For Love of Software

Articles I read in 2011

2012-01-02T01:04:00.000+02:00

Welcome 2012 !!
As usual, I'm interested in sharing articles I read during the past year, hopefully you'll find something useful in it.

2010's list can be found here
2009's list can be found here
2008's list can be found here
2007's list can be found here

Linking using history.back()

2011-08-15T03:00:00.000+02:00

Easy navigation and discoverability are very important aspects of website or application usability. The user should be able to easily dig deep in the content as well as to navigate to broader category of content.

I stumbled upon a site that has a "Back" link in a leaf level content page. As a user I thought it might be a way to go up and visit the related content or the parent category. Unfortunately, history.back() was used. And since I reached this page using a search or a direct link and opened it in a new tab, it led me nowhere.

The designer / developer assumed that the user will reach this page using a specific path and will open it in the same window and tab.

That was a very bad idea as the most of the content nowadays is reached using search engines. And you don't your visitors to leave your site back to Google after getting him to your site!!. Also the design is broken in case the page is opened in a new window or tab.

Maybe the rationale behind using this technique was to get the user back (within the site) to the location from where he got to the current page, be it direct navigation links or internal site search page. If so, a better implementation should be using HTTP referrer, or explicit query string value to indicate the source of the request.

Small details can make big changes in usability. Different scenarios should be taken into consideration when designing the navigation.

PowerShell 32 and 64 bit have different execution policy settings

2011-07-09T14:46:00.007+02:00

I use PowerShell to automate many repetitive tasks. And build automation is one of the areas I like most.
I faced a stiuation when I get this error with a PowerShell script running in visual studio project post build event:

File XXX.ps1 cannot be loaded because the execution of scripts is disabled on this system. Please see "get-help about _signing" for more details.

I know this error is usually caused by an execution policy that denies execution of scripts. So I made sure the Execution Policy is set to RemoteSigned. But This did not work!!

I added this to the batch file that calls the PowerShell script:

powershell "Get-ExecutionPolicy -List"

And the result was:

Scope                         ExecutionPolicy

-----                         ---------------

MachinePolicy                               Undefined

UserPolicy                               Undefined

Process                               Undefined

CurrentUser                               Undefined

LocalMachine                               Undefined

After some research I found that since the machine is 64-bit, there were 2 versions of PowerShell, 32 and 64-bit. Again I edited the batch file adding :

Powershell.exe "Get-Variable PSHOME"

And ran a build from visual studio, the result was:


Name      Value                                      
----      -----                                     
PSHOME    C:\Windows\SysWOW64\WindowsPowerShell\v1.0

This shows that the version invoked was the 32-bit version, while the version I used to Set-ExecutionPolicy was the 64-bit version. I determined the paths from start menu shortcuts to Powershell:

Windows PowerShell (x86): %SystemRoot%\syswow64\WindowsPowerShell\v1.0\powershell.exe
And Windows PowerShell: %SystemRoot%\system32\WindowsPowerShell\v1.0\powershell.exe

So I opened Windows PowerShell (x86) and executed:

Set-ExecutionPolicy RemoteSigned

And it worked.

So, did Visual Studio post build event call the 32-bit version because it's a 32-bit application?

SSMS and Deleting all records from a self referencing table

2011-07-06T02:11:00.005+02:00

Clearing all data from a self referencing table using SSMS can be tricky. Selecting all records in SSMS results grid and pressing DEL just won't work in many times.

Take the common example of Employee - Manager relationship:

If you try using SSMS results grid to delete all records, you may get this error:

No rows were deleted.

A problem occurred attempting to delete row 1.
Error Source: .Net SqlClient Data Provider.
Error Message: The DELETE statement conflicted with the SAME TABLE REFERENCE constraint "FK_Employee_Employee". The conflict occurred in database "test", table "dbo.Employee", column 'ManagerId'.

The statement has been terminated.

Using the Query window and running:
DELETE employee
Will work however.

The difference is that SSMS actually tries to delete row by row, which will violate the constraint. But when using a query to delete all records. SQL Serve is smart enough to clear all data.

Failed to extract the cab file in the solution error in SharePoint

2011-01-07T21:11:00.000+02:00

When deploying SharePoint solutions, you may encounter an error message like:

Failed to extract the cab file in the solution.

