This is my first post on Episerver development best practices. In this one, I'll present the best practices that should be followed while working with Episerver - ASP.Net MVC framework. 

This post focuses on content and object cache invalidation inside Optimizely CMS. It does not cover CDN cache, browser cache, or output cache.

Background

Most of the time, Optimizely handles cache invalidation automatically and very well. In normal publishing workflows, you usually do not need to clear cache manually.

That said, manual cache invalidation can still be useful in a few special cases: diagnostics, custom integrations, load-balanced proof-of-concepts, or controlled troubleshooting where you want to force content or object cache to refresh.

Below I have reordered the examples so the latest Optimizely CMS 12 and CMS 13 approach comes first, followed by older EPiServer examples for legacy projects.

Optimizely CMS 12 and CMS 13

For modern Optimizely CMS solutions, the recommended APIs are:

• IContentCacheRemover for content cache invalidation
• ISynchronizedObjectInstanceCache for general object cache invalidation

Clear the full content cache

using EPiServer;

public class CacheService
{
    private readonly IContentCacheRemover _contentCacheRemover;

    public CacheService(IContentCacheRemover contentCacheRemover)
    {
        _contentCacheRemover = contentCacheRemover;
    }

    public void ClearContentCache()
    {
        _contentCacheRemover.Clear();
    }
}

Clear the cache for a specific page or content item

using EPiServer;
using EPiServer.Core;

public class CacheService
{
    private readonly IContentCacheRemover _contentCacheRemover;

    public CacheService(IContentCacheRemover contentCacheRemover)
    {
        _contentCacheRemover = contentCacheRemover;
    }

    public void ClearSpecificContentCache()
    {
        _contentCacheRemover.Remove(ContentReference.StartPage);
    }
}

Clear a specific language branch

_contentCacheRemover.RemoveLanguage(ContentReference.StartPage, "en");

Clear the general object cache

If you previously used CacheManager, the newer approach is to use ISynchronizedObjectInstanceCache.

using EPiServer.Framework.Cache;

public class CacheService
{
    private readonly ISynchronizedObjectInstanceCache _synchronizedCache;

    public CacheService(ISynchronizedObjectInstanceCache synchronizedCache)
    {
        _synchronizedCache = synchronizedCache;
    }

    public void ClearObjectCache()
    {
        _synchronizedCache.Clear();
    }
}

Example endpoint for internal diagnostics

using EPiServer;
using EPiServer.Core;
using EPiServer.Framework.Cache;
using Microsoft.AspNetCore.Mvc;

[Route("util/cache")]
public class CacheController : Controller
{
    private readonly IContentCacheRemover _contentCacheRemover;
    private readonly ISynchronizedObjectInstanceCache _synchronizedCache;

    public CacheController(
        IContentCacheRemover contentCacheRemover,
        ISynchronizedObjectInstanceCache synchronizedCache)
    {
        _contentCacheRemover = contentCacheRemover;
        _synchronizedCache = synchronizedCache;
    }

    [HttpPost("content/clear")]
    public IActionResult ClearContentCache()
    {
        _contentCacheRemover.Clear();
        return Content("Ok, content cache cleared.");
    }

    [HttpPost("content/{id:int}")]
    public IActionResult ClearContent(int id)
    {
        _contentCacheRemover.Remove(new ContentReference(id));
        return Content("Ok, content cache cleared.");
    }

    [HttpPost("object/clear")]
    public IActionResult ClearObjectCache()
    {
        _synchronizedCache.Clear();
        return Content("Ok, object cache cleared.");
    }
}

Legacy EPiServer Examples

The examples below are for older EPiServer projects and are mainly useful when working with legacy CMS 6 to CMS 11 solutions.

Invalidate the cache for a specific EPiServer page

EPiServer 9.9+

var contentCacheRemover = ServiceLocator.Current.GetInstance<EPiServer.IContentCacheRemover>();
contentCacheRemover.Remove(ContentReference.StartPage);

EPiServer 6

DataFactoryCache.RemovePage(ContentReference.StartPage);

Invalidate the cache for an EPiServer site on a server

EPiServer 7

EPiServer.CacheManager.Clear();

EPiServer 6

EPiServer.DataFactoryCache.Clear();

A small web service to invalidate site cache:

public void ProcessRequest(HttpContext context)
{
    EPiServer.CacheManager.Clear();
    context.Response.ContentType = "text/plain";
    context.Response.Write("Ok, site cache cleared.");
}

Final Note

For most projects, it is best to let Optimizely manage cache invalidation automatically. Manual cache clearing should be reserved for troubleshooting, controlled utilities, and special integration scenarios.

