Re: function usage in a .config file - C# Discussion Boards

Re: KeyDown event is not firing while draging node on treeview

leppie19-Jul-04 6:59

leppie

19-Jul-04 6:59

u have to select it before dragging perhaps, seems more natural IMO

^{top secret xacc-ide 0.0.1}

Re: KeyDown event is not firing while draging node on treeview

Heath Stewart19-Jul-04 9:10

Heath Stewart

19-Jul-04 9:10

It's not supposed to. If you want to detect if the CTRL key was pressed while dragging, handle the Control.QueryContinueDrag event. See the documentation for that event as well as the QueryContinueDragEventArgs in the .NET Framework SDK for more information and examples.

Microsoft MVP, Visual C#
My Articles

Returning a filtered DataTable

Andy H19-Jul-04 6:44

Andy H

19-Jul-04 6:44

I have a class in my middle tier in which I hold a DataTable object, which has already been filled out with data. All I want to do is filter the DataRows based on one of the Fields in the DataRow, and return a DataTable object with the filtered Rows.

This is my code:

public DataTable FilteredData( int clientType )<br />
		{<br />
			string filter = "CLIENT_TYPE = " + clientType;<br />
			// It seems can only Select into an array of DataRows!<br />
			DataRow[] foundRows = _myData.Select( filter );<br />
			// Create a fresh DataTable obj and copy foundRows into it.<br />
			DataTable dt = new DataTable();<br />
			foreach( DataRow row in foundRows )<br />
			{<br />
				dt.Rows.Add( row );<br />
			}<br />
			return dt;<br />
		}

The problem is that when it reaches the line dt.Rows.Add( row ); I get an exception saying that "This row already belongs to another table". It doesn't matter that I will be returning a seperate DataTable, as the Rows won't be updated in the interface. Can someone please tell me how I can return my data in a DataTable object, I would rather not have to return an array of DataRows.

Re: Returning a filtered DataTable

Werdna19-Jul-04 8:58

Werdna

19-Jul-04 8:58

Why don't you use DataView instead of DataTable?
You won't have to duplicate the data.
With your approach, you have to clone each row as DataRow can only belong to one datatable.

Re: Returning a filtered DataTable

Heath Stewart19-Jul-04 9:04

Heath Stewart

19-Jul-04 9:04

I would recommend you don't disconnect the DataRow from its DataTable. There is a lot of other information in there, such as change information - not to mention that the column definitions are part of the DataTable (which you didn't replicated; FWIW, the easiest way to do that is DataTable.Clone, which replicates only the structure).

An array of DataRows is still bindable, as well. An array implements IList so that you can bind it to a DataTable, ComboBox, or any other control. In this case, it's much better to bind against a DataView instantiated with the DataTable as the ctor parameter, then use the RowFilter property. This will track all changes.

If you don't need binding or change tracking and just want a simple list of DataRows, then you'll either have to cope with the array, remove them from the old table and add them to the new table (and use a reverse for loop - changing the enumerable will throw an exception), or use Array.CopyTo (inheritted from your DataRow[]) which will perform a deep copy (a shallow copy won't solve your problem, since it only copies a reference to the original (same) object).

Microsoft MVP, Visual C#
My Articles

Re: Returning a filtered DataTable

Andy H19-Jul-04 22:40

Andy H

19-Jul-04 22:40

Thanks for all your help. I have decided to continue with using DataTable objects as the interface between the DataLayer and the Business layer and to now use DataView objects as the interface between the BusinessLayer and the GUI.

function usage in a .config file

Ryan@Salamandertechnologies.com19-Jul-04 6:39

Ryan@Salamandertechnologies.com

19-Jul-04 6:39

I must profess my config file ignorance. I'm required to use an open source logging project. The project deposits logs into text files based on a path attribute in my application config file. I would really like to dynamically set that path, say to, the application path. Can you use functions in a config file, or do I have to just do it manually?

