1
# Copyright (C) 2005 Canonical Ltd
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.
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.
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
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
24
from bzrlib.tests import TestCase
25
from bzrlib.inventory import Inventory, InventoryEntry
26
from bzrlib.xml4 import serializer_v4
29
_working_inventory_v4 = """<inventory file_id="TREE_ROOT">
30
<entry file_id="bar-20050901064931-73b4b1138abc9cd2" kind="file" name="bar" parent_id="TREE_ROOT" />
31
<entry file_id="foo-20050801201819-4139aa4a272f4250" kind="directory" name="foo" parent_id="TREE_ROOT" />
32
<entry file_id="bar-20050824000535-6bc48cfad47ed134" kind="file" name="bar" parent_id="foo-20050801201819-4139aa4a272f4250" />
36
_revision_v4 = """<revision committer="Martin Pool <mbp@sourcefrog.net>"
37
inventory_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
38
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
39
revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
40
timestamp="1125907235.212"
42
<message>- start splitting code for xml (de)serialization away from objects
43
preparatory to supporting multiple formats by a single library
46
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92" revision_sha1="7bdf4cc8c5bdac739f8cf9b10b78cf4b68f915ff" />
51
_revision_v5 = """<revision committer="Martin Pool <mbp@sourcefrog.net>"
52
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
53
revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
54
timestamp="1125907235.212"
56
<message>- start splitting code for xml (de)serialization away from objects
57
preparatory to supporting multiple formats by a single library
60
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92"/>
65
_revision_v5_utc = """\
66
<revision committer="Martin Pool <mbp@sourcefrog.net>"
67
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
68
revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
69
timestamp="1125907235.212"
71
<message>- start splitting code for xml (de)serialization away from objects
72
preparatory to supporting multiple formats by a single library
75
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92"/>
80
_committed_inv_v5 = """<inventory>
81
<file file_id="bar-20050901064931-73b4b1138abc9cd2"
82
name="bar" parent_id="TREE_ROOT"
83
revision="mbp@foo-123123"/>
84
<directory name="subdir"
85
file_id="foo-20050801201819-4139aa4a272f4250"
87
revision="mbp@foo-00"/>
88
<file executable="yes" file_id="bar-20050824000535-6bc48cfad47ed134"
89
name="bar" parent_id="foo-20050801201819-4139aa4a272f4250"
90
revision="mbp@foo-00"/>
94
_basis_inv_v5 = """<inventory revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92">
95
<file file_id="bar-20050901064931-73b4b1138abc9cd2"
96
name="bar" parent_id="TREE_ROOT"
97
revision="mbp@foo-123123"/>
98
<directory name="subdir"
99
file_id="foo-20050801201819-4139aa4a272f4250"
100
parent_id="TREE_ROOT"
101
revision="mbp@foo-00"/>
102
<file file_id="bar-20050824000535-6bc48cfad47ed134"
103
name="bar" parent_id="foo-20050801201819-4139aa4a272f4250"
104
revision="mbp@foo-00"/>
109
# DO NOT REFLOW THIS. Its the exact revision we want.
110
_expected_rev_v5 = """<revision committer="Martin Pool <mbp@sourcefrog.net>" format="5" inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41" revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9" timestamp="1125907235.212" timezone="36000">