(FYI, the wsp solution file is a cab file. just try to rename *.wsp to *.cab)
There are two reasons that I found causing this kind of error:

Having some special characters in files names in the solution, in my case it was "(" and ")".
Having two DLLs with the same name even if one is in the GAC folder and the other in bin folder, this happened in my case by mistake, but it's good to know anyway.

Have a nice deployment, without uninformative error messages.

Articles I read in 2010

2011-01-01T20:05:00.001+02:00

New year, new list :)
Seems I'm good at keeping old habits, I've even created the draft of 2011's list sending this post.

2009 list can be found here
2008 list can be found here
2007 list can be found here

LINQ to entities unncecessary casts

2010-04-04T23:43:00.007+02:00

Using LINQ with Entity Framework can give a big productivity boost. It removes the burden of O/R mapping and writing SQL queries. But Adding this layer can come with costs.
I was writing a desktop application that uses a SQL CE database and I used LINQ to entities to talk to the database.
One of the methods was slow and I suspected that the reason can be a slow query. In normal, hand-written SQL queries, I inspect query plans to check the reason for the bad performance. So I started to log queries generated by LINQ statements using:

(myQuery as ObjectQuery).ToTraceString()

Then I took the SQL query from log file to SSMS and I started to make some enhancements. One of them was adding an index to a smallint column used in a foreign key. But I discovered that the index was not used by SQL CE. That was strange, but looking to the where statement generated I found something like:

WHERE ( CAST( [Extent1].[PageNumber] AS int)) = ( CAST( @p__linq__468 AS int))

And the index on the PageNumber column (smallint) was not used as I found from execution plan that filtering is made after a table scan (not index seek).
I suspected that the cast used in the above query is what caused the optimizer not to recognize that the index should be used, and I verified this by removing the cast and running the query in SSMS again. This time the index is being used.
I started to Google about this issue. But found that SQL Server is smart enough not to be tricked by the silly cast. But since I'm using SQL CE. I had to find a way to fix this issue.
Removing the unnecessary cast was out of my hands, so I decided to change the data type of the column to int and sacrifice some space to gain some speed. I did that and updated the model. And the where clause became:

WHERE [Extent1].[PageNumber] = ( CAST( @p__linq__468 AS int))

No casts for the PageNumber column and the index is used.

When you use a high level technology, you take it with its pros and cons. But there are no excuses, you are responsible for what you ship to your client.

Splitting csv file based on content in one line using PowerShell

2010-04-01T16:06:00.007+02:00

Problem:
you have a csv file that contains department employees in a format like this:

Department,Employee
Sales,emp1
HR,emp2
Sales,emp3
Finance,emp4
Finance,emp5
Security,emp6
Security,emp7
Security,emp8
HR,emp9

And you need to split this file contents to separate files based on department name. So for the above example, we should get four files, Sales.csv, HR.csv, Finance.csv, and Security.csv. Each file contains only it's employees.
And the solution is really shows the power of PowerShell pipelining:

Import-Csv file.csv | Group-Object -Property "department" | Foreach-Object {$path=$_.name+".csv" ; $_.group | Export-Csv -Path $path -NoTypeInformation}

Dissecting the above commands:
Import-Csv file.csv:
Parses the csv file and returns an array of objects.

| Group-Object -Property "department":
Since we need to split by department, it makes sense to group objects by the department property.

| Foreach-Object {...}:
We need to apply an action for each group (department). So we pipeline the resulted groups to Foreach-Object

$path=$_.name+".csv":
Within the foreach, we need to create a temporary variable ($path) to be passed to the next pipeline responsible for the actual saving. Note that I use the semicolon ";" to separate this part from the next. And I used the name property of the group (which maps to department name in our case) to format the file name.

$_.group | Export-Csv -Path $path -NoTypeInformation:
Then for each group we have, we need to export its contents (csv file rows) to the file path created in the past step. So we again pipeline the group property of the group item (which is an ArrayList of original objects) to the Export-CSV Cmdlt.

And the result should be files like:

Finance.csv:
"Department","Employee"
"Finance","emp4"
"Finance","emp5"

CodeProject

Mounting file systems on Ubuntu

2010-04-01T13:54:00.008+02:00

I'm not an experienced Linux user. But I use Ubuntu (after trying redhat and Fedora) since it's supposed to be (Linux for human beings).
I have both Vista yes :( and Ubuntu on the same machine. And I regularly need to access ntfs file system with windows files when using Ubuntu.
When I open the file system from nautilus, I get this message, and I have to enter the root password.

