traverseda
/
cpython
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
cpython/Lib/test/test_gettext.py

582 lines
23 KiB
Python
Raw Normal View History

Test suite for new gettext.py module.
2000-08-25 16:14:25 -03:00
import os
Set this test up so that we don't have to create xx/LC_MESSAGES in the cvs tree. It creates the directory and gettext.mo file on the fly, from the base64 encode binary data.
2000-08-25 16:50:38 -03:00
import base64
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
import shutil
Test suite for new gettext.py module.
2000-08-25 16:14:25 -03:00
import gettext
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
import unittest
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
Standardize on test.test_support.run_unittest() (as opposed to a mix of run_unittest() and run_suite()). Also, add functionality to run_unittest() that admits usage of unittest.TestLoader.loadTestsFromModule().
2007-04-25 14:29:52 -03:00
from test import test_support
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
Make reindent.py happy (convert everything to 4-space indents!).
2000-10-23 14:22:08 -03:00
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# TODO:
# - Add new tests, for example for "dgettext"
# - Remove dummy tests, for example testing for single and double quotes
# has no sense, it would have if we were testing a parser (i.e. pygettext)
# - Tests should have only one assert.
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
GNU_MO_DATA = '''\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==
'''
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
UMO_DATA = '''\
3hIElQAAAAACAAAAHAAAACwAAAAFAAAAPAAAAAAAAABQAAAABAAAAFEAAAAPAQAAVgAAAAQAAABm
AQAAAQAAAAIAAAAAAAAAAAAAAAAAAAAAYWLDngBQcm9qZWN0LUlkLVZlcnNpb246IDIuMApQTy1S
ZXZpc2lvbi1EYXRlOiAyMDAzLTA0LTExIDEyOjQyLTA0MDAKTGFzdC1UcmFuc2xhdG9yOiBCYXJy
eSBBLiBXQXJzYXcgPGJhcnJ5QHB5dGhvbi5vcmc+Ckxhbmd1YWdlLVRlYW06IFhYIDxweXRob24t
ZGV2QHB5dGhvbi5vcmc+Ck1JTUUtVmVyc2lvbjogMS4wCkNvbnRlbnQtVHlwZTogdGV4dC9wbGFp
bjsgY2hhcnNldD11dGYtOApDb250ZW50LVRyYW5zZmVyLUVuY29kaW5nOiA3Yml0CkdlbmVyYXRl
ZC1CeTogbWFudWFsbHkKAMKkeXoA
'''
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
Added a test for the fix of SF bug #658233, where continuation lines in .po metadata caused a crash. Also, removed some unnecessary code. Backport candidate.
2003-05-20 14:28:54 -03:00
MMO_DATA = '''\
3hIElQAAAAABAAAAHAAAACQAAAADAAAALAAAAAAAAAA4AAAAeAEAADkAAAABAAAAAAAAAAAAAAAA
UHJvamVjdC1JZC1WZXJzaW9uOiBObyBQcm9qZWN0IDAuMApQT1QtQ3JlYXRpb24tRGF0ZTogV2Vk
IERlYyAxMSAwNzo0NDoxNSAyMDAyClBPLVJldmlzaW9uLURhdGU6IDIwMDItMDgtMTQgMDE6MTg6
NTgrMDA6MDAKTGFzdC1UcmFuc2xhdG9yOiBKb2huIERvZSA8amRvZUBleGFtcGxlLmNvbT4KSmFu
ZSBGb29iYXIgPGpmb29iYXJAZXhhbXBsZS5jb20+Ckxhbmd1YWdlLVRlYW06IHh4IDx4eEBleGFt
cGxlLmNvbT4KTUlNRS1WZXJzaW9uOiAxLjAKQ29udGVudC1UeXBlOiB0ZXh0L3BsYWluOyBjaGFy
c2V0PWlzby04ODU5LTE1CkNvbnRlbnQtVHJhbnNmZXItRW5jb2Rpbmc6IHF1b3RlZC1wcmludGFi
bGUKR2VuZXJhdGVkLUJ5OiBweWdldHRleHQucHkgMS4zCgA=
'''
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
LOCALEDIR = os.path.join('xx', 'LC_MESSAGES')
MOFILE = os.path.join(LOCALEDIR, 'gettext.mo')
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
UMOFILE = os.path.join(LOCALEDIR, 'ugettext.mo')
Added a test for the fix of SF bug #658233, where continuation lines in .po metadata caused a crash. Also, removed some unnecessary code. Backport candidate.
