TFS Alternative to Cloaking – shorten build base path

I was facing yet another [ProjectName] “contains more than the allowed 259 characters. Type or select a shorter path.”

I have learned how to cloak these entries when they are not needed for the build I am doing, and we have also resorted to shortening the path names of some so that we can include them. But our base folders were pretty long so I decided to try and change them. I googled for the $(SourceDir) entry and found a couple of helpful articles:

tfs – How do I change the build directory that MSBuild uses under Team Foundation Build? – Stack Overflow

TFS Top Tip #16 – Limit What you Build – Martin Woodward

From these I was able to change the entries in the config and Build Agent settings that would help me to shorten the path to something that would stop the madness of cloaking and avoid the above error (until we go above the limit again).

Here is what I did (this is all VS and TFS 2008):

I went to the Team Explorer in Visual Studio and right clicked on Builds and selected

I then inside the the manage selected my (only) build agent and got the window below:

Changing this from e:\teambuild\projects (which was a decent naming convention) to the shortened E:\Bld saved us 15 characters which solved a bunch of path length issues.

In addition, I changed the default “Sources” in the file

c:\Program Files\Microsoft Visual Studio 9.0\Common7\IDE\PrivateAssemblies\tfsbuildservice.exe.config

from <add key=”SourcesSubdirectory” value=”Sources” />

to <add key=”SourcesSubdirectory” value=”Src” />

saving me an additional 4 characters for a full 19 characters of shortening.

This seems very silly in a world where we are trying to be more descriptive and I hope that MS will eventually let us do as we please with our file and folders, but it saved us some hassle and maybe it will save you some too.

Worker Thread Class

In an article by Juval Löwy, he does a great job of describing and providing an example of a worker thread class. Since many of us are not yet able to pick up .Net 4, we do not yet have the threading beauty that apparently awaits us there. So I adopted this worker thread and made a few changes that fill a few gaps for me. The listing is below along with comments where I made changes to Juval’s original work.

 /// <summary>
 /// Thread mgmt wrapper class - base code taken from devx.com article by Juval Löwy
 /// Added mutex and exception string to allow exception info to be stored and passed out from thread.
 /// Also added ability to pass in delegate that is run as the worker thread function, and some
 /// additional ways to construct.
 /// Added a mutex and wrapped it in a contained class to pass into the worker so that they can monitor
 /// the EndLoop value. I also added a passthrough object to this class so that the worker can take in
 /// an object when it starts and pass one back out when it is done.
 /// Also found a race condition for a test that Juval was doing on the IsAlive when it asserted that
 /// the handle could not be same value as isalive, but when function is done and closing,
 /// this value could be set right at the end and isalive not yet unset. I changed it to a definite
 /// test - that the handle is set when isalive is false. This must be true, since the thread has to exit
 /// that way (unless someone crawls in and kills it which this wrapper is supposed to allow you to avoid
 /// completely).
 /// </summary>
public class WorkerThread : IDisposable
 {
 /// <summary>
 /// Thread function passing ref to Thread Objects so we can:
 /// 1. check to see if we're done, and
 /// 2. Pass in or out values to / from the thread.
 /// </summary>
 public delegate void ThreadFunc(ThreadObjects threadObjects);

 /// <summary>The <see cref="ThreadFunc"/> delegate.</summary>
 protected ThreadFunc threadFunc = null;

 /// <summary>The <see cref="ManualResetEvent"/> m_WorkerStartedEvent - that is reset when the worker thread starts its task.</summary>
 protected ManualResetEvent m_WorkerStartedEvent;
 /// <summary>ManualResetEvent m_WorkerDoneEvent - reset when the worker thread has completed its task.</summary>
 protected ManualResetEvent m_WorkerDoneEvent;
 /// <summary>The <see cref="ManualResetEvent"/> m_WorkerExceptionEvent - reset when an exception occurs</summary>
 protected ManualResetEvent m_WorkerExceptionEvent;

 /// <summary>The System thread that we are using to implement our worker thread</summary>
 protected Thread m_ThreadObj;

 /// <summary>holds the objects used to initialize, communicate with, and pull data back from the worker thread.</summary>
 protected ThreadObjects m_mutexedObjects;

