set the bash pipefail option in the test script to return the error

[monkeysphere.git] / tests / basic

diff --git a/tests/basic b/tests/basic
index 8e7f88a2daf83d52aa57dcb57843d752e9660911..d86388665492ba6e3336df8a8a2b4a478ef05a70 100755 (executable)
--- a/tests/basic
+++ b/tests/basic
@@ -13,11 +13,12 @@
 
 # all subcommands in this script should complete without failure:
 set -e
+set -o pipefail
 
 ## make sure that the right tools are installed to run the test.  the
 ## test has *more* requirements than plain ol' monkeysphere:
 
-[ -x $(which socat) ] || { echo "You must have socat installed to run this test." && exit 1 }
+which socat || { echo "You must have socat installed to run this test." ; exit 1; }
 
 ## FIXME: other checks?