C# Discussion Boards - CodeProject

11-May-04 4:29

Because it's faster than loading your own literal "" when needed. There's still uses for it. Perhaps you want to initialize a string member as an empty string instead of null. Just because something may not be useful in one case doesn't mean it's not useful in another.

Microsoft MVP, Visual C#
My Articles

Convert char* to string

peraonline11-May-04 2:09

11-May-04 2:09

Hello! what can I convert a char* to string?
The cast isn't permitted!
Thanks

bertcox11-May-04 3:26

bertcox

11-May-04 3:26

public void ConvertStringChar(string stringVal) {
char charVal = 'a';

// A string must be one character long to convert to char.
try {
charVal = System.Convert.ToChar(stringVal);
System.Console.WriteLine("{0} as a char is {1}",
stringVal, charVal);
}
catch (System.FormatException) {
System.Console.WriteLine(
"The string is longer than one character.");
}
catch (System.ArgumentNullException) {
System.Console.WriteLine("The string is null.");
}

// A char to string conversion will always succeed.
stringVal = System.Convert.ToString(charVal);
System.Console.WriteLine("The character as a string is {0}",
stringVal);
}

Judah Gabriel Himango11-May-04 3:49

Judah Gabriel Himango

11-May-04 3:49

char* someCharPtr;
string myString = new string(someCharPtr);

#include "witty_sig.h"

peraonline11-May-04 20:32

11-May-04 20:32

ok, I converted a char* into a string as:

string descr= new String(obj[1].descr);
but the problem isn't solved:

//obj[] is:

public struct node {
public int IdObj;
public char* descr;
}

node* obj=getObj();

//getObj is the function included into a DLL

//the string "descr" contain only simbols!!!

thanks for your help!

Heath Stewart12-May-04 2:55

12-May-04 2:55

Why are you using char* in a .NET struct anyway? Stay away from unmanaged types when possible. A char* or char[] is just a string in any language. Even when P/Invoking native functions, you still typically use a string in place of any character array (char*, wchar_t*, TCHAR*, etc.). It will marshal correctly so long as you marshal using the correct character set (see the CharSet property of the MarshalAsAttribute or StructLayoutAttribute if you're doing this to P/Invoke some native function).

Microsoft MVP, Visual C#
My Articles

peraonline12-May-04 3:10

12-May-04 3:10

If i try to replace the char* to string, the compiler make me the following error:
"Impossible accept address/dimension of variable of managed object."
So the only not managed object are char,int,ecc but not string! (independently of "unsafe" environment)

ps: the language used to build the function into a DLL is the C++, so the "string" is a char[].

Heath Stewart12-May-04 3:57

12-May-04 3:57

A string is also a char*. The difference is how the character array (string) is allocated.

There are properties you can set using the MarshalAsAttribute that I mentioned earlier) that specify the maximum size (in characters, not bytes) or a parameter index that does the same within a struct (if one exists). Read the documentation for the MarshalAsAttribute in the .NET Framework SDK for more information. Trust me, my forté is unmanaged interop. I've never had to use a single unsafe context to P/Invoke anything from basic to complex structs and function calls. They're really only necessary when you need to boost performance of a block of code; for example, iterating through the bits of a bitmap - using an unsafe context is much faster.

You should also see the Marshal class, which includes methods that further prevent you from having to use an unsafe context (because it does it for you, like alloc'ing memory).

Microsoft MVP, Visual C#
My Articles

peraonline12-May-04 4:03

12-May-04 4:03

Ok, I'll read the MarshalAsAttribute documentation!
Thanks very much for help!

Character Encoding

gUrM33T11-May-04 1:45

gUrM33T

11-May-04 1:45

Do I need to know about the encoding of a file before opening it or is it determined automatically by the framework? If .NET framework cannot automatically determine, then is there any way I can find this out myself? I'm not clear about this character encoding thing, so if you can please explain in detail, that would be of great help.

Thx

Gurmeet

BTW, can Google help me search my lost pajamas?

