Use withs

utils/subprocess_output.py

@@ -18,19 +18,19 @@ def get_subprocess_output(command, log, shell=None, stdin=None):
     # docs warn that the data read is buffered in memory. They suggest not to
     # use subprocess.PIPE if the data size is large or unlimited.
     with tempfile.TemporaryFile('rw') as stdout_f:
-        stderr_f = tempfile.TemporaryFile('rw')
-        proc = subprocess.Popen(command, close_fds=True, shell=shell, stdin=stdin,
-                                stdout=stdout_f, stderr=stderr_f)
-        proc.wait()
-        stderr_f.seek(0)
-        err = stderr_f.read()
-        stderr_f.close()
-        if err:
-            log.debug("Error while running {0} : {1}".format(" ".join(command),
-                                                             err))
-
-        stdout_f.seek(0)
-        output = stdout_f.read()
+        with tempfile.TemporaryFile('rw') as stderr_f:
+            proc = subprocess.Popen(command, close_fds=True, shell=shell,
+                                    stdin=stdin, stdout=stdout_f,
+                                    stderr=stderr_f)
+            proc.wait()
+            stderr_f.seek(0)
+            err = stderr_f.read()
+            if err:
+                log.debug("Error while running {0} : {1}".format(" ".join(command),
+                                                                 err))
+
+            stdout_f.seek(0)
+            output = stdout_f.read()
     return output