On Mon, Aug 29, 2011 at 6:32 AM, Allan McRae <allan@archlinux.org> wrote:
Running the pacman test-suite in a non-English locale results in a few failures. Force the test-suite to run with LC_ALL=C.
Signed-off-by: Allan McRae <allan@archlinux.org> ---
Not sure if this is the best approach or whether we should just have the three tests that require it to run under the C locale. I got query002, query005, and sync1100 to break under zh_CN.utf8. This is due to PACMAN_OUTPUT usage; ideally we should find ways to do these tests without that fragile rule. I'm surprised more don't break.
ldconfig isn't one that breaks (because it checks --debug untranslated outout), but wouldn't it be better to actually look for a file that ldconfig would spit out if we engineer the test to be a bit more realistic? And then follow some similar pattern for as many of these as possible. The three that don't pass above do look like we need to force the locale, however, so maybe this is the best approach.
Makefile.am | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/Makefile.am b/Makefile.am index 75760bb..fb7a4d3 100644 --- a/Makefile.am +++ b/Makefile.am @@ -22,7 +22,7 @@ dist_pkgdata_DATA = \
# run the pactest test suite and vercmp tests check-local: test/pacman test/util src/pacman src/util - $(PYTHON) $(top_srcdir)/test/pacman/pactest.py --debug=1 \ + LC_ALL=C $(PYTHON) $(top_srcdir)/test/pacman/pactest.py --debug=1 \ --test $(top_srcdir)/test/pacman/tests/*.py \ -p $(top_builddir)/src/pacman/pacman $(SH) $(top_srcdir)/test/util/pacsorttest.sh \ -- 1.7.6.1