I'm curious if you know of any good documentation on how to convert a C# string to an MFC CString. A C# application I'm working on needs to call a C++ function exported by a dll, and the prototype of that function includes several MFC CString objects.
I suspect it's trivial to do, but that it involves some esoteric knowledge of the built-ins that I just don't have. I'm also hoping to circumvent the use of managed C++ to do this but simply preparing the correct byte-format structure that can sit in and be accepted and properly interpreted as a CString by the C++ function.
The first thing to do is read the documentation for the DllImportAttribute in the .NET Framework SDK, as well as the MarshalAsAttribute. The first is how you P/Invoke native functions. Since an MFC CString can be cast to an LPTSTR, you use use the UnmanagedType.LPTStr in the MarshalAsAttribute for the string parameter. There's also several articles about P/Invoke and interoperability in the .NET Framework SDK that should help.
I have any unusual question for you all. My piqued interest in the Thai lanugage/culture has prompted me to develop my own Thai Learning System. For the various consonants and vowels I want to animate how to draw each symbol. I was thinking of using a Thai True Type font and for each character animate along it to demostrate how it is drawn.
My problem is how can I read the true type fonts coordinates and animating along the symbol? Perhaps I need to hook the event that actually paints the symbols on the canvas then with my own event actually paint the symbol in some sort of time lapse fashion to present it in an animated fashion.
My other thought was to create a bunch of methods that draw each symbol as lines/arcs/Bezier Curves, but I figured it would be easier tracing the true type font in a generic fashion. With the lines/arcs/Bezier Curves I don't know how to get teh GDI+ engine to animate the points. Is there a way?
If anyone has any other suggestions or a better solution, which I am sure exist, please let me know.
First of all, .NET doesn't contain such low-level hooks. All you could do is draw the fonts yourself. I'm not even aware of any way to control fonts to such a degree exposed in the Win32 APIs and I've spent over a decade getting headaches from them (though such headaches might have lead to temporary blindness, causing me to miss such functionality )! This would be a function of GDI so you could look into the documentation.
Besides, true type fonts don't really contain any data that would allow you to show how each character is drawn. They contain points that help with resizing but those aren't in any order that is representational to how a human would draw them.
Your best bet will be to - as you didn't want to do - use lines, beziers, etc., to draw the characters.
Might I recommend Macromedia Flash, though? I've seen a lot of examples of similar presentations that actually draw things in steps. You would also be able to easily host this in a web page (you can with .NET controls, but it requires the ~20 MB framework and a pre-installed security policy to allow your code form a specific site to run), any COM client, or even a .NET application (which can be a COM client).
If you wanted to use GDI+ in a .NET application, you would have to worry about all the animation. For example, if you use Graphics.DrawLine to draw a line, the line is just there. It won't be animated. You'd instead have to call Graphics.DrawLine repeatedly to draw longer and longer lines until you reached the desired length! That's easy, though, compared to curves - you would have to calculate each point in the spline or the start and sweep angles for arcs with each iteration of the animation!
You could use Managed DirectX - which will help you with graphics a little - but you'd still need to worry about creating a package to handle the character animation. Things like Flash and even DHTML with vector graphics already have this support - you just have to tell it how to animate what you want.
While the GraphicsPath and Matrix classes aren't supported in the Compact Framework, the Graphics class is. It won't be easy since you'll have to do a lot of calculates instead of harnessing the power of the Matrix (and I thought the "Matrix" was a bad thing! ), but you could perform all the drawing using the methods of the Graphics class.
I've stumbled upon a behavior of the .NET RichTextBox that I think is a bug, and I wanted somebody to do a sanity check before I report this to MS.
Below is a code listing. It displays a form with a RichTextBox control in it. Ctrl++ and Ctrl+- increase/decrease zoom. Space and Backspace keys change content of the text box.
Currently, zoom step is 0.5. Notice that when zoom factor is 2.0 (you can make it that by pressing Ctrl++ twice), the text box doesn't retain the zoom factor when its content changes. When you uncomment MessageBox statement in DisplayText (uses Clear() method of the text box class) method, you can more glimpse of the weird behavior. It is actually different with the MessageBox there. Increase zoom to 2.0 by pressing Ctrl++ twice and then press Space or Backspace repeatedly and notice zoom factor change from 2.0 to 1.0 to 2.0 to 1.0, …
After that, change the zoom step to, say, 0.3 (zoom factor doesn't become 2.0 with this step), and notice that the box behaves as expected (the zoom factor remains as set).
Is this a bug or reflection of my wrong expectations?
#region Windows Form Designer generated code
/// Required method for Designer support - do not modify
/// the contents of this method with the code editor.
private void InitializeComponent()
this.richTextBox1 = new System.Windows.Forms.RichTextBox();
this.richTextBox1.Anchor = ((System.Windows.Forms.AnchorStyles)
I'm seeing this behavior too, odd. I don't see anything wrong with your code right off hand (hey, just have to check before filing a bug ) and looking at the IL for the RichText.set_ZoomFactor (and a subsequent call to a private method) doesn't show any calculation bugs either. I'm guessing this is actually a problem with the Rich Edit common control somewhere in the EM_SETZOOM message handler.
So, googling for "Rich Edit" EM_SETZOOM[^] shows you're not the only one! Feel free to report, though. Maybe someone hasn't yet. Never know.
Problem: I need to call a function when either a combo box selection is changed or some new text is typed into the combo box.
