C / C++ / MFC

Debugger problems with run-time information

26-Jul-06 10:56

Zac Howland wrote:
It would have started...

True, in a sense.

Zac Howland wrote:
...but as soon as any access to the control was made, it would have given an assertion (in debug) or just crashed (in release).

Neither of which happened. It just went away quietly.

adamshuv26-Jul-06 5:13

adamshuv

26-Jul-06 5:13

Hi,
I'm trying to debug my code, and each time I run into a dynamically-linked function-call, the debugger cannot step-into the function, and does not stop at breakpoints inside it.
e.g.

class ISender {
public:
virtual void Send() = 0;
};

class XSender : public ISender {
...
virtual void Send();
};

ISender *m_sender = new XSender;

m_sender->Send()

Instead, it suggests that it shows the disassembly, as if it cannot find the matching code. I've tried that, and I still can't go inside these functions.
Static calls to functions of the same class work properly, e.g.
ISender m_sender = new XSender;

In addition, The debugger does not show the value of the object in the monitor, unless I specifically cast it to the run-time type of the object, i.e.:
(XSender *)m_sender

I suspect that thesse problems are related but I'm not sure.

Any ideas Confused | :confused:

Thanks.

while (i == live) {
have(fun);
}

Re: Debugger problems with run-time information

Matt Godbolt26-Jul-06 6:08

Matt Godbolt

26-Jul-06 6:08

Just a quick thing to check; you say the code's dynamically-linked (ie in a DLL); check the output pane of Visual Studio for the line that loads the DLL the call's in. It should say something like:

'StyleManager.exe': Loaded 'C:\Codebase\codedb\PF\Products\StyleManager\SMCmdLine\Debug\StyleManager.dll', Symbols loaded.

If it doesn't say 'Symbols loaded' for your DLL (instead it'll say 'No symbols loaded.'), then the debugger won't know how to debug into the DLL (other than in assembly view). You'll need to ensure that the DLL is built in debug mode and that the relevant PDB is around. Most of the time this happens automatically, but if you have a custom post-build step that copies the DLL somewhere else but not the PDB then you might get this kind of problem.

Matt Godbolt
Engineer, ProFactor Software
StyleManager project

Create a bitmap [modified]

#realJSOP26-Jul-06 5:04

#realJSOP

26-Jul-06 5:04

I have two bitmaps (loaded from resources).

I want to combine thse two bitmaps side-by-side into a 3rd bitmap.

I want to put the 3rd (combined) bitmap into an image list.

I tried using the BitmapDC class from the GDI articles section, but I'm missing something (obviously, otherwise I wouldn't be here).

Here's the code I'm using now:

CBitmap bmp1;
CBitmap bmp2;
bmp1.LoadBitmap(nImgConsult);
bmp2.LoadBitmap(nImgLabs);
CBitmapDC bmpDC(34, 21, GetDC(), RGB(255,255,255));
DrawBitmap(&bmp1, &bmpDC, CPoint(0, 0));
DrawBitmap(&bmp2, &bmpDC, CPoint(17, 0));
CBitmap* pBMP = bmpDC.Close();
if (!m_ResultsImages.Replace(0, pBMP, NULL))
{
    TRACE("ImageList.Replace failed\n");
}
delete pBMP;

Any hints, clues, or alternative code?

-- modified at 11:36 Wednesday 26th July, 2006

Re: Create a bitmap

Zac Howland26-Jul-06 5:40

Two Documents in an Application

26-Jul-06 5:40

Why not just use the GetBitmapBits/SetBitmapBits (and CGdiObject::GetObject) to get the bits from images 1 and 2, create a byte array that is the size of the sum of both those arrays (which should be double the size of either one if the bitmaps are the same size), copy the bytes from image 1 to the first part of the new array, and image 2 to the last part of the new array and then use SetBitmapBits on a newly created CBitmap object?

If you decide to become a software engineer, you are signing up to have a 1/2" piece of silicon tell you exactly how stupid you really are for 8 hours a day, 5 days a week

Zac

Re: Create a bitmap

valikac26-Jul-06 6:47

valikac

26-Jul-06 6:47

another approach is create a bitmap sized to fit both source bitmaps. arrange the bitmaps via bitblt()

Kuphryn

narayanagvs26-Jul-06 4:55

narayanagvs

26-Jul-06 4:55

Hi
Can we have two CDocument derived classes in an MDI application??

Re: Two Documents in an Application

led mike26-Jul-06 5:46

led mike

26-Jul-06 5:46

Yes

Re: Two Documents in an Application

BlitzPackage26-Jul-06 7:36

BlitzPackage

26-Jul-06 7:36

Thus the name MDI = Multiple Document

Strange error i cannot fathom

eeyor66626-Jul-06 4:33

eeyor666

26-Jul-06 4:33

hey can anyone help me sort this? the weird thing is my program ran at first but now if i try and run i get this debug error!

*********************************************************************************************
*": warning C4267: 'argument' : conversion from 'size_t' to 'DWORD', possible loss of data"
*********************************************************************************************

Sniff | :^)

If i'm not dead enough for life, am i alive enough for death?

