Brain Spew - Neville Mehta's Blog

Tech Ed 09

2009-09-06T14:49:00.002+10:00

So, it has been quite a while since I have blogged last, but I am planning on starting up again.
Im really looking forward to attending Tech Ed Australia 09 this week, being held on the Gold Coast. There are quite a few Readify guys presenting, so will be great to see them on stage doing what they do best!
I look forward to giving a few reviews of the presentations I watch this year. I am particularily keen to see a few of the .NET 4 ones!

Custom Actions in WSS 3 / MOSS 2007 - Part 1

2007-06-19T22:21:00.000+10:00

Something that really frustrates my about WSS 3 and MOSS 2007 is the lack of documentation on alot of the object model as well as some of the CAML. One of the areas that really wasn't very helpful at all was the section on custom actions, so I thought I would write this post.

To begin with, just a quick introduction...custom actions in SharePoint simply refer to the buttons (such as in the "Site Actions" menu) and some of the menus across the site (such as in "Site Settings"). The best way to deploy these custom actions is definitley through a Feature. This post will show the simple way to display these custom actions. In the next post I will show how to use custom controls for the menu buttons.

Anyways, onto the code (or XML in this case)!

In the first example, I will create a custom action for the "Site Actions" menu.
To do this (and for the sake of the rest of the examples) we will begin with creating our feature manifest. Here it is:

<?xml version="1.0" encoding="utf-8"?>
<Feature Id="65245A4A-5D65-45ea-896F-6A6A808F3C80"
Title="MyCustomActionFeature"
Description="This feature adds custom actions"
Version="1.0.0.0"
Scope="Site"
xmlns="http://schemas.microsoft.com/sharepoint/">
<ElementManifests>
<ElementManifest Location="MyCustomActions.xml"/>
</ElementManifests>
</Feature>

Here is what the MyCustomActions.xml file looks like to add an extra item to the Site Actions menu (Note that in most of the examples below, I have just left the Url blank because these are just examples, but this would be where you put the Url of where you want to link to):

<?xml version="1.0" encoding="utf-8" ?>
<Elements xmlns="http://schemas.microsoft.com/sharepoint/">
<CustomAction
Id="PeopleAndGroupsAction"
GroupId="SiteActions"
Location="Microsoft.SharePoint.StandardMenu"
Sequence="10"
Title="People and Groups"
ImageUrl="/_layouts/images/Actionscreate.gif">
<UrlAction
Url="/_layouts/people.aspx" />
</CustomAction>
</Elements>

Here is what the MyCustomActions.xml file looks like to add another custom actions menu item to the Actions menu of a list or document library.

<?xml version="1.0" encoding="utf-8" ?>
<Elements xmlns="http://schemas.microsoft.com/sharepoint/">
<CustomAction
Id="MyListActionsCustomAction"
GroupId="ActionsMenu"
Location="Microsoft.SharePoint.StandardMenu"
RegistrationType="List"
Sequence="10"
Title="My Actions Custom Action">
<UrlAction
Url="~site" />
</CustomAction>
</Elements>

Here is what the MyCustomActions.xml file looks like to add another section to the Site Settings page.

<?xml version="1.0" encoding="utf-8" ?>
<Elements xmlns="http://schemas.microsoft.com/sharepoint/">
<CustomActionGroup
Id="NewSiteSettingsGroup"
Location="Microsoft.SharePoint.SiteSettings"
Title="New Site Settings Group"
Sequence="10"
Description="" />
<CustomAction
Id="NewSiteSettingsGroupItem"
GroupId="NewSiteSettingsGroup"
Location="Microsoft.SharePoint.SiteSettings"
Sequence="10"
Title="My New Site Settings Item">
<UrlAction
Url="~site" />
</CustomAction>
</Elements>

Here is what the MyCustomActions.xml file looks like to add another section and custom action to the Central Administration page.