111
<message>- start splitting code for xml (de)serialization away from objects
112
preparatory to supporting multiple formats by a single library
115
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92" />
121
# DO NOT REFLOW THIS. Its the exact inventory we want.
122
_expected_inv_v5 = """<inventory format="5">
123
<file file_id="bar-20050901064931-73b4b1138abc9cd2" name="bar" revision="mbp@foo-123123" />
124
<directory file_id="foo-20050801201819-4139aa4a272f4250" name="subdir" revision="mbp@foo-00" />
125
<file executable="yes" file_id="bar-20050824000535-6bc48cfad47ed134" name="bar" parent_id="foo-20050801201819-4139aa4a272f4250" revision="mbp@foo-00" />
130
_expected_inv_v5_root = """<inventory file_id="f<" format="5" revision_id="mother!">
131
<file file_id="bar-20050901064931-73b4b1138abc9cd2" name="bar" parent_id="f<" revision="mbp@foo-123123" />
132
<directory file_id="foo-20050801201819-4139aa4a272f4250" name="subdir" parent_id="f<" revision="mbp@foo-00" />
133
<file executable="yes" file_id="bar-20050824000535-6bc48cfad47ed134" name="bar" parent_id="foo-20050801201819-4139aa4a272f4250" revision="mbp@foo-00" />
137
_expected_inv_v7 = """<inventory format="7" revision_id="rev_outer">
138
<directory file_id="tree-root-321" name="" revision="rev_outer" />
139
<directory file_id="dir-id" name="dir" parent_id="tree-root-321" revision="rev_outer" />
140
<file file_id="file-id" name="file" parent_id="tree-root-321" revision="rev_outer" />
141
<symlink file_id="link-id" name="link" parent_id="tree-root-321" revision="rev_outer" />
142
<tree-reference file_id="nested-id" name="nested" parent_id="tree-root-321" revision="rev_outer" reference_revision="rev_inner" />
146
_revision_utf8_v5 = """<revision committer="Erik Bågfors <erik@foo.net>"
147
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
148
revision_id="erik@bågfors-02"
149
timestamp="1125907235.212"
151
<message>Include µnicode characters
154
<revision_ref revision_id="erik@bågfors-01"/>
159
_inventory_utf8_v5 = """<inventory file_id="TREé_ROOT" format="5"
160
revision_id="erik@bågfors-02">
161
<file file_id="bår-01"
162
name="bår" parent_id="TREé_ROOT"
163
revision="erik@bågfors-01"/>
164
<directory name="sµbdir"
165
file_id="sµbdir-01"
166
parent_id="TREé_ROOT"
167
revision="erik@bågfors-01"/>
168
<file executable="yes" file_id="bår-02"
169
name="bår" parent_id="sµbdir-01"
170
revision="erik@bågfors-02"/>
175
class TestSerializer(TestCase):
176
"""Test XML serialization"""
177
def test_canned_inventory(self):
178
"""Test unpacked a canned inventory v4 file."""
179
inp = StringIO(_working_inventory_v4)
180
inv = serializer_v4.read_inventory(inp)
181
self.assertEqual(len(inv), 4)
182
self.assert_('bar-20050901064931-73b4b1138abc9cd2' in inv)
184
def test_unpack_revision(self):
185
"""Test unpacking a canned revision v4"""
186
inp = StringIO(_revision_v4)
187
rev = serializer_v4.read_revision(inp)
188
eq = self.assertEqual
190
"Martin Pool <mbp@sourcefrog.net>")
192
"mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9")
193
eq(len(rev.parent_ids), 1)
194
eq(rev.parent_ids[0],
195
"mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
197
def test_unpack_revision_5(self):
198
"""Test unpacking a canned revision v5"""
199
inp = StringIO(_revision_v5)
200
rev = bzrlib.xml5.serializer_v5.read_revision(inp)
201
eq = self.assertEqual
203
"Martin Pool <mbp@sourcefrog.net>")
204
eq(len(rev.parent_ids), 1)
205
eq(rev.timezone, 36000)
206
eq(rev.parent_ids[0],
207
"mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
209
def test_unpack_revision_5_utc(self):
210
inp = StringIO(_revision_v5_utc)
211
rev = bzrlib.xml5.serializer_v5.read_revision(inp)
212
eq = self.assertEqual
214
"Martin Pool <mbp@sourcefrog.net>")
215
eq(len(rev.parent_ids), 1)
217
eq(rev.parent_ids[0],
218
"mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
220
def test_unpack_inventory_5(self):
221
"""Unpack canned new-style inventory"""
222
inp = StringIO(_committed_inv_v5)
223
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
224
eq = self.assertEqual
226
ie = inv['bar-20050824000535-6bc48cfad47ed134']
228
eq(ie.revision, 'mbp@foo-00')
230
eq(inv[ie.parent_id].kind, 'directory')
232
def test_unpack_basis_inventory_5(self):
233
"""Unpack canned new-style inventory"""
234
inp = StringIO(_basis_inv_v5)
235
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
236
eq = self.assertEqual
238
eq(inv.revision_id, 'mbp@sourcefrog.net-20050905063503-43948f59fa127d92')
239
ie = inv['bar-20050824000535-6bc48cfad47ed134']
241
eq(ie.revision, 'mbp@foo-00')
243
eq(inv[ie.parent_id].kind, 'directory')
245
def test_repack_inventory_5(self):
246
inp = StringIO(_committed_inv_v5)
