Multi-dimension coordinate variables are not picking up the deflate level of a collection in History

[bena-nasa]

A user was asking why the cubed-sphere files produced by History were so big compared to a lat-lon output file given a similar number of variables and comparable number of total grid points. The answer is that the cubed-sphere files have both the centers and corners in the file, in double precision.

But one thing we did notice, and the point of the issue, is that if the collection set a non-zero deflate level, while the variables were compressed in the output file, the center and corner coordinates were not being compressed.

So this was clearly an oversight on my part and should be fixed. There seems no reason that if the variables are being compressed, that the coordinates should not either for consistency. I've already tested this so just making an issue for the PR.