tests: fix per-test core detection

Commit 9933c5ab in glusterfs-patch-acceptance-tests broke the code here
to count cores after each test, with two bad effects:

 * Tests continue to run after the job is already guaranteed to fail,
   tying up resources and delaying jobs for other patches.

 * Cores aren't detected until the end of the job, long after it might
   have been possible to figure out what was going on at the time the
   process died.

The current check here works for the current code in the other repo, but
could break if the two repos are changed without coordination again.

Change-Id: Ica9222c3474fddf8e3b40074da2bdfd4dabb79d7
Signed-off-by: Jeff Darcy <jdarcy@redhat.com>
Reviewed-on: http://review.gluster.org/13921
Smoke: Gluster Build System <jenkins@build.gluster.com>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
CentOS-regression: Gluster Build System <jenkins@build.gluster.com>

1 files changed, 2 insertions, 2 deletions

diff --git a/run-tests.sh b/run-tests.sh
index db956cf46fe..1487f30d832 100755
--- a/run-tests.sh
+++ b/run-tests.sh
@@ -239,7 +239,7 @@ function run_tests()
 
     for t in $(find ${regression_testsdir}/tests -name '*.t' \
                | LC_COLLATE=C sort) ; do
-        old_cores=$(ls /core.* 2> /dev/null | wc -l)
+        old_cores=$(ls /*-*.core 2> /dev/null | wc -l)
         total_tests=$((total_tests+1))
         if match $t "$@" ; then
             selected_tests=$((selected_tests+1))
@@ -287,7 +287,7 @@ function run_tests()
                 RES=${TMP_RES}
                 FAILED="${FAILED}${t} "
             fi
-            new_cores=$(ls /core.* 2> /dev/null | wc -l)
+            new_cores=$(ls /*-*.core 2> /dev/null | wc -l)
             if [ x"$new_cores" != x"$old_cores" ]; then
                 core_diff=$((new_cores-old_cores))
                 echo "$t: $core_diff new core files"