|
Perhaps this will do the trick:
<br />
StreamReader streamReader = new StreamReader("MyFileList.txt");<br />
<br />
while (streamReader.ReadLine() != null)<br />
{<br />
ListViewItem listViewItem = new ListViewItem(streamReader.ReadLine().Split(new char[] { ',' }));<br />
listView4.Items.Add(listViewItem);<br />
}<br />
<br />
streamReader.Close();<br />
|
|
|
|
|
OK, forget the Dog/Vet metaphor I posted yesterday.[^]. It was the easiest way to describe my original question which was general in nature and not specific to my problem. Here is the actual scenario.
I have an editor (type of data being edited is irrelevant). In this editor, I want to support undo/redo operations. I have employed the Command pattern for this purpose. All changes to underlying data must be encapsulated in a command object. For example, to set text you would create a SetText command and call its “Execute” method. To undo this operation, you would call the SetText object’s “Unexecute” method.
This architecture is working really well but it has one flaw. Every time an engineer executes a command he is responsible for managing command history (i.e. the undo/redo stacks). If the execution is successful, he MUST add the command to the undo stack. If the execution fails, he MUST NOT add the command to the undo stack. If the execution throws an exception, he MUST CLEAR THE STACK as we cannot allow the undo stack at any time to be out of sync with the application state (this logic is somewhat simplified for the purposes of this post).
If the engineer forgets to do any of these things he poses a risk of undefined behavior in the application. Furthermore, the logic will be the same for every command that gets executed, so there’s no need to write this command-history-management-logic every time a command is executed.
In steps the factory pattern. The factory encapsulates the creation of commands. Because the factory controls the creation of the commands, it can setup a relationship between the UndoManager class (manages command history) and each command that gets created. Voila!
Now here’s my problem (and perhaps it’s not a problem at all. Perhaps this is where I’m rubegoldberging). Any engineer can create a command and execute it. If the UndoManager class has no knowledge of this command, he cannot manage it and the undo history could get out of sync with the application state. Obviously this is really bad.
To solve this problem, I have nested a private UndoManager class in the command factory. This class implements the IUndoManager interface. All commands require an IUndoManager at construction. So, because the UndoManager is private to the factory, only the factory can create commands. For example, to create a generic command:
public ICommand GetGenericCommand()
{ return new GenericCommand(this.undoManager); }
It is true that any engineer could implement IUndoManager and then begin creating rogue commands, but this is unlikely.
I would love to hear comments on this solution. Am I being overly cautious? Is this a hack? Or have I implemented a reasonable safeguard in a reasonable fashion? Is there a pattern that matches this solution? (Hopefully not an antipattern… )
Thanks!
|
|
|
|
|
Edmundisme wrote: I would love to hear comments on this solution.
Sounds reasonable to me. 
led mike
|
|
|
|
|
 Edmundisme wrote: This architecture is working really well but it has one flaw. Every time an engineer executes a command he is responsible for managing command history (i.e. the undo/redo stacks). If the execution is successful, he MUST add the command to the undo stack. If the execution fails, he MUST NOT add the command to the undo stack. If the execution throws an exception, he MUST CLEAR THE STACK as we cannot allow the undo stack at any time to be out of sync with the application state (this logic is somewhat simplified for the purposes of this post).
If the engineer forgets to do any of these things he poses a risk of undefined behavior in the application. Furthermore, the logic will be the same for every command that gets executed, so there’s no need to write this command-history-management-logic every time a command is executed.
In steps the factory pattern. The factory encapsulates the creation of commands. Because the factory controls the creation of the commands, it can setup a relationship between the UndoManager class (manages command history) and each command that gets created. Voila!
I was about to suggest that the logic for managing the undo stack be placed in the commands, i.e. when a command is executed, it's responsible for placing itself on the undo stack (or not depending on the results). But it sounds like you've solved this problem?
Edmundisme wrote: Now here’s my problem (and perhaps it’s not a problem at all. Perhaps this is where I’m rubegoldberging). Any engineer can create a command and execute it. If the UndoManager class has no knowledge of this command, he cannot manage it and the undo history could get out of sync with the application state. Obviously this is really bad.
Underlying the commands is a document of some type that is being acted upon, correct? You're concern is that rogue command objects can be created that execute an action against a document. Being rogue command objects, they're not being managed, so they're never placed on the undo stack? As a result, a document's history could become corrupted.
The question is how could a rogue command object obtain access to the document in the first place?