Thanks to Wałdis Iljuczonok for previously highlighting the newer public API approach using IContentCacheRemover instead of older cache APIs.

Problem

I was importing a BACPAC generated on another server into my local development environment using SQL Server Management Studio and ran into the following errors:

TITLE: Microsoft SQL Server Management Studio
------------------------------

Could not import package.
Warning SQL0: A project which specifies Microsoft Azure SQL Database v12 as the target platform may experience compatibility issues with SQL Server 2014.
Warning SQL72012: The object [databaseXYZ_Data] exists in the target, but it will not be dropped even though you selected the 'Generate drop statements for objects that are in the target database but that are not in the source check box.
Warning SQL72012: The object [databaseXYZ_Log] exists in the target, but it will not be dropped even though you selected the 'Generate drop statements for objects that are in the target database but that are not in the source check box.
Error SQL72014: .Net SqlClient Data Provider: Msg 12824, Level 16, State 1, Line 5 The sp_configure value 'contained database authentication' must be set to 1 in order to alter a contained database. You may need to use RECONFIGURE to set the value_in_use.
Error SQL72045: Script execution error. The executed script:
IF EXISTS (SELECT 1
FROM   [master].[dbo].[sysdatabases]
WHERE  [name] = N'$(DatabaseName)')
BEGIN
ALTER DATABASE [$(DatabaseName)]
SET CONTAINMENT = PARTIAL
WITH ROLLBACK IMMEDIATE;
END

Error SQL72014: .Net SqlClient Data Provider: Msg 5069, Level 16, State 1, Line 5 ALTER DATABASE statement failed.
Error SQL72045: Script execution error. The executed script:
IF EXISTS (SELECT 1
FROM   [master].[dbo].[sysdatabases]
WHERE  [name] = N'$(DatabaseName)')
BEGIN
ALTER DATABASE [$(DatabaseName)]
SET CONTAINMENT = PARTIAL
WITH ROLLBACK IMMEDIATE;
END

(Microsoft.SqlServer.Dac)

Fix

Run the following T-SQL on the target SQL Server instance before importing the BACPAC:

EXEC sp_configure 'contained database authentication', 1;
GO
RECONFIGURE;
GO

Once that setting is enabled, rerun the import.

Why This Happens

The import process is trying to set the target database to partial containment:

ALTER DATABASE [YourDatabaseName]
SET CONTAINMENT = PARTIAL;

If contained database authentication is disabled on the SQL Server instance, that step fails and the import stops with SQL72014 and SQL72045.

This often happens when the BACPAC comes from Azure SQL Database and is being imported into a local SQL Server environment, where contained database authentication may be turned off by default.

Explanation

At first I suspected the issue might be caused by a corrupt BACPAC or Transparent Data Encryption (TDE), but the real cause was much simpler: the import required support for a contained database.

A partially contained database reduces dependencies on the master database and allows authentication and configuration to live more at the database level rather than relying entirely on server-level logins.

Because this has security implications, SQL Server does not always enable it by default on local or on-premises instances.

Notes

• SQL72014 and SQL72045 are the main errors causing the import failure.
• The SQL72012 warnings about data and log objects are not the root cause here.
• For most local development environments, enabling contained database authentication is enough to complete the import successfully.

Further Reading

• Contained database authentication server configuration option
• SqlPackage import documentation
• Import a BACPAC file to create a new database

Overview

If you need to upload files or entire folder structures to Amazon S3 from PowerShell, you no longer need to write your own recursive upload function unless you have very specific custom logic.

Today, there are two practical approaches:

• AWS CLI v2 – the simplest and most flexible option for most developers.
• AWS Tools for PowerShell – a good choice if you want to stay fully inside PowerShell cmdlets.

For most cases, I recommend AWS CLI v2.

Important Note About “Folders” in S3

Amazon S3 does not store folders the same way a normal file system does. S3 stores objects by key, and folder-like paths are created through key prefixes such as images/logo.png or releases/v1/app.zip.

That means you do not need to create folders manually before uploading files. If you upload an object to a key like photos/abc.png, S3 will display that path as a folder structure in the console.

Option 1. Use AWS CLI v2 (Recommended)

The modern way to upload files and folders to S3 from PowerShell is to use AWS CLI version 2.

This approach is faster, easier to maintain, and much better than building your own recursive upload script.

Install AWS CLI

Install the latest AWS CLI v2 on your Windows machine before continuing.

