Sorry your browser is not supported!

Hiyes ppl,

Working on a routine to perform a partial fade on areas of the screen.

screenPtr = dbGetPixelsPointer();

gives me what looks like the correct mem address but when i try to

pixColour = *screenPtr;

i get access violations. The help manual suggests indirection should work and would be quicker than dbPoint() and dbDot().

I have used dbLockPixels() prior to screen access and dbUnlockPixels afterwards although the code isnt running though to that point, (otherwise dbGetPixelsPtr would return NULL)

I also have two separate versions of this function, (not that it should affect this problem but ...) One where pixColour is defined as DWORD for 32bit screen depth and consequently one where it is defined as a WORD for 16 bit screen depth.

I just wanted to avoid having nested loops (x direction and y direction) with the overhead of what is essentially two unnecessary function calls.

Suggestions please folks ?

Maybe you are talking about something like this:



Of course you have to provide screen positions and RGB components.

You shouldn't get the pixel pointer without locking first - DirectX doesn't guarantee that the address will stay the same unless you lock. That's what may be causing your access violations.

Yup, I had the following...
NB This is to play around with a square at an offset on the screen
NB and to make it grey but semi-trans



however as soon as i debug to:
pointCol = *writePtr;

I get an access violation. As i stated the mem address looks right for the screen, but if i replace:

pointCol = *writePtr;
with
pointCol = dbPoint();

and
*writePtr = pointCol;
with
dbDot( ... );

it all works fine but I am also getting an overhead of two unnecessary function calls within a nested loop 8/

Sorry - last post didnt work right - should have had this code attached. (frustrating - have to wait to have posts checked 8/ )

This was the code i was using:


dbLockPixels();

pixPtr = dbGetPixelsPointer();

offsetPtr = pixPtr + (posY * pixelPitch) + (posX * bytesPerPixel);

for (yLoop = some; yLoop < lots; yLoop++)
{
for (xLoop = some; xLoop < lots; xLoop++)
{
writePtr = offsetPtr + (yLoop * pixelPitch);
writePtr += (xLoop * bytesPerPixel);

// at this point I *should* be able to go
pointCol = *writePtr;

// play around with the colour and then ...
*writePtr = newPointColour;
}
}

dbUnlockPixels();




however as soon as i debug to:
pointCol = *writePtr;

I get an access violation. As i stated the mem address looks right for the screen, but if i replace:

pointCol = *writePtr;
with
pointCol = dbPoint();

and
*writePtr = pointCol;
with
dbDot( ... );

it all works fine but I am also getting an overhead of two unnecessary function calls within a nested loop 8/