I haven't looked at GOF in awhile regarding the Command pattern as it's applied to undo functionality, so I'm not sure how they implemented it. But how about the document itself being responsible for creating commands. The document has several private Command classes that implement an interface. And perhaps the document holds an UndoManager object for managing the commands.
When a client wants to execute an action against a document, they first ask for the appropriate command. The document gives them back a command it has created and is responsible for. The client, however, only sees the command object through some generic interface. This approach makes rogue commands impossible.
Edmundisme wrote: To solve this problem, I have nested a private UndoManager class in the command factory. This class implements the IUndoManager interface. All commands require an IUndoManager at construction. So, because the UndoManager is private to the factory, only the factory can create commands. For example, to create a generic command:
public ICommand GetGenericCommand()
{ return new GenericCommand(this.undoManager); }
This may be close to what I've suggested above, I'm not sure. I need to know more about how the actual document, or model, is being used by the commands.
|
|
|
|
|
I like your suggestion about encapsulating the document access - forcing modifications through a single access point. Unfortunately I'm adding Undo/Redo to an existing application and this solution would require a complete rework of the document management.
Basically, if I can get everyone to understand that all modifications must be done via command objects we'll be OK. However, any code at anytime could directly change the document... 
Thanks for your ideas!
|
|
|
|
|
Hi,
I need to develop a commercial use application and I found
some code source that is under GNU version 2, license term agreement.
For GNU, does it mean that I can incorporate the binary format (the dll)
of the code (I didn't modify the source code) into my application
without contributing back my application source code ?
Thanks
|
|
|
|
|
As far as I know using GNU code forces you to label your application GNU too.
protected internal static readonly ... and I wish the list could continue ...
|
|
|
|
|
I beleive that under both GPL and GNU open source liceneces, you must make your source code freely available.
Linksys found themselves having to publish the source code for thier router firmware:
http://www.wi-fiplanet.com/tutorials/article.php/3562391
Hope this helps
Regards
Wayne Phipps
____________
Time is the greatest teacher... unfortunately, it kills all of its students
View my Blog
|
|
|
|
|
But the case is that I doesn't modify their source code, just use it as a dll, and this will lead to open source my part of application code also ?
|
|
|
|
|
Yes. There might be a slight hope for you if and only if you only use the library only, and not the source code (including not using any header files), ie maybe if you were doing something like distributing a GPL plug-in dll for your application, you might be safe from open-sourcing your product. Even in that best case scenario, you might still lose, and you definitely would be open to legal problems.
More than likely (say you used a library like the GNU Scientific Library, which does things like solve equations, and which has a GPL license), you would have to release your entire product under the GPL license. That means two things. The first is that whenever you give your software to someone, you have to include all the source code (I dont know whether it would include things like associated media files). Secondly, whoever legally obtains your software/source code can redistribute it in original, or modified form (which themselves has to be GPL).
Basically, unless you want to have your application be put on Sourceforge, when the first person who buys your software decides to fork your project, (and as someone who makes my living writing OSS, I have to say there are some great, though not always applicable reasons for doing this) DO NOT INCLUDE THAT CODE IN YOUR APPLICATION, either in modified or unmodified form.
|
|
|
|
|
I'm trying to use the the following code to show pdf or xls file generated by ReportViewer.
After application closes, temporary files remain in disk.
How to delete PDF files after Adobe Reader is closed ?
Is the best way to save pdf files to temporary directory and delete whole directory if application exits or use filewatcher?
Any sample how to implement this ?
void ShowPDF( byte[] bytes ) {
static int i=0;
string outputfile = "temp"+(i++).ToString()+".pdf";
FileStream fs = new FileStream(outputFile, FileMode.Create);
fs.Write(bytes, 0, bytes.Length);
fs.Close();
System.Diagnostics.Process p = new System.Diagnostics.Process();
p = new System.Diagnostics.Process();
p.StartInfo.FileName = outputFile;
p.Start();
}
Andrus
|
|
|
|
|
Hi,
I have a preview form for several type of contents.
I want to adapt the form size to the contents size dynamicly.
For example, I have a mediaPlayer COM named mpPreview.
I tried to do this:
this.Size.Height = mpPreview.Size.Height;
But it gives me the following error:
Error 1 Cannot modify the return value of 'System.Windows.Forms.Form.Size' because it is not a variable D:\Os meus documentos\Visual Studio\Coordinator\Preview.cs 78 13 Coordinator
Why?
