Reorganize imports to make Python 3 support easier.

src/allmydata/dirnode.py

@@ -1,5 +1,5 @@
 """Directory Node implementation."""
-import time, unicodedata
+import time
 
 from zope.interface import implementer
 from twisted.internet import defer
@@ -18,7 +18,7 @@ from allmydata.check_results import DeepCheckResults, \
      DeepCheckAndRepairResults
 from allmydata.monitor import Monitor
 from allmydata.util import hashutil, base32, log
-from allmydata.util.encodingutil import quote_output
+from allmydata.util.encodingutil import quote_output, normalize
 from allmydata.util.assertutil import precondition
 from allmydata.util.netstring import netstring, split_netstring
 from allmydata.util.consumer import download_to_data
@@ -101,12 +101,6 @@ def update_metadata(metadata, new_metadata, now):
     return metadata
 
 
-# 'x' at the end of a variable name indicates that it holds a Unicode string that may not
-# be NFC-normalized.
-
-def normalize(namex):
-    return unicodedata.normalize('NFC', namex)
-
 # TODO: {Deleter,MetadataSetter,Adder}.modify all start by unpacking the
 # contents and end by repacking them. It might be better to apply them to
 # the unpacked contents.

src/allmydata/test/common_py3.py

@@ -13,11 +13,17 @@ from future.utils import PY2
 if PY2:
     from builtins import filter, map, zip, ascii, chr, hex, input, next, oct, open, pow, round, super, bytes, dict, list, object, range, str, max, min  # noqa: F401
 
+from past.builtins import unicode
+
 import os
 import time
 import signal
 
 from twisted.internet import reactor
+from twisted.trial import unittest
+
+from ..util.assertutil import precondition
+from ..util.encodingutil import unicode_platform, get_filesystem_encoding
 
 
 class TimezoneMixin(object):
@@ -65,3 +71,20 @@ class SignalMixin(object):
         if self.sigchldHandler:
             signal.signal(signal.SIGCHLD, self.sigchldHandler)
         return super(SignalMixin, self).tearDown()
+
+
+class ReallyEqualMixin(object):
+    def failUnlessReallyEqual(self, a, b, msg=None):
+        self.assertEqual(a, b, msg)
+        self.assertEqual(type(a), type(b), "a :: %r, b :: %r, %r" % (a, b, msg))
+
+
+def skip_if_cannot_represent_filename(u):
+    precondition(isinstance(u, unicode))
+
+    enc = get_filesystem_encoding()
+    if not unicode_platform():
+        try:
+            u.encode(enc)
+        except UnicodeEncodeError:
+            raise unittest.SkipTest("A non-ASCII filename could not be encoded on this platform.")

src/allmydata/test/common_util.py

@@ -9,24 +9,15 @@ from twisted.python import failure
 from twisted.trial import unittest
 
 from ..util.assertutil import precondition
-from allmydata.util.encodingutil import (unicode_platform, get_filesystem_encoding,
-
-                                         get_io_encoding)
+from allmydata.util.encodingutil import get_io_encoding
 from future.utils import PY2
 if PY2: # XXX this is a hack that makes some tests pass on Python3, remove
         # in the future
     from ..scripts import runner
-from .common_py3 import SignalMixin
+from .common_py3 import (
+    SignalMixin, skip_if_cannot_represent_filename, ReallyEqualMixin,
+)
 
-def skip_if_cannot_represent_filename(u):
-    precondition(isinstance(u, unicode))
-
-    enc = get_filesystem_encoding()
-    if not unicode_platform():
-        try:
-            u.encode(enc)
-        except UnicodeEncodeError:
-            raise unittest.SkipTest("A non-ASCII filename could not be encoded on this platform.")
 
 def skip_if_cannot_represent_argv(u):
     precondition(isinstance(u, unicode))
@@ -87,12 +78,6 @@ def flip_one_bit(s, offset=0, size=None):
     return result
 
 
-class ReallyEqualMixin(object):
-    def failUnlessReallyEqual(self, a, b, msg=None):
-        self.assertEqual(a, b, msg)
-        self.assertEqual(type(a), type(b), "a :: %r, b :: %r, %r" % (a, b, msg))
-
-
 class StallMixin(object):
     def stall(self, res=None, delay=1):
         d = defer.Deferred()

src/allmydata/test/test_encodingutil.py

@@ -64,15 +64,15 @@ from twisted.trial import unittest
 
 from twisted.python.filepath import FilePath
 
-from allmydata.test.common_util import ReallyEqualMixin
+from allmydata.test.common_py3 import (
+    ReallyEqualMixin, skip_if_cannot_represent_filename,
+)
 from allmydata.util import encodingutil, fileutil
 from allmydata.util.encodingutil import argv_to_unicode, unicode_to_url, \
     unicode_to_output, quote_output, quote_path, quote_local_unicode_path, \
     quote_filepath, unicode_platform, listdir_unicode, FilenameEncodingError, \
     get_io_encoding, get_filesystem_encoding, to_str, from_utf8_or_none, _reload, \
     to_filepath, extend_filepath, unicode_from_filepath, unicode_segments_from
-from allmydata.dirnode import normalize
-from .common_util import skip_if_cannot_represent_filename
 from twisted.python import usage
 
 
@@ -258,7 +258,7 @@ class EncodingUtil(ReallyEqualMixin):
         _reload()
         filenames = listdir_unicode(u'/dummy')
 
-        self.failUnlessEqual(set([normalize(fname) for fname in filenames]),
+        self.failUnlessEqual(set([encodingutil.normalize(fname) for fname in filenames]),
                              set(TEST_FILENAMES))
 
 
@@ -283,7 +283,7 @@ class StdlibUnicode(unittest.TestCase):
 
         # We only require that the listing includes a filename that is canonically equivalent
         # to lumiere_nfc (on Mac OS X, it will be the NFD equivalent).
-        self.failUnlessIn(lumiere_nfc + ".txt", set([normalize(fname) for fname in filenames]))
+        self.failUnlessIn(lumiere_nfc + ".txt", set([encodingutil.normalize(fname) for fname in filenames]))
 
         expanded = fileutil.expanduser(u"~/" + lumiere_nfc)
         self.failIfIn(u"~", expanded)

src/allmydata/util/encodingutil.py

@@ -6,6 +6,7 @@ unicode and back.
 from past.builtins import unicode
 
 import sys, os, re, locale
+import unicodedata
 
 from allmydata.util.assertutil import precondition, _assert
 from twisted.python import usage
@@ -358,3 +359,9 @@ def listdir_unicode(path):
 
 def listdir_filepath(fp):
     return listdir_unicode(unicode_from_filepath(fp))
+
+
+# 'x' at the end of a variable name indicates that it holds a Unicode string that may not
+# be NFC-normalized.
+def normalize(namex):
+    return unicodedata.normalize('NFC', namex)