Changing desired access rights to KEY_ALL_ACCESS solves the problem. Currently I do not know why, but it works in my case.

There were no any signs of errors. Ajax controls just were not functioning. Digging deeper, we found, that JavaScript is throwing exceptions:
SCRIPT5022: AjaxControlToolkit requires ASP.NET Ajax 4.0 scripts. Ensure the correct version of the scripts are referenced. If you are using an ASP.NET ScriptManager, switch to the ToolkitScriptManager in AjaxControlToolkit.dll.

This is the most common error with AJAX Control Toolkit — you need to use ToolkitScriptManager instead of ScriptManager. Read more here.

But in our case we was already using ToolkitScriptManager in the all places in our source code.

Digging deeper we found that there is a bug in Script Manager, that is trying to load all DLLs in the project’s bin folder. Script Manager is trying to load DLL files even if they are not used in the project. [I know, that it is not best practice, to keep files not used by project on the production server, but that's another story.]

The solution was plain and simple: remove all unused DLLs from production machine’s bin folder.

More info about bug and the same problem in different product:

• Custom Attribute Load Failure in GetAjaxFrameworkAssemblyAttribute from MS Connect
• Compatibility issue using ASP.Net 4.0, Extensions and ScriptManager from WebSync @ Google Groups

Below are details from Wikipedia:

Forward-declaring enums is also possible in C++11. Previously, enum types could not be forward-declared because the size of the enumeration depends on the definition of its members. As long as the size of the enumeration is specified either implicitly or explicitly, it can be forward-declared:

enum Enum1; // Illegal in C++03 and C++11; the underlying type cannot be determined.
enum Enum2 : unsigned int; // Legal in C++11, the underlying type is explicitly specified.
enum class Enum3; // Legal in C++11, the underlying type is int.
enum class Enum4 : unsigned int; // Legal C++11.
enum Enum2 : unsigned short; // Illegal in C++11, because Enum2 was previously declared with a different underlying type.

How to repeat this bug in C# .NET 2.0.

Open Visual Studio and create New Project – Windows Forms Project.

Create a new .NET 2.0 project

Add one TextBox and two Buttons to the project.

Form with 2 Buttons and TextBox

Add click handlers for two buttons and one helper function. Here is a code.

      private void button1k_Click(object sender, EventArgs e)
      {
         textBox.Text = GenerateString(1000);
      }
 
      private void button10k_Click(object sender, EventArgs e)
      {
         textBox.Text = GenerateString(10000);
      }
 
      private string GenerateString(int lenght)
      {
         string text = "";
         for (int i = 0; i < lenght; ++i)
            text += Convert.ToChar('0' + i % 10);
         return text;
      }

Now run the code! Clicking on the first button, TextBox shows some random numbers as expected.

Correct result, text is shown

Now click on the second button. Text disappears, and only blinking cursor is shown. Also note, that TextBox is functioning -- you can still delete / add characters, select all, use copy / paste, you just do not see the results on the screen.

Incorrect result, only cursor blinks

Workarounds? Currently none. Submitted this bug to Connect: .NET 2.0 TextBox fails to display long texts.

Tested on Windows 7 Pro 64-bit and Windows 7 Ultimate 64-bit with latest updates on 12/2/2011. Compiled with latest Visual Studio 2010 Pro 10.0.40219.1 SP1Rel w/ Microsoft .NET Framework Version 4.0.30319 SP1Rel.

P.S. One similar bug is reported on Microsoft Connect, however it seems like separate issue: Vanishing text bug.

P.S.S. My source code can be downloaded from MS Connect.

VS Pro w/o code analysis tool

This however will change with upcoming Visual Studio 11. See the SDL blog: Code Analysis for All.

This is the first time that Code Analysis has been made available in an Express edition of Visual Studio…

Searching Google for quick answer I found famous Microsoft employee and blogger Larry Osterman’s blog post about Hey, why am I leaking all my BSTR’s? He describes OLE caching of BSTR objects and tool called UMDH. OLE caching is not my case, however UMDH may be useful is you are able to figure out what to do.

Another link I found is from 1999 (VS6) Implementing a Simple MFC Client for an ATL Server. They have notes about memory leaks all over the place. Particularly interesting are two examples I have copied below.