This is very annoying. And although the prompting for the password can be suppressed (but this is another story). The real issue is that the file system is not automatically mounted when the system starts up. This means that when using an application that needs to access the hard drive (virtual box for example). I have to open the drive from nautilus first.
So the solution is to edit the /etc/fstab file that contains the information needed to mount volumes on startup.
This is OK, I opened a terminal, and ran sudo -i to run as root then made a folder to mount the volume under: mkdir /mount/DriveName
then gedit and opened /etc/fstab/ and added:

/dev/sda1 /media/DriveName ntfs rw,nosuid,nodev,allow_other,default_permissions,blksize=4096 0 0

restarted and I could access the file system without prompting for password. Then I started to test how I can access the file system (I'm a good developer and I do test my work). Everything looked OK. But when I try to delete a file:

Cannot move file to trash, do you want to delete immediately?

I checked the permissions and found that root is the owner and the group that has access. I also could write to .Trash-1000 (similar to $RECYCLE.BIN in windows)
The solution was to go back to fstab and adding the username I use as the owner:

/dev/sda1    /media/DriveName    ntfs    rw,nosuid,nodev,allow_other,default_permissions,blksize=4096,uid=username    0    0

And finally, I'm happy !!

Changing AJAXtoolkit:CalendarExtender day format

2010-03-08T23:06:00.005+02:00

AJAX toolkit has a set of awesome controls, and very customizable too. But one property that is not customizable OOB is how the day names are displayed in the Calendar extender.
The displayed format is two letters only. But what if you want it to be displayed in three letters?
Here a small cosmetic surgery (i.e. code changes).

Before	After

So, what does it take to do it successfully?
I opened CalendarBehavior.debug.js, which contains the debug version of the extender's JavaScript. Day names were obtained from an array named: dtf.ShortestDayNames:

dayCell.appendChild(document.createTextNode(dtf.ShortestDayNames[(i + firstDayOfWeek) % 7]));

Searching the library code, I found that this array is defined in MicrosoftAjax.debug.js. as:

ShortestDayNames":["Su","Mo","Tu","We","Th","Fr","Sa"]

And fortunately, another array was defined with three letters names:

AbbreviatedDayNames":["Sun","Mon","Tue","Wed","Thu","Fri","Sat"]

So all what it takes is to replace ShortestDayNames with AbbreviatedDayNames in both CalendarBehavior.debug.js and CalendarBehavior.js:

dayCell.appendChild(document.createTextNode(dtf.AbbreviatedDayNames[(i + firstDayOfWeek) % 7]));

And build.

It's great to have the source between your hands :)

My Twitter account

2010-02-06T10:18:00.002+02:00

I'm not a big fan of social networking sites. But I think twitter can be a great information sharing platform. Follow me, I don't intend to make so much noise anyway :)

Tip: turn auto close Off for you SQL Server databases

2010-01-29T23:52:00.004+02:00

On a development machine, I had a high CPU usage for the SQL server service. To understand the reason, I looked at the SQL Server logs (From SQL Server Management Studio - SQL Server Logs).

As you may notice from the screenshot above, the activity of starting the reporting service database was recurring again and again.

I checked the database options, and the noticed that the auto close property was set true. This explains why SQL Server tries to start the database at every connection.

If the Auto close property is set to on:

the database is closed and shut down cleanly when the last user of the database exits and all processes in the database complete.

As quoted from msdn.

Unfortunately, the property was set to true when the database was created during the SQL server installation.

Turning Auto close to false (off), and maybe restarting the service, returned the CPU usage of the SQL Server service to a normal level.

Articles I read in 2009

2010-01-01T01:00:00.134+02:00

A new year, and as a habit, I share ~~the~~ a list of articles I read in the past year, 2009.
It was a tough year on the personal, career and study levels, I wish I can contribute more to the community in the new year through blogging, article writing and open source software.

2008 list can be fond here
2007 list can be fond here

Getting started with SharePoint 2010

2009-11-12T06:53:00.002+02:00

SharePoint 2010 is coming with a lot of new features and if you are seriously interested in SharePoint, you should start build your 2010's knowledge. My colleague Mourad Askar posted on his blog about SharePoint 2010 Information and Tutorials which I find very informative.
On Microsoft's SharePoint 2010 site, you can find tutorials, videos and you can register for the beta.

