Hello Jay Le 16/05/2020 à 06:33, Jayathirth D v a écrit :
Apologies for late reply(mail got lost and I forgot about checking it).
No problem, thanks for the reply. In the meantime I have sent an updated version of the test case to Sergey, with test for arbitrary RenderedImage as a separated method.
I was under the impression that we have another implementation of RenderedImage which we can use, but looks like there are none.
I have found in Java2D source code that SunGraphics2D.getTileIndex(int, int, int) method is using the same formula than JAI. Consequently I think it is possible to demonstrate BufferedImage problem by wrapping it behind another RenderedImage (just delegating all method calls to the wrapped BufferedImage) for forcing SunGraphics2D to execute the general case instead that its optimizations for BufferedImage special case. That test would demonstrate that without the fix, attempt to draw a wrapped BufferedImage.subSubImage(…) with Graphics2D.drawRenderedImage(…) causes an IndexOutOfBoundsException to be thrown. Please let me know if such test is desired.
Martin
