#31 bug in LE32 abmp resources

[Alexander R. Pruss]

The endianness is wrong on -LE32 bitmap resource
palette entry counts. In fact, as far as I can tell, no
endianness conversion is done anywhere in the palette
generation code.

[Alexander R. Pruss]

Logged In: YES
user_id=756930

Moreover, there needs to be two bytes of padding after this
count.

[Alexander R. Pruss]

Logged In: YES
user_id=756930

A fix that worked for me was to change the
*tmpPtr++ = 0x01;
*tmpPtr++ = 0x00;
in bitmap.c:BMP_ConvertWindowsBitmap() to:
if ( vfLE32 ) {
*tmpPtr++ = 0x00;
*tmpPtr++ = 0x01;
*tmpPtr++ = 0x00;
*tmpPtr++ = 0x00;
}
else {
*tmpPtr++ = 0x01;
*tmpPtr++ = 0x00;
}

Unfortunately, I guess, this only fixes things for .bmp
inputs. I don't know enough about the pilrc code to change
it for other kinds of inputs.

I've tested this with the Clie Simulator in generating Graffiti
skins, and after this little fix (and the other fix in my other
bug report) 8-bit skins with palettes now work.

[chrilith]

Logged In: YES
user_id=637029

Hi,

Also :
BITMAP ID 3100 "keyCommand.bmp" DENSITY 2

doesn't seem to generate the DENSITY 2 bmp.

But defined as this, it works :
BITMAP ID 3100
BEGIN
BITMAP "keyCommand.bmp" BPP 1 DENSITY 2
END

But the bitmap data is wrong (seems to be shifted)