Call `gzip` in subprocess instead of gzipping in Python in http_server.

mojo/devtools/common/devtoolslib/http_server.py

Index: mojo/devtools/common/devtoolslib/http_server.py
diff --git a/mojo/devtools/common/devtoolslib/http_server.py b/mojo/devtools/common/devtoolslib/http_server.py
index d8a275770135205263026d3cba56a3aa733ddf77..7c04174fe34828381b67dfe988a6228f2f1649bb 100644
--- a/mojo/devtools/common/devtoolslib/http_server.py
+++ b/mojo/devtools/common/devtoolslib/http_server.py
@@ -6,15 +6,14 @@ import atexit
 import datetime
 import email.utils
 import errno
-import gzip
 import hashlib
 import logging
 import math
 import os.path
-import shutil
 import socket
-import threading
+import subprocess
 import tempfile
+import threading
 
 import SimpleHTTPServer
 import SocketServer
@@ -155,9 +154,8 @@ def _get_handler_class_for_path(mappings):
                 if not self.gzipped_file:
                   self.gzipped_file = tempfile.NamedTemporaryFile(delete=False)
                   self.original_file_name = candidate
-                  with open(candidate, 'rb') as source:
-                    with gzip.GzipFile(fileobj=self.gzipped_file) as target:
-                      shutil.copyfileobj(source, target)
+                  subprocess.check_call(['gzip', '-c', candidate],

[piotrt, 2015/08/25 09:23:36]

Maybe it would be good to handle errors here? the .check_call method 
will raise a CalledProcessError if the gzip fails.

[ppi, 2015/08/25 11:34:36]

Done.

+                                        stdout=self.gzipped_file)
                   self.gzipped_file.close()
                 return self.gzipped_file.name
               return candidate