On Wed, Aug 25, 2021 at 9:24 PM Taylor Blau <me@xxxxxxxxxxxx> wrote:
> On Wed, Aug 25, 2021 at 08:17:10PM -0700, Carlo Marcelo Arenas Belón wrote:
> > diff --git a/t/test-lib-functions.sh b/t/test-lib-functions.sh
> > index e28411bb75..2803c97df3 100644
> > --- a/t/test-lib-functions.sh
> > +++ b/t/test-lib-functions.sh
> > @@ -790,7 +790,7 @@ test_path_exists () {
> > test_dir_is_empty () {
> > test "$#" -ne 1 && BUG "1 param"
> > test_path_is_dir "$1" &&
> > - if test -n "$(ls -a1 "$1" | egrep -v '^\.\.?$')"
> > + if test -n "$(ls -a1 "$1" | grep -v '^\.$' | grep -v '^\.\.$')"
>
> This replacement is correct, but I'm not sure that I necessarily find it
> simpler. If we really are concerned about egrep usage, then
>
> if test -n "$(find "$1" | grep -v '^\.$')"
Interesting idea; if having a much simpler expression is so important
then we could do instead [^.], since all use cases will be covered
(nobody is going to create a three dotted file to workaround a test
IMHO)
> But it looks like we are fairly OK with egrep in t (`git
> grep 'egrep' -- t | wc -l` turns up 19 matches), so I'm not sure the
> change is necessary in the first place.
egrep (and also fgrep, which we intentionally support because it is
missing from some ancient AIX system[1]) will be removed in the
next[2] release of GNU grep.
Carlo
[1] 87539416fd (tests: grep portability fixes, 2008-09-30)
[2] https://git.savannah.gnu.org/cgit/grep.git/commit/?id=a9515624709865d480e3142fd959bccd1c9372d1
