On 2023-08-27 18:16, Pádraig Brady wrote:
On 27/08/2023 17:06, Bruno Haible wrote:
Pádraig Brady wrote:
Perhaps 'sleep' is a shell builtin here which is confusing things?
Yes:
$ type sleep
sleep is a shell builtin
Does the following help (I don't have easy access to an alpine
system):
diff --git a/tests/misc/usage_vs_getopt.sh
b/tests/misc/usage_vs_getopt.sh
index 3ffb7b0b1..f4f12136a 100755
--- a/tests/misc/usage_vs_getopt.sh
+++ b/tests/misc/usage_vs_getopt.sh
@@ -87,7 +87,7 @@ for prg in $built_programs; do
'[' | expr | stty )
continue;;
# Wrap some utilities known by the shell by env.
- echo | false | kill | printf | pwd | test | true )
+ echo | false | kill | printf | pwd | sleep | test | true )
prg="env $prg";;
esac
checkprg $prg
Yes. With this change, the failure disappears.
Cool. Applied (as well as all your other coreutils and gnulib changes).
Perhaps it would be a good idea to release an updated tarball in the
next few days, so we can test again with the latest patches applied,
just to make sure there are no regressions.