C / C++ / MFC

26-May-03 1:48

Hello everybody,

I hope you can help me with my problem.
At this moment I am developping a project in DirectSound. I need to play a wave file. Because of future enhancements the wave file is played with DirectSound.

I use 3D buffers and the extended waveformat. All initialisation goes well, but when I try to lock the buffer the error DSERR_PRIOLEVELNEEDED is returned. I checked the SetCooperativeLevel function and that returns DS_OK.

The code I use for those two functions look like this:

<br />
MainhWnd = GetSafeHwnd();<br />
hr = m_pDS->SetCooperativeLevel(MainhWnd, DSSCL_NORMAL);<br />
<br />
....<br />
<br />
VOID* pbData = NULL;<br />
VOID* pbData2 = NULL;<br />
DWORD dwLength;<br />
DWORD dwLength2;<br />
int lPos=0;<br />
DWORD m_dwBufferBytes;<br />
<br />
hr= m_pDSBuffer->Lock(lPos, m_dwBufferBytes, &pbData, &dwLength, &pbData2, &dwLength2, 0);<br />

Does anyone know what the problem is, or how I can solve this.

Thanks in advance,

Miepmuts

Rage26-May-03 2:11

26-May-03 2:11

Already tried with a higher level within SetCooperativeLevel ?

~RaGE();

Miepmuts26-May-03 2:15

26-May-03 2:15

Yes I tried to change the Cooperative Level to DSSCL_PRIORITY in combination with the SetFormat function. But that gives the same result, also another window handle doesn't change a thing.

Rage26-May-03 2:16

26-May-03 2:16

I suppose DSBLOCK_ENTIREBUFFER is equal to 0, since you are giving 0 as last parameter of your function. If not, then this will attempt to lock a buffer which can have any size, since m_dwBufferBytes is not initialised, causing the lock to fail randomly.

~RaGE();

Miepmuts26-May-03 2:54

26-May-03 2:54

It's not a random error but every time the same one: DSERR_PRIOLEVELNEEDED

Even when I fill in those two variables the function returns the same error.
It is a very strange error. I asked someone at my work, and even he didn't see a solution.

Rage26-May-03 3:44

26-May-03 3:44

It must be something you are doing inbetween the code lines you have posted.

~RaGE();

Miepmuts26-May-03 4:15

Communication problem between dialog window and frame window

26-May-03 4:15

Allright then. Here it goes:

<br />
hr = DirectSoundCreate(NULL, &m_pDS, 0);<br />
if(hr == DS_OK)<br />
{<br />
  HWND MainhWnd;<br />
  MainhWnd = GetSafeHwnd();<br />
  hr = m_pDS->SetCooperativeLevel(MainhWnd, DSSCL_NORMAL);<br />
  if(hr == DS_OK)<br />
  {<br />
    //Create Primary audio 3D Buffer<br />
    DSBUFFERDESC bufferDesc;<br />
    memset(&bufferDesc, 0, sizeof(DSBUFFERDESC));<br />
    bufferDesc.dwSize = sizeof(DSBUFFERDESC);<br />
    bufferDesc.dwFlags = DSBCAPS_PRIMARYBUFFER | DSBCAPS_GETCURRENTPOSITION2 ;<br />
    bufferDesc.dwBufferBytes = 0;<br />
    bufferDesc.lpwfxFormat = NULL;<br />
    hr = m_pDS->CreateSoundBuffer(&bufferDesc, &m_pPrimaryBuffer, NULL);<br />
    if(hr == DS_OK)<br />
    {<br />
	hr = m_pPrimaryBuffer->QueryInterface( IID_IDirectSound3DBuffer, (void**)    &m_pDS3DBuffer );<br />
	//Set primary internal extended format<br />
	WAVEFORMATEXTENSIBLE formatex;<br />
	memset(&formatex, 0, sizeof(WAVEFORMATEXTENSIBLE));<br />
	formatex.Format.wFormatTag = WAVE_FORMAT_EXTENSIBLE;<br />
	formatex.Format.nChannels = 4;<br />
	formatex.Format.nSamplesPerSec = 44100;<br />
	formatex.Format.nAvgBytesPerSec = 352800;<br />
	formatex.Format.nBlockAlign = 8;<br />
	formatex.Format.wBitsPerSample = 16;<br />
	formatex.Format.cbSize = 22;<br />
	formatex.Samples.wValidBitsPerSample = 16;<br />
	formatex.dwChannelMask = SPEAKER_FRONT_LEFT | SPEAKER_FRONT_RIGHT<br />
				 | SPEAKER_BACK_LEFT | SPEAKER_BACK_RIGHT;<br />
	formatex.SubFormat = KSDATAFORMAT_SUBTYPE_PCM;<br />
	//Now create a normal sound buffer<br />
	DSBUFFERDESC bufferSec;<br />
	memset(&bufferSec, 0, sizeof(bufferSec));<br />
	bufferSec.dwSize = sizeof(bufferSec);<br />
	bufferSec.dwFlags = DSBCAPS_GETCURRENTPOSITION2;<br />
	bufferSec.lpwfxFormat = &waveformat;<br />
	hr = m_pDS->CreateSoundBuffer(&bufferDesc, &m_pDSBuffer, NULL);<br />
    }<br />
}<br />
<br />
HRESULT hr;<br />
BYTE* pbWavData;<br />
UINT cbWavSize;<br />
VOID* pbData = NULL;<br />
VOID* pbData2 = NULL;<br />
DWORD dwLength;<br />
DWORD dwLength2;<br />
<br />
int nWaveFileSize = ckIn.cksize;<br />
pbWavData = new BYTE[nWaveFileSize];<br />
if(pbWavData == NULL)<br />
{<br />
    return 0;<br />
}<br />
MMIOINFO mmioinfoIn;<br />
<br />
cbWavSize = 0;<br />
if(mmioGetInfo(hmmioIn, &mmioinfoIn, 0) != 0)<br />
{<br />
    return 0;<br />
}<br />
UINT cbDataIn = nWaveFileSize;<br />
if(cbDataIn > ckIn.cksize)<br />
{<br />
    cbDataIn = ckIn.cksize;<br />
}<br />
ckIn.cksize -= cbDataIn;<br />
	<br />
for(DWORD ct = 0; ct < cbDataIn; ct++)<br />
{<br />
    if(mmioinfoIn.pchNext == mmioinfoIn.pchEndRead)<br />
    {<br />
	if(mmioAdvance(hmmioIn, &mmioinfoIn, MMIO_READ) != 0)<br />
	{<br />
	    return 0;<br />
	}<br />
	if(mmioinfoIn.pchNext == mmioinfoIn.pchEndRead)<br />
	{<br />
	    return 0;<br />
	}<br />
    }<br />
<br />
    *((BYTE*)pbWavData + ct) = *((BYTE*)mmioinfoIn.pchNext);<br />
    mmioinfoIn.pchNext++;<br />
}<br />
if(mmioSetInfo(hmmioIn, &mmioinfoIn, 0) != 0)<br />
{<br />
    return 0;<br />
}<br />
cbWavSize = cbDataIn;<br />
m_dwBufferBytes = 8 * 1024;<br />
<br />
StartDataRead(&m_ckIn);<br />
int lPos=0;<br />
hr= m_pDSBuffer->Lock(lPos, m_dwBufferBytes, &pbData, &dwLength, &pbData2,&dwLength2,    DSBLOCK_ENTIREBUFFER);<br />
if( hr != DS_OK)  //hr returns a strange number which equals to DSERR_PRIOLEVELNEEDED<br />
{<br />
   return 0;<br />
}<br />
<br />
memcpy(pbData, pbWavData, m_dwBufferBytes);<br />
m_pDSBuffer->Unlock(pbData, m_dwBufferBytes, NULL, 0);<br />
pbData = NULL;<br />
return 1;<br />
<br />

I hope you can do something with this. There is a part of code in between of initialising the mmio and the chunks.

GCC with MCVC IDE

Andrew Walker26-May-03 1:24

Andrew Walker

26-May-03 1:24

Is it possible to run GCC within Microsoft's IDE (6 or .NET, it doesn't matter which), I just want to test some template code.

I know that it is possible because comeau and the intel compiler do it, but has anyone seen it for GCC.

Cool | :cool:

thanks

Re: GCC with MCVC IDE

tareqsiraj26-May-03 9:42

tareqsiraj

26-May-03 9:42

Try the following links. Smile | :)

http://sources.redhat.com/ml/crossgcc/2001-05/msg00046.html
http://www.cygwin.com/ml/crossgcc/2001-09/msg00069.html

-Tareq

VC 5 <-> VC6

JensB26-May-03 0:04

JensB

26-May-03 0:04

Hello

Does anyone know a url where i can find an extensively comparison between Visual C++ 5.0 and the 6.0 version? Like for example: 6.0 version had this as advantage, this as disadvantage etc

Greetings
Jens

Scope Error

Paul Hooper25-May-03 23:45

Paul Hooper

25-May-03 23:45

I just spent the best part of today finding a nasty little bug and thought I might "share" with you.

