|
This returned null in my windows service.
|
|
|
|
|
string originalPath = System.Reflection.Assembly.GetExecutingAssembly().CodeBase;
string location = originalPath.Substring( 0, originalPath.LastIndexOf(@"/"));
this works
|
|
|
|
|
For a service running in a secured system under the 2.0 framework :
AppDomain.CurrentDomain.BaseDirectory
Thanks for your attempts to help.
|
|
|
|
|
I really would like to know...
how can Assembly.GetExecutingAssembly() return null?!
|
|
|
|
|
It may be due to looking for this at the OnStart implementation (in which I'm an event handler of a different process) but I really don't know. When I got tired of blowing up with each individual attempt, I finally did this:
In my OnStart event handler, I commented out all code and had a sequence of Debug.Trace messages that displayed the results of each item as follows:
Assembly.GetExecutingAssembly()==null?" GetExecutingAssembly null ":Assembly.GetExecutingAssembly();
AppDomain.CurrentDomain.BaseLocation == null?" CurrentDomain.BaseLocation null": ....
This way I got the results of everyone of the items that people suggested or that I had attempted. In that example everything came up null except for the AppDomain approach.
|
|
|
|
|
Huh, I'll have to see what I get when I do that.
|
|
|
|
|
I just tried it in the OnStart of one of my complex services, both debug and release builds.
System.Windows.Application.ExecutablePath and System.Windows.Application.StartupPath
gave the information about the executable that loaded the DLL, which is still good for simple services, but not here.
System.Reflection.Assembly.GetExecutingAssembly().CodeBase
gave the name of the actual file containing the executing code
However! My OnStart is in a base class in another DLL so I got information on that, including when in release mode the particular DLL in in the GAC!
So, I wrote an override of OnStart in my service class and I got the expected result.
But at no time did System.Reflection.Assembly.GetExecutingAssembly() return null, I am still most curious your statement that yours did.
At any rate, I doubt that OnStart() is the proper place for the code you're trying to run, it should be very small, my version is:
protected override void
OnStart
(
string[] args
)
{
# if DEBUG
System.Threading.Thread.Sleep ( 60000 ) ;
# endif
this.oktoproceed = true ;
base.OnStart ( args ) ;
if ( this.when != null )
{
this.when.Start() ;
}
LogMessage ( this.ServiceName + " started" ) ;
}
(when is a System.Timers.Timer that controls the cycling of the service)
|
|
|
|
|
I have a general service that starts up several WCF services. This all works great. Now I have two different thread processes I need to start. I would like to use a configuration file (XML) to define the classes and target methods and this requires that I use reflection to create the thread object correctly. So I basically have to recreate:
Thread newThread = new Thread(new ThreadStart(class.Method));
Any ideas or article someone can point me to? Is this even possible?
Thanks,
Michael
|
|
|
|
|
Well, a ThreadStart may be a zero-parameter delegate, but that method can contain anything...
<br />
_obj = Get_My_Object_Through_Reflection_Somehow<br />
<br />
Thread newThread = new Thread(new ThreadStart(RunObjMethod));<br />
newThread.Start();<br />
<br />
...<br />
<br />
private void RunObjMethod()<br />
{<br />
_obj.Method();<br />
}<br />
There are of course ways to clean this up... You could pass the object and a string with the method name to a private class, and have the class do something like the above code in its constructor, just to keep everything encapsulated. Make it nice and general, and hook up a static method to wrap the constructor, and you can make it look like:
<br />
StartInThread.Launch(myObj, "MethodName");<br />
<br />
...<br />
<br />
private class StartInThread<br />
{<br />
object _obj;<br />
string _methName;<br />
<br />
public static void Launch(object obj, string methName)<br />
{<br />
new StartInThread(obj, methName);<br />
}<br />
private StartInThread(object obj, string methName)<br />
{<br />
_obj = obj; _methName = methName;<br />
Thread thrd = new Thread(new ThreadStart(Run));<br />
thrd.Name = methName;<br />
thrd.IsBackground = true;<br />
thrd.Start();<br />
}<br />
<br />
private void Run()<br />
{<br />
...get a MethodInfo for _methName in _obj, and call it directly...<br />
}<br />
}<br />
(Pardon any syntax errors, since I just typed that up in this box)
|
|
|
|
|
Thank you. Your code really did not help me at all but it did get me to thinking towards a solution.
I just need to define an interface:
public interface IThreadProcess
{
bool TerminateThread{get; set;}
void StartThreadProcess();
}
Then I just have to create an instance cast to the interface and VOILA I now have addressability to the method implementation.
|
|
|
|
|
I was going to say:
May I suggest you look into System.ServiceProcess.ServiceBase.Run ( System.ServiceProcess.ServiceBase[] )
But I suppose your first sentence means you already do that.
Everything is possible.
|
|
|
|
|
Has anyone out there had any experiance creating 'Remote' windows hooks for applications using C#. These are thread specific hooks - NOT global and NOT local.
I want to load an application using CreateProcess API get the thread id and create a remote hook with filters for the CBT and Keyboard hooks for that application.
I believe to create remote hooks you have to create a DLL that contains your filter functions and reference them in the SetWindowsHooksEx API function. All the C# examples available seem to create local hooks using the thread for the current application and subsequently the filter functions can reside locally, within the same application.
If a C# DLL is indeed required can I still use delegates to reference the filter functions within it?