Thx,
Nuno
|
|
|
|
|
Try using this:
<br />
this.Size = new Size(this.Size.Width, mpPreview.Size.Height);<br />
|
|
|
|
|
To understand the reason why C# prevents you from doing this, you need to understand the basics of value types (or structures).
Unlike reference types, when value types are returned from methods (or set via an assignment), a copy of the value type is returned and not the orginal version.
In the situation above, Form.Size is a property (which is a method underneath) and it returns the Size value type.
So when you attempt to assign a value to Size.Height, you are actually attempting to assign it to a copy and not the underlying Size structure.
Although, it is technically valid to do this, the C# compiler makes the assumption that this is not what you want to do, and hence fires the compiler error.
Excelsior
Arjun Bahree
"By The Might of Mjolnir"
I Came! I Coded! I Conquered!
|
|
|
|
|
Use a this.Height instead of this.Size.Height and it will work.
Excelsior
Arjun Bahree
"By The Might of Mjolnir"
I Came! I Coded! I Conquered!
|
|
|
|
|
Thx guys!
This effort for now was useless.
My content isnt resizing, so i cannot see this working.
Basicly i have three kind of elements in this form:
- web browser
- media player
- picture box
I want this form to popup when i click previw in a list of element of several file types.
For now i'm just focusing in videos.
I want my media player to resize to current video's size.
Is there any easy way of doing that, without having to see the movies information in order to get its size?
Thx,
Nuno
|
|
|
|
|
Hi,
I'm building a MDI application which child forms will be running inside a splitContainer panel.
Well, when i maximize i have two problems:
- The form title bar seems weird (usually, in other applications, the title bar gets flat and not with that rounded blue thing)
- If i move the spliter, the windows doesnt get maximized automaticly.
Is there anything i can do to solve this?
Thx
Nuno
|
|
|
|
|
I am developing an xml playlist(in attribute style) based slide show. I am able to read all the information from xml assign them to variables and then to respective picture boxes and text box. The problem is it waits till the end of the loop to go back to the function that called this process in the form_load method and then display the last file in the play list and exit the loop without ever displaying the first two images
private void Form1_Load(object sender, EventArgs e)
{
CheckForFile(); }
public void CheckForFile() {
LoadPlaylist((System.Convert.ToString(@"C:\PlayerLibrary\config\ab.xml")));
}
public void LoadPlaylist(string strFileName)
{
string strItemName = "";
XmlTextReader playlistReader = new XmlTextReader(strFileName);
while (playlistReader.Read() && (playlistReader.NodeType == XmlNodeType.Element || playlistReader.NodeType == XmlNodeType.Whitespace))
{
if (playlistReader.HasAttributes && playlistReader.Name == "name")
{
strItemName = playlistReader.GetAttribute("location");
playFile(strItemName);
break;
} } playlistReader.Close(); }
public void playFile(string playFileName)
{
string fileExtention = "";
fileExtention = playFileName.Substring(playFileName.Length - 3, 3);
//check if the file is image type then display it using pBoxPlayer
if (fileExtention == "bmp" || fileExtention == "jpg" || fileExtention == "JPG" || fileExtention == "BMP" || fileExtention == "gif" || fileExtention == "Gif" || fileExtention == "PNG" || fileExtention == "png")
{
this.pBoxPlayer.Load(playFileName);
this.pBoxPlayer.Visible = true;
this.Show();
this.pBoxPlayer.Show();
} } }
sam
|
|
|
|
|
Hello,
If you really whant to stay with the loop, you have to work with:
this.Refresh();
Application.DoEvents();
Thread.Sleep(xxx);
I would rather get the informations out of the XML and then use a Timer to step threw the list.
All the best,
Martin
|
|
|
|
|
Dear CPians,
I have been working with the Xml.Serialization namespace in order to serialize C# objects to Xml. In some of my Classes I have long string members that I need to enclose within CDATA sections when serialized. I have been struggling to find a way to do this.
Any ideas?