Configure Credentials

The old approach of hardcoding AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY in scripts is no longer a good default.

Instead, use a named profile:

aws configure --profile my-s3-profile

If your organisation uses AWS IAM Identity Center (SSO), use:

aws configure sso --profile my-sso-profile
aws sso login --profile my-sso-profile

This keeps credentials out of your script and makes your setup easier to reuse.

Upload a Single File

aws s3 cp "C:\Files\logo.png" "s3://my-bucket/assets/logo.png" --profile my-s3-profile

Upload an Entire Folder

For whole directories, sync is usually better than cp --recursive because it only uploads new or changed files.

aws s3 sync "C:\MyDirectory" "s3://my-bucket/releases/v1/" --profile my-s3-profile

Upload Only Certain File Types

aws s3 sync "C:\MyDirectory" "s3://my-bucket/images/" --exclude "*" --include "*.jpg" --include "*.png" --profile my-s3-profile

Preview Before Uploading

If you want to see what will happen before making changes, use --dryrun:

aws s3 sync "C:\MyDirectory" "s3://my-bucket/releases/v1/" --dryrun --profile my-s3-profile

Mirror a Folder Exactly

If you want S3 to match your local folder exactly, add --delete:

aws s3 sync "C:\MyDirectory" "s3://my-bucket/releases/v1/" --delete --profile my-s3-profile

Be careful: --delete removes files in the target that no longer exist in the source.

Use a Profile in PowerShell

If you do not want to pass --profile on every command, set it in your current PowerShell session:

$env:AWS_PROFILE = "my-s3-profile"
aws s3 sync "C:\MyDirectory" "s3://my-bucket/releases/v1/"

Option 2. Use AWS Tools for PowerShell

If you prefer native PowerShell cmdlets instead of the AWS CLI, you can use AWS Tools for PowerShell.

Install the S3 Module

Install-Module AWS.Tools.S3 -Scope CurrentUser

Use an Existing AWS Profile

Set-AWSCredential -ProfileName my-s3-profile

Upload a Single File

Write-S3Object -BucketName "my-bucket" -Key "assets/logo.png" -File "C:\Files\logo.png"

Upload a Folder Recursively

Write-S3Object -BucketName "my-bucket" -Folder "C:\MyDirectory" -KeyPrefix "releases/v1/" -Recurse

Which Option Should You Use?

• Use AWS CLI v2 if you want the most common and portable approach.
• Use AWS Tools for PowerShell if you prefer PowerShell cmdlets and are already working in PowerShell-heavy automation.

Final Thoughts

If your goal is simply to upload or sync files and folders to S3, the built-in AWS tools are much better than writing your own recursive function. They are faster, safer, easier to read, and easier to maintain.

My recommendation today would be simple: use aws s3 sync for folders, use aws s3 cp for single files, and avoid storing access keys directly inside scripts unless you have a very specific reason to do so.

Here is a simple recursive function to display all file-names with full path in folders and sub-folders. You can only run this command in Windows PowerShell available since the release of Windows 7.

$folder = "c:\\MyDirectory\\"

 Function Upload($item) {
    foreach ($i in Get-ChildItem $item)
    {
        Try
        {
            if((Get-Item $i.FullName) -is [System.IO.DirectoryInfo]){
                  Write-Output $i.FullName
                  Upload($i.FullName)

            }else{
              Write-Output $i.FullName
            }
        }catch{
            Write-Output $i.FullName
        }
    }   
}
 
Upload($folder)

Copy the above code and save in .ps1 file.

If you are not interested in external programs to renames all files in a folder to lowercase, there is a simple command for you.

- Open Command Prompt (cmd.exe) in Windows
- Go to the directory and run the following command

for /f "Tokens=*" %f in ('dir /l/b/a-d') do (rename "%f" "%f")

Note: This is not a recursive function, it will rename files to lowercase only on the directory where you will run the command.