One important thing, SharePoint 2010 will no run on 32Bit machines. If you have a core 2 due processor, it supports 64 bit. And if in doubt and want to run SharePoint on a virtual machine, check VMWare's tool that tests if your processor is capable of running a 64bit guest OS.
http://download3.vmware.com/software/wkst/VMware-guest64check-5.5.0-18463.exe

Personally, I have a core 2 Due Laptop with Vista 32 bit installed. I installed the 64 bit edition of Linux Ubuntu 9.04 and can host Windows 2008 64bit OS using Virtual Box. What's good about this scenario is that Ubuntu's memory usage is light and gives a good space to host virtual machines.

Using PowerShell and SMO to change database columns collations

2009-07-11T13:30:00.006+02:00

Changing a SQL serve database columns collations manually can be a tedious task. I have a database that I want to change the collation of all its non system columns to "Arabic_CS_AS".
Here is a PowerShell script that uses SQL Server Management Objects (SMO) to do this task:
(note that I load the assemblies with version 10.0.0 which is the version of SQL server 2008 I have installed on my system)


[System.Reflection.Assembly]::Load("Microsoft.SqlServer.Smo, Version=10.0.0.0, Culture=neutral, PublicKeyToken=89845dcd8080cc91")
[System.Reflection.Assembly]::Load("Microsoft.SqlServer.ConnectionInfo, Version=10.0.0.0, Culture=neutral, PublicKeyToken=89845dcd8080cc91")

$con = New-Object Microsoft.SqlServer.Management.Common.ServerConnection

$con.ConnectionString="Data Source=.\SQLExpress;Integrated Security=SSPI;"

$con.Connect()

$srv = New-Object Microsoft.SqlServer.Management.Smo.Server $con
$db = $srv.Databases["test"]

foreach ($table in $db.Tables)
{
  if($table.IsSystemObject)
  { 
      continue
  }
   
  foreach($column in $table.Columns)
  {
      if(-not ([string]::IsNullOrEmpty($column.Collation)))
      {
          $column.Collation = "Arabic_CS_AS"
          $column.Alter()
      }
  }
}

$con.Disconnect()

Not calling Dispose can cause InvalidComObjectException

2009-07-02T17:05:00.003+02:00

During load testing an application that performs thousands of operations against Active Directory and under high load conditions. The process stopped working and our logs showed this error:

System.Runtime.InteropServices.InvalidComObjectException: COM object that has been
separated from its underlying RCW cannot be used.

Searching for this error, most answers on forums referred to trying to access a COM object from a thread other than the thread that created it. We use multithreading, but we did not use objects across threads.

Logs pointed us the location of the code where we should investigate. I made a review on a method that was called thousands of times and creates DirectoryEntry instances. The DirectoryEntry was not disposed!!

We were not sure that this can cause the above exception, but it was a bug and it needed to be fixed anyway. We fixed it and reapplied the scenarios that caused this exception, and it disappeared.

Other that understanding a new reason for that mysterious exception, there are some useful lessons:

Proper logging can help identifying errors quickly.
Failing to dispose disposable objects causes performance penalties that some developers underestimate their effect. It can make your application stop working!!
Early code review is important to spot these kinds of errors.
Test your application under real-life conditions.
Using memory profilers and dispose trackers is worth trying in some cases.

More great things about Visual Web Developer 2008 Express

2009-05-13T23:21:00.007+02:00

In case you have not installed it yet: In addition to the great features Visual Web Developer 2008 Express has. SP1 is even much better!!

It now has more project tamplates. The most important one is the web application template (I hate the Website template).
You can also add class library project to the same solution!!

Download and enjoy!!

Downloading files using Windows PowerShell with progress information

2009-05-02T19:53:00.005+02:00

Downloading a file using PowerShell is very easy, you just need to call WebClient.DownloadFile method from the .net framework. But let's make things more interesting.

I need a script to download a list of files whose URLs are specified in a file to local folder. And the script should check if the file already exists. If it is, it should skip to the next file. To give a better user experience, a (progress bar) will be displayed to notify the user about the progress.
Error handling and reporting is important. So we'll take care of it inside the script.

Let's start analyzing how to accomplish this.