2003-05-20 14:28:54 -03:00
MMOFILE = os.path.join(LOCALEDIR, 'metadata.mo')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
class GettextBaseTest(unittest.TestCase):
def setUp(self):
Don't try to create the directory if it already exists, otherwise the test fails
2004-07-18 21:08:59 -03:00
if not os.path.isdir(LOCALEDIR):
os.makedirs(LOCALEDIR)
Merged revisions 86236,86240,86332,86340,87271,87273,87447 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/py3k To comply with the 2.x doc style, the methods in trace.rst use brackets around optional arguments. The rest is a mostly straight merge, modulo support changed to test_support and use of the old super call style in test_tuple. ........ r86236 | eric.araujo | 2010-11-06 03:44:43 +0100 (sam., 06 nov. 2010) | 2 lines Make sure each test can be run standalone (./python Lib/distutils/tests/x.py) ........ r86240 | eric.araujo | 2010-11-06 05:11:59 +0100 (sam., 06 nov. 2010) | 2 lines Prevent ResourceWarnings in test_gettext ........ r86332 | eric.araujo | 2010-11-08 19:15:17 +0100 (lun., 08 nov. 2010) | 4 lines Add missing NEWS entry for a fix committed by Senthil. All recent modifications to distutils should now be covered in NEWS. ........ r86340 | eric.araujo | 2010-11-08 22:48:23 +0100 (lun., 08 nov. 2010) | 2 lines This was actually fixed for the previous alpha. ........ r87271 | eric.araujo | 2010-12-15 20:09:58 +0100 (mer., 15 déc. 2010) | 2 lines Improve trace documentation (#9264). Patch by Eli Bendersky. ........ r87273 | eric.araujo | 2010-12-15 20:30:15 +0100 (mer., 15 déc. 2010) | 2 lines Use nested method directives, rewrap long lines, fix whitespace. ........ r87447 | eric.araujo | 2010-12-23 20:13:05 +0100 (jeu., 23 déc. 2010) | 2 lines Fix typo in superclass method name ........
2011-02-02 20:12:18 -04:00
with open(MOFILE, 'wb') as fp:
fp.write(base64.decodestring(GNU_MO_DATA))
with open(UMOFILE, 'wb') as fp:
fp.write(base64.decodestring(UMO_DATA))
with open(MMOFILE, 'wb') as fp:
fp.write(base64.decodestring(MMO_DATA))
Make test.test_support.EnvironmentVarGuard behave like a dictionary. All changes are mirrored to the underlying os.environ dict, but rolled back on exit from the with block.
2009-05-01 14:35:37 -03:00
self.env = test_support.EnvironmentVarGuard()
self.env['LANGUAGE'] = 'xx'
Merged revisions 86239,86241 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/py3k ........ r86239 | eric.araujo | 2010-11-06 05:09:29 +0100 (sam., 06 nov. 2010) | 2 lines Fix caching error found by regrtest -R (#10229) ........ r86241 | eric.araujo | 2010-11-06 05:24:00 +0100 (sam., 06 nov. 2010) | 2 lines Add missing NEWS entry for r86239. ........
2010-11-06 01:36:28 -03:00
gettext._translations.clear()
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
def tearDown(self):
Make test.test_support.EnvironmentVarGuard behave like a dictionary. All changes are mirrored to the underlying os.environ dict, but rolled back on exit from the with block.
2009-05-01 14:35:37 -03:00
self.env.__exit__()
del self.env
Make tests clean up after themselves better. This means: * call tearDown when Setup is called * shutil.rmtree the root of the created directory instead of just the leaf directory * set the LANGUAGE environment variable to what it was originally and not assume 'en'.
2003-04-27 16:42:41 -03:00
shutil.rmtree(os.path.split(LOCALEDIR)[0])
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
#17898: reset k and v so that the loop doesn't use an old value
2015-04-14 11:03:35 -03:00
GNU_MO_DATA_ISSUE_17898 = b'''\
3hIElQAAAAABAAAAHAAAACQAAAAAAAAAAAAAAAAAAAAsAAAAggAAAC0AAAAAUGx1cmFsLUZvcm1z
OiBucGx1cmFscz0yOyBwbHVyYWw9KG4gIT0gMSk7CiMtIy0jLSMtIyAgbWVzc2FnZXMucG8gKEVk
WCBTdHVkaW8pICAjLSMtIy0jLSMKQ29udGVudC1UeXBlOiB0ZXh0L3BsYWluOyBjaGFyc2V0PVVU
Ri04CgA=
'''
Set this test up so that we don't have to create xx/LC_MESSAGES in the cvs tree. It creates the directory and gettext.mo file on the fly, from the base64 encode binary data.
2000-08-25 16:50:38 -03:00
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
class GettextTestCase1(GettextBaseTest):
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def setUp(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
GettextBaseTest.setUp(self)
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
self.localedir = os.curdir
self.mofile = MOFILE
gettext.install('gettext', self.localedir)
Set this test up so that we don't have to create xx/LC_MESSAGES in the cvs tree. It creates the directory and gettext.mo file on the fly, from the base64 encode binary data.
2000-08-25 16:50:38 -03:00
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_some_translations(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# test some translations
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(_('albatross'), 'albatross')
eq(_(u'mullusk'), 'bacon')
eq(_(r'Raymond Luxury Yach-t'), 'Throatwobbler Mangrove')
eq(_(ur'nudge nudge'), 'wink wink')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_double_quotes(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# double quotes
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(_("albatross"), 'albatross')
eq(_(u"mullusk"), 'bacon')
eq(_(r"Raymond Luxury Yach-t"), 'Throatwobbler Mangrove')
eq(_(ur"nudge nudge"), 'wink wink')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_triple_single_quotes(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# triple single quotes
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(_('''albatross'''), 'albatross')
eq(_(u'''mullusk'''), 'bacon')
eq(_(r'''Raymond Luxury Yach-t'''), 'Throatwobbler Mangrove')
eq(_(ur'''nudge nudge'''), 'wink wink')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_triple_double_quotes(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# triple double quotes
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(_("""albatross"""), 'albatross')
eq(_(u"""mullusk"""), 'bacon')
eq(_(r"""Raymond Luxury Yach-t"""), 'Throatwobbler Mangrove')
eq(_(ur"""nudge nudge"""), 'wink wink')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_multiline_strings(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# multiline strings
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(_('''This module provides internationalization and localization
Test suite for new gettext.py module.
