Norihiro Tanaka wrote:
I confirmed that the bug is not fixed in CentOS 5 yet.
Thanks for checking. Does the attached patch work around the problem?
>From 70690079165989c6df9b2c2b7e2a765dff9c7aa4 Mon Sep 17 00:00:00 2001 From: Paul Eggert <egg...@cs.ucla.edu> Date: Tue, 9 Dec 2014 23:01:16 -0800 Subject: [PATCH] tests: port fmbtest to CentOS 5 * tests/fmbtest: Skip test if sed does not work with the cs_CZ.UTF-8 locale. Problem reported by Norihiro Tanaka in: http://bugs.gnu.org/19293#11 --- tests/fmbtest | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/tests/fmbtest b/tests/fmbtest index c4f8c88..cc0f2c8 100755 --- a/tests/fmbtest +++ b/tests/fmbtest @@ -12,6 +12,10 @@ cz=cs_CZ.UTF-8 # If cs_CZ.UTF-8 locale doesn't work, skip this test. LC_ALL=$cz locale -k LC_CTYPE 2>/dev/null | grep -q charmap.*UTF-8 \ || skip_ this system lacks the $cz locale +test 1 -eq $(printf '\xf0\x9d\x87\x95\n' \ + | LC_ALL=cs_CZ.utf8 sed -n '/[^0123456789]/p' \ + | wc -l) \ + || skip_ sed does not work with the $cz locale # If matching is done in single-byte mode, skip this test too printf 'é\n' | LC_ALL=$cz grep -Eq '^[é]{2}$' -- 1.9.3