SelectedIndexChanged takes care of the first half, but I can't figure out the OnTextChanged part.
Does anyone know what event is triggered when a user types in text into the ComboBox?
I know there are Enter and Leave events... but then I will have to keep track of whether the text changed myself, and that's my very last resort, because I have a gut feeling (called laziness ) that there has to be a better way to do this
I don't see the advantage of using the KeyDown event instead of the Leave event. I got to store the value and do everytyhing manually anyway.
Plus neither KeyDOwn nor Leave will catch the cases when the ComboBox is set programmatically.
I guess there is no ComboBox.OnSelectedValueChaged silver bullet.
The TextChanged event fires if your ComboBox.DropDownStyle is set to ComboBoxStyle.DropDown. The unfortunate part is that is fires for every character typed, so you might want to handle the LostFocus event or the Validating event in order to add the new text or to call the function after the text is entered in its entirety.
For those interested, here is what I ended up doing.
But first a little mode details on my case: it's a login dialog, and whenever the value in ServerComboBox changes I have to go and pull a list of databases off the server and put it into the DBComboBox.
This might look like an overkill with ServerComboBox.Leave, ServerComboBox.SelectedIndexChanged, and DBComboBox.DropDown handlers.
I would skip Leave & SelectedIndexChanged and just do DropDown if I didn't have to make a call to the server. When I do the call the the server on DropDown, the GUI hangs a little, so it's better to do it right when the ServerComboBox is changed.
Why do I have the DropDown handler at all: there is one case when Leave & SelectedIndexChanged don't do it: on startup when there is no history in the registry and we use the hardcoded default value for the ServerComboBox.
I don't know if I am making sence or it's too specific to my problem. Here it is anyway:
Let me know if you see any way to improve the code.
So here it is:
/// The DbComboBox is dependent on the ServerComboBox and has to display the databases
/// available for the selected server.
class DbComboBox : ComboBox
string m_onEnterServerName = ""; //the name of the server we pulled databases from the last time
/// <param name="ServerComboBox">server combo box that we need to display Dbs for</param>
public DbComboBox(ComboBox ServerComboBox):base()
m_ServerComboBox = ServerComboBox;
//handles user changes but not programatic updates
//programatic updates are handled by SelectedIndexChange and a call to RefreshDbHistory
//from Server property in the SignIn dialog
m_ServerComboBox.Leave += new System.EventHandler(ServerChangedEventHandler);
//handles selection change, but not when the user types in something new
m_ServerComboBox.SelectedIndexChanged += new System.EventHandler(ServerChangedEventHandler);
//in case there is no history and login defaults we need to load DBs for the
this.DropDown += new System.EventHandler(ServerChangedEventHandler);
/// Event handler that is called when we suspect that the selection for the server combo box has changed
/// <param name="sender"></param>
/// <param name="e"></param>
private void ServerChangedEventHandler(object sender, System.EventArgs e)
/// Goes to the server, pulls databaselist, populates combobox
public void RefreshVaultDbHistory()
string newServer = m_ServerComboBox.Text;
if ((!newServer.Equals("")) && (!newServer.Equals(m_onEnterServerName)))
I have a software package that gets installed on a local sql server and the customer can run as many clients as they need. What I need to be able to do is license the package by securing the sql server. For each "sub-office" there should be a license. Each sub-office has its own database on the sql server.
I know that this isn't the right forum. I just though maybe I can get a quick answer since this one has a larger audience. If anything I can implement a custom solution not based on the sql server. Suggestions welcomed.
So basically you want to license the database itself? We have a similar setup with our flagship product where a single database manages the accounts that can access each of the company databases (or sub-office database in your case). Similar to you, we allow unlimited user accounts. You could even forego this and just use Windows Authentication using SSPI. The thing that maintains how many people can access each database at a time is a .NET Remoting object that caches the current license count for each company (just in case the Remoting object shuts down or something). It hands out licenses and takes care of decrementing those counts when a client disconnects. The latter part is problematic, however, since if your client application crashes there's not really any good way to decrement the count since the clean-up code didn't run. If you don't have to expose your Remoting object through IIS in order to hang-off port 80, you could always poll clients by implementing your own ILease and returning that in GetLifetimeService.
As far as limiting connections in SQL Server, all you can do is limit licenses for a SQL Server instance, but that's nothing that a DBO couldn't change.
Awesome! That is exactly what I was thinking about doing.
One little question. I have my clients automatically find the sql server (multiple sql servers are possible also, I do searching using SQLDMO.ListAvailableSQLServers). What would be the best way of storing the URI? I was thinking maybe store it on the SQL database and have the client look it up. It URI doesn't work or license check fails then client will shutdown.
You have to split this up into two part. First convert then encode and send to server.
Take a look at this site: http://www.verypdf.com/ and see PDFcamp/PDFwriter - it looks like this is the least expensive way of converting html to pdf. If you are looking for more pro components then around at: www.pdfstore.com (Some of the prices out there are scare if you are a startup and dont have much money.)
To put the pdf in database just read the file in and use Convert.ToBase64String() and send it to the database.
You could also - depending on the RDBMS - store the PDF as a binary stream, which some ADO.NET classes (like those in System.Data.SqlClient) support. IMO, you should store the PDF on the filesystem and either store the path in the DB or use the primary key as part of the filename, which is what we do in our flagship product.