|
Also be aware that if it is not Unicode, the name will become CreateServiceA.
MVP 2010 - are they mad?
|
|
|
|
|
In the MSDN article on "C++ Stack Semantics for Reference Types" it says that:
"A compiler-generated assignment operator will follow the usual standard C++ rules with the following additions:
*
Any non-static data members whose type is a handle to a reference type will be shallow copied (treated like a non-static data member whose type is a pointer).
*
Any non-static data member whose type is a value type will be shallow copied.
*
Any non-static data member whose type is an instance of a reference type will invoke a call to the reference type’s copy-constructor."
Regarding the last item:
1) First it implies that a CLR reference type can be a non-static data member of a native C++ class since, unless I am mistaken, a compiler-generated assignment operator never occurs for CLR classes. Yet elsewhere I also read that a native C++ class having a member that is a ref class type is ill-formed.
2) Secondly, given that 1) is allowable I would have intuited that the assignment operator for the instance of the reference type would calll the user-defined assignment operator for the reference type and not the user-defined copy constructor for the reference type.
It seems I am misreading badly something in this topic, even though I understand what C++ stack semantics for reference types are about, or else the explanation above is badly mangled. I hope somebody can straighten me out on the wording above.
Edward Diener
|
|
|
|
|
|
| The topic just finishes saying that the compiler will not generate a default copy constructor or assignment operator for a reference type and then it starts in on what I quoted in my OP. That entire section seems out of place in the topic and I have never been able to understand what it means, whether it is referring to a an assignment operator or, as you suppose, a copy constructor. If it is referring to a native C++ class, as I suppose, I was under the strong impression that a native C++ class can not have a data member which is an instance of a CLR class so I can not understand what it means. Edward Diener
|
|
|
|
|
I have some problem in this simple passage:
String ^string_num="9,99";
double d_num1=Double::Parse(string_num);
double d_num2=System::Convert::ToDouble(string_num);
i expect d_num1 and d_num2 to be 9,99....but from the watcher i clearly see they both are 9.9900000000000002
any help? 
|
|
|
|
|
Floating point (and by extension double) numbers in computers can only ever be approximations of their decimal value, as the number is stored in binary scientific notation, i.e. exponent and mantissa, rather than decimal numeric. If you want accuracy then you need to stick to integers.
See here[^] for a full description.
MVP 2010 - are they mad?
|
|
|
|
|
ok this make sense , what i'll do is truncate after 2 decimal unit when visualizing the data to video
thank to both of you
|
|
|
|
|
barbetto80 wrote: what i'll do is truncate after 2 decimal unit when visualizing the data
Before taking this decision, make sure this will give you the results you need. For example using float/double for financial data is not a good idea as the roundings may cause loss of accuracy. This may not seem important with an example such as you showed above but it gets worse if you use the data in calculations and you can end up with significant errors in your results.
MVP 2010 - are they mad?
|
|
|
|
|
i know...but that is only a simple representation of a zoom factor application on an image....so i think i can show a truncated value because the user don't need the full value.
But thanks for the advice
|
|
|
|
|
... or use the Decimal type.
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
I only read code that is properly formatted, adding PRE tags is the easiest way to obtain that. [The QA section does it automatically now, I hope we soon get it on regular forums as well]
|
|
|
|
|
try as following:
String ^str="9.99";
System::Single d=Single::Parse(str);
|
|
|
|
|
This won't help, it will still have the accuracy problem.
MVP 2010 - are they mad?
|
|
|
|
|
The following code is from a Visual C++ app to copy a file from one directory to another. It has been simplified, so I can get it working, from the objective where button and textbox controls will populate the String^ variables. I think its pretty clear what the code is trying to do, however I can't get it to work. I would appreciate any help in fixing it, any suggestions of a better approach if this one seems wrong or even some code/advice on how to trap and read any errorcodes that the CopyFile function is trowing instead of the simple messagebox I'm using.
Thanks in advance,
<br />
String^ source = "C:\Temp\Firefox.pcv";<br />
String^ dest = "E:\Test\Firefox.pcv";<br />
char* file1 = (char*)(Marshal::StringToHGlobalAnsi(source + "\0")).ToPointer();<br />
char* file2 = (char*)(Marshal::StringToHGlobalAnsi(dest + "\0")).ToPointer();<br />
<br />
if(CopyFile((LPCTSTR)file1, (LPCTSTR)file2, TRUE) == 0){<br />
MessageBox::Show("ERROR");<br />
}else{<br />
MessageBox::Show("OKAY");<br />
}<br />
<br />
Marshal::FreeHGlobal(IntPtr((void*)file1));<br />
Marshal::FreeHGlobal(IntPtr((void*)file2));<br />
|
|
|
|
|
Because you are using managed code why don't you just use:
File::Copy( path, path2 );
Here is the MSDN article.
Greetings
Covean
|
|
|
|
|
"I can't get it to work" isn't real helpful to us.
Did you check the error code?
Also, for what it's worth, your casts of char* to LPCTSTR in
the CopyFile() call are wrong.
I always recommend, if you're going to use a cast,
ask yourself why you need a cast. Because it won't
compile without it? If so, why?
Then if you do really need the cast, use the right types.