Thanks,
Ryan

Re: function usage in a .config file

Heath Stewart19-Jul-04 8:38

Heath Stewart

19-Jul-04 8:38

If you want the application path, simply use Application.StartupPath. Don't force a user to put the path of the executable into the .config file in that very path - it's redundent, requires extra steps (always faulty), and is easy to get from code (even with native executables).

If you want simple string settings, see ConfigurationSettings.AppSettings in the .NET Framework SDK, which also includes examples.

If you want to make your own configuration section, implement IConfigurationSectionHandler, which is documented in the .NET Framework SDK as well and includes samples. There are also several articles here on CodeProject that discuss section handlers that you should search for.

Microsoft MVP, Visual C#
My Articles

Re: function usage in a .config file

RB@Emphasys19-Jul-04 9:02

RB@Emphasys

19-Jul-04 9:02

Hey Heath,

As I stated in my orginial post, I am forced to use an open source project that requires me to use the .config file. Is there a way to use COnfigurationSettings.AppSettings in the .config file?

Thanks,
Ryan

Re: function usage in a .config file

Heath Stewart19-Jul-04 9:07

Heath Stewart

19-Jul-04 9:07

What does that have to do with open source? ConfigurationSettings.AppSettings and Application.StartupPath should both be available for any project that is run against an implement of the CLI (.NET, Mono, and Portable.NET). Read about the <appSettings> section in the .NET Framework SDK.

Open source just means the source is available for anyone to use and modify according to the license agreement. What language, framework, platform, etc. is not a factor of whether or not something is called "open source". Technically, every project posted on this site that provides source code is open source.

Microsoft MVP, Visual C#
My Articles

Re: function usage in a .config file

Andy Brummer19-Jul-04 9:12

Andy Brummer

19-Jul-04 9:12

you only store data in the config file. There isn't any environment variable type of mapping unless the library provides it.

You probably have to set the logging directory programatically whenever the application directory changes.

I can imagine the sinking feeling one would have after ordering my book,
only to find a laughably ridiculous theory with demented logic once the book arrives - Mark McCutcheon

Re: function usage in a .config file

RB@Emphasys19-Jul-04 9:48

RB@Emphasys

19-Jul-04 9:48

Thanks Andy, thats what I needed to know. I wasn't sure if there might be some sort of environment variables.

Ryan

ApplicationException VS Exception

Gimpy198319-Jul-04 5:00

Gimpy1983

19-Jul-04 5:00

Hi!

I was wondering... Is it possible that ApplicationException (or any other Exception class for that matter) catch some errors that Exception cannot catch? ApplicationException inherits from Exception so I assume Exception would catch every errors ApplicationException would.

Thanks a lot
P.S.: btw, I haven't seen such a thing... yet Poke tongue | ;-P

----------------------
Be proud. Be yourself.

Re: ApplicationException VS Exception

Heath Stewart19-Jul-04 5:30

Heath Stewart

19-Jul-04 5:30

Exceptions are caught in order:

try
{
}
catch (TargetException e) // Derivative of ApplicationException
{
}
catch (ApplicationException e) // Derivative of Exception
{
}
catch (Exception e)
{
}

If an exception was thrown and was a TargetException, then it would be caught by the first catch block. If it was any other derivative of ApplicationException (or ApplicationException itself), then the second target block would catch it. If it was any other Exception, the last line will catch it. If you don't plan on using the variables, you can drop them (just use catch (Exception). If you don't plan on using the exception information, you can simply use catch to catch all exceptions.

If you reverse the order of the first two catch blocks, the first catch block will catch all derivatives of ApplicationException, including an ApplicationException itself. In fact, this generates a compiler error.

You should take a look at the C# Language Specification[^].

Microsoft MVP, Visual C#
My Articles

Called but not executed :confused:

Gimpy198319-Jul-04 4:54

Gimpy1983

19-Jul-04 4:54

Does anyone has ever encountered this problem?
I called a method but it doesn't execute WTF | :WTF:

(btw it's executed most of the time but sometimes it's not)

Here's what I've done to test this:
I call the method normally and the first line in the method is a Trace.WriteLine instruction.
So I guess either of two things happen:
1. Trace is not working properly (which would be pretty bad)
2. My method really is not called (which I think is even worse)

but everything after the call seems normal...

