22 more fails in gdi32:bitmap after 5d5e5d1 and PR#8348 commit

After the commit of 5d5e5d1 which fixed CORE-15002 there are 22 more failures in the gdi32:bitmap test.
This was a result of the commit of PR#8348.
I have created a minimized gdi32:bitmap test that only includes the changed results here gdi32_winetest_bitmap-07.zip
For these tests, I left all of the surrounding tests in place and added "expected" values as well.
Here is the compiled version of this test gdi32_winetest.exe

Here are the results of this test before the commit causing the extra failures:

And here are the results of this test after the commit causing the extra failures:

Here it is easy to see the difference of after (38) minus before (16) for a total of 22 more fails.
These involve two of the basic bitmap tests as follows:
1) test_SetDIBitsToDevice and 2) test_SetDIBitsToDevice_RLE8.
In each case these involve bitmaps that are Top Down and these have negative heights.
The more common bitmaps use positive heights and are Bottom Up.

I have created a Visual Basic 6 program compiled as an EXE here: SetDIBitsToDevice0.4.4.exe .
This allows a visual way to test the function of SetDIBitsToDevice in Windows to better understand its functionality.