The code:
The first line of code defines the script parameters which are inputFile: the path of the file that contains the list of URLs and folder where we'll download files to.

param ([string] $inputFile,[string]$folder)

We make some basic input validation to check if parameters are set:

trap {Write-Host "Error: $_" -Foregroundcolor Red -BackGroundColor Black;exit}

if([string]::IsNullOrEmpty($folder))
{
throw "folder parameter not set";
}

if([string]::IsNullOrEmpty($inputFile))
{
throw "inputFile parameter not set";
}

The above code starts with defining an error handler that will run when a stopping error occurs in the script (in the current scope). It simply say: When an error occurs, write it to the host and exit the script.
Note that I pass Foregroundcolor and BackGroundColor parameter to the Write-Host Cmdlet so the user gets the same experience he gets with other PowerShell errors.
The next validation code simply check if the parameters are passed. If not, an error is thrown.

Next, we read the contents of the input file:

$files = Get-Content $inputFile -ErrorAction Stop

I use the Get-Content Cmdlet specifying the ErrorAction parameter = Stop. ErrorAction is a common parameter for PowerShell Cmdlets. The Stop value asks PowerShell to consider any errors fatal errors that will stop the script. But as I defined the trap handler as shown above, the same error handling will apply. The error message will be displayed and the script will exit.

The next line creates a WebClient object to be used in the download porcess:

$web = New-Object System.Net.WebClient

Next, I initialize a counter to be used in progress display based on the number of files downloaded so far. The a for loop is used to iterate on files.
Note that I define another trap handler that will act whenever an error is thrown within the for loop scope. It calls a specific function that handles download errors then continues the loop:

foreach($file in $files)
{
trap {ErrorHandler($_);continue}
.
.
.
}

Inside the loop, I create the download file path and display the progress using Write-Progress Cmdlet.

$path = [IO.Path]::Combine($folder,$file.SubString($file.LastIndexOf("/")+1));

Write-Progress -Activity "downloading" -Status $path -PercentComplete (($i / $files.Length)*100)

And if the file does not exit, DownloadFile is called.

if([System.IO.File]::Exists($path) -eq $False )
{
    $web.DownloadFile($file,$path)
}

I used the script to download presentations from the MIX 2009 conference. The attached ZIP file includes both the PowerShell script and the file that contains the URLs)

Complete code listing:


param ([string] $inputFile,[string]$folder)

function ErrorHandler($error)
{
Write-Host "Error while downloading file:$file" -Foregroundcolor Red -BackGroundColor Black
Write-Host $error -Foregroundcolor Red -BackGroundColor Black
Write-Host ""
}

trap {Write-Host "Error: $_" -Foregroundcolor Red -BackGroundColor Black;exit}

if([string]::IsNullOrEmpty($folder))
{
throw "folder parameter not set";
}

if([string]::IsNullOrEmpty($inputFile))
{
throw "inputFile parameter not set";
}


$files = Get-Content $inputFile -ErrorAction Stop

$web = New-Object System.Net.WebClient
$i = 0
foreach($file in $files)
{
trap {ErrorHandler($_);continue}

$path = [IO.Path]::Combine($folder,$file.SubString($file.LastIndexOf("/")+1));

Write-Progress -Activity "downloading" -Status $path -PercentComplete (($i / $files.Length)*100)



if([System.IO.File]::Exists($path) -eq $False )
{
 $web.DownloadFile($file,$path)
}

$i = $i+1
}

Oracle buys Sun !!

2009-04-22T23:17:00.009+02:00

OK, I think you already know this. But it's not the news I'm talking about. I'm talking about the implications.

I did not use Java technology a lot, I admire it, but I cannot consider myself as a Java developer. However, having Java technology in our world and it's competition with Microsoft's .net framework is a healthy thing. It gives us options.
Some Java developers were worried after news about the deal has spread. I think that Oracle + Java has been and will stay a good choice to build enterprise applications. Maybe the future will carry news about better and better RIA applications built with Java platform.

What I really care about and made me worried is the future of MySQL, which is the most popular open source database that has a free community edition.
How is this related to MySQL? Here is the story:

MySQL has a pluggable database engine architecture. So it has many database engines to choose from, and you deal with them all using the same SQL dialect.
One famous engine is the InnoDB engine, which is transactional and high performance database engine that was widely used by MySQL users. InnoDB was developed by Innobase, a Finnish company.
Oracle acquired Innobase, which caused worries about the future of the transactional engine. Sun bought MySQL. And MySQL released the Falcon storage engine.

