Some of the tests, especially wgrep-bom-with-{multibyte,unibyte} require a UTF-8 locale to work correctly.
This patch ensures LC_ALL is set to C.UTF-8 in the fixture so that it is applied before running all tests. (C.UTF-8 should be available for all Glibc based system.)