<?xml version="1.0" encoding="utf-8" ?>
<Elements xmlns="http://schemas.microsoft.com/sharepoint/">
<CustomActionGroup
Id="CustomCentralAdminSection"
Location="Microsoft.SharePoint.Administration.ApplicationManagement"
Title="Custom Central Admin Section"
Sequence="10" />
<CustomAction
Id="DoSomethingInCentralAdmin"
GroupId="CustomCentralAdminSection"
Location="Microsoft.SharePoint.Administration.ApplicationManagement"
Sequence="10"
Title="Do Something In Central Admin">
<UrlAction
Url="" />
</CustomAction>
</Elements>

As you can see, adding these actions is very easy. If you want to add a group like in the Site Settings and Central Admin examples, all you have to do is add the CustomActionGroup element and then use that ID for the GroupID in your CustomAction element.

There are also attributes such as RequiresSiteAdministrator and Rights which all some security trimming.

Also, if you are trying to find the appropriate value for the Location attribute, or GroupID attribute, there are a few listed in the SDK documentation. If you want more, you can simply look/search through the "Features" directory (%Program Files%\Common Files\Microsoft Shared\web server extensions\12\TEMPLATE\FEATURES\) for the CustomActions written by Microsoft. After all, they used the same method Ive shown above for alot of the menus and buttons you see.

Anyway, I hope you have got the idea of custom actions. In one of the posts I have coming up I will show how to use custom controls (and hence use the ControlAssembly and ControlClass attributes) to render your custom actions.

Setting up Reporting Services on MOSS 2007

2007-05-06T19:55:00.000+10:00

Liam Cleary has an excellent series of articles(1, 2 and 3, ) on using the Reporting Services add on for MOSS 2007. I finally got the chance to play around with this today and was following Liam's instructions. They were excellent, but I did come accross a few things I thought I would mention quickly.
Firstly, you wont see the "SharePoint Integration" option in the Reporting Services Configuration tool until you install SQL Server SP 2.
Secondly, if you get an error saying "Exception of type 'Microsoft.ReportingServices.RsProxy.AccessDeniedException' was thrown." when trying to deploy your report project, make sure your ReportServer virtual directory in IIS does not allow anonymous access (not sure how mine got that way because I dont think its the default, and I definitley didnt enable anonymous access).

Cannot get Membership Provider with name [MembershipProvider name] error in SharePoint 2007 / WSS 3

2007-01-26T12:01:00.000+11:00

Ive been doing some work with MOSS 2007 lately and came across this error:

Cannot get Membership Provider with name [MembershipProvider name]. The membership provider for this process was not properly configured. You must configure the membership provider in the .config file for every SharePoint process.

Just to save someone else the time of figuring out what the problem is, I thought I would let you know that the reason this is happening is because when you use a membership provider in SharePoint 2007 (or WSS 3) you must add the details about the membership provider in the web.config of the site you have created AS WELL AS in Central Admin so you can create site collection administrators etc.

Uninstalling SQL Embedded Edition after a WSS 3 installation

2007-01-16T10:17:00.000+11:00

If you ever try to install WS 3.0 on a single machine, an instance of SQL Server Embedded Edition is created, and this is will be your default for config and content databases. I wanted to use SQL Server 2005 Dev Edition. To do this, you must install WSS 3.0 with the farm option, not just a stand alone (single) option. But, the SQL Server Embedded databases dont uninstall. I looked around alot for info on uninstalling this, but eventually found this. Take a look at issue 33.

Reporting Services "rsReportServerDatabaseUnavailable" error when installing TFS

2006-12-31T14:17:00.000+11:00

I decided to install TFS today and after following the simple instructions on Cliff's notes for a team system install I came across a reporting services error halfway through the installation.
The error said something like "Cannot contact Reporting Services, make sure the database is installed and reachable blah, blah, blah" (obviously not a direct quote :-) ).