If you need to read more about Yozio Sublink API, here is the documentation http://docs.yozio.com/articles/sublink-apis
Below is a code to create Yozio Sublink.

using System;
using System.Net;
using Newtonsoft.Json;

namespace Yozio
{
    public class YozioResult
    {
        public string status { get; set; }
        public Body body { get; set; }
    }
 
    public class Body
    {
        public string sub_link { get; set; }
        public object link_alias { get; set; }
        public MetaData meta_data { get; set; }
        public long timestamp { get; set; }
    }
 
    public class MetaData
    {
        public string utm_source { get; set; }
        public string utm_medium { get; set; }
        public string utm_campaign { get; set; }
    }
 
    public static class YozioApi
    {
        public static YozioResult SubLink
        {
            get
            {
                const string apiKey = "YOUR-YOZIO-API-KEY";
                const string yozioSuperLink = "YOZIO-SUPER-LINK"; //e.gk7.k.cf

                const string urlString = "http://api.yozio.com/v2.0/?app_key={0}&" +
                                         "method=sub.link.create&" +
                                         "short_url={1}&" +
                                         "reassign_old_alias_to_this_link=true&" +
                                         "meta_data[utm_source]=web&" +
                                         "meta_data[utm_medium]=link&" +
                                         "meta_data[utm_campaign]=blog";

                var url = string.Format(urlString, apiKey, yozioSuperLink);

                var client = new WebClient();
                client.Headers.Add("user-agent", "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.2; .NET CLR 1.0.3705;)");
                var json = client.DownloadString(new Uri(url));
                var result = JsonConvert.DeserializeObject<yozioresult>(json);
                return result;
            }
        }
    }
}

Usage: Getting Sublink

Yozio.YozioApi.SubLink.body.sub_link;

/Adnan

Follow the steps to batch convert Photoshop Files (PSD) into PNG files.

I was in the process of migrating over to Visual Studio 2015 from Visual Studio 2013. When I executed a local command line build, I received the following error.

Microsoft Visual Studio 2015 Version 14.0.23107.0.
Copyright (C) Microsoft Corp. All rights reserved.
Solution file 'xyz.sln' is from a previous version of this application and must 
be migrated in order to build in this version of the application. 
To migrate the solution, open the solution in this version of the application.
Migration completed successfully, but some warnings were detected during migration.
For more information, see the migration report:  UpgradeLog06.htm

I opened the solution using Visual Studio 2015, and got the migration report. I performed the solution clean followed by solution rebuild and everything works fine. On closing the solution, and performing a local command line build again I got the same error message.
Every time I open the solution using VS 2015 a new "Backup" folder is created with iteration and a new migration report is displayed.
The migration report shows 8 projects that have the following 5 warnings.

Visual Studio needs to make non-functional changes to this project in 
order to enable the project to open in Visual Studio 2015, Visual Studio 2013, 
Visual Studio 2012, and Visual Studio 2010 SP1 without impacting project behavior

The solution I was opening in Visual studio 2015 has around 35 projects and every-time a backup folder is created, it takes space in disk and the make changes in solution file. It was quite a hassle to check-in solution file to source control with backup folder information. So I started digging around how to stop this backup folder nonsense, and after some soul searching I finally fixed it.

Here is the solution:
- Firstly, open your project file (.csproj) in text editor.
- Find the following two lines as shown in the image and delete them.

- Then save the solution file.
Now, open your file in visual studio 2015 again. No more new backup folders and migration reports. It is safe to delete the old backup folders(s) and associated HTML file(s).

/Adnan

The simple way to get random item from an Array is to use the return value from random.next(0, array.length) as index to get value from the array.

var randomIndex = random.Next(0, Array.Length);
Console.Write(Array[randomIndex]);

The downside of the above code is it might return you item multiple times (repetition) as we don't keep track of items that we are getting from the source Array.

The easy approach is to consider Array as a deck of cards. We want the items to be 'shuffled' similar to a deck of cards, meaning avoiding any repetition. So we will use  a List<> for the source items, grab them at random and push them to a Stack<> to create the deck of items.

You can create a Stack from anything that is IEnumerable

var stack = new Stack(myList);

See MSDN: http://msdn.microsoft.com/en-us/library/76atxd68.aspx

However, the stack constructor will be using a loop internally, you just don't see it. So for understanding the purpose I will use an example to create Stack with loop.

public static Stack CreateShuffledDeck(IEnumerable values) {

var random = new Random();  var list = new List(values); var stack = new Stack();  while (list.Count > 0) {  // Get the next item at random. var randomIndex = random .Next(0, list.Count); var randomItem = list[randomIndex];  // Remove the item from the list and push it to the top of the deck. list.RemoveAt(randomIndex); stack.Push(randomItem ); }  return stack; } 

Now we have a solution to create a Shuffled Deck. We can now get random items out using Stack.Pop method . Popping something from the stack means "taking the top 'thing'" off the stack.

public static string[] RandomArrayEntries(string[] arrayItems, int count) {
var listToReturn = new List();

if (arrayItems.Length != count) {
var deck = CreateShuffledDeck(arrayItems);

for (var i = 0; i < count; i++) {
var arrayItems= deck.Pop();
listToReturn .Add(item);
}

return listToReturn .ToArray();
}

return arrayItems;
}

We can execute the above code as following:

var countriesArray = new string[] { "Sweden", "Pakistan", "United Kingdom", "Denmark", "Norway", "Finland" };

var newRandomAraay = RandomArrayEntries(countriesArray, 3);

/Adnan

This fix applies to legacy ASP.NET Web Forms or ASP.NET MVC applications running on .NET Framework and using DotNetOpenAuth. If you are building a new application on ASP.NET Core, use the built-in external authentication providers instead of this older pattern.