 /// <summary>
 /// This object is passed into the worker so it can monitor when to stop,
 /// and to get / set the passthrough object
 /// </summary>
 public class ThreadObjects
 {
 /// <summary>if true tells thread to stop processing (must be processed by the worker thread user code)</summary>
 protected bool _mEndLoop;
 /// <summary>controls access to m_endloop</summary>
 protected Mutex _endLoopMutex;

 /// <summary>used to pass data into and out of the worker thread</summary>
 protected object _mPassThrough;
 /// <summary>controls access to the passthrough data</summary>
 protected Mutex _mPassThroughMutex;

 /// <summary>
 /// Initializes a new instance of the <see cref="ThreadObjects"/> class.
 /// Creates the mutexes.
 /// </summary>
 public ThreadObjects()
 {
 _endLoopMutex = new Mutex();
 _mPassThroughMutex = new Mutex();
 }
 /// <summary>
 /// Gets or sets a value indicating whether the worker thread should stop processing.
 /// This is initially set false and then set to true by the caller to tell the worker
 /// thread to stop. The worker thread function should check this value periodically so
 /// that the thread can be shutdown gracefully.
 /// </summary>
 /// <value><c>true</c> if processing should stop; otherwise, <c>false</c>.</value>
 public bool EndLoop
 {
 set
 {
 _endLoopMutex.WaitOne();
 _mEndLoop = value;
 _endLoopMutex.ReleaseMutex();
 }
 get
 {
 bool result = false;
 _endLoopMutex.WaitOne();
 result = _mEndLoop;
 _endLoopMutex.ReleaseMutex();
 return result;
 }
 }

 /// <summary>
 /// Gets or sets an object value that the worker thread can use or set.
 /// It is recommended that this only be set before starting the thread by the caller
 /// and that it be checked by the caller only after the thread is complete. Otherwise the
 /// values are possibly unreliable and require careful management.
 /// For instance, this could be used to hold an array of strings indicating the status of
 /// a thread that cyclically performs a task. The Thread could keep it updated and
 /// delete old info out of it when it has new and keep just the last 20 events in there
 /// so a status monitor could check out whats happening from time to time without
 /// forcing the thread to stop and restart.
 /// </summary>
 /// <value>Any value that the caller / worker wishes to pass through.</value>
 public object PassThrough
 {
 set
 {
 _mPassThroughMutex.WaitOne();
 _mPassThrough = value;
 _mPassThroughMutex.ReleaseMutex();
 }
 get
 {
 object result = "";
 _mPassThroughMutex.WaitOne();
 result = _mPassThrough;
 _mPassThroughMutex.ReleaseMutex();
 return result;
 }
 }

 internal void Close()
 {
 _endLoopMutex.Close();
 _mPassThroughMutex.Close();
 }

 private string _name = "";
 /// <summary>Gets or sets the name of the thread.</summary>
 public string Name { get { return _name; }  set { _name = value; } }
 }

 /// <summary>
 /// Gets or sets a value indicating whether the worker thread should stop processing.
 /// This is initially set false and then set to true by the caller to tell the worker
 /// thread to stop. The worker thread function should check this value periodically so
 /// that the thread can be shutdown gracefully.
 /// </summary>
 /// <value><c>true</c> if processing should stop; otherwise, <c>false</c>.</value>
 public bool EndLoop
 {
 get { return m_mutexedObjects.EndLoop; }
 set { m_mutexedObjects.EndLoop = value; }
 }

 /// <summary>
 /// Gets or sets an object value that the worker thread can use or set.
 /// It is recommended that this only be set before starting the thread by the caller
 /// and that it be checked by the caller only after the thread is complete. Otherwise the
 /// values are possibly unreliable and require careful management.
 /// For instance, this could be used to hold an array of strings indicating the status of
 /// a thread that cyclically performs a task. The Thread could keep it updated and
 /// delete old info out of it when it has new and keep just the last 20 events in there
 /// so a status monitor could check out whats happening from time to time without
 /// forcing the thread to stop and restart.
 /// </summary>
 /// <value>Any value that the caller / worker wishes to pass through.</value>
 public object PassThrough
 {
 get { return m_mutexedObjects.PassThrough; }
 set { m_mutexedObjects.PassThrough = value; }
 }

