2052.3.2
by John Arbash Meinel
Change Copyright .. by Canonical to Copyright ... Canonical |
1 |
# Copyright (C) 2005 Canonical Ltd
|
1185.16.70
by Martin Pool
- improved handling of non-ascii branch names and test |
2 |
#
|
3 |
# This program is free software; you can redistribute it and/or modify
|
|
4 |
# it under the terms of the GNU General Public License as published by
|
|
5 |
# the Free Software Foundation; either version 2 of the License, or
|
|
6 |
# (at your option) any later version.
|
|
7 |
#
|
|
8 |
# This program is distributed in the hope that it will be useful,
|
|
9 |
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
10 |
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
11 |
# GNU General Public License for more details.
|
|
12 |
#
|
|
13 |
# You should have received a copy of the GNU General Public License
|
|
14 |
# along with this program; if not, write to the Free Software
|
|
4183.7.1
by Sabin Iacob
update FSF mailing address |
15 |
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
1185.16.70
by Martin Pool
- improved handling of non-ascii branch names and test |
16 |
|
17 |
"""Test that various operations work in a non-ASCII environment."""
|
|
18 |
||
19 |
import os |
|
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
20 |
import sys |
21 |
from unicodedata import normalize |
|
1185.16.70
by Martin Pool
- improved handling of non-ascii branch names and test |
22 |
|
1830.3.1
by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms |
23 |
from bzrlib import osutils |
24 |
from bzrlib.osutils import pathjoin |
|
25 |
from bzrlib.tests import TestCase, TestCaseWithTransport, TestSkipped |
|
1534.4.26
by Robert Collins
Move working tree initialisation out from Branch.initialize, deprecated Branch.initialize to Branch.create. |
26 |
from bzrlib.workingtree import WorkingTree |
27 |
||
28 |
||
29 |
class NonAsciiTest(TestCaseWithTransport): |
|
1185.16.70
by Martin Pool
- improved handling of non-ascii branch names and test |
30 |
|
31 |
def test_add_in_nonascii_branch(self): |
|
32 |
"""Test adding in a non-ASCII branch."""
|
|
33 |
br_dir = u"\u1234" |
|
1185.16.71
by Martin Pool
- try to avoid test failure on platforms with ascii-only filesystems |
34 |
try: |
1534.4.26
by Robert Collins
Move working tree initialisation out from Branch.initialize, deprecated Branch.initialize to Branch.create. |
35 |
wt = self.make_branch_and_tree(br_dir) |
1185.12.91
by Aaron Bentley
Fixed exception (No such thing as EncodingError that I can see) |
36 |
except UnicodeEncodeError: |
1534.4.26
by Robert Collins
Move working tree initialisation out from Branch.initialize, deprecated Branch.initialize to Branch.create. |
37 |
raise TestSkipped("filesystem can't accomodate nonascii names") |
1185.16.71
by Martin Pool
- try to avoid test failure on platforms with ascii-only filesystems |
38 |
return
|
1534.4.26
by Robert Collins
Move working tree initialisation out from Branch.initialize, deprecated Branch.initialize to Branch.create. |
39 |
file(pathjoin(br_dir, "a"), "w").write("hello") |
40 |
wt.add(["a"], ["a-id"]) |
|
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
41 |
|
42 |
||
43 |
a_circle_c = u'\xe5' |
|
1830.3.5
by John Arbash Meinel
make_entry refuses to create non-normalized entries. |
44 |
a_circle_d = u'a\u030a' |
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
45 |
a_dots_c = u'\xe4' |
46 |
a_dots_d = u'a\u0308' |
|
47 |
z_umlat_c = u'\u017d' |
|
48 |
z_umlat_d = u'Z\u030c' |
|
3201.1.1
by jameinel
Fix bug #185458, switch from NFKC to NFC and add tests for filenames that would be broken under NFKC |
49 |
squared_c = u'\xbc' # This gets mapped to '2' if we use NFK[CD] |
50 |
squared_d = u'\xbc' |
|
51 |
quarter_c = u'\xb2' # Gets mapped to u'1\u20444' (1/4) if we use NFK[CD] |
|
52 |
quarter_d = u'\xb2' |
|
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
53 |
|
54 |
||
1830.3.1
by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms |
55 |
class TestNormalization(TestCase): |
56 |
"""Verify that we have our normalizations correct."""
