Weblog Ton Stegeman [MVP]

I have written some weblog articles (ECMAScript and Managed models) and an article for the Dutch .NET Magazine on the new SharePoint 2010 client object models. Last week I did a presentation for the SDE event of the SDN on this subject. While presenting the Silverlight demo described in this blog post, one of the attendees was paying more attention than I did when writing the code. When I explained the code snippet below, the question was why I was using ExecuteQuery and I just explained the Silverlight API is asynchronous.

   1: ClientContext context = ClientContext.Current;

   2: Web web = context.Web;

   3: String listName = "CommentsList";

   4: ListCollection lists = web.Lists;

   5: IEnumerable<List> resultLists = context.LoadQuery(lists.Where(list => list.RootFolder.Name == listName));

   6: context.ExecuteQuery();

   7: List commentsList = resultLists.FirstOrDefault();

   8: if (commentsList != null)

   9: {

  10:     CamlQuery camlQuery = new CamlQuery();

  11:     camlQuery.ViewXml = "<View><Query><Where><IsNotNull><FieldRef Name='City'/></IsNotNull></Where></Query></View>";

  12:     _comments = commentsList.GetItems(camlQuery);

  13:     context.Load(_comments,

  14:          items => items.Include(

  15:             item => item.Id,

  16:             item => item["Title"],

  17:             item => item["Date"],

  18:             item => item["City"]));

  19:  

  20:     ClientRequestSucceededEventHandler success = new ClientRequestSucceededEventHandler(SuccesHandler);

  21:     ClientRequestFailedEventHandler failure = new ClientRequestFailedEventHandler(FailureHandler);

  22:     context.ExecuteQueryAsync(success, failure);

  23: }

In the presentation I did not have a very good answer to that question (other than ‘uhhhmmmm’). Time to sort it out and get a better answer. This MSDN article describes the asynchronous way of querying SharePoint 2010. This is why I am using this method to submit the CAML query. To get a reference to the SharePoint list, I copied a code snippet from another demo. I am using the RootFolder name to find the list that I want to use, and therefore I need an extra call to SharePoint. I could have used Lists.GetByTitle instead. So I am using an extra roundtrip to SharePoint, in a synchronous call, to get the reference to the list. So apparently the Silverlight client object model is also synchronous. After some more digging in the SDK I found the answer in the article Client Context as Central Object. This is a quote from that page:

“The Silverlight client object model provides both an ExecuteQuery() method, which can be called synchronously from threads that do not modify the user interface (UI), and an asynchronous ExecuteQueryAsync() method for cases where threads do modify the UI.”

Conclusion is the Silverlight client object model can handle both scenarios. If your code does not modify the UI directly, you can use the synchronous method. If your application directly updates the interface, you need to use ExecuteQueryAsync. In my application the snippet above is running in a background thread that does not update the UI, so both methods work. There was no real need to use the asynchronous method.

Thanks for paying attention in the session and asking the question!