Without leak:
//Update static text with new value
BSTR tmpBStr;
 
m_pObject1->get_ObjectString(&tmpBStr);
_bstr_t tmpbstr(tmpBStr, FALSE); //necessary to avoid a memory leak
 
SetDlgItemText(IDC_CURPROPVAL, tmpbstr);

With leak:
//Update static text with new value
BSTR tmpBStr;
 
m_pObject1->get_ObjectString(&tmpBStr);
_bstr_t tmpbstr;
 
tmpbstr= tmpBStr; //Caution: Memory leak occurs
SetDlgItemText(IDC_CURPROPVAL, tmpbstr);

And explanation:

The leak occurs when the tmpbstr variable is initialized. A call to SysAllocString (reference available in the Component Services section of the Platform SDK) is automatically made when creating the tmpbstr variable. This new allocation is never freed later, resulting in a memory leak. Using this version of the _bstr_t constructor avoids the issue by attaching the BSTR object to tmpbstr without a call to SysAllocString. For more information on this issue, see _bstr_t::_bstr_t (reference available in the C++ Language Reference section of the Visual C++ Documentation).

The problem is that explanation is wrong. My speculation: Most probably one person (senior programmer?) found memory leak in documentation, fixed it and told another person (junior programmer?) to update documentation. And that second person got it wrong. First: _bstr_t constructor _bstr_t::_bstr_t() does not allocate memory. Second: if it were allocating, then it would be freed, see the code below:
_bstr_t& _bstr_t::operator=(const wchar_t* s)
{
    if (s == NULL || ...)
    {
        _Free();
        m_Data = new Data_t(s);


And yes, _Free does exactly that. It frees m_Data if it contains something, in our case as I noted above, it points to 0×00000000;

operator=(const wchar_t* s) is exactly what is called in my case – Unicode compile.

Again, the code in Microsoft article leaks memory as comment says, but it does not leak as it is described. It leaks because memory allocated by m_pObject1->get_ObjectString is not freed.

In the end I found my memory leak. It was similar problem, I needed to call _bstr_t constructor with FALSE as second argument. From MSDN: If false, the bstr argument is attached to the new object without making a copy by calling SysAllocString.

So be careful playing with BSTRs.

The ABA problem occurs during synchronization, when a location is read twice, has the same value for both reads, and “value is the same” is used to indicate “nothing has changed”. However, another thread can execute between the two reads and change the value, do other work, then change the value back, thus fooling the first thread in to thinking “nothing has changed” even though the second thread did work that violates that assumption.

An Idiot’s Guide to C++ Templates by Ajay Vijayvargiya from The Code Project.

C++0x Finally Becomes a Standard discussed in Slashdot.

What is Cloud Programming and why should we care By GanesanSenthilvel at The Code Project.

Interesting discussion in StackOverflow about Creating a memory leak with Java.

Why is subtracting two times (in 1927) is giving a strange result? or in Shanghai at midnight at the end of 1927, the clocks went back 5 minutes and 52 seconds. From StackOverflow.

$300M To Save 6 Milliseconds in high frequency trading discussed at Slashdot.

Age bias in IT: Some consider it IT’s dirty little secret, or even IT’s big open secret – from computerworld.com by Tam Harbert.

Am I Too Old To Learn New Programming Languages? from Slashdot and Become a Good Programmer in Six Really Hard Steps from ApochPiQ in GameDev.net.

InfoWorld writes about The nine circles of IT hell by Dan Tynan.

Biggest Mistakes in Web Design 1995-2015 from webpagesthatsuck.com.

From glyph recognition to augmented reality with video demos and source code By Andrew Kirillov in The Code Project.

Introduction to Google App Inventor by VinayakIyer from The Code Project.

Guide to Image Composition with Win32 MsImg32.dll By Paul Watt from The Code Project.

JavaScript Jargon (Coming to JavaScript from C#) By Jonathan Cardy.

Learn to code: Codecademy is the easiest way to learn how to code. It’s interactive, fun, and you can do it with your friends.

Minimalist Coding Guidelines By gggustafson.