2000-08-25 16:14:25 -03:00
support for your Python programs by providing an interface to the GNU
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
gettext message catalog library.'''),
'''Guvf zbqhyr cebivqrf vagreangvbanyvmngvba naq ybpnyvmngvba
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
fhccbeg sbe lbhe Clguba cebtenzf ol cebivqvat na vagresnpr gb gur TAH
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
trggrkg zrffntr pngnybt yvoenel.''')
Set this test up so that we don't have to create xx/LC_MESSAGES in the cvs tree. It creates the directory and gettext.mo file on the fly, from the base64 encode binary data.
2000-08-25 16:50:38 -03:00
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_the_alternative_interface(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# test the alternative interface
Merged revisions 86236,86240,86332,86340,87271,87273,87447 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/py3k To comply with the 2.x doc style, the methods in trace.rst use brackets around optional arguments. The rest is a mostly straight merge, modulo support changed to test_support and use of the old super call style in test_tuple. ........ r86236 | eric.araujo | 2010-11-06 03:44:43 +0100 (sam., 06 nov. 2010) | 2 lines Make sure each test can be run standalone (./python Lib/distutils/tests/x.py) ........ r86240 | eric.araujo | 2010-11-06 05:11:59 +0100 (sam., 06 nov. 2010) | 2 lines Prevent ResourceWarnings in test_gettext ........ r86332 | eric.araujo | 2010-11-08 19:15:17 +0100 (lun., 08 nov. 2010) | 4 lines Add missing NEWS entry for a fix committed by Senthil. All recent modifications to distutils should now be covered in NEWS. ........ r86340 | eric.araujo | 2010-11-08 22:48:23 +0100 (lun., 08 nov. 2010) | 2 lines This was actually fixed for the previous alpha. ........ r87271 | eric.araujo | 2010-12-15 20:09:58 +0100 (mer., 15 déc. 2010) | 2 lines Improve trace documentation (#9264). Patch by Eli Bendersky. ........ r87273 | eric.araujo | 2010-12-15 20:30:15 +0100 (mer., 15 déc. 2010) | 2 lines Use nested method directives, rewrap long lines, fix whitespace. ........ r87447 | eric.araujo | 2010-12-23 20:13:05 +0100 (jeu., 23 déc. 2010) | 2 lines Fix typo in superclass method name ........
2011-02-02 20:12:18 -04:00
with open(self.mofile, 'rb') as fp:
t = gettext.GNUTranslations(fp)
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
# Install the translation object
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
t.install()
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(_('nudge nudge'), 'wink wink')
# Try unicode return type
t.install(unicode=True)
eq(_('mullusk'), 'bacon')
Patch #1349274: gettext.install() now optionally installs additional translation functions other than _() in the builtin namespace.
2006-02-19 09:26:36 -04:00
# Test installation of other methods
import __builtin__
t.install(unicode=True, names=["gettext", "lgettext"])
eq(_, t.ugettext)
eq(__builtin__.gettext, t.ugettext)
eq(lgettext, t.lgettext)
del __builtin__.gettext
del __builtin__.lgettext
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
Set this test up so that we don't have to create xx/LC_MESSAGES in the cvs tree. It creates the directory and gettext.mo file on the fly, from the base64 encode binary data.
2000-08-25 16:50:38 -03:00
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
class GettextTestCase2(GettextBaseTest):
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def setUp(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
GettextBaseTest.setUp(self)
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
self.localedir = os.curdir
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
# Set up the bindings
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
gettext.bindtextdomain('gettext', self.localedir)
gettext.textdomain('gettext')
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
# For convenience
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
self._ = gettext.gettext
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_bindtextdomain(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
self.assertEqual(gettext.bindtextdomain('gettext'), self.localedir)
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_textdomain(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
self.assertEqual(gettext.textdomain(), 'gettext')
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_some_translations(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# test some translations
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(self._('albatross'), 'albatross')
eq(self._(u'mullusk'), 'bacon')
eq(self._(r'Raymond Luxury Yach-t'), 'Throatwobbler Mangrove')
eq(self._(ur'nudge nudge'), 'wink wink')
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_double_quotes(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# double quotes
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(self._("albatross"), 'albatross')
eq(self._(u"mullusk"), 'bacon')
eq(self._(r"Raymond Luxury Yach-t"), 'Throatwobbler Mangrove')
eq(self._(ur"nudge nudge"), 'wink wink')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_triple_single_quotes(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# triple single quotes
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(self._('''albatross'''), 'albatross')
eq(self._(u'''mullusk'''), 'bacon')
eq(self._(r'''Raymond Luxury Yach-t'''), 'Throatwobbler Mangrove')
eq(self._(ur'''nudge nudge'''), 'wink wink')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_triple_double_quotes(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# triple double quotes
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(self._("""albatross"""), 'albatross')
eq(self._(u"""mullusk"""), 'bacon')
eq(self._(r"""Raymond Luxury Yach-t"""), 'Throatwobbler Mangrove')
eq(self._(ur"""nudge nudge"""), 'wink wink')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_multiline_strings(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# multiline strings
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(self._('''This module provides internationalization and localization
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
support for your Python programs by providing an interface to the GNU
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
gettext message catalog library.'''),
'''Guvf zbqhyr cebivqrf vagreangvbanyvmngvba naq ybpnyvmngvba
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
fhccbeg sbe lbhe Clguba cebtenzf ol cebivqvat na vagresnpr gb gur TAH
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
trggrkg zrffntr pngnybt yvoenel.''')