My Articles: HTML Reader C++ Class Library, Numeric Edit Control

Re: Character Encoding

Mike Dimmick11-May-04 3:33

Mike Dimmick

11-May-04 3:33

StreamReader has several constructors, some of which take an Encoding and/or a boolean value to indicate whether the encoding should be detected or not.

A bit of poking around in Reflector reveals that if you don't provide an encoding, it uses UTF8, and if you don't say otherwise, it tries to detect the encoding rather than use the default UTF8.

When trying to detect an encoding, it uses the Byte Order Mark character. The Unicode standard indicates that this character, U+FEFF, should appear at the beginning of the text in whatever encoding is used. In UTF-16 little-endian, this becomes the byte sequence 0xFF 0xFE; in UTF-8, it's (IIRC) 0xEF 0xBB 0xBF. If there's no Byte Order Mark, it simply uses the encoding specified in the constructor, unless you didn't use one of those variants, in which case it uses UTF-8. .NET can also detect UTF-16BE, or big-endian, where the bytes of UTF-16 are the other way round.

If you use File.OpenText or FileInfo.OpenText, you don't get to specify an encoding.

Unfortunately very few of us have files encoded as UTF-8. They're far more likely to be encoded using our default code page. For most Western European and North American users, this is going to be Windows 1252 (Windows Western). You can get hold of an encoding for the user's configured ANSI code page using Encoding.Default.

Western users, particularly UK, US and Canada, may not notice at first that the encoding is different, because the first 256 code points of Unicode are the same as ISO Latin 1 (a little, though not a lot, different from 1252). Due to the way it's encoded, the first 128 code points of UTF-8 are also the same as Latin 1 and ASCII (ISO-646-US). Any UTF-8 code byte greater than 127 indicates that one or more following bytes needs to be interpreted along with this one to get the full character.

There's no reliable way to detect which encoding is used by a random sample of text in a byte-oriented character stream (which isn't UTF-8). The concept of Byte Order Marks is relatively new. You either have to know or ask the user.

More information links:

Microsoft Global Development Portal[^]
Code Page reference tables[^]
The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode and Character Sets (No Excuses!)
[^]
http://www.unicode.org/[^]
Character Sets[^] (from my blog).

Stability. What an interesting concept. -- Chris Maunder

Re: Character Encoding

Paul Watson11-May-04 4:11

Paul Watson

11-May-04 4:11

Great information, thanks Mike.

regards,
Paul Watson
Bluegrass
South Africa

Chris Maunder wrote:
"I'd rather cover myself in honey and lie on an ant's nest than commit myself to it publicly."

Jon Sagara replied:
"I think we've all been in that situation before."

Crikey! ain't life grand?

Re: Character Encoding

Heath Stewart11-May-04 4:25

11-May-04 4:25

Great information!

I just wanted to add that BOMs (byte order marks) aren't always present in a text file as well. There no requirement for BOMs.

While there's no reliable way to detect encoding - like you said - web browers and other applications (like Word) do try to detect the encoding. If you - the original poster - needs to do something like that, a simple (but probably not the most efficient way) is to take a random sampling of strings within the text file and use StringInfo.GetTextElementEnumerator and enumerate the text elements. With either all of those or a random sample, call TextElementEnumerator.GetTextElement (returns a String) and check the Length. If it's greater than one, you at least know you're dealing with a multi-byte character set (MBCS), like UTF-8. If all of them were 2 bytes, then it's likely it's a double-byte character set (DBCS), like UTF-16 (there's also 4-byte characters, known as UTF-32!). If they're all 1 byte, then you've probably got an ASCII (or other single-byte encoding) file. From there you can make certain assumptions. You see browsers doing this when they start displaying question marks for chracters in odd places (this also happens when the specified encoding is wrong).

