I came across a need for this when dealing with Guid columns in SQL Server. In code, Guid.Empty isn’t NULL, it’s “00000000-0000-0000-0000-000000000000”. This doesn’t go too well with the unique index of the table, which of course ignores NULL values but considers “00000000-0000-0000-0000-000000000000” a real value – which it is! So the challenge was to alter the LINQ query depending on one of the input parameters of the method. If the Guid parameter is equal to Guid.Empty, I want to query the database for NULL values in that column. If the Guid has a valid value, query the database using it the way it is.

I made this work by adding a conditional IF, but for readability I have split it up in an old fashion IF / ELSE. Define the LINQ query, and based on the input parameter “parentFolderId” define the WHERE statement using lambda :

var existingFolder = from f in db.Folders
             select f;

if (Guid.Empty.Equals(parentFolderId))
{
    // parent folder is empty, which is NULL in DB (Guid.Empty = '00000000-0000-0000-0000-000000000000')
    existingFolder = existingFolder.Where(p => p.FolderName == folderName && p.ParentFolderID == null);
}
else
{
    existingFolder = existingFolder.Where(p => p.FolderName == folderName && p.ParentFolderID == parentFolderId);
}

if (existingFolder.FirstOrDefault() != null)
    return false; // folder already exists

// create new folder

Remember that a LINQ query isn’t executed until the result set is accessed. In this case the database isn’t queried until the line where “existingFolder.FirstOrDefault()” is executed, and the desired result is achieved with minimal code.

Site path: C:\inetpub\wwwroot\Citrix\XenApp.
The Citrix XML Service object was not found: 404 Object not found. This message was reported from the XML Service at address http://xenapp.internaldomain.local:8080/scripts/ctxsta.dll  http://com.citrix.xml.CtxSTAProtocol.RequestTicket. The specified Secure Ticket Authority could not be contacted and has been temporarily removed from the list of active services.

Site path: C:\inetpub\wwwroot\Citrix\XenApp.
All the configured Secure Ticket Authorities failed to respond to this XML transaction.

Mostly, google will give you the answer. What I didn't see anywhere, is that the configuration is case sensitive.

So when you set up the web interface and Secure Gateway:

Wrong: http://servername01.domain.local:8080/scripts/ctxsta.dll
Correct: http://servername01.domain.local:8080/Scripts/CtxSTA.dll

When first initialized, the emulator creates a database in Microsoft SQL Server Express (which you can change to your local SQL Server if you want). It also sets up three services to emulate the Blob, Queue and Table types on different ports:

If you receive the following error message when attempting to start one of the services (typically the Blob service)

Process cannot access the file because it is being used by another process

it actually means that the port is taken, not a file. Run “Netstat /a” from your command prompt to find the culprit that’s taking up your precious port (which will probably be your torrent client – don’t tell RIAA!).

This blogpost’s main focus is not database optimization specifically, but more the importance of thinking and doing something about it even when using ORMs.

When creating Views in the database or writing SELECTs in general most developers remember to add INDEXES on the WHERE clauses, but when going through a O/R Mapper we sometimes forget. We have abstracted all of the database and our mindset is not "there".

Some examples:

var customer = _session.Query<Customer>()
                .Where(x => x.Email == “myemail@tempuri.org”);

or a more advanced one:

var linq = (from product in session.Query<Product>()
join category in session.Query<Category>()
on product.Category.Id equals category.Id
where category.CategoryName == "Condiments"
            && !product.Discontinued
select product).ToList();

The first example above is very simple and normally would take milliseconds to run. But what if we had millions of records in the database, then the query might take seconds. Particularly if we don’t have the Email column indexed in the database. In the second example maybe you don’t have a clue what kind of SELECT that was actually executed and then its hard to know what to optimize in the SQL Server…

Last example:

var blogpost = _session.Query<BlogPost>()
                .Where(x => x.Content.Contains("nHibernate"));

What is so special about this example? I guess you have used the Contains method hundred of times when working with Strings. But in this case it is on the ORM and the SELECT that is executed is most likely a ‘LIKE '”%nHibernate%’” who might trigger full table scan. (Depends on the scenarios, but a custom Full-Text Search might be vice here..)