247
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
249
bzrlib.xml5.serializer_v5.write_inventory(inv, outp)
250
self.assertEqualDiff(_expected_inv_v5, outp.getvalue())
251
inv2 = bzrlib.xml5.serializer_v5.read_inventory(StringIO(outp.getvalue()))
252
self.assertEqual(inv, inv2)
254
def assertRoundTrips(self, xml_string):
255
inp = StringIO(xml_string)
256
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
258
bzrlib.xml5.serializer_v5.write_inventory(inv, outp)
259
self.assertEqualDiff(xml_string, outp.getvalue())
260
inv2 = bzrlib.xml5.serializer_v5.read_inventory(StringIO(outp.getvalue()))
261
self.assertEqual(inv, inv2)
263
def tests_serialize_inventory_v5_with_root(self):
264
self.assertRoundTrips(_expected_inv_v5_root)
266
def check_repack_revision(self, txt):
267
"""Check that repacking a revision yields the same information"""
269
rev = bzrlib.xml5.serializer_v5.read_revision(inp)
271
bzrlib.xml5.serializer_v5.write_revision(rev, outp)
272
outfile_contents = outp.getvalue()
273
rev2 = bzrlib.xml5.serializer_v5.read_revision(StringIO(outfile_contents))
274
self.assertEqual(rev, rev2)
276
def test_repack_revision_5(self):
277
"""Round-trip revision to XML v5"""
278
self.check_repack_revision(_revision_v5)
280
def test_repack_revision_5_utc(self):
281
self.check_repack_revision(_revision_v5_utc)
283
def test_pack_revision_5(self):
284
"""Pack revision to XML v5"""
285
# fixed 20051025, revisions should have final newline
286
rev = bzrlib.xml5.serializer_v5.read_revision_from_string(_revision_v5)
288
bzrlib.xml5.serializer_v5.write_revision(rev, outp)
289
outfile_contents = outp.getvalue()
290
self.assertEqual(outfile_contents[-1], '\n')
291
self.assertEqualDiff(outfile_contents, bzrlib.xml5.serializer_v5.write_revision_to_string(rev))
292
self.assertEqualDiff(outfile_contents, _expected_rev_v5)
294
def test_empty_property_value(self):
295
"""Create an empty property value check that it serializes correctly"""
296
s_v5 = bzrlib.xml5.serializer_v5
297
rev = s_v5.read_revision_from_string(_revision_v5)
299
props = {'empty':'', 'one':'one'}
300
rev.properties = props
301
txt = s_v5.write_revision_to_string(rev)
302
new_rev = s_v5.read_revision_from_string(txt)
303
self.assertEqual(props, new_rev.properties)
305
def test_roundtrip_inventory_v7(self):
306
inv = Inventory('tree-root-321', revision_id='rev_outer')
307
inv.add(inventory.TreeReference('nested-id', 'nested', 'tree-root-321',
308
'rev_outer', 'rev_inner'))
309
inv.add(inventory.InventoryFile('file-id', 'file', 'tree-root-321'))
310
inv.add(inventory.InventoryDirectory('dir-id', 'dir',
312
inv.add(inventory.InventoryLink('link-id', 'link', 'tree-root-321'))
313
inv['tree-root-321'].revision = 'rev_outer'
314
inv['dir-id'].revision = 'rev_outer'
315
inv['file-id'].revision = 'rev_outer'
316
inv['link-id'].revision = 'rev_outer'
317
txt = xml7.serializer_v7.write_inventory_to_string(inv)
318
self.assertEqualDiff(_expected_inv_v7, txt)
319
inv2 = xml7.serializer_v7.read_inventory_from_string(txt)
320
self.assertEqual(5, len(inv2))
321
for path, ie in inv.iter_entries():
322
self.assertEqual(ie, inv2[ie.file_id])
324
def test_wrong_format_v7(self):
325
"""Can't accidentally open a file with wrong serializer"""
326
s_v6 = bzrlib.xml6.serializer_v6
327
s_v7 = xml7.serializer_v7
328
self.assertRaises(errors.UnexpectedInventoryFormat,
329
s_v7.read_inventory_from_string, _expected_inv_v5)
330
self.assertRaises(errors.UnexpectedInventoryFormat,
331
s_v6.read_inventory_from_string, _expected_inv_v7)
333
def test_tree_reference(self):
334
s_v5 = bzrlib.xml5.serializer_v5
335
s_v6 = bzrlib.xml6.serializer_v6
336
s_v7 = xml7.serializer_v7
337
inv = Inventory('tree-root-321')
338
inv.add(inventory.TreeReference('nested-id', 'nested', 'tree-root-321',
339
'rev-outer', 'rev-inner'))
340
self.assertRaises(errors.UnsupportedInventoryKind,
341
s_v5.write_inventory_to_string, inv)
342
self.assertRaises(errors.UnsupportedInventoryKind,
343
s_v6.write_inventory_to_string, inv)
344
txt = s_v7.write_inventory_to_string(inv)
345
inv2 = s_v7.read_inventory_from_string(txt)
346
self.assertEqual('tree-root-321', inv2['nested-id'].parent_id)
347
self.assertEqual('rev-outer', inv2['nested-id'].revision)
348
self.assertEqual('rev-inner', inv2['nested-id'].reference_revision)
349
self.assertRaises(errors.UnsupportedInventoryKind,
350
s_v6.read_inventory_from_string,
351
txt.replace('format="7"', 'format="6"'))
352
self.assertRaises(errors.UnsupportedInventoryKind,
353
s_v5.read_inventory_from_string,
354
txt.replace('format="7"', 'format="5"'))
356
def test_revision_ids_are_utf8(self):
357
"""Parsed revision_ids should all be utf-8 strings, not unicode."""
