C / C++ / MFC

27-Jun-02 0:29

Hi all, I am in a bit of a Dilemma. My app currently takes the format:

Declare Globals.
function InitApp().
function UpdateFrame().
function CallBack WindowProc().
function WinMain().

I would like to incorporate a CD3DApp class which WinMain() would create, initialise and set the Run() function. The CD3DApp class would then handle its own messages in a similar style to the code:

while(GetMessage(&msg, NULL, 0, 0))
{
TranslateMessage(&msg);
ProcessMessage(&msg);
}

which obviously uses the WindowProc CallBack function (so I could specify which buttons do what etc.). I DO NOT want to use MFC. I would like to make my InitApp(), UpdateFrame() and Run() a part of the CD3DApp class and also the global variables as member variables (I think the code is much cleaner then). However, if I put the messaging code listed above in the Run() the CallBack WindowProc cannot access the member variables of the CD3DApp class because the object hasn't been passed through to it (whereas before they were global which allowed for access). There is a timer in the WindowProc which is used to create an animation, which is why I would like access to the CD3DApp class.

What I would like to know is:

1.) Is there any way of creating a class that has its own WindowProc function, so that it can access the member variables?

2.) If not, is there any better way to go about solving this problem of getting the code cleaner (losing the globals)?

"When I left you I was but the learner, now I am the master" - Darth Vader

Michael P Butler27-Jun-02 0:38

Michael P Butler

27-Jun-02 0:38

Have a look at this, it should help you build a mini-framework.

http://www.dilascia.com/wpp.htm

Michael Smile | :)

Logic, my dear Zoe, merely enables one to be wrong with authority. - The Doctor

Alan Chambers27-Jun-02 0:52

27-Jun-02 0:52

Thanks for the link Michael, I've downloaded the resource and am going to try and rip it apart to get what I need. Essentially, I'd like to do my own framework and just need a little guidance, so hopefully this will be it.

Many Thanks,
Alan.

"When I left you I was but the learner, now I am the master" - Darth Vader

Michael P Butler27-Jun-02 1:07

Michael P Butler

27-Jun-02 1:07

I have a copy of the book and I used it to build my own little DirectX framework. The book is well worth reading if you can find a copy.

Michael Smile | :)

Logic, my dear Zoe, merely enables one to be wrong with authority. - The Doctor

Alan Chambers27-Jun-02 1:36

How strange are these problems?

27-Jun-02 1:36

Cheers Michael, will keep my eyes peeled for it when I'm in Waterstones or Blackwells next Big Grin | :-D

"When I left you I was but the learner, now I am the master" - Darth Vader

Alan Chambers27-Jun-02 0:28

27-Jun-02 0:28

OK, I have the following code which starts my app rolling:

int WINAPI WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, LPSTR szCommandLine, int nCmdShow)
{
if (FAILED(InitApp(hInstance, nCmdShow)))
return 0;

//*

MSG msg;
while(GetMessage(&msg, NULL, 0, 0))
{
TranslateMessage(&msg);
ProcessMessage(&msg);
}

return 0;
}

I have set up my own WindowProc (Fullscreen Exclusive) and initialised DirectDraw, a Primary Surface and Direct3D (global variables initialised in InitApp). OK, if I then load a bitmap into a plain offscreen surface (same res as screen), and bltfast it to the Primary Surface back buffer at the //* point and call the primary buffer flip() routine, the bitmap appears, magic. However, if I then bltfast the same picture to the backbuffer (that was the primary surface before the flip) but at a ddiferent position and then call the primary buffer flip() routine again, it is blank (sometimes I get a flicker of the picture before it goes blank). So something must be clearing the data of the primary buffer, I assume? Now for the strange things:

1.) If I set up a timer and handle it in my window proc, with the exact same code, it works, I get a small animation where the picture moves from point a to point b.

2.) Every time I start the app the initial Primary Surface gets cleared and is blank (Is this perhaps to do with my WindowClass which sets the screen to blackness - even though the primary surface hasn't been created at this point?), but the back buffer retains any previous image that was drawn into it. Yes thats right, if I eliminate all drawing code and merely flip() the Primary Surface, up comes the image I displayed in a previous debug initiation, but if I flip again its blank etc.

3.) Now for the one that will have you rocking. Assuming the bltfast code etc. is by the //* position of my code above. Insert the line MessageBox(GetActiveWindow(), "YAY", "YAY", MB_OK); at that point, but above the bltfast code, IT WORKS! You get a message box and then once you click ok, the primary surface flips a couple of times and leaves you with a beautiful bitmap (no blank surface).

If anyone can answer any of these strange problems, I would well appreciate it. I have had to buy head and shoulders shampoo cos I've scratched my head so much.

"When I left you I was but the learner, now I am the master" - Darth Vader

ScrollView Problem

LittleYellowBird27-Jun-02 0:09

LittleYellowBird

27-Jun-02 0:09

Hi,

I've got an MDI app with a ScrollView which acts strangely when I load a file with more than 2060 lines.

The problem manifests its self like this .... if I click above or below the 'Thumb' on the Vertical Scroll Bar it all works OK, but if I grab the 'Thumb' and drag it downwards it jumps back to the top of the file at around line 2060.

My code to set the ScrollView Size is like this:-

CSize sizeTotal;

// get the number of lines
CSQMDoc *thisDoc = GetDocument();
int nLines = thisDoc->GetCount();

// get the height for 1 line of text
TEXTMETRIC tm;
CDC *pDC = GetDC();
pDC->GetTextMetrics(&tm);
int nLineHeight = tm.tmHeight + tm.tmExternalLeading;

// set the size of the 'view' area
sizeTotal.cx = 400;
sizeTotal.cy = nLines * nLineHeight;
SetScrollSizes(MM_TEXT, sizeTotal);
ResizeParentToFit( );

Anyone got any ideas?

Cheers, Ali.

Ali

Rage27-Jun-02 1:58

Rage

27-Jun-02 1:58

I know this kind of thing is hard to debug. Did you try to save the values of sizeTotal.cy in a text file in order to see if it is your calculation that is wrong or if it is the SetScrollSizes() that fails ?

~RaGE();

Roger Allen27-Jun-02 2:11

Roger Allen

27-Jun-02 2:11

The scroll bars in a scroll view have a scroll range of 32767 counts.

My bet is that the line

sizeTotal.cy = nLines * nLineHeight;

With nLines being 2060 majes sizeTotal.cy > 32767 causing all the vertical scroll ranges to be come messed up.

I have seen an artile that covers how to get around this. I will try and find the reference.

Roger Allen
Sonork 100.10016

If I had a quote, it would be a very good one.

LittleYellowBird27-Jun-02 4:19

LittleYellowBird

27-Jun-02 4:19

Roger Allen wrote:
With nLines being 2060 majes sizeTotal.cy > 32767 causing all the vertical scroll ranges to be come messed up.

Yes, I've had a look in the debugger and it is going past 32767! So it looks very much like that is the problem.

Thanks for the article reference - I'll go and have a look. Smile | :)

Ali