Now, what is the future of MySQL under the control of Oracle, the database giant? The problem is not only that Oracle may kill MySQL gradually (I think it won't do it directly). But is that MySQL engineers started to leave Sun after Oracle's deal. (more about this in this article).

So. Big fishes eat small fishes. It's the turn of open source advocates and MySQL users to have their word.

Calling a PowerShell script in a path with a white space from command line

2009-04-18T19:37:00.005+02:00

I stuck in this problem once, so here is a solution in case you face it.
First, how to call a script from PowerShell console when the script file path contains white space? because executing this:
PS C:\> c:\new folder\myscript.ps1 param1
will give this error:
The term 'c:\new' is not recognized as a cmdlet, function, operable program, or script file. Verify the term and try again.

Putting the path between quotations like this:
PS C:\> "c:\new folder\myscript.ps1" param1
Will lead to:
Unexpected token 'param1' in expression or statement.

And the solution is to use the Invoke Operator "&", which is used to run script blocks
PS C:\> & 'c:\new folder\myscript.ps1' param1

So farm so good. Now coming to the next part which is calling this from command line.
Executing a PowerShell script from command line is as easy as:
C:\Documents and Settings\Hesham>powershell c:\MyScript.ps1 param1

This is fine as long as the script path has no spaces. For example, executing:
C:\Documents and Settings\Hesham>powershell c:\new folder\MyScript.ps1 param1
Again gives:
The term 'c:\new' is not recognized as a cmdlet, function, operable program, or script file. Verify the term and try again.

With the help of PowerShell -?, here is a solution:
C:\Documents and Settings\Hesham>powershell -command "& {&'c:\new folder\MyScript.ps1' param1}"

Tada!!

Why should you learn PowerShell?

2009-03-07T10:20:00.006+02:00

Whether you are a software developer, a tester, a system administrator, or even a regular user, PowerShell has something for you to offer.

It's amazing capabilities open many possibilities for you and it can be used in several scenarios:

For system administrators: quick and easy way to deal with the system in a consistent way. You'll have the power of many built-in commands, the .net CLR. Using it, you can perform several tasks like managing the file system and permissions, monitor even log, work with Active Directory. And so much more.
As a Developer: you can use PowerShell commands to automate some systems like exchange server from your application. Most Microsoft server products support or will support PowerShell as a programmable interface to automate the product. You can also use it to check environment issues if you have a production or testing environment issues that you suspect it's root cause to be environment related.
As a tester: PowerShell can be used for testing automation. With its easy to use commands and the simple syntax, I believe it's very suitable for this purpose. Have a look at: Why Should I Test With PowerShell?

If you want to get a high level image about what PowerShell can do for you and the flexibility it provides, you can watch this video by Jeffrey Snover the architect of PowerShell:

How to know Active Directory attribute names

2009-02-02T17:03:00.002+02:00

When dealing programmaically with active directory objects using .net code, VBScript or PowerShell, you need to set values of attribues you find in the "Active Directory Users and Computers" Snap-in (run dsa.msc). But these names are not always the same as the names used when setting attribute values in code. So how to know the attribute names?

I stumbled upon a nice msdn page that has the Mappings for the Active Directory Users and Computers Snap-in. It has has links to object type specific UI labels to attribute names.
For example, the User Object User Interface Mapping page shows that the Office UI label maps to physicalDeliveryOfficeName. How could you guess it?

Working with Active Directory using PowerShell

2009-01-30T10:45:00.005+02:00

Working with Active Directory is one of the important administrative tasks. VBSctipt was the most used language for administartors to automate repetitive tasks.
Now, windows PowerShell is the future, so it's important to know how to use it to work with Active Directory.
I'll provide a simple example that should clarify some concepts. In this scenario, it's required to set the email attribute of all users under a certain OU (Organaizational Unit) in the format: sAMAccountname@domainname.com and output the results to a text file.
PowerShell 1.0 does not have specific built in Cmdlets to handle active directory objects. But it has a basic support for [ADSI]. This will not limit us as we still can use the .net class library easly in PowerShell.
Here is how the code works:

First we decclare a variable that holds the output file path:
$filePath = "c:\MyFile.txt"

Then, create the root directory entry which represents the OU that we need to modify users under it. Note the LDAP: it tells: get the OU named "MyOU" from the domain "win2008.demo"
$rootOU=[ADSI]"LDAP://ou=MyOU,dc=win2008,dc=demo"