Thanking you in advance.
|
|
|
|
|
Hi,
I'm just knocking up a quick program which works out how much savings I can save by entering a deposit, a number of weeks i want to save and then outputting me a total. I have got the program working fine, but if my total saving were £150.50 my total field only displays £150.5
Is there a currency format i'm missing? Here's my code:
namespace Savings_Calculator
{
public partial class Form1 : Form
{
public Form1()
{
InitializeComponent();
}
double deposit;
int weeks;
double total;
private void btnCalculate_Click(object sender, EventArgs e)
{
deposit = Convert.ToDouble(txtDeposit.Text);
weeks = Convert.ToInt32(txtWeeks.Text);
total = deposit * weeks;
txtTotal.Text = "£" + Convert.ToString(total);
}
Many thanks.
|
|
|
|
|
I guess it would be best to use the 'decimal' type, and you can print it nicely with .ToString("N2") (2 decimals displayed)
Visual Studio can't evaluate this, can you?
public object moo<br />
{<br />
__get { return moo; }<br />
__set { moo = value; }<br />
}
|
|
|
|
|
Hi all, I'm hoping someone ca help me out here as I've been stuck on this one for a while now.
I have a DLL (written in C++) that exposes the following function:
int sendAPDU (byte* TxAPDU, int TxLen, byte* RxAPDU, int* RxLen)
Essentially, this function takes an array of bytes (TxAPDU) with TxLen elements and returns an array of byte (RxAPDU) with RxLen elements. The key issue here which is throwing me is the fact that the last two parameters are reference parameters (or strictly speaking output parameters).
My problem is that, in C#, I can't seem to figure out how to import that function from the DLL and call it. I thought it would be something like:
[DllImport("MyLibrary.dll", EntryPoint = "sendAPDU")]<br />
public static extern int sendAPDU(Byte[] TxAPDU, int TxLen, ref Byte[] RxAPDU, ref int RxLen);
And then when calling the function it'd be something like:
Byte [] TxArray = new Byte[26];<br />
Byte [] RxArray = new Byte[MAX_APDU_LEN];<br />
int RxLength;<br />
<br />
int retval = sendAPDU (TxArray, 26, ref RxArray, ref RxLen);
So what is it that I'm doing wrong? Forgive me if this is a straight forward issue but I'm ew to C#.
|
|
|
|
|
Hi..
I don't know if i can help you on this because i'm not very experienced about PInvoke but just a few thoughts:
1. You have to possibility to Marshal Function Parameters in .Net so maybe the MarshalAs - Attribute could help you.
something like:
<br />
[DllImport("MyLibrary.dll", EntryPoint = "sendAPDU")]<br />
public static extern int sendAPDU([MarshalAs(UnmanagedType.LPArray)] Byte[] TxAPDU, int TxLen, [MarshalAs(UnmanagedType.LPArray)]ref Byte[] RxAPDU, ref int RxLen);<br />
the MarshalAs Attribute can be found under System.Runtime.InteropServices
2. i don't know if this makes a difference in PInvoke but did you try to use out instead of ref for your output parameters?
greets
m@u
|
|
|
|
|
Thanks m@u. I was afraid it might have something to do with Marshaling because I bumped into that once before and never really figured it out. I suppose I still don't understand this whole managed memory versus unmanaged memory thing.
At any rate, I tried your suggestion and read up on the MarshalAs attribute but it still didn't solve my problem and I still don't think I really understand how to use this MarshalAs attribute but to my great relief, the solution to my problem popped out of virtually nowhere.
The C++ programmer in me still pondered on this principle of passing pointers as funtion parameters and it occured to me that, if in C++ passing an array and passing a pointer is one and the same thing, doesn't the same apply for C# and reference parameters?
So I changed my function declaration which follows the DLLImport attribute from:
public static extern int sendAPDU(Byte[] TxAPDU, int TxLen, ref Byte[] RxAPDU, ref int RxLen);
to:
public static extern int sendAPDU(Byte[] TxAPDU, int TxLen, Byte[] RxAPDU, ref int RxLen);
and would you believe it, it works!
I'm still unsure about the memory management issue that this introduces. In C++ this would have left room for serious memory overflow issues, so what happens in C# if you call a function from a DLL written in C++? Let's say, I have a function like this:
public static extern int myFunc(byte[] ByteArray);
which, in the DLL, does something like this:
ByteArray[10] = 10;
Now, if you call this from C# like this:
byte[] sendArray = { 1, 2, 3 };<br />
int retval = myFunc(sendArray);
Clearly in C++ this would have written the value 10 into memory, 7 bytes past the end of the memory allocated to sendArray, causing havoc. So what happens when you call the DLL function from C#?
Sorry guys, I realise that this is one of the central issues to C# and you can't claim to be a C# programmer if you don't understand managed vs. unmanaged memory. But I am a noob.
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