I'm thinking maybe
ThreadPool.QueueUserWorkItem(new WaitCallback(SendData), DataObj);
causes the problem... but it doens't seem to get there since the Trace.WriteLine instruction mentioned above is not executed (well no text is in my log!)

----------------------
Be proud. Be yourself.

Re: Called but not executed :confused:

Heath Stewart19-Jul-04 5:35

Heath Stewart

19-Jul-04 5:35

Tracing is not a replacement for debugging. If you really want to know if your method is being called, insert a breakpoint at the beginning of the method and click Debug->Start, or hit F5.

Trace messages are only visible if you set up trace listeners. By default, the only trace listener calls the native OutputDebugString API, which will show up in the VS.NET Output window (or any other registered debug window) when you debug your application (using the steps above). If you want to add additional trace listeners, use either the Debug.TraceListeners or Trace.TraceListeners to add instances of a TraceListener (there are some built-in, documented in the .NET Framework SDK documentation for these properties and types I'm mentioning) or the >lsystem.diagnostics/trace> configuration section (also documented in the .NET Framework SDK.

Also keep in mind that Debug method are only compiled into your assembly when the DEBUG symbol is defined at compile-time (default for Debug builds). Trace methods are only compiled into your assembly when the TRACE symbol is defined at compile-time (default for Debug and Release builds).

Microsoft MVP, Visual C#
My Articles

Re: Called but not executed :confused:

Gimpy198319-Jul-04 9:27

Gimpy1983

19-Jul-04 9:27

Ok thanks for the advice about trace not replacing debug Smile | :)

The thing is I'm trying to debug a network application (3-4 instances running locally or on computers in my LAN for now) and it's kinda hard to know what's really happening if I interrupt the normal course of execution by using debug. Do you have a suggestion other than tracing since it doesn't seem to be too good apparently.

Thanks

----------------------
Be proud. Be yourself.

Re: Called but not executed :confused:

Heath Stewart19-Jul-04 9:32

Heath Stewart

19-Jul-04 9:32

Tracing is fine as long as you implement it correctly (but still will never give the sort of state information debugging can). As I mentioned in my previous post, make sure you actually have trace listeners installed that will work correctly. The DefaultTraceListener won't work since debugging messages are sent only to the local computer (though remote debugging services may dispatch them elsewhere - I honestedly never tried this approach). You could always write your own TraceListener derivative that sends messages using MSMQ, Web Services, or some other standard or proprietary mechanism.

Microsoft MVP, Visual C#
My Articles

Going Unmanaged doesn't free Memory

Guilio karadanais19-Jul-04 4:24

Guilio karadanais

19-Jul-04 4:24

I've got a memory issue with an unmanaged call to GDI.

I've followed the CreateDC, releaseDC, DeleteDC step outlined in the platform SDK,these methods all return true or 1, indicating that they succeeded. However after each call to the method it consumes 3 mb of RAM and does not release the memory, the issue is that method will be called many times and builds up to about 60mb memory used.

I store a bitmap in a member variable, which i dispose of.

iWidth = nWidth;
iHeight = nHeight;
hdcSrc = User32.GetWindowDC(User32.GetDesktopWindow());
hdcDest = GDI32.CreateCompatibleDC(hdcSrc);
hBitmap = GDI32.CreateCompatibleBitmap(hdcSrc,iWidth,iHeight);
image = new Bitmap(iWidth,iHeight);

GDI32.SelectObject(hdcDest,hBitmap);
GDI32.BitBlt(hdcDest,0,0,iWidth,iHeight,hdcSrc,iStartX,iStartY,SRCCOPY);
image = Image.FromHbitmap(new IntPtr(hBitmap));

User32.ReleaseDC(User32.GetDesktopWindow(),hdcSrc);
GDI32.DeleteDC(hdcDest);
GDI32.DeleteObject(hBitmap);

if I call GC.Collect() it clears the memory and brings it back down but this is oviously not a good idea.