We need to get all users under this OU, so we create a .net directory searcher instance using New-Object Cmdlet
$searcher= New-Object System.DirectoryServices.DirectorySearcher

Setting the root of the search to the OU and the filter to find users only. and start to find all objects that match the filter:
$searcher.searchroot=$rootOU $searcher.Filter = "objectclass=user" $res=$searcher.FindAll()

Initializing the output file by writing the string "Emails"
"Emails:" Out-File -FilePath $filePath

Iterating on the results:
foreach($u in $res)

Getting the user object and setting the mail attribute, and committing:
$user = $u.GetDirectoryEntry() $name=$user.sAMAccountname $user.mail="$name@win2008.demo" $user.SetInfo()

Appending the mail to the output file (note the append parameter):
$user.mail Out-File -FilePath $filePath -append

You can save these commands to a .ps1 file and execute from PowerShell, for example:
c:\filename.ps1
note that you need to execute Set-ExecutionPolicy RemoteSigned first.

And here is the complete code listing, note that no error checking or handling is included for simplicity.


#Set-ExecutionPolicy RemoteSigned

$filePath = "c:\MyFile.txt"

$rootOU=[ADSI]"LDAP://ou=MyOU,dc=win2008,dc=demo"

$searcher= New-Object System.DirectoryServices.DirectorySearcher

$searcher.searchroot=$rootOU
$searcher.Filter = "objectclass=user"

$res=$searcher.FindAll()

"Emails:" Out-File -FilePath $filePath
foreach($u in $res)
{
$user = $u.GetDirectoryEntry()

$name=$user.sAMAccountname
$user.mail="$name@win2008.demo"
$user.SetInfo()

$user.mail Out-File -FilePath $filePath -append

$user.psbase.Dispose()


}
$rootOU.psbase.Dispose()
$res.Dispose()
$searcher.Dispose()

Articles I read in 2008

2009-01-01T01:00:00.000+02:00

It's a new year, and it's time to share a list of most articles I read the last year. Hopefully you find it interesting and useful.

2007 list can be fond here

Handling errors when calling PowerShell using C#

2008-12-05T21:18:00.002+02:00

In a previous post, I explained how to call PowerShell Commands from C# code in an easy way. But what about error handling?
First, we should know about the types of errors that need to be handled:

Exceptions due to error in syntax of the command or script.
Errors generated from commands themselves due to logical errors or invalid parameters.

Handling Exceptions:
Several exceptions can be thrown. The base for PowerShell exceptions is the System.Management.Automation.RuntimeException class. Exceptions can originate from bad syntax or calling invalid commands.
Handling this kind of errors follows the common exception handling in .net as this example shows:


       try
       {
           using (RunspaceInvoke invoke = new RunspaceInvoke())
           {
               result = invoke.Invoke("dir " + "c:\\" + " -recurse -Filter *.exe");
           }
       }
       catch (System.Management.Automation.RuntimeException ex)
       {
           Console.WriteLine(ex.Message);
           //Specific handling for PowerShell errors
       }
       catch (Exception ex)
       {
           //general handling and logging
       }

Command Errors:
These errors that commands return. For example, when you run this command in PowerShell while you don't have a Q drive:
dir Q:
You'll get this error:

Get-ChildItem : Cannot find drive. A drive with name 'q' does not exist.
At line:1 char:4

Which makes sense. But how to check for this kind of errors?
The Invoke method of the RunspaceInvoke class has an overload that accepts an IList as the 3rd parameter. This out parameter will contain a list of errors that has occurred.


              IList errors;
              using (RunspaceInvoke invoke = new RunspaceInvoke())
              {
                  result = invoke.Invoke("dir " + "Q:\\" + " -recurse -Filter *.exe", null, out errors);
              }

              if (errors.Count > 0)
              {
                  PSObject error = errors[0] as PSObject;
                  if (error != null)
                  {
                      ErrorRecord record = error.BaseObject as ErrorRecord;
                      Console.WriteLine(record.Exception.Message);
                      Console.WriteLine(record.FullyQualifiedErrorId);
                  }

                  return;
              }

In the above code, I cast the error to PSObject and get its BaseObject and cast it to ErrorRecord which contains the error information.

An interesting part here, is that you can check the FullyQualifiedErrorId property which is a string to distinguish errors and create logic to handle specific errors. ErrorDetails property can also be checked. But be careful because it can be null.

I hope this post can make your programming with PowerShell easier.