First thing I did was open the Reporting Services site in IE and found that there was an error saying that it couldnt access the location c:/inetpub/wwwroot/web.config "Access is Denied"...so I gave the NETWORK SERVICE account access to that location.
Then, I tried opening the site again and got a Reporting Services error saying the usual SQL Server error message...something like..."Cannot access SQL Server, make sure the server access remote connections blah, blah, blah"...but it did have the all important word: "rsReportServerDatabaseUnavailable" on it. I decided to Google that and found this post that fixed it.

What is the UseSynchronizationContext property in WCF

2006-12-02T15:08:00.000+11:00

The UseSynchronizationContext from the ServiceBehavior attribute in WCF is used to determine which thread your service will execute on.

Basically, if you have something like this on your service class at the time of creating your ServiceHost:

[ServiceBehavior(UseSynchronizationContext = true)]
class MyServiceClass: IMyServiceContract
{
// Implementation
}

then the value from the System.Threading.SynchronizationContext.Current is read and cached so that when a request comes to the service, the host can marshal the request onto the thread that the host was created on using the cached SynchronizationContext.

If you are familiar with the concept behind methods like Control.Invoke, Control.BeginInvoke() and Control.InvokeRequired() then you will be familiar with what the System.Threading.SynchronizationContext class is trying to achieve. Basically the SynchronizationContext.Send() and SynchronizationContext.Post() methods are similar to the Control.Invoke() and Control.BeginInvoke() methods.

If you try to host a service in, for example, a WPF application and also call that service from the same thread in the WPF application, you will notice that you get a deadlock when the client tries to call the service. The reason for this is that the default value of the UseSynchronizationContext is true and so when you create the ServiceHost on the UI thread of the WPF application, then the current synchronization context is a DispatcherSynchronizationContext which holds a reference to a System.Windows.Threading.Dispatcher object which then holds a reference to the current thread. The DispatcherSynchronizationContext will then be used when a request comes in to marshal requests onto the UI thread. But if you are calling the service from the UI thread then you have a deadlock when it tries to do this!!

Ive types this up pretty quick while its still in my head so hopefully it makes sense :-)

WCF Masterclass presented by Juval Lowy

2006-12-01T14:13:00.000+11:00

Juval Lowy is coming to Sydney and Melbourne to present IDesign's WCF Masterclass course. If you are serious at all about getting up to speed and ahead of the rest in WCF then you cant miss this opportunity!

Details can be found here and you can book here

Output Caching in ASP.NET 2

2006-11-29T11:54:00.000+11:00

Scott Guthrie has put up a great post on output caching in ASP.NET 2 and what he calls donut caching.

Controls.Clear() Doesnt Dispose In Win Forms!

2006-11-29T10:58:00.000+11:00

We are currently doing memory leak testing on the project I am on that uses Windows Forms in .NET 1.1, and an interesting thing has come up. We noticed that when calling Controls.Clear() none of the controls in the collection were being disposed and hence causing handles to be leaked.
If you open Reflector and see the System.Windows.Forms.Control.ControlCollection.Clear() method, you will notice that it simply calls RemoveAt(int index) on each control in the collection.
So how do we get around this problem of controls not disposing?
Well, do we simply have to iterate through the collection and dispose each control like below to clear the controls collection?

foreach (Control control in this.Controls)
{
if (control != null)
control.Dispose();
}

NO! This wont work.
What will happen is that only about half the controls will actually be disposed.
The reason for this is that the Dispose() method removes the control from its collection and shifts all the controls down one index in the collection, and so when the foreach operator trys to get the next control, it goes to the next index which is really two infront of what it started on before disposing the previous control.
So, unfortunatley what we need to do is something like:

for (int i = 0; i < Controls.Count; i++)
{
if (Controls[0] != null)
Controls[0].Dispose();
}

Now, I dont know if this has remained the same in .NET 2 or 3 or event ASP.NET...I will leave that to you to play around with.