358
s_v5 = bzrlib.xml5.serializer_v5
359
rev = s_v5.read_revision_from_string(_revision_utf8_v5)
360
self.assertEqual('erik@b\xc3\xa5gfors-02', rev.revision_id)
361
self.assertIsInstance(rev.revision_id, str)
362
self.assertEqual(['erik@b\xc3\xa5gfors-01'], rev.parent_ids)
363
for parent_id in rev.parent_ids:
364
self.assertIsInstance(parent_id, str)
365
self.assertEqual(u'Include \xb5nicode characters\n', rev.message)
366
self.assertIsInstance(rev.message, unicode)
368
# ie.revision should either be None or a utf-8 revision id
369
inv = s_v5.read_inventory_from_string(_inventory_utf8_v5)
370
rev_id_1 = u'erik@b\xe5gfors-01'.encode('utf8')
371
rev_id_2 = u'erik@b\xe5gfors-02'.encode('utf8')
372
fid_root = u'TRE\xe9_ROOT'.encode('utf8')
373
fid_bar1 = u'b\xe5r-01'.encode('utf8')
374
fid_sub = u's\xb5bdir-01'.encode('utf8')
375
fid_bar2 = u'b\xe5r-02'.encode('utf8')
376
expected = [(u'', fid_root, None, None),
377
(u'b\xe5r', fid_bar1, fid_root, rev_id_1),
378
(u's\xb5bdir', fid_sub, fid_root, rev_id_1),
379
(u's\xb5bdir/b\xe5r', fid_bar2, fid_sub, rev_id_2),
381
self.assertEqual(rev_id_2, inv.revision_id)
382
self.assertIsInstance(inv.revision_id, str)
384
actual = list(inv.iter_entries_by_dir())
385
for ((exp_path, exp_file_id, exp_parent_id, exp_rev_id),
386
(act_path, act_ie)) in zip(expected, actual):
387
self.assertEqual(exp_path, act_path)
388
self.assertIsInstance(act_path, unicode)
389
self.assertEqual(exp_file_id, act_ie.file_id)
390
self.assertIsInstance(act_ie.file_id, str)
391
self.assertEqual(exp_parent_id, act_ie.parent_id)
392
if exp_parent_id is not None:
393
self.assertIsInstance(act_ie.parent_id, str)
394
self.assertEqual(exp_rev_id, act_ie.revision)
395
if exp_rev_id is not None:
396
self.assertIsInstance(act_ie.revision, str)
398
self.assertEqual(len(expected), len(actual))
401
class TestEncodeAndEscape(TestCase):
402
"""Whitebox testing of the _encode_and_escape function."""
405
# Keep the cache clear before and after the test
406
bzrlib.xml5._ensure_utf8_re()
407
bzrlib.xml5._clear_cache()
408
self.addCleanup(bzrlib.xml5._clear_cache)
410
def test_simple_ascii(self):
411
# _encode_and_escape always appends a final ", because these parameters
412
# are being used in xml attributes, and by returning it now, we have to
413
# do fewer string operations later.
414
val = bzrlib.xml5._encode_and_escape('foo bar')
415
self.assertEqual('foo bar"', val)
416
# The second time should be cached
417
val2 = bzrlib.xml5._encode_and_escape('foo bar')
418
self.assertIs(val2, val)
420
def test_ascii_with_xml(self):
421
self.assertEqual('&'"<>"',
422
bzrlib.xml5._encode_and_escape('&\'"<>'))
424
def test_utf8_with_xml(self):
426
utf8_str = '\xc2\xb5\xc3\xa5&\xd8\xac'
427
self.assertEqual('µå&ج"',
428
bzrlib.xml5._encode_and_escape(utf8_str))
430
def test_unicode(self):
431
uni_str = u'\xb5\xe5&\u062c'
432
self.assertEqual('µå&ج"',
433
bzrlib.xml5._encode_and_escape(uni_str))
added
removed
bzrlib/tests/test_xml.py