To find out the SELECTs really executed on the SQL server you might have used the SQL Service Profiler and Database Engine Tuning Advisor. But most times we don’t. If you are using NHibernate they log all the queries on the DEBUG level using log4net.  But another perfect tool for this profiling is Ayendes NHProfiler/EFProfiler ++. The tools won’t just display the SELECTs but also help you doing them better.

As he writes on Efprof.com

"Entity Framework Profiler is a real-time visual debugger allowing a development team to gain valuable insight and perspective into their usage of Entity Framework. The product is architected with input coming from many top industry leaders within the OR/M community. Alerts are presented in a concise code-review manner indicating patterns of misuse by your application. To streamline your efforts to correct the misuse, we provide links to the problematic code section that triggered the alert."

So, when you are abstracting the database using various frameworks you should always know and care about your queries. If you don’t, you probably get issues in the production environments.

An operation failed because the following certificate has validation errors: [….] Errors: The root of the certificate chain is not a trusted root authority..

This might seem strange, because this will appear even if you have the CA certificate in your certificate store. The reason is that Sharepoint has it’s own registry of certificates, and you will have to add the CA there as well.

1. Open “ADFS 2.0 Management”
2. Expand Service – Certificates
3. Right click the primary (if more than one) certificate under Token-signing, and select View Certificate
4. Choose the Details tab, and click “Copy to file…”
5. Complete the wizard, saving the certificate as “DER encoded binary” (name it ADFSRoot.cer or something)
6. Copy the .cer file over to your Sharepoint server

Now you have to add this certificate to Sharepoints list of root authorities. You’ll be using the Sharepoint 2010 Management Shell for this operation:
1. Start Sharepoint 2010 Management Shell
2. Run the following two commands (change the path to where your .cer is located):

$root = New-Object System.Security.Cryptography.X509Certificates.X509Certificate2("C:\TEMP\ADFSroot.cer")

New-SPTrustedRootAuthority -Name "ADFS Token SigningRoot Authority" -Certificate $root

3. The certificate properties will be listed as a confirmation that the certificate has been added.

If your certificate chain contains more than one certificate, you will have to do this with each one giving them a unique name.

Thanks to Steve Peschka for providing an excellent guide.

When setting up XenApp 6.5 for the first time, there was absolutely no information on the database requirements to be found. The config-tool won't create the DB for you, so you'll have to do it manually, but info is hard to find.

I ended up booting my EVA that i downloaded for trial from Citrix and checking the collation on that database. Turns it was set to Latin1_General_CP1_CI_AS. So create your DB with that collation, and you should be good.
(If I remember correctly, this is the same collation as XenDesktop uses, which has equal little info. I remember the only place I found it was in the comments of the generated scripts during setup)

Other than collation, default settings should be fine. I set Recovery model to Full, but either way, it shouldn't have an impact.

If you don't know how to set collation, create the DB like this:
CREATE DATABASE CitrixXenAppDB COLLATE SQL_Latin1_General_CP1_CI_AS

Add a static method for handling the validation logic (or to skip it entirely):

// callback for validating SSL certificate during handshake
private static bool CustomCertificateValidatior(object sender,
    X509Certificate certificate, X509Chain chain,
    SslPolicyErrors policyErrors)
{
    // anything goes!
    return true;

    // PS: you could put your own validation logic here, 
    // through accessing the certificate properties:
    // var publicKey = certificate.GetPublicKey();

}

You then just add a RemoteCertificateValidationCallback before you make the SOAP / HTTP request, and your custom validator will take care of the validation.

// method where request is made

// ensure SSL certificate validation uses custom method
ServicePointManager.ServerCertificateValidationCallback +=
       new RemoteCertificateValidationCallback(CustomCertificateValidatior);

// initiate a SOAP or HTTP request like normal
// and your custom method will be used for validation

// ..

• at what time
• for how long
• how often
• the browser type
• http status
• and last but definately not least: by who.

And not only did you assume this would be easily accessible, you also told a client you’d have it done in a couple of hours. I assume some of you that ended up here either got lucky straight away, or you have already spent way more time than expected and added yet another element to your list of reasons why Sharepoint will make you loose your hair prematurely.