Functional Programming Concepts and C# 3

2006-11-25T15:18:00.000+11:00

Readify collegue Andrew Matthews brought my attention to a fantastic article on functional programming with languages such as C-omega and F# and thier influence on some of the features found in C# 3.0. Definitley worth a read!

PathGeometry and StreamGeometry Mini-Language Syntax

2006-11-09T12:42:00.000+11:00

I have been playing around with geometry in WPF lately and came accross the following syntax for drawing a curve:



<Path
Data="M10,100 C 100,0 200,200 300,100 z"
Fill="Blue"
Stroke="Black"
StrokeThickness="4" />



This is confusing because the Data property of the System.Windows.Shapes.Path object takes an object dervied from System.Windows.Media.Geometry.

It turns out that this is a special syntax for defining StreamGeometry objects or Figures in a PathGeometry object.

in the above example the "M" is saying is saying start at the absoloute position of (10, 100) then the "C" is saying draw a cubic bezier curve starting at (100, 0) then go to (200, 200) and ending at (300, 100). The "z" simply says to end the curve here and draw a line from the starting position to the current point.

For a more complete explaination, go to:
Path Marup Syntax

C# 3.0

2006-11-06T17:33:00.000+11:00

I have also recently been quite interested to see whats coming up in C# 3 and came accross Sahil Malik's blog and his excellent series of posts on C# 3.0 which are short and to the point.

C Omega

2006-11-06T10:42:00.000+11:00

Daniel Crowley-Wilson has recently been playing around with C Omega, which has got me a little curious to know more about it...you can have a look at what its all about here.

Sound/Midi Not Working in Adobe Flash, YouTube etc. using CMedia sound card

2006-11-05T13:03:00.000+11:00

About two months ago, I was playing around with the codecs on my Windows XP machine...I cant exactly remember why, but I think I was trying to install a new DivX one. I installed a few differnet players and codecs until I found the one I wanted. Everything seemed to be fine...until I went to YouTube and noticed that I was not getting any sound. I thought it was an Adobe Flash problem, so reinstalled it, but still no luck. I found that I could play MP3s normally but couldnt play MIDIs or any Flash sounds. It appeared to be playing as normal but no sound came out. I thought it may be a sound card problem (C Media) so I reinstalled the drivers...no luck.

I noticed in Control Panel under "Sounds and Audio Devices" (Windows XP) that my "default device" for MIDI Music Playback was blank and also my "default device" was blank. Even if I selected an option from the drop down list and pressed "Apply" then "OK", the settings would not be saved and next time I go back in, it would be blank again.

I also noticed that the speaker icon in the bottom right hand corner of the screen (in the task tray) had disappeared as well.

I was getting really frustrated and googled the problem and found that this was a very old problem with no resolution.

After playing around and tweaking alot of settings, I finally found the solution with a little help from an old friend RegMon. I found that Windows was looking for two keys in the registry and couldnt find them. These were:

HKEY_LOCAL_MACHINE/SOFTWARE/Microsoft/Windows NT/CurrentVersion/drivers32/midimapper

HKEY_LOCAL_MACHINE/SOFTWARE/Microsoft/Windows NT/CurrentVersion/drivers32/wavemapper

It seems that one of the programs I installed then uninstalled when looking for the codec/player that I wanted must have deleted these keys.

Anyway, the solution is that these two registry keys should be set to the following string values:

HKEY_LOCAL_MACHINE/SOFTWARE/Microsoft/Windows NT/CurrentVersion/drivers32/midimapper should be set to midimap.dll.

HKEY_LOCAL_MACHINE/SOFTWARE/Microsoft/Windows NT/CurrentVersion/drivers32/wavemapper should be set to msacm32.drv.

If you are still having problems, you may need to reinstall some default audio codecs the instructions found here.

Finally, everything is back to normal again! I can now select my voice playback and midi playback default device and the speaker icon is back in the system tray...and I can watch Flash videos and listen to MIDIs.

