Hi gdx, I just confirmed my ideas with this test.
It remains to check the color table and the sprites indexes.
Hi gdx, the index of left routine is not correct and I confirm it with this test:
BB Dragon Index Test
I have create two ROM files:
A) Green Sprite Index for 4000h-41FFh (left) / 4400h-45FFh (right) --> ID 0-15
B) White Sprite Index for 4200h-43FFh (left) / 4600h-47FFh (right) --> ID 16-31
The ID numbers are shown during the game to replace the dragons.
The static position in dragon-left will be ID 30 and 31 but in White Sprite Index these ID are 16-17.
You can also use this method for other sprites. ;-)
PS.
I wish you a perfect patch for this timeless game!
I reconstructed the colored sprites with your DRGTEST.PNG.
The cause of the bug is clear now. The dragon sprite bug is caused because this is the next sprite that is displayed instead.
The XOR function of the sprites is not used. When I modified the pattern table of sprites I thinked it affects another but it was because I had modified really the next one.
We can see also that the three sprites of the dragon when him disappears are shifted and the bigger is not displayed correctly. That is another bug.