Penrose Tiling in Obfuscated Python or Who says you can’t write obfuscated Python? from Jeff Preshing.

Learn Perl in about 2 hours 30 minutes By Sam Hughes.

PowerShell is powerful but can it handle gaming? PowerShell Falling Blocks By Lasse W.

Sudoku solver via a webcam: A nice computer vision application with source code from Bojan Banko.

Windows API: Slim Reader/Writer (SRW) Locks from MSDN (available from Vista and Server 2008).

The Five Levels of ISP Evil from Dane Jasper.

spin.js – cool open source JavaScript spinner hosted in GitHub.

Starting up inside the box (Vista and above) from Raymond Chen.

Stuxnet Malware Analysis Paper By AmrThabet. This article will focus on Stuxnet’s windows infection methods and spreading methods.

The Power of Rooting on Android By David Magnotti.

The Principles of Good Programming by Christopher Diggins.

.NET ThreadPool vs. Tasks by Paul Stovell.

Why does creating a shortcut to a file change its last-modified time… sometimes? from Raymond Chen.

Writing a Multiplayer Game (in WPF) By Paulo Zemek. This article will explain some concepts of game development and how to apply and adapt them for multiplayer development.

To install it, go to Control Panel – Programs – Turn Windows features on or off – Check ‘Microsoft .NET Framework 3.5.1′ – OK.

Note, that .NET Framework 2.0 is included into version 3.5.

Below are some screenshots from the installation process.

The opposite – the feature that no one have used is: Feedback form, that is located on all pages.

To be exact, the Feedback feature is used exactly 0 (zero, null, none) times. However SilverBench is used many, many times each day. If you look at the SilverBench results page, you can see, that from 5 – 10 people submit their benchmark scores every day.

So no one can predict, that after 22 weeks, the one of the most prominent features will not be used even single time.

P.S. Forgot to mention, that 36 days ago we launched HTTP Response header tool. An online tool, that can be used to view HTTP header stats and information for any site in the Internet.

• It happens only on Windows XP (not on Vista, Server 2003, 2008, Win 7);
• It does not happens on all Windows XP, because it is Race condition;
• I was unable to reproduce it with physical glass/monitor attached to the computer; It only happened using Remote Desktop;
• It does not happen when debugger is attached and breakpoint is being hit;

Latter I found people that have similar issue:

• CreateProcessAsUser from service fails with 233 (FUS, Windows XP) on microsoft.public.platformsdk.security;
• CreateProcessAsUser fails with 233 (No process is on the other end of the pipe), XP;
• In comments from the article;
• CreateProcessAsUser fails on XP x64 on Microsoft forums;

The code is the following:
WTSQueryUserToken(..., &hToken);
SetTokenInformation(hToken, ...);
SomeVistaAndWin7ElevatedTokenStuff(hToken);
CreateProcessAsUser(hToken, ...);


And CreateProcessAsUser fails with GetLastError() = 233. Looking in System Error Codes (0-499) – ERROR_PIPE_NOT_CONNECTED 233 (0xE9) – No process is on the other end of the pipe.

The confusing part is about Pipe, because you didn’t expect to get pipe error here – you have not created any pipe.

In the first post I have linked, Thomas Graefenhain writes:

I’ve debugged a little bit with the kernel debugger, and have found the
problem: CreateProcessAsUser uses internally, when creating a process in an
other session, the function CreateRemoteProcessW from ADVAPI32.DLL. This
function opens a pipe with the name
\\.\Pipe\TerminalServer\SystemExecSrvr\%d where %d is the SessionID and
sending the request over to csrss.exe. …

In another post someone mentioned that Sleep(2000) fixed the problem. This explains why under debugger everything works without an error.

The good news is that it happens only under Windows XP and under Remote Desktop, so small group of users are affected. The bad new is that there are no elegant workarounds. Windows XP is in the Extended Support Phase, so I am not counting on fix from Microsoft.

Currently I use the following workaround (simplified version, see below). If you have something better or more elegant, please let me know in the comments below.
{
     Sleep(100);
     CreateProcessAsUser(hToken, ...);
} while (wasError && GetLastError() == 233 && IsWindowsXP());