We had some code that basically went

{<br />
   CPen apen;<br />
   apen.CreatePen(...);<br />
   dc->SelectObject(&apen);<br />
<br />
   //Do lots of work with dc<br />
<br />
   dc->SelectStockObject(BLACK_PEN);  //Select apen out so it can clean up<br />
}

Last version, the code was changed to something like

{<br />
   if (option)<br />
   {<br />
      CPen apen;<br />
      apen.CreatePen(...);<br />
      dc->SelectObject(&apen);<br />
<br />
      //Do lots of work with dc<br />
   }<br />
   else<br />
   {<br />
      CPen bpen;<br />
      bpen.CreatePen(...);<br />
      dc->SelectObject(&bpen);<br />
<br />
      //Do lots of work with dc<br />
   }<br />
   dc->SelectStockObject(BLACK_PEN);  //Select pens out so they can clean up<br />
}

I am sure you can see the error... apen and bpen are now being destroyed before selected out. Resource leak.

Running this code on Win2000/XP is fine because they clean up the mess. Running on 95,98,ME caused massive resource leaks (obviously!) The error is obvious when simplified as above but it was not the least bit obvious in the production code.

It's always the simple things that bite you!

Paul Hooper

If you spend your whole life looking over your shoulder, they will get you from the front instead.

Re: Scope Error

John R. Shaw26-May-03 9:51

John R. Shaw

26-May-03 9:51

CPen* pOldPen = dc->SelectObject(&apen);
//Do lots of work with dc
dc->SelectObject(pOldPen);

or you could just use SaveDC() and RestoreDC() so you do not have to keep track of what objects are selected.

int nSavedDC = dc->SaveDC();
dc->SelectObject(&apen);
//Do lots of work with dc
dc->RestoreDC(nSavedDC);

Trust in the code Luke. Yea right!

Re: Scope Error

John R. Shaw26-May-03 10:04

John R. Shaw

26-May-03 10:04

Oh! One more thing, do not do this:

if(..)
{
    CPen apen;
    ....
    dc->SelectObject(&apen);
    //Do lots of work with dc
}

Unless you are going to restore the old pen before leaving the 'if' block, because (of course) apen goes out of scope as soon as you leave the block.

CPen apen;
if(..)
{
    ....
    dc->SelectObject(&apen);
    //Do lots of work with dc
}

Also it you must use CPen::CreatePen(), instead of the constructor, to create your pens then you must call CPen::DeleteObject() to destroy it or you will get memory leaks.

Trust in the code Luke. Yea right!

chen25-May-03 23:38

chen

25-May-03 23:38

Hi,

In a dialog based application (main application window), a frame window
can be drived by click a menu item. Now my problem is that i want use some functions defined at the appDlg.cpp, how should i call these functions in myframe.cpp? How the communications are constructed ?

Thanks in advance!

chen

Re: Communication problem between dialog window and frame window

Rage25-May-03 23:52

How Do I Drag&Drop an ActiveX Control

25-May-03 23:52

Three solutions :

1. Use user defined messages with SendMessage or PostMessage.

2. Overwrite the Create (Or DoModal) function of the frame window to add a pointer to a CYourDialog (where CYourDialog is the name of your dialog) as parameter. Put a member variable m_pMainDlg in your Frame window class, and store the pointer in it. Call your base class functions with

m_pMainDlg->Function1();

3.[MFC] Use AfxGetMainWnd() in a type cast :

CYourDialog *m_pMainDialog = (CYourDialog*)AfxgetMainWnd();
m_pMainDlg->Function1();

~RaGE();

ccdai25-May-03 23:11

ccdai

25-May-03 23:11

Hello....

I am having problems when I am trying to drag and drop one of my activeX control from an application to MS-WORD.

The problem is that I am not sure what format it is for an ActiveX Control in clipboard (ie. CF_TEXT is for text) so I dont know what to put as the first parameter in the function when I need to cashe my data into the global memory created ....

"opOleDataSource->CacheGlobalData(??????, hGlobal);"

Does anyone knows how to do this and Please advise me !!