 /// <summary>
 /// Gets the ThreadState of the thread.
 /// </summary>
 /// <value>The <see cref="System.Threading.ThreadState"/> of the underlying thread.</value>
 public ThreadState State { get { return m_ThreadObj.ThreadState; } }

 /// <summary>Holds the exception details (if any) when an exception is generated in the thread so
 /// that it can be picked up by the thread that started this worker.</summary>
 protected string m_ExceptionInfo;

 /// <summary>controls access to the exception info</summary>
 protected Mutex m_ExceptionInfoMutex;

 /// <summary>Gets the thread being managed.</summary>
 public Thread Thread{get{return m_ThreadObj;}}

 /// <summary>
 /// Gets or sets the exception info. A holding place for any exception that occurs in the thread.
 /// It will contain whatever string the worker function puts in it (i.e. message, stack, etc.).
 /// Generally this should be an empty string unless there is an exception, thus allowing the
 /// caller to be able to determine if a thread processed successfully.
 /// </summary>
 /// <value>The exception info, or an empty string if there was no exception.</value>
 public string ExceptionInfo
 {
 set
 {
 m_ExceptionInfoMutex.WaitOne();
 m_ExceptionInfo = value;
 m_ExceptionInfoMutex.ReleaseMutex();
 }
 get
 {
 string result = "";
 m_ExceptionInfoMutex.WaitOne();
 result = m_ExceptionInfo;
 m_ExceptionInfoMutex.ReleaseMutex();
 return result;
 }
 }

 /// <summary>
 /// Gets the generated thread name using the namespace and the current system
 /// ticks. It does not need to be unique, but we make it so in order to tell
 /// threads apart (since they were not explicitly named).
 /// </summary>
 /// <value>The name of the generated thread.</value>
 static string GenThreadName { get { return typeof(WorkerThread).FullName + DateTime.Now.Ticks.ToString(); } }

 /// <summary>
 /// Initializes a new instance of the <see cref="WorkerThread"/> class
 /// using a <see cref="ThreadFunc"/> delegate and a thread name.
 /// </summary>
 /// <param name="threadFuncIn"><see cref="ThreadFunc"/> delegate
 /// passes an object that allows the worker function to get
 /// input and/or set a result, and to monitor if the caller
 /// has requested that the worker stop processing.
 /// </param>
 /// <param name="threadName">Name of the thread. If this is made meaningful, then
 /// it makes debugging easier, so it is recommended that it be set accordingly.</param>
 public WorkerThread(ThreadFunc threadFuncIn, string threadName)
 {
 m_mutexedObjects = new ThreadObjects();
 m_mutexedObjects.Name = threadName;
 m_mutexedObjects.EndLoop = false;
 m_ThreadObj = null;
 m_ExceptionInfo = "";
 m_ExceptionInfoMutex = new Mutex();
 m_WorkerStartedEvent = new ManualResetEvent(false);
 m_WorkerDoneEvent = new ManualResetEvent(false);
 m_WorkerExceptionEvent = new ManualResetEvent(false);
 threadFunc = threadFuncIn;
 ThreadStart threadStart = new ThreadStart(Worker);
 m_ThreadObj = new Thread(threadStart);
 m_ThreadObj.Name = threadName;
 }

 /// <summary>
 /// Initializes a new instance of the <see cref="WorkerThread"/> class
 /// using a <see cref="ThreadFunc"/> delegate. The thread name is generated
 /// using <see cref="GenThreadName"/> which grabs the full name of the WorkerThread type and
 /// appends the system clock ticks.
 /// </summary>
 /// <param name="threadFunc"><see cref="ThreadFunc"/> delegate
 /// passes an object that allows the worker function to get
 /// input and/or set a result, and to monitor if the caller
 /// has requested that the worker stop processing.
 /// </param>
 public WorkerThread(ThreadFunc threadFunc) : this(threadFunc, GenThreadName)
 {
 }

