On 02/24/2014 11:31 AM, Jim Meyering wrote:
If you find the time, would you please also add a test case like the
one Vincent posted?
I'm afraid that'd be iffy, as there was no bug -- in the test case,
'diff' had correct behavior. In general, there are a lot of correct
outputs for 'diff', and I'm leery of test cases that would require a
particular behavior when the exact choice isn't obvious.I suppose we
could use his test case and make sure that 'diff' doesn't generate the
obviously-suboptimal output, but the value of such a test case would be
relatively small, that is, almost surely the test case wouldn't catch
similar bugs in the future.