Any Idea's?

Re: Going Unmanaged doesn't free Memory

Heath Stewart19-Jul-04 4:37

Heath Stewart

19-Jul-04 4:37

nEgAtIvE cReEp wrote:
if I call GC.Collect() it clears the memory and brings it back down but this is oviously not a good idea

That should be your clue right there. Unmanaged resources are called "unmanaged" because they are not managed by the GC. If you call GC.Collect and it makes a difference, then unmanaged reosurces are not your problem.

The problem here is that you must dispose those images you're creating (the image) variable. When finished (and be re-assigning image to something else), call image.Dispose(); otherwise, the Image is not cleaned-up until the GC gets around to it or you call GC.Collect.

Microsoft MVP, Visual C#
My Articles

Re: Going Unmanaged doesn't free Memory

Guilio karadanais19-Jul-04 5:07

Guilio karadanais

19-Jul-04 5:07

Thanks for the reply.

I do dispose of the image variable, I just didn't include the code on the post. Although I realize my comment on the GC doesn't make sense.

I will have to investigate if any .NET objects are still hanging around, that acompany the Call to GDI) i.e the form.

thanks again

Hex to Dec conversion (big numbas!)

Blubbo19-Jul-04 4:06

Blubbo

19-Jul-04 4:06

I'm working on the hex conversion software for myself and engineering use. I'm trying to get the int value of oxFFFFFFFFFFFFFFFF as the windows XP calculator can figure this.

I'm ok for 15 characters long in Hex but not 16.

I have been getting the result of -1 dec value.

What I did was doing by:
(note that the editbox in windows application are: edtHex (editbox for Hex values) and edtDec (editbox for Dec))

long tempDec;
tempDec = long.Parse(this.edtHex.Text, System.Globalization.NumberStyles.HexNumber);
edtDec.Text = tempDec;

Anyone have any suggestion on how I can correct this?
Am I correct that this "long" is 64bit in XP?

Re: Hex to Dec conversion (big numbas!)

Heath Stewart19-Jul-04 4:30

Heath Stewart

19-Jul-04 4:30

A long (C# alias for Int64) is 64 bits on any platform.

16 F's for a long is -1. 16 F's for a ulong, however, is 18,446,744,073,709,551,615. It's important to understand the difference between signed (like int and long) and unsigned (like uint and ulong) types: signed can be negative, and a negative value is when the most significant bit is 1.

It's also important to note that unsigned primatives are not CLS-compliant, meaning that only CLS-compliant languages (i.e., no additional support for the CLI) like JScript.NET cannot use them. Both C# and VB.NET support unsigned types, though.

Microsoft MVP, Visual C#
My Articles

EAP Certificate requests...

Vodstok19-Jul-04 4:03

Vodstok

19-Jul-04 4:03

I dont even know if this is possible , but...

The company i work for uses 802.11x for wireless authentication. We have an issue regularly where the certificates expire or become invalid, because usually because the user has changed their password recently, and a bug in Microsoft's software causes it to fail on the automated request.

What i am looking to do is create a web page that will sit on our intranet that will allow users to request a new certificate based on who they are. We currently have to walk the user through going into MMC and walking through the certificate request procedure.

This is particularly timely since we had an issue recently that expired every certificate on the network....

i do know there is a certificate namespace, and since MMC allows for the request, there MUST be a way to use the object for this procedure, I am just at a loss where to start tocreate a page for it.

_________________________________________
Yes, I am the highly suggestable type.

Re: EAP Certificate requests...

Heath Stewart19-Jul-04 4:42

Heath Stewart

19-Jul-04 4:42

Microsoft Certificate Services actually already has this functionality, and you can set up a CA site easily. I suggest you read the product documentation for Microsoft Certificate Services. I set up such a site with very few steps some time back on our network and it was quite easy, but I don't remember all that I did.

Microsoft MVP, Visual C#
My Articles

Use Ctrl+Left/Right to switch messages, Ctrl+Up/Down to switch threads, Ctrl+Shift+Left/Right to switch pages.