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
class PluralFormsTestCase(GettextBaseTest):
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def setUp(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
GettextBaseTest.setUp(self)
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
self.mofile = MOFILE
Set this test up so that we don't have to create xx/LC_MESSAGES in the cvs tree. It creates the directory and gettext.mo file on the fly, from the base64 encode binary data.
2000-08-25 16:50:38 -03:00
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_plural_forms1(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
x = gettext.ngettext('There is %s file', 'There are %s files', 1)
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(x, 'Hay %s fichero')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
x = gettext.ngettext('There is %s file', 'There are %s files', 2)
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(x, 'Hay %s ficheros')
Set this test up so that we don't have to create xx/LC_MESSAGES in the cvs tree. It creates the directory and gettext.mo file on the fly, from the base64 encode binary data.
2000-08-25 16:50:38 -03:00
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_plural_forms2(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Merged revisions 86236,86240,86332,86340,87271,87273,87447 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/py3k To comply with the 2.x doc style, the methods in trace.rst use brackets around optional arguments. The rest is a mostly straight merge, modulo support changed to test_support and use of the old super call style in test_tuple. ........ r86236 | eric.araujo | 2010-11-06 03:44:43 +0100 (sam., 06 nov. 2010) | 2 lines Make sure each test can be run standalone (./python Lib/distutils/tests/x.py) ........ r86240 | eric.araujo | 2010-11-06 05:11:59 +0100 (sam., 06 nov. 2010) | 2 lines Prevent ResourceWarnings in test_gettext ........ r86332 | eric.araujo | 2010-11-08 19:15:17 +0100 (lun., 08 nov. 2010) | 4 lines Add missing NEWS entry for a fix committed by Senthil. All recent modifications to distutils should now be covered in NEWS. ........ r86340 | eric.araujo | 2010-11-08 22:48:23 +0100 (lun., 08 nov. 2010) | 2 lines This was actually fixed for the previous alpha. ........ r87271 | eric.araujo | 2010-12-15 20:09:58 +0100 (mer., 15 déc. 2010) | 2 lines Improve trace documentation (#9264). Patch by Eli Bendersky. ........ r87273 | eric.araujo | 2010-12-15 20:30:15 +0100 (mer., 15 déc. 2010) | 2 lines Use nested method directives, rewrap long lines, fix whitespace. ........ r87447 | eric.araujo | 2010-12-23 20:13:05 +0100 (jeu., 23 déc. 2010) | 2 lines Fix typo in superclass method name ........
2011-02-02 20:12:18 -04:00
with open(self.mofile, 'rb') as fp:
t = gettext.GNUTranslations(fp)
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
x = t.ngettext('There is %s file', 'There are %s files', 1)
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(x, 'Hay %s fichero')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
x = t.ngettext('There is %s file', 'There are %s files', 2)
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(x, 'Hay %s ficheros')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
Issue #28563: Fixed possible DoS and arbitrary code execution when handle plural form selections in the gettext module. The expression parser now supports exact syntax supported by GNU gettext.
2016-11-08 15:15:55 -04:00
# Examples from http://www.gnu.org/software/gettext/manual/gettext.html
def test_ja(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
f = gettext.c2py('0')
s = ''.join([ str(f(x)) for x in range(200) ])
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(s, "00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_de(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
f = gettext.c2py('n != 1')
s = ''.join([ str(f(x)) for x in range(200) ])
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(s, "10111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_fr(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
f = gettext.c2py('n>1')
s = ''.join([ str(f(x)) for x in range(200) ])
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(s, "00111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111")
Set this test up so that we don't have to create xx/LC_MESSAGES in the cvs tree. It creates the directory and gettext.mo file on the fly, from the base64 encode binary data.
2000-08-25 16:50:38 -03:00
Issue #28563: Fixed possible DoS and arbitrary code execution when handle plural form selections in the gettext module. The expression parser now supports exact syntax supported by GNU gettext.
2016-11-08 15:15:55 -04:00
def test_lv(self):
eq = self.assertEqual
f = gettext.c2py('n%10==1 && n%100!=11 ? 0 : n != 0 ? 1 : 2')
s = ''.join([ str(f(x)) for x in range(200) ])
eq(s, "20111111111111111111101111111110111111111011111111101111111110111111111011111111101111111110111111111011111111111111111110111111111011111111101111111110111111111011111111101111111110111111111011111111")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_gd(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
f = gettext.c2py('n==1 ? 0 : n==2 ? 1 : 2')
s = ''.join([ str(f(x)) for x in range(200) ])
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(s, "20122222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_gd2(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# Tests the combination of parentheses and "?:"
f = gettext.c2py('n==1 ? 0 : (n==2 ? 1 : 2)')
s = ''.join([ str(f(x)) for x in range(200) ])
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(s, "20122222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
Issue #28563: Fixed possible DoS and arbitrary code execution when handle plural form selections in the gettext module. The expression parser now supports exact syntax supported by GNU gettext.