If you assumed that this data should be available somewhere you were right. The trick is to know how to access it from your custom web part. The data is stored in the WSS_Logging database, in a table called RequestUsage. You can run a simple query just to see what’s in there and which fields you have access to:

SELECT *
FROM [RequestUsage] 

Then you can start getting a bit more creative. You see, the RequestUsage table contains statistics for all sorts of requests – like requests to the MetadataWebService. So depending on your needs, filter out the type of document types / extensions / paths you are interested in. I needed just to know the hit count of .aspx pages in a specific sub site, so my query looked somewhat like this:

SELECT distinct(DocumentPath), COUNT(*) as DocCount
    FROM [RequestUsage]
    WHERE DocumentPath LIKE '/Marketing%'
      AND ( DocumentPath LIKE '%.aspx'   )
GROUP BY DocumentPath      

This is not a lesson in SQL, but that query gives me a nice list of all .aspx documents accessed under the Marketing sub site with the corresponding hit count. To get the top 5 out of that result, you’ll end up with something like this:

with cte as (
    SELECT distinct(DocumentPath), COUNT(*) as DocCount
    FROM [RequestUsage]
    WHERE DocumentPath LIKE '/Marketing%'
      AND ( DocumentPath LIKE '%.aspx'   )
    GROUP BY DocumentPath )
select top 5 DocumentPath, DocCount from cte order by DocCount desc

So we have a query we would like to run, and then you can display the result in a web part. There are numerous ways to do that, and I won’t cover it here but I’ll show the C# code for running this query. The following code snippet is put together from my actual project, and should give you enough information to piece together a working method. Take note of how the connection string for the WSS_Logging database is retrieved:

// run with elevated privileges
SPSecurity.RunWithElevatedPrivileges(delegate()
{
    string connectionString = SPUsageApplication.DefaultApplication.UsageDatabase.DatabaseConnectionString;
    SqlConnection connection = new SqlConnection(connectionString);

    // build your query
    StringBuilder sb = new StringBuilder();
    sb.Append("with cte ");
    sb.Append("as ");
    sb.Append("( ");
    sb.Append("SELECT distinct(DocumentPath), COUNT(*) as DocCount ");
    sb.Append("FROM [RequestUsage] ");

    // ...
    // add all the WHERE filters you want here
    // ...

    sb.Append(" GROUP BY DocumentPath ");
    sb.Append("");
    sb.Append(") select top " + ItemCount + " DocumentPath, DocCount from cte order by DocCount desc");
    string sql = sb.ToString();

    SqlCommand command = new SqlCommand();
    command.Connection = connection;
    command.CommandText = string.Format(sql, SPContext.Current.Web.ServerRelativeUrl);
    command.CommandType = System.Data.CommandType.Text;
    command.Connection.Open();

    SqlDataReader reader = command.ExecuteReader();
    while (reader.Read())
    {
        // access the fields, for example document name
        string docName = reader["DocumentPath"].ToString().Substring(reader["DocumentPath"].ToString().LastIndexOf('/') + 1);
        docName = HttpUtility.UrlDecode(docName);

        // build an HTML table, add items to a list or whatever you prefer
        // ..
    }

    // close the db connection
    command.Connection.Close();

}

The magic lies in how you build and define the filter to suit your requirements.

Using Automapper from Jimmy Bogard can in some best-cases reduce the code needed to only two lines of code. No matter how many properties exist in the object.

Automapper has moved to github and you can read more about it there, and it is also available as an nuget pacakge.

I will not go in-depth, but let me give you some examples:

I have a Customer object with Firstname and Lastname. I also have an CustomerViewModel object with the same properties. To verify i have wrote the test below. Linenumber 5 and 6 are the only lines of code needed to perform the map.

Basic

 1: [Test]

 2: public void WhenMappingCustomerToViewModel_ThenAllPropertiesShouldBeMapped()

 3: {

 4:     var customer = new Customer{Firstname = "John",Lastname = "Johnsen"};;

 5:     Mapper.CreateMap<Customer, CustomerViewModel>();

 6: 

 7:     var customerViewModel = Mapper.Map<Customer>(customer);

 8: 

 9:     Assert.AreEqual(customer.Firstname,customerViewModel.Firstname);

 10:     Assert.AreEqual(customer.Lastname,customerViewModel.Lastname);

 11: }

AssertConfigurationIsValid

I will give you some more advanced examples later on in this posts, I will just start telling you about Automappers method “AssertConfigurationIsValid”. A very simple test that you always should add in your tests suite:

 1: [Test]

 2: public void WhenMappingCustomerToViewModel_ConfigurationShouldBeValid()

 3: {

 4:     Mapper.CreateMap<Customer, CustomerViewModel>();

 5:     Mapper.AssertConfigurationIsValid();

 6: }

AssertConfiguraitonIsValid will raise an exception if some of the properties in CustomerViewModel is NOT mapped from the Customer object. If for example only the CustomerViewModel has an Age property the given exception will be raised:

AutoMapper.AutoMapperConfigurationException : The following 1 properties on ns+CustomerViewModel are not mapped:
Age
Add a custom mapping expression, ignore, or rename the property on ns+Customer.