Mark Salsbery
Microsoft MVP - Visual C++
![Java | [Coffee]](https://codeproject.global.ssl.fastly.net/script/Forums/Images/coffee.gif)
|
|
|
|
|
code as following,
String^ source = "C:\\Temp\\Firefox.pcv";
String^ dest = "E:\\Test\\Firefox.pcv";
System::IO::File::Copy(source,dest);
using the clr base as much as better,it's simple and clear.
the system::io::file is a static class .
|
|
|
|
|
I have create application using vc++ 2005 clr form application n trying to deploy my application using package and deploy , n when I install my pc on another computer it show msgbox
"this application has failed to start because the configuration is incorect . reintaling the application may fix the problem"
I already reinstal my application n still this msgbox shown up , n I try many thins like instaling .ne framwork 2.0 and instaling all necesary vc++ 2005 distrubutable file but still o the same result
any one can help me ?? please
|
|
|
|
|
Make sure you're using the right build configuration for the target platform. I've had some apps where the default Win32 coniguration gave me this error message, but compiling with the x86 config worked fine.
Dybs
The shout of progress is not "Eureka!" it's "Strange... that's not what i expected". - peterchen
|
|
|
|
|
Hi
use "depends.exe" to see which .dll could not be found. Maybe you also installed one of the last security updates from Microsoft. This gives the problem that the runtime dll's (MFC, CLR) changes. you can copy depends.exe to the client you have installed your app and load your application. You should see whats going on. A really helpful tool...
|
|
|
|
|
Hi
I have a simple Win32 application with one window. It implements the IAccessible interface to expose some functionality to screen readers.
If I compile it for 32-bit, everything works fine. AccExplorer32, UISpy and Narrator all recognize my interface and the data it provides.
If I compile for 64-bit, AccExplorer32 works fine, but UISpy and Narrator don't. The difference is that AccExplorer32 is native code and UISpy/Narrator are managed. BTW, I believe both UISpy and Narrator are 64-bit applications.
They successfully request my IAccessible object, but after that only QueryInterface, AddRef and Release are getting called. None of the IAccessible or IDispatch methods are being accessed.
So my current theory is that there is some trick to getting 64-bit COM interfaces accessed by managed code. Am I on the right track? What can I do to fix this problem or at least understand what is happening?
Thanks
Ivo
|
|
|
|
|
I'm trying to take existing Bitmaps images in memory and save (or copy)the bits data to an HGLOBAL. Is this possible using .NET classes? The Bitmap images in memory use the System::Drawing::Bitmap classes not the GDI+ version. I can use some of the GDI+ methods if needed but wanted to state the Images originate from the non GDI+ kind.
I've tried the LockBits function but the data that is extracted does not appear proper when viewed in the debugger. I think it's missing the BITMAPINFOHEADER data.
Has anyone done anything like this? It's sort of a reverse of creating an image starting from HGLOBAL and going through the steps to the resulting image.
modified on Tuesday, January 19, 2010 4:28 PM
|
|
|
|
|
Actually the lockbits and it's methods give you the opportunity to get the raw information of the bitmap stored sequentially from position Scan0, each line length have a certain Stride dimension and the information is stored according to PixelFormat.
http://www.bobpowell.net/lockingbits.htm[^]
Usually this method is used for fast access to the bitmap so you can easyle manipulate the data within it or copy the data to other portion of memory (marshall copy).
If you need also the header you can easily create it by yourself...see http://en.wikipedia.org/wiki/BMP_file_format[^]
modified on Wednesday, January 20, 2010 3:07 AM
|
|
|
|
|
Thanks for replying.
The problem with this is that I don't need any access to the bit's except to make a reference to it for calling a Save method. There seems to me to be more utility in the GDI+ functions than there are in the .NET classes. Since I am starting out with a .NET image, I would like to stay with the appropriate classes. I know the Global functions are deprecated, but the means of which the images are acquired, are with legacy code in ActiveX controls so Global functions are appropriate. If need be, I'll use GDI+, but not sure how to get there.
|
|
|
|
|
maybe i don't get the problem but this is waht i understand from your explanation
you want to have a pointer to a memory segment which contains bmp header + bmp data row right?
if yes you can do so
1) create a pointer of dimension 54+RowSize*Height*3 bytes (a classic 24 bit rgb)
example:
array<byte> ^memoryData=gcnew array<Byte>(54 + RowSize*Height*3);
pin_ptr<unsigned char>tmpP1 = &memoryData[0];
unsigned char *np1 = tmpP1;
2) create you personal 54 byte header (http://en.wikipedia.org/wiki/BMP_file_format[^]) or find googling someone that have already done it
3) copy the header in the first part of the memory...use marshall copy and copy to the location memoryData[0]
4) lockbit the image and copy the raw information using marshall copy and copy to the location memoryData[54]
the trick is done if i have understand your real problem
|
|
|
|
|
I guess I failed to explain better.
I have one HGLOBAL. This is used to create an image of type System::Drawing::Bitmap.
HGLOBAL hdl = GloabalAlloc(BitmapImageSize);
PBITMMAPINFOHEADER bmih = (PBITMAPINFOHEADER)GlobalLock(hdl);
LPDWORD ImgBits = (LPDWORD)(bmih + sizeof(BITMAPINFOHEADER));
....
....
....
StretchDIBits()......
That gives me an image from an HGLOBAL
If I had an image already in memory, from some other means (loaded from file maybe), I want to then point that image to the HGLOBAL.
Hope that was clearer
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