2016-11-08 15:15:55 -04:00
def test_ro(self):
eq = self.assertEqual
f = gettext.c2py('n==1 ? 0 : (n==0 || (n%100 > 0 && n%100 < 20)) ? 1 : 2')
s = ''.join([ str(f(x)) for x in range(200) ])
eq(s, "10111111111111111111222222222222222222222222222222222222222222222222222222222222222222222222222222222111111111111111111122222222222222222222222222222222222222222222222222222222222222222222222222222222")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_lt(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
f = gettext.c2py('n%10==1 && n%100!=11 ? 0 : n%10>=2 && (n%100<10 || n%100>=20) ? 1 : 2')
s = ''.join([ str(f(x)) for x in range(200) ])
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(s, "20111111112222222222201111111120111111112011111111201111111120111111112011111111201111111120111111112011111111222222222220111111112011111111201111111120111111112011111111201111111120111111112011111111")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_ru(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
f = gettext.c2py('n%10==1 && n%100!=11 ? 0 : n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2')
s = ''.join([ str(f(x)) for x in range(200) ])
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(s, "20111222222222222222201112222220111222222011122222201112222220111222222011122222201112222220111222222011122222222222222220111222222011122222201112222220111222222011122222201112222220111222222011122222")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
Issue #28563: Fixed possible DoS and arbitrary code execution when handle plural form selections in the gettext module. The expression parser now supports exact syntax supported by GNU gettext.
2016-11-08 15:15:55 -04:00
def test_cs(self):
eq = self.assertEqual
f = gettext.c2py('(n==1) ? 0 : (n>=2 && n<=4) ? 1 : 2')
s = ''.join([ str(f(x)) for x in range(200) ])
eq(s, "20111222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_pl(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
f = gettext.c2py('n==1 ? 0 : n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2')
s = ''.join([ str(f(x)) for x in range(200) ])
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(s, "20111222222222222222221112222222111222222211122222221112222222111222222211122222221112222222111222222211122222222222222222111222222211122222221112222222111222222211122222221112222222111222222211122222")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_sl(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq = self.assertEqual
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
f = gettext.c2py('n%100==1 ? 0 : n%100==2 ? 1 : n%100==3 || n%100==4 ? 2 : 3')
s = ''.join([ str(f(x)) for x in range(200) ])
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
eq(s, "30122333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333012233333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
Issue #28563: Fixed possible DoS and arbitrary code execution when handle plural form selections in the gettext module. The expression parser now supports exact syntax supported by GNU gettext.
2016-11-08 15:15:55 -04:00
def test_ar(self):
eq = self.assertEqual
f = gettext.c2py('n==0 ? 0 : n==1 ? 1 : n==2 ? 2 : n%100>=3 && n%100<=10 ? 3 : n%100>=11 ? 4 : 5')
s = ''.join([ str(f(x)) for x in range(200) ])
eq(s, "01233333333444444444444444444444444444444444444444444444444444444444444444444444444444444444444444445553333333344444444444444444444444444444444444444444444444444444444444444444444444444444444444444444")
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
def test_security(self):
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
raises = self.assertRaises
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# Test for a dangerous expression
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
raises(ValueError, gettext.c2py, "os.chmod('/etc/passwd',0777)")
Issue #28563: Fixed possible DoS and arbitrary code execution when handle plural form selections in the gettext module. The expression parser now supports exact syntax supported by GNU gettext.
2016-11-08 15:15:55 -04:00
# issue28563
raises(ValueError, gettext.c2py, '"(eval(foo) && ""')
raises(ValueError, gettext.c2py, 'f"{os.system(\'sh\')}"')
# Maximum recursion depth exceeded during compilation
raises(ValueError, gettext.c2py, 'n+'*10000 + 'n')
self.assertEqual(gettext.c2py('n+'*100 + 'n')(1), 101)
# MemoryError during compilation
raises(ValueError, gettext.c2py, '('*100 + 'n' + ')'*100)
# Maximum recursion depth exceeded in C to Python translator
raises(ValueError, gettext.c2py, '('*10000 + 'n' + ')'*10000)
self.assertEqual(gettext.c2py('('*20 + 'n' + ')'*20)(1), 1)
def test_chained_comparison(self):
# C doesn't chain comparison as Python so 2 == 2 == 2 gets different results
f = gettext.c2py('n == n == n')
self.assertEqual(''.join(str(f(x)) for x in range(3)), '010')
f = gettext.c2py('1 < n == n')
self.assertEqual(''.join(str(f(x)) for x in range(3)), '100')
f = gettext.c2py('n == n < 2')
self.assertEqual(''.join(str(f(x)) for x in range(3)), '010')
f = gettext.c2py('0 < n < 2')
self.assertEqual(''.join(str(f(x)) for x in range(3)), '111')
def test_decimal_number(self):
self.assertEqual(gettext.c2py('0123')(1), 123)
def test_invalid_syntax(self):
invalid_expressions = [
'x>1', '(n>1', 'n>1)', '42**42**42', '0xa', '1.0', '1e2',
'n>0x1', '+n', '-n', 'n()', 'n(1)', '1+', 'nn', 'n n',
]
for expr in invalid_expressions:
with self.assertRaises(ValueError):
gettext.c2py(expr)
def test_nested_condition_operator(self):
self.assertEqual(gettext.c2py('n?1?2:3:4')(0), 4)
self.assertEqual(gettext.c2py('n?1?2:3:4')(1), 2)
self.assertEqual(gettext.c2py('n?1:3?4:5')(0), 4)
self.assertEqual(gettext.c2py('n?1:3?4:5')(1), 1)
def test_division(self):
f = gettext.c2py('2/n*3')
self.assertEqual(f(1), 6)
self.assertEqual(f(2), 3)
self.assertEqual(f(3), 0)
self.assertEqual(f(-1), -6)
self.assertRaises(ZeroDivisionError, f, 0)
def test_plural_number(self):
Issue #28563: Make plural form selection more lenient and accepting non-integer numbers. Django tests depend on this.
