I noticed today that a piece of code I wrote for Jaguar/Panther no
longer works right in Tiger. I didn't change anything, and the
exact same code still works flawlessly on Panther.
to create the NSBitmapImageRep, and then I fetch the bitmap base
address and fill it with data. On Tiger, the resulting image looks
like garbage when displayed or serialized to TIFF. It looks like
there's something wrong with the data length calculations,
depending on the target width the rows are skewed. It seems to work
correctly when the image width in pixels is a multiple of 24 which
is the number of bits per sample times the number of channels.
The test case below produces correct results on Panther, and the
exact same binary run on Tiger will show the problem. For
comparison, the ".raw" file can be opened in Photoshop.
Is this a bug in Tiger or have I been doing something wrong all
along without getting into trouble?
I see there's a new, similar method in Tiger but I'd rather not use
that one because my code needs to run on Jaguar and Panther too.
This is a change to the semantics of -initWithBitmapDataPlanes:...
It is now possible for the bitmap to not have packed bytesPerRow in
order to speed up drawing (something that was always possible, but
just didn't happen). Your code assumes that the bitmap data is packed
and so as you fill up the bitmap, you start getting skewed rows.
Either pass in an explicit bytesPerRow value instead of zero to -
initWithBitmapDataPlanes:... to explicitly indicate packed values or
change your loop to be something like:
unsigned char* row = destbuffer;
int i, j;
for (i = 0; i < height; i++, row += [imageRep bytesPerRow]) {
unsigned char* p = row;
for (j = 0; j < width; j++) {
*p++ = c;
*p++ = c;
*p++ = c;
c++;
}
}
Note that in Panther compiled apps, we should still alloc packed
bytesPerRow for compatibility. I assume you recompiled on Tiger.
Andrew Platzer
Application Frameworks
Apple Computer, Inc.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Cocoa-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/cocoa-dev/email@hidden
