Jim Meyering wrote:
> Kamil Dudka wrote:
>> From d889021cebb7bf798d1b7bf24149c354627e9553 Mon Sep 17 00:00:00 2001
>> From: Kamil Dudka <kdu...@redhat.com>
>> Date: Fri, 20 Mar 2009 11:20:12 +0100
>> Subject: [PATCH] filevercmp: extension for simple and numbered backups
>
> Hi Kamil,
>
> Thanks again.
> Complete patch below.
> I'm merging this incremental, to retain (per Bruno's suggestion)
> the two test cases you removed.
> I'll add the ChangeLog entry, of course.
Pushed.
http://git.savannah.gnu.org/cgit/gnulib.git/commit/?id=1721cf06d9
Just after doing that I went to add a slightly different test of
ls -v in coreutils, but was surprised by this:
$ printf '%s\n' a a~ a.~1~ a.0 |src/sort -V
a~
a
a.~1~
a.0
I would have seen this if I'd looked more carefully at your test case,
once I restored the two removed entries.
I do see how this makes sense, and will test for this behavior.
Just wanted to note it here.
