ND2 scanline pad fix

[sbesson]

See https://trello.com/c/TaqXNGco/152-nd2-pixel-offset.

From the Nikon ND2 samples in our repository, it seems like having an odd size both alongside the X dimension and the C dimension is necessary and sufficient to have the scanline padded.

This PR simplifies the NativeND2Reader.getScanlinePad() method to use this simpler calculation. To test this PR, check the automated tests keep passing against all ND2 samples. Additionally, test the sample file provided in the Fiji forum thread and check that without this PR, the image display with some horizontal lines while with this PR, it matches the expectation from NIS Elements.

[dgault]

Testing the sample file from the forum thread shows that the scanlines are present without this PR but with the PR applied the file is read and displayed correctly. I additionally tested a number of other nd2 files from our repo and all continued to read and display correctly also.

The subset jobs are green and the new config file also passes testing - https://ci.openmicroscopy.org/job/BIOFORMATS-DEV-merge-repository-subset/444/

I have been trying to locate the addition of this logic originally and although the getScanlinePad function was added back in 2014 the logic itself was being used dating further back still.

Based on the sample files and information that we have this relaxation of the getScanlinePad rules appears to work as expected. Once the merge full job is back to normal this should be good to merge.

[dgault]

Daily builds and tests are all back to normal and previous issues were unrelated to this PR.