Lower- UpperCase does not matter

In Javscript and Json it is a standard to have lowercase properties. So if my CustomerViewModel had properties firstname and lastname and Customer has FirstName and LastName everything will just work. I do not have to change the mappingclass.

Flattening

If the Customer has an object of type Address, and Address has the property Line1, then this value will be automatically mapped if the CustomerViewModel has an property called AddressLine1. Eks Customer.Address.Line1 will be automatically mapped to CustomerViewModel.AddressLine1.

Projection

The Customer might have a property DateOfBirth of type DateTime, but in CustomerViewModel I have a property called Birthdate of type string. And I want this string to be in the format ddMMyyyy. In these cases i have to add a line when calling Mapper.CreateMap()

Mapper.CreateMap<Customer, CustomerViewModel>()
    .ForMember(d => d.BirthDate, o => o.MapFrom(s => s.DateOfBirth.ToString("ddMMyyyy")));

Mapping a IEnumerable<T> (added later because of specific request in email)

The framework support mapping of IEnumerable<T> as long as you have created a map for T.

 1: [Test]

 2: public void WhenMappingAListOfCustomersToViewModel_ThenAListWithAllTheViewModelsShouldBeMapped()

 3: {

 4:     var customer = new Customer{Firstname = "John",Lastname = "Johnsen"};

 5:     var customer2 = new Customer{Firstname = "hans",Lastname = "hansen"};

 6:     var customList = new List<Customer> { customer, customer2 };

 7: 

 8:     Mapper.CreateMap<Customer, CustomerViewModel>()

 9:         .ForMember(d => d.BirthDate, o => o.MapFrom(s => s.DateOfBirth.ToString("ddMMyyyy")));

 10: 

 11:     var customerViewModelList = Mapper.Map<IEnumerable<Customer>>(customList);

 12: 

 13:     Assert.AreEqual(customList.Count(), customerViewModelList.Count());

 14: }

So to summarize, Automapper is really powerful and can save you many hours of boring mapping implementation.

As Webstep guru Thor Halvor explained in the this excellent blogpost – there are security restrictions to prevent/limit cross domain access of XMLHttpRequest’s – the cornerstone of AJAX.

Flash and silverlight has the same restrictions – and solves this by using crossdomain.xml and clientaccesspolicy.xml. These files are placed on the server you want to communicate with – and must contain * or the domain you want to contact the server from.

Anyways – there is a similar mechanism that XMLHttpRequest supports. This mechanism is called CORS – Cross-origin resource sharing. It is a newer(2004) and preferred alternative to JSONP – and works more or less like the xml files mentioned above. The only difference is that it isn’t implemented as a file – it’s part of the HTTP Header. This makes it a bit more difficult to set up than the others.

When a javascript on siteA wants to make a request to siteB – then the script first makes an initial OPTIONS request to site B – and looks at the HTTP Header it receives.

Access-Control-Allow-Origin: *

If the value is * – then it means that XmlHttpRequests can communicate with that site – from any other server – and a regular XMLHttpRequest can be made just like you were communication with your own server. You can of course type in domain names here to prevent everybody from using your API.

Here’s how to configure this on an Microsoft IIS Server – Web.Config – under the <configuration> node

<system.webServer>
  <httpProtocol>
    <customHeaders>
       <add name="Access-Control-Allow-Origin" value="*" />
    </customHeaders>
  </httpProtocol>
</system.webServer>

So to sum it up: use CORS whenever possible, instead of hacking your way around with JSONP. You’ll have prettier code, better error handling and it’s safer to use with regard to XSS Attacks as far as I have understood. Also – CORS supports all types of HTTP requests (Get/Post/Put,Delete), while JSONP only supports Get.

Read more about CORS here:
http://my.opera.com/core/blog/2011/10/28/cors-goes-mainline

A third (and the newest) alternative is UMP – I might blog more about this some other time.