2016-11-14 13:25:44 -04:00
f = gettext.c2py('n != 1')
self.assertEqual(f(1), 0)
self.assertEqual(f(2), 1)
self.assertEqual(f(1.0), 0)
self.assertEqual(f(2.0), 1)
self.assertEqual(f(1.1), 1)
self.assertRaises(TypeError, f, '2')
self.assertRaises(TypeError, f, b'2')
self.assertRaises(TypeError, f, [])
self.assertRaises(TypeError, f, object())
Issue #28563: Fixed possible DoS and arbitrary code execution when handle plural form selections in the gettext module. The expression parser now supports exact syntax supported by GNU gettext.
2016-11-08 15:15:55 -04:00
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
#17898: reset k and v so that the loop doesn't use an old value
2015-04-14 11:03:35 -03:00
class GNUTranslationParsingTest(GettextBaseTest):
def test_plural_form_error_issue17898(self):
with open(MOFILE, 'wb') as fp:
fp.write(base64.decodestring(GNU_MO_DATA_ISSUE_17898))
with open(MOFILE, 'rb') as fp:
# If this runs cleanly, the bug is fixed.
t = gettext.GNUTranslations(fp)
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
class UnicodeTranslationsTest(GettextBaseTest):
def setUp(self):
GettextBaseTest.setUp(self)
Merged revisions 86236,86240,86332,86340,87271,87273,87447 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/py3k To comply with the 2.x doc style, the methods in trace.rst use brackets around optional arguments. The rest is a mostly straight merge, modulo support changed to test_support and use of the old super call style in test_tuple. ........ r86236 | eric.araujo | 2010-11-06 03:44:43 +0100 (sam., 06 nov. 2010) | 2 lines Make sure each test can be run standalone (./python Lib/distutils/tests/x.py) ........ r86240 | eric.araujo | 2010-11-06 05:11:59 +0100 (sam., 06 nov. 2010) | 2 lines Prevent ResourceWarnings in test_gettext ........ r86332 | eric.araujo | 2010-11-08 19:15:17 +0100 (lun., 08 nov. 2010) | 4 lines Add missing NEWS entry for a fix committed by Senthil. All recent modifications to distutils should now be covered in NEWS. ........ r86340 | eric.araujo | 2010-11-08 22:48:23 +0100 (lun., 08 nov. 2010) | 2 lines This was actually fixed for the previous alpha. ........ r87271 | eric.araujo | 2010-12-15 20:09:58 +0100 (mer., 15 déc. 2010) | 2 lines Improve trace documentation (#9264). Patch by Eli Bendersky. ........ r87273 | eric.araujo | 2010-12-15 20:30:15 +0100 (mer., 15 déc. 2010) | 2 lines Use nested method directives, rewrap long lines, fix whitespace. ........ r87447 | eric.araujo | 2010-12-23 20:13:05 +0100 (jeu., 23 déc. 2010) | 2 lines Fix typo in superclass method name ........
2011-02-02 20:12:18 -04:00
with open(UMOFILE, 'rb') as fp:
UnicodeTranslationsTest.setUp(): Removed the coerce flag to the GNUTranslations constructor.
2003-04-24 15:08:13 -03:00
self.t = gettext.GNUTranslations(fp)
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
self._ = self.t.ugettext
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
def test_unicode_msgid(self):
convert usage of fail* to assert*
2009-06-30 19:57:08 -03:00
unless = self.assertTrue
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
unless(isinstance(self._(''), unicode))
unless(isinstance(self._(u''), unicode))
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
def test_unicode_msgstr(self):
eq = self.assertEqual
eq(self._(u'ab\xde'), u'\xa4yz')
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
Added a test for the fix of SF bug #658233, where continuation lines in .po metadata caused a crash. Also, removed some unnecessary code. Backport candidate.
