Formatting inconsistency -- yapf tests

[Wackerbarth]

I observe that PR#169 includes a trivial change to the last line of particular files that does not affect the parsing of that file.

In one version, the line ends with a newline. In the other, it ends with the end-of-file and no terminating new-line.

Basically, this is the difference in having the newline character be a line separator or having it be the terminal character of each line.

I would prefer that one, or the other, convention be used throughout.
In particular, I suggest that we not have a filter that allows both, thus failing to flag trivial changes from one form to the other.