DataGrid / DataGridView DataBinding

2006-10-14T19:25:00.000+10:00

I was recently playing around with binding to Windows Forms grids. I have never had to do anything really compilated with them before and have always just assumed that when I do, I will easily be able to figure it out on the spot.

Anyway, I cant remember why, but I wanted to bind a simple string array to a Win Forms DataGridView. I had always assumed that you would simply need to do something like this:

string customerNames = new string[] { "Jill", "Jack", "Johnny", "Jackson", "Jim" };

DataGridView dataGridView = new DataGridView();
dataGridView.DataSource = customerNames;

But, I was wrong! All that appeared in the grid was one column with a header that says "Length" and the rows would only contain the length of the strings, not the values of them....so I thought Ill try a list of strings and see what happens...

List listOfStrings = new List();

Same story...I got a column with the lengths of the string. By now I had caught onto what was happening. All the grid was doing was determining that the DataSource was an IList and so it simply got the objects from the list, reflected over the type and saw that Length was the only property and so displayed a column of length.
So how do you display an array of strings in a grid? The answer is, you cant just set the DataSource to the array like I did above, you have to create an object and then expose a property of type string and use that...something like this:

public class Customer
{
private string _customerName;

public void Customer()
{
}

public void Customer(string NewCustomerName)
{
_customerName = NewCustomerName;
}

public string CustomerName
{
get
{
return customerName;
}
set
{
_customerName = value;
}
}

Customer[] customers = new Customer[] { new Customer("Jack"), new Customer("Jill"), new Customer("Jim") };

DataGridView dataGridView = new DataGridView();
dataGridView.DataSource = customers;

What will happen here is that the grid will look through the array, pull out the Customer objects, reflect over them and see that there is a CustomerName property and display a column of customer names.

Ok, thats fine...but then out of curiosity I pulled out Reflector and decided to have a look at what happens when a DataSet or DataTable is set as the DataSource of the grid...something like this (note that the code below is only to show you something, you should probably be using a stringly typed dataset, and always check for nulls and that there are more than zero tables in the dataset etc.):

DataSet myDataSet = new DataSet();

/**** Populate dataset and do other stuff here ****/

DataGridView dataGridView = new DataGridView();
dataGridView.DataSource = myDataSet;
dataGridView1.DataMember = myDataSet.Tables[0];

Say the first table had two columns on it like "CustomerID" and "CustomerName" then what you would see in the grid are two columns named "CustomerID" and "CustomerName"...wait a second...how did the grid know that? There are no properties on the DataSet, DataTable, DataRow, DataView, DataRowView objects that have those names. So how did the grid just figure out that it had to display those columns when it couldnt figure out how to display my string array?

Well, lets take a look at some of the objects involved. The DataSet and DataTable both implement System.ComponentModel.IListSource. This interface has a property on it called ContainsListCollection and more importantly contains a method called GetList() which returns an IList.

When GetList() is called for the DataSet, it returns a DataViewManager, when GetList() is called for the DataTable, it returns a DataView. Both these classes implement the System.Collections.IList interface. Ok, we are getting closer...The DataView is actually a list of DataRowView object among other things.

Now what happens is, when the grid calls GetList() on the say...DataTable which is set to the DataMember property of the grid, it returns a DataView object. Since this implements IList, the grid knows it can iterate through the list and get a bunch of objects, in this case DataRowView objects.

Now, the DataRowView objects are whats returned, but when the grid reflects over these objects to create columns etc. there are no CustomerID properties or CustomerName properties, so how did it know to create these columns. Well, the DataRowView object implements an interface called System.ComponentModel.ICustomTypeDescriptor which has a GetProperties method. This is very important as the grid looks for this interface before reflecting over the type normally (using Reflection). If the object implements this interface then the grid will call its GetProperties method to obtain its properties and hence the columns. So, the DataRowView is smart enough to return the column names of the underlying DataTable to the grid so it will create those columns in the DataGridView instead of actually returning its properties like Row, RowVersion etc. (which can be returned using normal reflection anyway).