Thanks
Regards
Phillip.
|
|
|
|
|
If I understand your question correctly, you are probably wanting to do something that in not 'legal' within the context of C# managed code. Think of it. If you can do that with your application, what stops someone from writing a cookoos egg that launches the login window and hooks into the keyboard to steal you id and password.
That is probably why the examples are in the same application context.
|
|
|
|
|
Yes, that is a valid point and I have noticed that some of the examples out there are for ‘legitimate’ password retrieval utilities. However, I can assure you my intentions are honorable and focused on extending the functionality of an ‘off the shelf’ business application for which no API exists.
For remote hooks it is necessary to have the filter functions in a DLL that is injected into the address space of the application at run time. Seemingly, given the nature of .NET and the CLR this allocation is not permissible as such. Would it be possible to write unmanaged C# code to achieve this objective? Can I still utilize the Win32 API’s using the P/Invoke layer if the code is managed? If not, what other means are there to utilize them?
Thanks for your help
Phillip.
|
|
|
|
|
The only option is unmanaged C++. So if this is going to be a published application, let me know what it is so that I never buy it.
|
|
|
|
|
say Triangle is a class and the following has been declared:
<br />
Triangle[] triangles = new Triangle[4];<br />
for (int i = 0; i != triangles.Length; i++)<br />
{<br />
triangles[i] = new Triangle();<br />
}<br />
with respect to what's happening in the heap, are two codes below the same?
thanks!
<br />
Triangle[] copy = new Triangle[triangles.Length];<br />
for (int i = 0; i != copy.Length; i++)<br />
{<br />
copy[i] = triangles[i];<br />
}<br />
Triangle[] copy = triangles;
I am a SysAdmin, I battle my own daemons.
|
|
|
|
|
They might be functionally identical, depending on how you use them, but they aren't the same.
Basically, you have:
triangles -> arrayObject: [ triangle1, triangle2, triangle3 ]
The first block of code would create:
copy -> copyOfArray: [ triangle1, triangle2, triangle3 ]
(copy is a pointer to a new array containing references to the same objects)
The second block would create:
copy -> arrayObject: [ triangle1, triangle2, triangle3 ]
(copy is a pointer to the original array)
Subtle difference, but in the first case, you could sort the copied array independently, and even add and remove elements without touching the original array. In the second case, sorting "copy" would sort "triangles".
(I'm assuming that Triangle is a class, not a struct. If it's a struct, then there's even more of a difference, because the first case would make a completely separate array of independent objects)
EDIT: Silly me, you said it was a class. Last paragraph is irrelevant
|
|
|
|
|
I don't think you have actually created a copy yet. I think you just have a reference to the existing objects. If you truely wanted a copy you would need to do this:
Triangle[] copy = new Triangle[triangles.Length];
for (int i = 0; i != copy.Length; i++)
{
copy[i] = new Triangle();
copy[i] = triangles[i];
}
<./pre>
Otherwise you would have two objects that point to the same thing.
Ben
|
|
|
|
|
Even that wouldn't make an independent copy. You would need some sort of Clone() method, so the inner statement would be:
copy[i] = triangles[i].Clone() as Triangle;
|
|
|
|
|
Right, the code should be:
copy[i] = new Triangle();
copy[i].property1 = triangles[i].property1;
copy[i].property2 = triangles[i].property2;
copy[i].property3 = triangles[i].property3;
...
for as many properties that exist.
If you implement a Triangle.Clone() then you could just do
copy[i] = triangle[i].Clone();
Ben
|
|
|
|
|
This isn't an answer to your question, but this code creates two different types of arrays. The first (triangles) are a list of objects and the second (copy) is a list of reference pointers to the first collection. To do what you are showing there you need to so something like this:
public class Triangle:ICloneable
{
public Triangle(){}
protected Triangle(Triangle clone){ clone.Value = this.Value; }
public object Clone(){return new Triangle(this);}
}
Then you change your loop as such :
for ( blah blah)
{
copy[i] = triangles[i].Clone();
}
As to what things look like on the heap.....I have no idea.
|
|
|
|
|
No, they are not the same.
DenClancy wrote: Triangle[] copy = new Triangle[triangles.Length];
for (int i = 0; i != copy.Length; i++)
{
copy[i] = triangles[i];
}
This will give you a new array, but it contains references to the same objects as the original array.
DenClancy wrote: Triangle[] copy = triangles;
This will only give you another reference to the same array.
---
single minded; short sighted; long gone;
|
|
|
|
|
Hi,
I am using Windows Indexing services via the OleDB Driver.
Now everything works good, till i am to pass (") double quotes into the query
Sample query to pass
SELECT path FROM SCOPE() WHERE Matches("test*");
Now after all the process the VS 2005 just appends \ to my query
SELECT path FROM SCOPE() WHERE Matches(\"test*\");
"Mess with the Best, Die like the rest"
|
|
|
|
|
Passing queries is a real bad idea to being with because of sql injection issues. Still to fix your problems you just need more double quotes. I believe your query will work if it is like this:
SELECT path FROM SCOPE() WHERE Matches("""test*""");
Hope that helps.
Ben
|
|
|
|
|
Thx for the replying
Below is the original c# code
queryString = queryString + " MATCHES(""" + strExactPhrase.ToString() + """)";
* *
*'s are the points where i am getting an error "; expected" in VS 2005
Thank you!
"Mess with the Best, Die like the rest"
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
