Pass extra env variables also to the symbolizer script.
In particular LLVM_SYMBOLIZER_PATH is needed by asan_symbolize.py.
Also explicitly wait for test launcher to finish running and use its return
code as an overall exit code of the test. asan_symbolize.py always returns
0 (if it doesn't crash), not very useful.
BUG=424280
[email protected]
Review URL: https://codereview.chromium.org/639713009
Cr-Commit-Position: refs/heads/master@{#300252}
diff --git a/testing/test_env.py b/testing/test_env.py
index 1a41790..4c19427 100755
--- a/testing/test_env.py
+++ b/testing/test_env.py
@@ -174,10 +174,11 @@
p1 = subprocess.Popen(cmd, env=env, stdout=subprocess.PIPE,
stderr=sys.stdout)
p2 = subprocess.Popen(["../tools/valgrind/asan/asan_symbolize.py"],
- stdin=p1.stdout)
+ env=env, stdin=p1.stdout)
p1.stdout.close() # Allow p1 to receive a SIGPIPE if p2 exits.
+ p1.wait()
p2.wait()
- return p2.returncode
+ return p1.returncode
else:
return subprocess.call(cmd, env=env)
except OSError:
