoder[0] ?? string.Empty;
        return false;
    }
}

Blog4Umbraco and AJAX Comments

Harry Jennerway — Tue, 19 Oct 2010 12:58:48 +0000

Blog4Umbraco is probably one of the only ways to get a decent blog for Umbraco. I had a site which used some other JavaScript and lots of jQuery UI which I neeed to integrate, but as soon as I applied my MasterPage, the AJAX comment form broke. A quick search shows a few people have had this problem, so I want to post my fix. It’s almost embarrassingly simple

In /scripts/Blog4Umbraco/blog4umbraco.js, there’s a SetGravatar method, which makes what as far as I can tell is an invalid request. Instead of calling “/base/blog4umbraco/GetGravatarImage/” + s_email + “/80.aspx”, changing the path to “/base/blog4umbraco/GetGravatarImage/” + s_email + “.aspx” fix the problem in my case. Ie:

function SetGravatar(s_imgId, s_email, i_size){
    jQuery.get("/base/blog4umbraco/GetGravatarImage/" + s_email + ".aspx",
       function(data){
        var avatar = jQuery("#" + s_imgId );
        if(data != ""){
              var src = "http://www.gravatar.com/avatar/" + data + "?s=" + i_size;
              avatar.src(src).show();
          }else{
              avatar.hide();
          }
       });
}

Maybe it’ll work for you too.

Umbraco: Breadcrumb Trail

Harry Jennerway — Sun, 30 May 2010 14:00:18 +0000

I’ve been using Umbraco recently, and always find it hard to find XSL templates or spend hours tweaking one when countless other people must have done the same thing, but not published the finished stylesheet.

Below is a navigation breadcrumb template which prints a trail to the current page, starting with level 1.

 version="1.0" encoding="UTF-8"?>
 

	
    " ">
    "»">
]>
 

	version="1.0"
	xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
	xmlns:msxml="urn:schemas-microsoft-com:xslt"
	xmlns:msxsl="urn:schemas-microsoft-com:xslt"
	xmlns:umbraco.library="urn:umbraco.library"
	exclude-result-prefixes="msxml umbraco.library">
	 method="xml" omit-xml-declaration="yes" />
	 name="currentPage" />
 
	 match="/">
         class="breadcrumb">
             select="$currentPage/ancestor::node [string(data [@alias='umbracoNaviHide']) != '1']">
                 href="{umbraco.library:NiceUrl(@id)}">
                     name="title">
                         select="@nodeName" />
                    >
                     select="@nodeName" />
                >
                 test="@level != 0">
                    »
                >
            >
             select="$currentPage/@nodeName" />
        >
    >
>

Entity Framework in Mednium Trust

Harry Jennerway — Mon, 08 Mar 2010 18:47:56 +0000

ASP.NET and medium trust typically don’t get along well once you start using third party libraries. It really should, however, be easier to get the Entity Framework working on medium trust, and this had me baffled for a few hours trying to split the EDMX file and other fruitless solutions.

To get this working I needed to:

Move the EDMX to a separate assembly.
Set the EDMX metadata artifact processing to ‘Embed in output assembly’.
Add a reference in the site.

Change the connection string to

 name="ShopEntities" connectionString="metadata=res://*/;provider=System.Data.SqlClient;provider connection string="Data Source=servername;Initial Catalog=dbname;User Id=username;Password=P@55word;MultipleActiveResultSets=True"" providerName="System.Data.EntityClient" />

Which every guide claimed would work. There’s one more step if you have custom partial entity classes (in App_Code). If so, these also need to be moved to your output assembly, as you can’t split partial classes across assemblies. Lastly, change the namespace in your partial classes to the same as your output assembly, and everything should work.

ASP.NET: Remove from the cache using

Harry Jennerway — Sat, 25 Jul 2009 10:09:50 +0000

The Cache object in ASP.NET isn’t like other collections, probably because it needs to be specialized so it doesn’t make use of generics. Unfortunately, that makes it more difficult to work with. This small utility method removes an object from the cache by key using a predictate. You could easily modify it to compare either of the IDictionaryEnumerator’s Key or Value properties.

/// 
/// Removes an item from the cache using a Predictate to match the key.
/// 
/// The Predictate to use for each key to determine whether the entry should be removed.
/// 
/// Removes an item from the cache using a Predictate to match the key.
/// 
/// The Predictate to use for each key to determine whether the entry should be removed.
public static void RemoveFromCache(Predicate<string> keyCriteria) {
	IDictionaryEnumerator enuma = HttpContext.Current.Cache.GetEnumerator();
	while(enuma.MoveNext()) {
		if(keyCriteria(enuma.Key.ToString())) {
			HttpContext.Current.Cache.Remove(enuma.Key.ToString());
		}
	}
}