Thomas Rast wrote
Judging from the test, this parses as
Change git-diff's whitespace-ignoring modes to generate
output only if a non-(empty patch, which git-apply
rejects) results.
which is a bit weird, isn't it? :-)
Yes, rotten wording. I meant git-apply rejects empty patches,
so let's not do that!
I chose this route rather than making git-apply more
forgiving because as currently implemented, git-diff is not
entirely ignoring whitespace when commanded to do so.
Greg