2003-05-20 14:28:54 -03:00
class WeirdMetadataTest(GettextBaseTest):
def setUp(self):
GettextBaseTest.setUp(self)
Merged revisions 86236,86240,86332,86340,87271,87273,87447 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/py3k To comply with the 2.x doc style, the methods in trace.rst use brackets around optional arguments. The rest is a mostly straight merge, modulo support changed to test_support and use of the old super call style in test_tuple. ........ r86236 | eric.araujo | 2010-11-06 03:44:43 +0100 (sam., 06 nov. 2010) | 2 lines Make sure each test can be run standalone (./python Lib/distutils/tests/x.py) ........ r86240 | eric.araujo | 2010-11-06 05:11:59 +0100 (sam., 06 nov. 2010) | 2 lines Prevent ResourceWarnings in test_gettext ........ r86332 | eric.araujo | 2010-11-08 19:15:17 +0100 (lun., 08 nov. 2010) | 4 lines Add missing NEWS entry for a fix committed by Senthil. All recent modifications to distutils should now be covered in NEWS. ........ r86340 | eric.araujo | 2010-11-08 22:48:23 +0100 (lun., 08 nov. 2010) | 2 lines This was actually fixed for the previous alpha. ........ r87271 | eric.araujo | 2010-12-15 20:09:58 +0100 (mer., 15 déc. 2010) | 2 lines Improve trace documentation (#9264). Patch by Eli Bendersky. ........ r87273 | eric.araujo | 2010-12-15 20:30:15 +0100 (mer., 15 déc. 2010) | 2 lines Use nested method directives, rewrap long lines, fix whitespace. ........ r87447 | eric.araujo | 2010-12-23 20:13:05 +0100 (jeu., 23 déc. 2010) | 2 lines Fix typo in superclass method name ........
2011-02-02 20:12:18 -04:00
with open(MMOFILE, 'rb') as fp:
Added a test for the fix of SF bug #658233, where continuation lines in .po metadata caused a crash. Also, removed some unnecessary code. Backport candidate.
2003-05-20 14:28:54 -03:00
try:
self.t = gettext.GNUTranslations(fp)
except:
self.tearDown()
raise
def test_weird_metadata(self):
info = self.t.info()
self.assertEqual(info['last-translator'],
'John Doe <jdoe@example.com>\nJane Foobar <jfoobar@example.com>')
Merged revisions 85223 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/py3k ........ r85223 | eric.araujo | 2010-10-05 01:52:37 +0200 (mar., 05 oct. 2010) | 3 lines Fix interaction of custom translation classes and caching (#9042) ........
2010-10-04 20:59:35 -03:00
class DummyGNUTranslations(gettext.GNUTranslations):
def foo(self):
return 'foo'
class GettextCacheTestCase(GettextBaseTest):
def test_cache(self):
self.localedir = os.curdir
self.mofile = MOFILE
self.assertEqual(len(gettext._translations), 0)
t = gettext.translation('gettext', self.localedir)
self.assertEqual(len(gettext._translations), 1)
t = gettext.translation('gettext', self.localedir,
class_=DummyGNUTranslations)
self.assertEqual(len(gettext._translations), 2)
self.assertEqual(t.__class__, DummyGNUTranslations)
# Calling it again doesn't add to the cache
t = gettext.translation('gettext', self.localedir,
class_=DummyGNUTranslations)
self.assertEqual(len(gettext._translations), 2)
self.assertEqual(t.__class__, DummyGNUTranslations)
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
def test_main():
Standardize on test.test_support.run_unittest() (as opposed to a mix of run_unittest() and run_suite()). Also, add functionality to run_unittest() that admits usage of unittest.TestLoader.loadTestsFromModule().
2007-04-25 14:29:52 -03:00
test_support.run_unittest(__name__)
Make reindent.py happy (convert everything to 4-space indents!).
2000-10-23 14:22:08 -03:00
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
if __name__ == '__main__':
test_main()
# For reference, here's the .po file used to created the GNU_MO_DATA above.
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
#
# The original version was automatically generated from the sources with
# pygettext. Later it was manually modified to add plural forms support.
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
'''
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
# Dummy translation for the Python test_gettext.py module.
Quick and dirty fix for test_extcall failures trigged by Guido's recent classobject.c change. When calling an unbound method with no instances as first argument, the error message has changed. The message now contains the class name, but the output text being compared to is too generic, so skip printing it.