Problem

I was helping a friend wire up Facebook OAuth login in an older ASP.NET application using the DotNetOpenAuth extensions installed from NuGet.

The code looked like this:

Uri ui = new Uri("~/Login.aspx", UriKind.Relative);

var fbClient = new DotNetOpenAuth.AspNet.Clients.FacebookClient("***", "***********");
fbClient.RequestAuthentication(context, ui);

The problem is that RequestAuthentication expects an instance of HttpContextBase.

If you try to pass HttpContext.Current directly, it fails because HttpContext.Current is a HttpContext, not a HttpContextBase.

Why This Happens

HttpContextBase was introduced as an abstraction over HttpContext. This makes ASP.NET code easier to test and easier to work with in components that should not depend directly on the concrete runtime context.

To bridge the gap between the two types, ASP.NET provides HttpContextWrapper.

Solution

Wrap HttpContext.Current in a HttpContextWrapper before calling RequestAuthentication:

var httpContextBase = new HttpContextWrapper(HttpContext.Current);
fbClient.RequestAuthentication(httpContextBase, ui);

Explanation

HttpContextWrapper acts as an adapter. It takes the current ASP.NET request context and exposes it as a HttpContextBase, which is exactly what the DotNetOpenAuth API expects.

So if you are maintaining a legacy ASP.NET application and run into a type mismatch between HttpContext and HttpContextBase, this wrapper is the correct fix.

Modern Note

For new applications, this is no longer the recommended approach. In modern ASP.NET Core applications, external login providers such as Facebook are configured through the built-in authentication middleware, and System.Web, HttpContextBase, and HttpContextWrapper are not part of that model.

A common need I have in my ASP.NET MVC based projects is to render a complete "View" or "PartialView" to string instead of the HTTP response and then present it or embed it in another rendered view.

You can implement the following code in shared controller or preferably in base Controller so that you can access this function in all controllers across your project. You can access the ControllerContext within controller and pass it to the function. It will return the rendered view in HTML string, the usage is self-explanatory.

public static string RenderViewToString(string viewName, object model) 
{
 if (string.IsNullOrEmpty(viewName)) 
     viewName = ControllerContext.RouteData.GetRequiredString("action");

 ViewData.Model = model;
 using(StringWriter sw = new StringWriter()) 
 {
  ViewEngineResult viewResult = ViewEngines.Engines.FindPartialView(ControllerContext, viewName);
  ViewContext viewContext = new ViewContext(ControllerContext, viewResult.View, ViewData, TempData, sw);
  viewResult.View.Render(viewContext, sw);
  return sw.GetStringBuilder().ToString();
 }
}

ControllerContext can be access using following method.

ControllerContext.RouteData.GetRequiredString("action");

If you want to put the function in a helper class you have to pass the ControllerContext from controller to the function.

public static string RenderViewToString(ControllerContext context, string viewName, object model) 
{
 if (string.IsNullOrEmpty(viewName)) 
     viewName = context.RouteData.GetRequiredString("action");
 
 var viewData = new ViewDataDictionary(model);
 using(var sw = new StringWriter()) 
 {
  var viewResult = ViewEngines.Engines.FindPartialView(context, viewName);
  var viewContext = new ViewContext(context, viewResult.View, viewData, new TempDataDictionary(), sw);
  viewResult.View.Render(viewContext, sw);
  return sw.GetStringBuilder().ToString();
 }
}

Call the function in your Action Method

//Somewhere in HomeController
public ActionResult Index() 
{
 //Second parameter(model) can be null
 var context = ControllerContext.RouteData.GetRequiredString("action");
 var content = RenderViewToString(context, "profile", new ProfileModel());
 //var content = RenderViewToString("profile", new ProfileModel());

 
        //Do something with the content, e.g.get profile specific template and send it to e-mail


 //This does nothing to do with rendered string
 return View();
}

/Adnan