Merge lp:~lool/ubuntu-dev-tools/fix-non-ascii-name into lp:~ubuntu-dev/ubuntu-dev-tools/trunk

[Loïc Minier]

Properly decode non-ascii developer name from environment or gecos.

[Stefano Rivera (stefanor)]

Looks good, but that test won't pass at package build time, with the C locale.

We'll need to run the tests with LANG=C.UTF-8 in debian/rules

[lp:~lool/ubuntu-dev-tools/fix-non-ascii-name]

1391. By Loïc Minier on 2013-05-15

Default to UTF-8 encoding when locale doesn't specify one (as Python defaults
to ascii); thanks Stefano Rivera.

1392. By Loïc Minier on 2013-05-15

Encode the developer name in the current locale to run the test; thanks
Stefano Rivera.

[Loïc Minier (lool)]

(missed your review for some reason; sorry)

Good catch; I've pushed r1391 to default to utf-8 in the config when locale doesn't specify an encoding and r1392 to convert the string from utf-8 into the current locale in the test.

I didn't want to require a specific locale to run the tests or add heavy build-deps; the only requirement now is that the current locale's encoding allows representing "Jöe Déveloper"; alternatively we could set the locale in the test but that seemed more invasive.

[Loïc Minier (lool)]

BTW I've tested r1392 with fr_FR.UTF-8, C and fr_FR.ISO-8859-1 locales; this last one was still failing with r1391.

=== modified file 'ubuntutools/config.py'
--- ubuntutools/config.py	2013-03-18 23:18:02 +0000
+++ ubuntutools/config.py	2013-05-15 00:21:27 +0000
@@ -21,6 +21,7 @@
 import shlex
 import socket
 import sys
+import locale
 
 from ubuntutools.logger import Logger
 
@@ -173,4 +174,10 @@
     if export:
         os.environ['DEBFULLNAME'] = name
         os.environ['DEBEMAIL'] = email
+
+    # decode env var or gecos raw string with the current locale's encoding
+    encoding = locale.getdefaultlocale()[1]
+    if not encoding:
+        encoding = 'utf-8'
+    name = name.decode(encoding)
     return name, email
 
=== modified file 'ubuntutools/test/test_config.py'
--- ubuntutools/test/test_config.py	2013-03-18 23:18:02 +0000
+++ ubuntutools/test/test_config.py	2013-05-15 00:21:27 +0000
@@ -1,4 +1,5 @@
 # test_config.py - Test suite for ubuntutools.config
+# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2010, Stefano Rivera <[e-mail]>
 #
@@ -17,6 +18,7 @@
 import __builtin__
 import os
 import sys
+import locale
 from StringIO import StringIO
 
 import mox
@@ -210,3 +212,12 @@
         os.environ['DEBEMAIL'] = orig = '%s <%s>' % (name, email)
         self.assertEqual(ubu_email(), (name, email))
         self.assertEqual(os.environ['DEBEMAIL'], orig)
+
+    def test_unicode_name(self):
+        encoding = locale.getdefaultlocale()[1]
+        if not encoding:
+            encoding = 'utf-8'
+        name = 'Jöe Déveloper'.decode('utf-8')
+        os.environ['DEBFULLNAME'] = name.encode(encoding)
+        os.environ['DEBEMAIL']    = email = '[e-mail]'
+        self.assertEqual(ubu_email(), (name, email))