2001-08-24 16:11:57 -03:00
# Copyright (C) 2001 Python Software Foundation
# Barry Warsaw <barry@python.org>, 2000.
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
#
msgid ""
msgstr ""
"Project-Id-Version: 2.0\n"
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
"PO-Revision-Date: 2003-04-11 14:32-0400\n"
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
"Last-Translator: J. David Ibanez <j-david@noos.fr>\n"
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
"Language-Team: XX <python-dev@python.org>\n"
"MIME-Version: 1.0\n"
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
"Content-Type: text/plain; charset=iso-8859-1\n"
"Content-Transfer-Encoding: 8bit\n"
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
"Generated-By: pygettext.py 1.1\n"
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
"Plural-Forms: nplurals=2; plural=n!=1;\n"
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
#: test_gettext.py:19 test_gettext.py:25 test_gettext.py:31 test_gettext.py:37
#: test_gettext.py:51 test_gettext.py:80 test_gettext.py:86 test_gettext.py:92
#: test_gettext.py:98
msgid "nudge nudge"
msgstr "wink wink"
#: test_gettext.py:16 test_gettext.py:22 test_gettext.py:28 test_gettext.py:34
#: test_gettext.py:77 test_gettext.py:83 test_gettext.py:89 test_gettext.py:95
msgid "albatross"
msgstr ""
#: test_gettext.py:18 test_gettext.py:24 test_gettext.py:30 test_gettext.py:36
#: test_gettext.py:79 test_gettext.py:85 test_gettext.py:91 test_gettext.py:97
msgid "Raymond Luxury Yach-t"
msgstr "Throatwobbler Mangrove"
#: test_gettext.py:17 test_gettext.py:23 test_gettext.py:29 test_gettext.py:35
#: test_gettext.py:56 test_gettext.py:78 test_gettext.py:84 test_gettext.py:90
#: test_gettext.py:96
msgid "mullusk"
msgstr "bacon"
#: test_gettext.py:40 test_gettext.py:101
msgid ""
"This module provides internationalization and localization\n"
"support for your Python programs by providing an interface to the GNU\n"
"gettext message catalog library."
msgstr ""
"Guvf zbqhyr cebivqrf vagreangvbanyvmngvba naq ybpnyvmngvba\n"
"fhccbeg sbe lbhe Clguba cebtenzf ol cebivqvat na vagresnpr gb gur TAH\n"
"trggrkg zrffntr pngnybt yvoenel."
Patch #633547: Support plural forms. Do TODOs in test suite.
2002-11-21 17:45:32 -04:00
# Manually added, as neither pygettext nor xgettext support plural forms
# in Python.
msgid "There is %s file"
msgid_plural "There are %s files"
msgstr[0] "Hay %s fichero"
msgstr[1] "Hay %s ficheros"
Expand the test suite to test both the GNU gettext and translation class-based APIs.
2000-08-30 00:32:07 -03:00
'''
From http://mail.python.org/pipermail/i18n-sig/2003-April/001557.html - Expose NullTranslations and GNUTranslations to __all__ - Set the default charset to iso-8859-1. It used to be None, which would cause problems with .ugettext() if the file had no charset parameter. Arguably, the po/mo file would be broken, but I still think iso-8859-1 is a reasonable default. - Add a "coerce" default argument to GNUTranslations's constructor. The reason for this is that in Zope, we want all msgids and msgstrs to be Unicode. For the latter, we could use .ugettext() but there isn't currently a mechanism for Unicode-ifying msgids. The plan then is that the charset parameter specifies the encoding for both the msgids and msgstrs, and both are decoded to Unicode when read. For example, we might encode po files with utf-8. I think the GNU gettext tools don't care. Since this could potentially break code [*] that wants to use the encoded interface .gettext(), the constructor flag is added, defaulting to False. Most code I suspect will want to set this to True and use .ugettext(). - A few other minor changes from the Zope project, including asserting that a zero-length msgid must have a Project-ID-Version header for it to be counted as the metadata record.
2003-04-11 15:36:43 -03:00
# Here's the second example po file example, used to generate the UMO_DATA
# containing utf-8 encoded Unicode strings
'''
# Dummy translation for the Python test_gettext.py module.
# Copyright (C) 2001 Python Software Foundation
# Barry Warsaw <barry@python.org>, 2000.
#
msgid ""
msgstr ""
"Project-Id-Version: 2.0\n"
"PO-Revision-Date: 2003-04-11 12:42-0400\n"
"Last-Translator: Barry A. WArsaw <barry@python.org>\n"
"Language-Team: XX <python-dev@python.org>\n"
"MIME-Version: 1.0\n"
"Content-Type: text/plain; charset=utf-8\n"
"Content-Transfer-Encoding: 7bit\n"
"Generated-By: manually\n"
#: nofile:0
msgid "ab\xc3\x9e"
msgstr "\xc2\xa4yz"
'''
Added a test for the fix of SF bug #658233, where continuation lines in .po metadata caused a crash. Also, removed some unnecessary code. Backport candidate.
2003-05-20 14:28:54 -03:00
# Here's the third example po file, used to generate MMO_DATA
'''
msgid ""
msgstr ""
"Project-Id-Version: No Project 0.0\n"
"POT-Creation-Date: Wed Dec 11 07:44:15 2002\n"
"PO-Revision-Date: 2002-08-14 01:18:58+00:00\n"
"Last-Translator: John Doe <jdoe@example.com>\n"
"Jane Foobar <jfoobar@example.com>\n"
"Language-Team: xx <xx@example.com>\n"
"MIME-Version: 1.0\n"
"Content-Type: text/plain; charset=iso-8859-15\n"
"Content-Transfer-Encoding: quoted-printable\n"
"Generated-By: pygettext.py 1.3\n"
'''
#17898: reset k and v so that the loop doesn't use an old value
2015-04-14 11:03:35 -03:00
#
# messages.po, used for bug 17898
#
'''
# test file for http://bugs.python.org/issue17898
msgid ""
msgstr ""
"Plural-Forms: nplurals=2; plural=(n != 1);\n"
"#-#-#-#-# messages.po (EdX Studio) #-#-#-#-#\n"
"Content-Type: text/plain; charset=UTF-8\n"
'''