Microsoft MVP, Visual C#
My Articles

beginInvoke()

sreejith ss nair11-May-04 1:32

sreejith ss nair

11-May-04 1:32

hai there,
i have a collection of label control in my panel control.On button click or based on some events i need to dispose the label controls in Panel (asynchronous ) and need to add few new label controls.
how it is possible /

hai, enjoy coding
Sreejith SS Nair

Re: beginInvoke()

Heath Stewart11-May-04 4:14

Re: HTTP request,response

11-May-04 4:14

You don't need to dispose the labels - just remove them from the Controls collection property of your panel. If they are no longer being referenced (i.e., there's nothing else referencing them like any class fields or something), the GC will pick them up and dispose of them in a (usually) timely fashion.

If you aren't using fields to refer to these labels, then you'll have to iterate (not enumerate) through the Controls collection and find the label controls like so:

for (int i=0; i< Controls.Count; i++)
{
  Control c = Controls[i];
  if (c is Label) Controls.Remove(c);
  i--; // Don't increment the index since you've now shifted the collection
}

If you are using fields, you can do something like this:

Controls.Remove(myLabel1);
myLabel1 = new Label(); // Initialize...
Controls.Remove(nyLabel2);
myLabel2 = new Label(); // Initialize...

Notice that with this way, new label controls are created that the field refer to, meaning that they are no longer refering to the old labels. So long as nothing else is (and they were removed from the Controls collection, so it's not referencing them anymore either), they will be GC'd.

For non-component classes, calling Dispose is important, like instances of the Graphics class you create manually, or an Image or Bitmap. Controls are good about disposing themselves when necessary, but you could still call Dispose if you wanted to, like so:

Controls.Remove(myLabel1);
myLabel1.Dispose();
myLabel1 = new Label(); // Initialize...
Controls.Remove(myLabel2);
myLabel2.Dispose();
myLabel2 = new Label(); // Initialize...

All this does is free up a little more memory immediately, which would've been freed when necessary by the GC or when the GC got around to it.

Microsoft MVP, Visual C#
My Articles

Re: beginInvoke()

Paul Watson11-May-04 4:14

Paul Watson

11-May-04 4:14

Do you really need to dispose them, what about just removing? You should be able to use panel.Controls.Remove(controltoremove) for that.

regards,
Paul Watson
Bluegrass
South Africa

Chris Maunder wrote:
"I'd rather cover myself in honey and lie on an ant's nest than commit myself to it publicly."

Jon Sagara replied:
"I think we've all been in that situation before."

Crikey! ain't life grand?

HTTP request,response

dcronje11-May-04 1:05

dcronje

11-May-04 1:05

I've got a client that wants to send me files automatically to a aspx webpage. Ive got to accept the stream, and save it to the local disk.

Doing manually its easy, but this is automatic, I dont know when the stream will come in or from where. Can anyone help or get me into the right direction with this one?

Is HttpRequest and HttpResponse the right way?
Or TcpListener on the HTTP port?

Will realy appreciate any help, thanks
Dries

Heath Stewart11-May-04 4:05

Re: help on wmi and registry

11-May-04 4:05

ASP.NET already supports this and wraps it in a nice, neat API for you. All you need to do is include file input controls on your .aspx page (<input type="file" ...>) inside a <form>. The client portion is done.

For the server portion, see the Files Property[^] of the HttpRequest class, which you can get from Page.Request. This returns you an HttpFileCollection, which contains a collection of HttpPostFile instance. This takes the difficulties out of MIME-multipart that is used to upload files with an HTTP POST.

The documentation for the HttpRequest.Files property does contain an example.

And next time post this in the ASP.NET forum.

Microsoft MVP, Visual C#
My Articles

help on wmi and registry

chettu11-May-04 0:54

chettu

11-May-04 0:54

hi i have problem, i have the list of softwares names form the registy. and i am capturing the event opening of any software using wmi, which gives me the .exe name and i use FileVersionInfo to get the software name of that .exe. but this software name doesnt match with the software name from the registry. i have to some how map this software name with the software name from the registry. or i have to find a way to get the software names using an Win32_API, is there an API which will give me all the softwares installed. plz help me in this regard

thanks and regards

deepak

Heath Stewart11-May-04 3:59