On Thu, 4 Jan 2024 18:04:59 GMT, Sergey Bylokhov <s...@openjdk.org> wrote:
> > I thought about it… I thought that the filter itself should exercise the > > code path that's modified but it doesn't seem like it does… I mean creating > > images with regular ICC profile and using the wrapper TestColorSpace in > > conversion: new ColorConvertOp(new > > TestColorSpace(ColorSpace.getInstance(CS_LINEAR_RGB), null); but it didn't > > work, the test passes without the fix. > > It is passed since it blindly checks the difference between src and dst, and > since with proper implementation of TestColorSpacewe we always do "some" > conversion -> the images are different -> the test passed. Still, the current test catches the typo in the code `srcColorSpace` → `dstColorSpace` that's now replaced. From this point view, the stated bug is fixed. At the same time, I agree there seem to be more bugs in the area. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/16895#discussion_r1442124080