|
|
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
57 |
|
1830.3.1
by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms |
58 |
def test_normalize(self): |
3201.1.1
by jameinel
Fix bug #185458, switch from NFKC to NFC and add tests for filenames that would be broken under NFKC |
59 |
self.assertEqual(a_circle_d, normalize('NFD', a_circle_c)) |
60 |
self.assertEqual(a_circle_c, normalize('NFC', a_circle_d)) |
|
61 |
self.assertEqual(a_dots_d, normalize('NFD', a_dots_c)) |
|
62 |
self.assertEqual(a_dots_c, normalize('NFC', a_dots_d)) |
|
63 |
self.assertEqual(z_umlat_d, normalize('NFD', z_umlat_c)) |
|
64 |
self.assertEqual(z_umlat_c, normalize('NFC', z_umlat_d)) |
|
65 |
self.assertEqual(squared_d, normalize('NFC', squared_c)) |
|
66 |
self.assertEqual(squared_c, normalize('NFD', squared_d)) |
|
67 |
self.assertEqual(quarter_d, normalize('NFC', quarter_c)) |
|
68 |
self.assertEqual(quarter_c, normalize('NFD', quarter_d)) |
|
1830.3.1
by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms |
69 |
|
70 |
||
1830.3.2
by John Arbash Meinel
normalized_filename is a much better name |
71 |
class NormalizedFilename(TestCaseWithTransport): |
72 |
"""Test normalized_filename and associated helpers"""
|
|
73 |
||
74 |
def test__accessible_normalized_filename(self): |
|
75 |
anf = osutils._accessible_normalized_filename |
|
1830.3.8
by John Arbash Meinel
unicodedata.normalize requires unicode strings |
76 |
# normalized_filename should allow plain ascii strings
|
77 |
# not just unicode strings
|
|
78 |
self.assertEqual((u'ascii', True), anf('ascii')) |
|
1830.3.2
by John Arbash Meinel
normalized_filename is a much better name |
79 |
self.assertEqual((a_circle_c, True), anf(a_circle_c)) |
80 |
self.assertEqual((a_circle_c, True), anf(a_circle_d)) |
|
81 |
self.assertEqual((a_dots_c, True), anf(a_dots_c)) |
|
82 |
self.assertEqual((a_dots_c, True), anf(a_dots_d)) |
|
83 |
self.assertEqual((z_umlat_c, True), anf(z_umlat_c)) |
|
84 |
self.assertEqual((z_umlat_c, True), anf(z_umlat_d)) |
|
3201.1.1
by jameinel
Fix bug #185458, switch from NFKC to NFC and add tests for filenames that would be broken under NFKC |
85 |
self.assertEqual((squared_c, True), anf(squared_c)) |
86 |
self.assertEqual((squared_c, True), anf(squared_d)) |
|
87 |
self.assertEqual((quarter_c, True), anf(quarter_c)) |
|
88 |
self.assertEqual((quarter_c, True), anf(quarter_d)) |
|
1830.3.2
by John Arbash Meinel
normalized_filename is a much better name |
89 |
|
90 |
def test__inaccessible_normalized_filename(self): |
|
91 |
inf = osutils._inaccessible_normalized_filename |
|
1830.3.8
by John Arbash Meinel
unicodedata.normalize requires unicode strings |
92 |
# normalized_filename should allow plain ascii strings
|
93 |
# not just unicode strings
|
|
94 |
self.assertEqual((u'ascii', True), inf('ascii')) |
|
1830.3.2
by John Arbash Meinel
normalized_filename is a much better name |
95 |
self.assertEqual((a_circle_c, True), inf(a_circle_c)) |
96 |
self.assertEqual((a_circle_c, False), inf(a_circle_d)) |
|
97 |
self.assertEqual((a_dots_c, True), inf(a_dots_c)) |
|
98 |
self.assertEqual((a_dots_c, False), inf(a_dots_d)) |
|
99 |
self.assertEqual((z_umlat_c, True), inf(z_umlat_c)) |
|
100 |
self.assertEqual((z_umlat_c, False), inf(z_umlat_d)) |
|
3201.1.1