Thank You !! Smile | :)

Parsing in C

Rickard Andersson²⁰25-May-03 22:57

Rickard Andersson²⁰

25-May-03 22:57

I have some problems with my code here. You can see where the error appears in the code:

char szCommand[] = "luc 192.168.244.151 5000 50 mypass";

void HotIceParser(const char* command, PHIPD phipd)
{
	char buffert[20+1] = {0};
	char* p1 = NULL;
	char* p2 = NULL;

#ifdef DEBUG_MODE
	assert(phipd != NULL);
#endif
	
	p1 = &buffert[0];
	p2 = &command[0];

	while((*p2 != ' ') && (*p2 != '\0'))
		*p1++ = *p2++;

	MessageBox(NULL, buffert, "1",MB_OK);

	if(strcmp(buffert,"luc") == 0)
	{
		phipd->type = LUC;
		p1 = &buffert[0];
		p2++;
		while((*p2 != ' ') && (*p2 != '\0'))
			*p1++ = *p2++;

		MessageBox(NULL, buffert, "2",MB_OK); // STACK AROUND buffert IS CORRUPTED!!! WHY??
		strcpy(phipd->luc.pszIP, buffert);
	}
}

Can you help me to solve this?

Rickard Andersson⁸
Here is my card, contact me later!
UIN: 50302279
E-Mail: nikado@pc.nu
Interests: C++, ADO, SQL, Winsock, 0s and 1s

jhwurmbach25-May-03 23:26

jhwurmbach

25-May-03 23:26

Do you really NEED to use C? Your line
char buffert[20+1] = {0};
seems to be C++ to me: I C, you would use zeromem() or memset().

But if using plain, ugly C is not of prime importance, because you use a C++-compiler anyway, you can just as well use std::string to do the parsing:

void HotIceParser(const char* cmd, PHIPD phipd)

std::string command( cmd );
std::string::iterator p1 = command.find_first_not_of(' \0');

..and so on..

My opinions may have changed, but not the fact that I am right.

Rickard Andersson²⁰26-May-03 0:24

Rickard Andersson²⁰

26-May-03 0:24

jhwurmbach wrote:
Do you really NEED to use C

Hm... in fact no. But I want to learn it.

jhwurmbach wrote:
char buffert[20+1] = {0};
seems to be C++ to me:

I C, you would use zeromem() or memset().

Thanks! Smile | :)

I want perfect C code!

Rickard Andersson⁸
Here is my card, contact me later!
UIN: 50302279
E-Mail: nikado@pc.nu
Interests: C++, ADO, SQL, Winsock, 0s and 1s

John R. Shaw26-May-03 10:13

John R. Shaw

26-May-03 10:13

char buffert[20+1] = {0};

This is C. When you initialise a buffer like this the buffer is initialised to all zeros.

char buffert[20+1] = {'\0'};
or
char buffert[20+1] = {(char)0};
is more correct.

char buffer[4] = {'1'} produces a buffer containing the follwing:

'1','\0','\0','\0'

Trust in the code Luke. Yea right!

Paul Hooper25-May-03 23:29

Paul Hooper

25-May-03 23:29

I am presuming that you are sending szCommand to the function. szCommand is longer than 20 characters and buffert will only hold 20. Your while loop will keep filling buffert past the 21 limit of your array and corrupt whatever is around.

You need to check that your command parameter is not bigger than buffert.

Paul Hooper

If you spend your whole life looking over your shoulder, they will get you from the front instead.

Rickard Andersson²⁰26-May-03 0:30

Rickard Andersson²⁰

26-May-03 0:30

Paul Hooper wrote:
I am presuming that you are sending szCommand to the function

Yes.

Paul Hooper wrote:
szCommand is longer than 20 characters and buffert will only hold 20.

Yes, but it's strange when Rage could get it to work... and me too... Unsure | :~

Paul Hooper wrote:
You need to check that your command parameter is not bigger than buffert.

Thanks!

Rickard Andersson⁸
Here is my card, contact me later!
UIN: 50302279
E-Mail: nikado@pc.nu
Interests: C++, ADO, SQL, Winsock, 0s and 1s

Paul Hooper26-May-03 1:51

Paul Hooper

26-May-03 1:51

Rickard Andersson18 wrote:
Yes, but it's strange when Rage could get it to work... and me too...

Not strange at all. Once you overrun a buffer ANYTHING is possible - depending on what has been overwritten. Anywhere from complete system compromise, pain, suffering, earthquakes, volcanoes (oops getting a little carried away) to ... absolutely nothing. These are really nasty faults BECAUSE the results can be variable. Just ask Microsoft!

Paul Hooper

If you spend your whole life looking over your shoulder, they will get you from the front instead.

Rage25-May-03 23:46