[gjs: 2/18] build: Suppress LSan on System.exit() test
- From: Philip Chimento <pchimento src gnome org>
- To: commits-list gnome org
- Cc:
- Subject: [gjs: 2/18] build: Suppress LSan on System.exit() test
- Date: Tue, 31 Dec 2019 23:15:10 +0000 (UTC)
commit ccaea1ef445bc332b96895a7e10bd310d043d3c2
Author: Philip Chimento <philip chimento gmail com>
Date: Thu Oct 31 22:14:24 2019 -0700
build: Suppress LSan on System.exit() test
The System.exit() test currently leaks memory if you are running a
debug-enabled SpiderMonkey, because it will abort when you call
System.exit() due to not calling JS_ShutDown(). Therefore, suppress LSan
when running this test to avoid bogus reports of memory leaks.
installed-tests/scripts/testCommandLine.sh | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
---
diff --git a/installed-tests/scripts/testCommandLine.sh b/installed-tests/scripts/testCommandLine.sh
index aaada2d7..f5003f9b 100755
--- a/installed-tests/scripts/testCommandLine.sh
+++ b/installed-tests/scripts/testCommandLine.sh
@@ -99,10 +99,10 @@ report "System.exit(42) should exit with the correct exit code"
# FIXME: should check -eq 42 specifically, but in debug mode we will be
# hitting an assertion. For this reason, skip when running under valgrind
-# since nothing will be freed.
+# since nothing will be freed. Also suppress LSan for the same reason.
echo "# VALGRIND = $VALGRIND"
if test -z $VALGRIND; then
- $gjs exit.js
+ ASAN_OPTIONS=detect_leaks=0 $gjs exit.js
test $? -ne 0
report "System.exit() should still exit across an FFI boundary"
else
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]