I hope this post gives you a basic insight into the way the DataGrid / DataGridView binds to objects.

Windows XP Shortcut Keys

2006-09-01T20:40:00.000+10:00

Here is a great article on shortcut keys in Windows XP. I never knew about Windows Button + U with the cool magnifying glass tool.

Windows XP Shortcuts

Tech Ed Gone For The Year...

2006-08-31T22:11:00.000+10:00

This time last week, I was at Tech Ed - Sydney. I know its pretty late now, but I thought I would still write something about it...

This was my first Tech Ed. I had heard about the previous years and that it was a great event, but had never got the chance to experience it. Im happy to say that it did live upto my expectations! The presentations I saw were great and were on things like TFS, WCF, Atlas, BizTalk, WF, SQL Server etc. ... I even managed to squeeze in an infrastructure presentation on Windows Vista/Longhorn Server Core (although, I have to admit that because of my lack of knowledge in this area, some of it went over my head).

Something that I did not anticipate about Tech Ed was the amount of people that I would run into. I have only been working in this industry for just over one and a half years, but its amazing how many people you make friends with. It goes to show that the community around this stuff is really pretty small. Not to mention the people I met for the first time.
I finally had the pleasure of meeting Frank Arrigo and Scott Guthrie, both of who's blog I have been reading for a while, so it was quite strange to meet them in person because you feel like you already know them so well.

Intro to System.Transactions

2006-06-20T14:03:00.000+10:00

The .NET Framework v2 has been around for a while now. Since I only started working in the world of .NET about a year and a half ago I have spent alot of time playing catch up and getting up to speed with v1.1 before I move onto v2, so there are still quite a few things in v2 I havent had a chance to look at in as much detail as I would have liked to.
So, recently I thought I would take a look at what the System.Transactions namespace is all about.
Here are the three links that were pretty much all I needed to get started, they talk about System.Transactions, Resource Managers, Transaction Managers etc.:

System.Transactions: Let's get transactional! Mitch Denny's Blog

System.Transactions and ADO.NET 2.0

Introducing System.Transactions in the .NET Framework 2

After reading the stuff above, one thing that confused me a little was...how many records would you expect to be written to the database with the code below:

Well, the answer is one. Only the second row (the one in the outerScope) will be written. Why? I hear you ask, because the innerScope is using the same transaction that the outerScope is using.

What's happening is, since we havent given any arguments in the constructor of the innerScope, the default behaviour is to either use a current transaction or to create a new one if none exists. This is the same as passing in TransactionScopeOption.Required to the innerScope constructor. The current transaction is created by the outerScope and so that is used. Again, this can be controlled by the constructors of the TransactionScope class.
When the end of the innerScope's using statement is reached, the status of the current transaction changes to Aborted and so, the second insert statement is no longer running within a transaction and so is written to the database even though outerScope.Complete() is not called.

You can check the status of the current transaction by using the
System.Transactions.Transaction.Current.TransactionInformation.Status property.

What would happen if I called outerScope.Complete() before the end of the outerScope using statement? Well, you would get a TransactionAbortedException and the second insert statement would still be written to the database.

First Post

2006-06-19T20:39:00.000+10:00

This is my first post ever...not quite sure what I should write to it, so how about a little about myself to start with.

Well...Ive been consulting for a pretty brilliant company called Readify for approximately the last 10 months.

At the time of writing this post, my main reason for starting this blog is so that I can write about all things .NET, SQL Server, BizTalk...Im sure you get where Im going with this.

Most of this stuff is still quite new to me so Im sure a lot of the posts you'll be seeing up here will be titled "Intro to ... ".
My reason for this is that when I am trying to learn about something, the first thing I type into a search engine is "Intro to ...". Hopefully, Ill be able to save someone having to search for hours and hours just trying to get started with something.