 /// <summary>
 /// Initializes a new instance of the <see cref="WorkerThread"/> class.
 /// </summary>
 /// <param name="autoStart">if set to <c>true</c> if the thread should start
 /// as soon as it is set up. The thread name is generated
 /// using <see cref="GenThreadName"/> which grabs the full
 /// name of the WorkerThread type and appends the system clock ticks.</param>
 /// <param name="threadFunc"><see cref="ThreadFunc"/> delegate
 /// passes an object that allows the worker function to get
 /// input and/or set a result, and to monitor if the caller
 /// has requested that the worker stop processing.
 /// </param>
 public WorkerThread(bool autoStart, ThreadFunc threadFunc)
 : this(autoStart, threadFunc, GenThreadName)
 {
 }

 /// <summary>
 /// Initializes a new instance of the <see cref="WorkerThread"/> class.
 /// </summary>
 /// <param name="autoStart">if set to <c>true</c> if the thread should start
 /// as soon as it is set up. The thread name is generated
 /// using <see cref="GenThreadName"/> which grabs the full
 /// name of the WorkerThread type and appends the system clock ticks.</param>
 /// <param name="threadFunc"><see cref="ThreadFunc"/> delegate
 /// passes an object that allows the worker function to get
 /// input and/or set a result, and to monitor if the caller
 /// has requested that the worker stop processing.
 /// </param>
 /// <param name="threadName">Name of the thread. If this is made meaningful, then
 /// it makes debugging easier, so it is recommended that it be set accordingly.</param>
 public WorkerThread(bool autoStart, ThreadFunc threadFunc, string threadName)
 : this(threadFunc, threadName)
 {
 if (autoStart)
 {
 Start();
 }
 }

 /// <summary>
 /// This <see cref="WaitHandle"/> is used to indicate that the thread worker has started.
 /// It is set at the beginning of the thread function inside the WorkerThread class, and
 /// verified in the IsAlive, that if the thread is dead, that this handle got set.
 /// If not, an exception is thrown.
 /// See also <seealso cref="WorkerDoneHandle"/> and <seealso cref="WorkerExceptionHandle"/>
 /// </summary>
 /// <value>The handle.</value>
 public WaitHandle WorkerStartedHandle
 {
 get
 {
 return m_WorkerStartedEvent;
 }
 }

 /// <summary>
 /// This <see cref="WaitHandle"/> is used to indicate that the thread worker has finished.
 /// It is set at the end of the thread function inside the WorkerThread class, and
 /// verified in the IsAlive, that if the thread is dead, that this handle got set.
 /// If not an exception is thrown.
 /// See also <seealso cref="WorkerStartedHandle"/> and <seealso cref="WorkerExceptionHandle"/>
 /// </summary>
 /// <value>The handle.</value>
 public WaitHandle WorkerDoneHandle
 {
 get
 {
 return m_WorkerDoneEvent;
 }
 }

 /// <summary>
 /// This <see cref="WaitHandle"/> is used to indicate that the thread worker terminated in an exception state.
 /// It is set if an exception is caught in the wrapper worker function.
 /// See also <seealso cref="WorkerStartedHandle"/> and <seealso cref="WorkerDoneHandle"/>
 /// </summary>
 /// <value>The handle.</value>
 public WaitHandle WorkerExceptionHandle
 {
 get
 {
 return m_WorkerExceptionEvent;
 }
 }

 /// <summary>
 /// Starts the thread after checking that it is currently instantiated and
 /// is not alive. This is called automatically if autostart is set in the
 /// call to the constructor.
 /// </summary>
 public void Start()
 {
 if(m_ThreadObj == null)
 throw new NullReferenceException("Start called on null thread object.");

 if(m_ThreadObj.IsAlive)
 throw new ThreadStateException("Start called on thread that is already running.");

 m_ThreadObj.Start();
 }

 /// <summary>
 /// If object is being disposed, this function will ensure that the thread is killed.
 /// Moved the close handles to this function from Kill, because we may still want to use these
 /// values after killing the thread to check on what happened.
 /// </summary>
 public void Dispose()
 {
 Kill();
 m_mutexedObjects.Close();
 m_WorkerStartedEvent.Close();
 m_WorkerDoneEvent.Close();
 m_WorkerExceptionEvent.Close();
 }