RChin26-Jul-06 4:46

RChin

26-Jul-06 4:46

Smells like an stl issue.
You will get this type of message if you do something like this:

<br />
std::vector<long> vecLong;<br />
int n = vecLong.size();<br />

A lot of the stl classes returns a size_t numeric data type (instead of say, int, long etc).
Its not an error, but you should be aware of any overflow issues. If you are sure that your DWORD variable has enough range to accomodate the stl's size_t number, then just do a simple type cast.

<br />
std::vector<long> vecLong;<br />
int n = static_cast<int>(vecLong.size());<br />

I Dream of Absolute Zero

David Crow26-Jul-06 4:48

Re: Strange error i cannot fathom [modified]

26-Jul-06 4:48

The compiler is telling you exactly what is wrong. Show the line of code in question.

"Money talks. When my money starts to talk, I get a bill to shut it up." - Frank

"Judge not by the eye but by the heart." - Native American Proverb

toxcct26-Jul-06 5:02

toxcct

26-Jul-06 5:02

this is not an error. it is just a warning.

the compiler warns you about the fact that you copy some value of a particuliar type into a smaller type, hence the "possible loss of data".

what you can do to avoid this is :
- make the two types identical
- explicitely cast while assigning the value
- disable temporarily/totally the warning using the #pragma warning(disable:4267) statment...

Zac Howland26-Jul-06 5:06

26-Jul-06 5:06

toxcct wrote:
- disable temporarily/totally the warning using the #pragma comment(disable:4267) statment...

I believe you meant #pragma warning(disable:4267).

If you decide to become a software engineer, you are signing up to have a 1/2" piece of silicon tell you exactly how stupid you really are for 8 hours a day, 5 days a week

Zac

toxcct26-Jul-06 5:09

toxcct

26-Jul-06 5:09

Zac Howland wrote:
believe you meant #pragma warning(disable:4267).

yup, typo

TOXCCT >>> GEII power

[VisualCalc 3.0 updated ][Flags Beginner's Guide new! ]

Zac Howland26-Jul-06 5:03

26-Jul-06 5:03

Are you compiling as 64-bit by chance? I believe size_t is a 64-bit integer on 64-bit platforms whereas DWORD (which is just an unsigned long) is always 32-bits. Thus, if you try to implicitly cast a size_t to a DWORD, it will give you that warning. To get around it, either ignore it, use bitshift operations to get the lower 32-bits, or use static_cast to explicitly cast it to a DWORD.

If you decide to become a software engineer, you are signing up to have a 1/2" piece of silicon tell you exactly how stupid you really are for 8 hours a day, 5 days a week

Zac

conversion

RockyJames26-Jul-06 4:32

RockyJames

26-Jul-06 4:32

hi
i have funtion named change(long lngNo,unsigned char* pData);

where some times i send intger data,float data ,BOOL data as second parameter.How do i convert integer ,float ,BOOL to unsigned char*.

thanks in before

james

David Crow26-Jul-06 4:45

26-Jul-06 4:45

RockyJames wrote:
How do i convert integer ,float ,BOOL to unsigned char*.

Use the address-of operator (&). For example:

int x;
change(5L, &x);

"Money talks. When my money starts to talk, I get a bill to shut it up." - Frank

"Judge not by the eye but by the heart." - Native American Proverb

RockyJames26-Jul-06 4:53

RockyJames

26-Jul-06 4:53

error C2664: : cannot convert parameter 2 from 'int *__w64 ' to 'unsigned char *'

David Crow26-Jul-06 5:03

26-Jul-06 5:03

You'll need to cast it, since change() is expecting an unsigned char *.

Why are you trying to make a one-size-fits-all function?

"Money talks. When my money starts to talk, I get a bill to shut it up." - Frank

"Judge not by the eye but by the heart." - Native American Proverb

toxcct26-Jul-06 5:05

toxcct

26-Jul-06 5:05

can't you write several overloads of the function, which every func receive the right type ?

TOXCCT >>> GEII power

[VisualCalc 3.0 updated ][Flags Beginner's Guide new! ]

Zac Howland26-Jul-06 5:13

26-Jul-06 5:13

RockyJames wrote:
i have funtion named change(long lngNo,unsigned char* pData);

where some times i send intger data,float data ,BOOL data as second parameter.How do i convert integer ,float ,BOOL to unsigned char*.

It is very bad practice to write a one-size-fits-all function. If you are trying to do something with the bytes (i.e. send them across a network using sockets), that is one thing, but if you are trying to decrease the number of methods you need to write for a given operation that doesn't involve direct byte transmissions, then this is the wrong approach.

A cleaner way of doing it would be to:

typedef unsigned long ulong;
typedef unsigned char byte;
// hide this method so that it isn't exposed
void change(ulong size, byte* data)
{
	// do something with data
}

// this is the exposed method
template<class T> void change(const T& data)
{
	change(sizeof(T), (byte*)&data);
}

If you decide to become a software engineer, you are signing up to have a 1/2" piece of silicon tell you exactly how stupid you really are for 8 hours a day, 5 days a week

Zac

RockyJames26-Jul-06 16:00

RockyJames

26-Jul-06 16:00

The function is not written by me,it is one of the function in control.that`s the problem..so how can i send integer ,float ,BOOL in BYTE or unsinged char format.....?

Zac Howland26-Jul-06 19:14