by jameinel
Fix bug #185458, switch from NFKC to NFC and add tests for filenames that would be broken under NFKC |
101 |
self.assertEqual((squared_c, True), inf(squared_c)) |
102 |
self.assertEqual((squared_c, True), inf(squared_d)) |
|
103 |
self.assertEqual((quarter_c, True), inf(quarter_c)) |
|
104 |
self.assertEqual((quarter_c, True), inf(quarter_d)) |
|
1830.3.1
by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms |
105 |
|
106 |
def test_functions(self): |
|
107 |
if osutils.normalizes_filenames(): |
|
1830.3.2
by John Arbash Meinel
normalized_filename is a much better name |
108 |
self.assertEqual(osutils.normalized_filename, |
109 |
osutils._accessible_normalized_filename) |
|
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
110 |
else: |
1830.3.2
by John Arbash Meinel
normalized_filename is a much better name |
111 |
self.assertEqual(osutils.normalized_filename, |
112 |
osutils._inaccessible_normalized_filename) |
|
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
113 |
|
114 |
def test_platform(self): |
|
1711.7.36
by John Arbash Meinel
Use different filenames to avoid path collisions on win32 w/ FAT32 |
115 |
# With FAT32 and certain encodings on win32
|
116 |
# a_circle_c and a_dots_c actually map to the same file
|
|
117 |
# adding a suffix kicks in the 'preserving but insensitive'
|
|
118 |
# route, and maintains the right files
|
|
119 |
files = [a_circle_c+'.1', a_dots_c+'.2', z_umlat_c+'.3'] |
|
1685.1.74
by Wouter van Heyst
fix nonascii tests to run properly under LANG=C |
120 |
try: |
1711.7.36
by John Arbash Meinel
Use different filenames to avoid path collisions on win32 w/ FAT32 |
121 |
self.build_tree(files) |
1685.1.74
by Wouter van Heyst
fix nonascii tests to run properly under LANG=C |
122 |
except UnicodeError: |
123 |
raise TestSkipped("filesystem cannot create unicode files") |
|
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
124 |
|
125 |
if sys.platform == 'darwin': |
|
1711.7.36
by John Arbash Meinel
Use different filenames to avoid path collisions on win32 w/ FAT32 |
126 |
expected = sorted([a_circle_d+'.1', a_dots_d+'.2', z_umlat_d+'.3']) |
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
127 |
else: |
1711.7.36
by John Arbash Meinel
Use different filenames to avoid path collisions on win32 w/ FAT32 |
128 |
expected = sorted(files) |
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
129 |
|
130 |
present = sorted(os.listdir(u'.')) |
|
131 |
self.assertEqual(expected, present) |
|
132 |
||
1830.3.1
by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms |
133 |
def test_access_normalized(self): |
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
134 |
# We should always be able to access files created with
|
1830.3.16
by John Arbash Meinel
NEWS about fixing #43689 |
135 |
# normalized filenames
|
1711.7.36
by John Arbash Meinel
Use different filenames to avoid path collisions on win32 w/ FAT32 |
136 |
# With FAT32 and certain encodings on win32
|
137 |
# a_circle_c and a_dots_c actually map to the same file
|
|
138 |
# adding a suffix kicks in the 'preserving but insensitive'
|
|
139 |
# route, and maintains the right files
|
|
3201.1.1
by jameinel
Fix bug #185458, switch from NFKC to NFC and add tests for filenames that would be broken under NFKC |