 /// <summary>
 /// Kills this thread. First it checks to see if we have a thread and if it is currently
 /// alive. If not, the function returns without doing anything. Otherwise, EndLoop is set,
 /// and it waits using a Join for the thread to die. If the thread won't die after the allotted time,
 /// then the thread is aborted. THe called thread should never require this of the caller, but
 /// in some cases, since we are calling code that is unmanaged and outside of our control,
 /// we may have threads that are hung. Thus the thread should have resource freeing coded into
 /// the catch handler of the thread function.
 /// </summary>
 public void Kill(TimeSpan timeout)
 {
 //Kill is called on client thread - must use cached object
 if (m_ThreadObj == null) // thread is not there - leave it
 return;

 if (IsAlive == false)
 return;

 m_mutexedObjects.EndLoop = true;
 //Wait for thread to die
 Join(timeout);

 if (m_ThreadObj.IsAlive == true)
 {    // this thread refuses to die and we gave it 2 minutes - abort.
 m_ThreadObj.Abort(); // send abort Exception
 }
 }

 /// <summary>
 /// Kills this thread. First it checks to see if we have a thread and if it is currently
 /// alive. If not, the function returns without doing anything. Otherwise, EndLoop is set,
 /// and it waits using a Join for the thread to die.
 /// </summary>
 public void Kill()
 {
 Kill(new TimeSpan(0, 2, 0));
 }

 /// <summary>
 /// Joins thread - if thread is already dead, it returns, otherwise it waits on the thread to die.
 /// </summary>
 public void Join()
 {
 if (m_ThreadObj == null) // thread is not there - leave it
 return;

 if (IsAlive == false)
 return;

 Debug.Assert(Thread.CurrentThread.GetHashCode() !=
 m_ThreadObj.GetHashCode());

 m_ThreadObj.Join();
 }

 /// <summary>
 /// Joins with the specified milliseconds timeout.
 /// </summary>
 /// <param name="millisecondsTimeout">The milliseconds timeout.</param>
 /// <returns></returns>
 public bool Join(int millisecondsTimeout)
 {
 TimeSpan timeout = TimeSpan.FromMilliseconds(millisecondsTimeout);

 return Join(timeout);
 }

 /// <summary>
 /// Joins with the specified timeout.
 /// </summary>
 /// <param name="timeout">The timeout.</param>
 /// <returns></returns>
 public bool Join(TimeSpan timeout)
 {
 if (m_ThreadObj == null) // thread is not there - leave it
 return true;

 if (IsAlive == false)
 return true;

 if (Thread.CurrentThread.GetHashCode() == m_ThreadObj.GetHashCode())
 throw new Exception("Attempting to Join a thread that has the same identity as the current thread.");

 return m_ThreadObj.Join(timeout);
 }

 /// <summary>
 /// Gets or sets the name of the thread.
 /// </summary>
 /// <value>The name of the thread.</value>
 public string Name { get { return m_ThreadObj.Name; } set { m_ThreadObj.Name = value; } }

 /// <summary>
 /// Gets a value indicating whether the worker thread is alive.
 /// </summary>
 /// <value><c>true</c> if the worker thread is alive; otherwise, <c>false</c>.</value>
 public bool IsAlive
 {
 get
 {
 Debug.Assert(m_ThreadObj != null);
 bool isAlive = m_ThreadObj.IsAlive;
 return isAlive;
 }
 }

 /// <summary>
 /// Worker runs the function that was passed in wrapped by an exception handler.
 /// If the function being called has any issues the exception is caught
 /// and the exception message is put into a string along with the thread name,
 /// call stack and any inner exceptions for access by the caller / creator of the thread.
 /// </summary>
 protected void Worker()
 {
 try
 {
 if (threadFunc != null)
 {
 m_WorkerStartedEvent.Set(); // indicate that we have started
 threadFunc.Invoke(m_mutexedObjects);
 }
 else
 throw new NullReferenceException("WorkerThread function expected.");
 }
 catch(Exception ex)
 {
 string exInfo = ex.Message;
 ExceptionInfo = Thread.CurrentThread.Name + " exception:" + exInfo;
 m_WorkerExceptionEvent.Set();
 }
 finally
 {
 m_WorkerDoneEvent.Set();
 }
 }
 }