140 |
files = [a_circle_c+'.1', a_dots_c+'.2', z_umlat_c+'.3', |
141 |
squared_c+'.4', quarter_c+'.5'] |
|
1685.1.74
by Wouter van Heyst
fix nonascii tests to run properly under LANG=C |
142 |
try: |
2204.3.1
by Alexander Belchenko
fix win32 selftest regression introduced by bzr.dev.revno2198 |
143 |
self.build_tree(files, line_endings='native') |
1685.1.74
by Wouter van Heyst
fix nonascii tests to run properly under LANG=C |
144 |
except UnicodeError: |
145 |
raise TestSkipped("filesystem cannot create unicode files") |
|
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
146 |
|
147 |
for fname in files: |
|
148 |
# We should get an exception if we can't open the file at
|
|
149 |
# this location.
|
|
1830.3.2
by John Arbash Meinel
normalized_filename is a much better name |
150 |
path, can_access = osutils.normalized_filename(fname) |
1830.3.1
by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms |
151 |
|
152 |
self.assertEqual(path, fname) |
|
153 |
self.assertTrue(can_access) |
|
154 |
||
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
155 |
f = open(path, 'rb') |
1711.7.36
by John Arbash Meinel
Use different filenames to avoid path collisions on win32 w/ FAT32 |
156 |
try: |
157 |
# Check the contents
|
|
158 |
shouldbe = 'contents of %s%s' % (path.encode('utf8'), |
|
159 |
os.linesep) |
|
160 |
actual = f.read() |
|
161 |
finally: |
|
162 |
f.close() |
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
163 |
self.assertEqual(shouldbe, actual, |
2204.3.3
by Alexander Belchenko
Don't formats unicode strings in test error messages with '%s': it's unsafe |
164 |
'contents of %r is incorrect: %r != %r' |
1711.7.36
by John Arbash Meinel
Use different filenames to avoid path collisions on win32 w/ FAT32 |
165 |
% (path, shouldbe, actual)) |
1185.85.75
by John Arbash Meinel
Adding bzrlib.osutils.unicode_filename to handle unicode normalization for file paths. |
166 |
|
1830.3.1
by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms |
167 |
def test_access_non_normalized(self): |
168 |
# Sometimes we can access non-normalized files by their normalized
|
|
1830.3.2
by John Arbash Meinel
normalized_filename is a much better name |
169 |
# path, verify that normalized_filename returns the right info
|
1830.3.16
by John Arbash Meinel
NEWS about fixing #43689 |
170 |
files = [a_circle_d+'.1', a_dots_d+'.2', z_umlat_d+'.3'] |
1830.3.1
by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms |
171 |
|
172 |
try: |
|
173 |
self.build_tree(files) |
|
174 |
except UnicodeError: |
|
175 |
raise TestSkipped("filesystem cannot create unicode files") |
|
176 |
||
177 |
for fname in files: |
|
178 |
# We should get an exception if we can't open the file at
|
|
179 |
# this location.
|
|
1830.3.2
by John Arbash Meinel
normalized_filename is a much better name |
180 |
path, can_access = osutils.normalized_filename(fname) |
1830.3.1
by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms |
181 |
|
182 |
self.assertNotEqual(path, fname) |
|
183 |
||
184 |
# We should always be able to access them from the name
|
|
185 |
# they were created with
|
|
186 |
f = open(fname, 'rb') |
|
187 |
f.close() |
|
188 |
||
1830.3.2
by John Arbash Meinel
normalized_filename is a much better name |
189 |
# And normalized_filename sholud tell us correctly if we can
|
1830.3.1
by John Arbash Meinel
Change the return value of unicode_filename, and make it testable on all platforms |
190 |
# access them by an alternate name
|
191 |
if can_access: |
|
192 |
f = open(path, 'rb') |
|
193 |
f.close() |
|
194 |
else: |
|
195 |
self.assertRaises(IOError, open, path, 'rb') |