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
25
from bzrlib.tests import TestCase
26
from bzrlib.inventory import Inventory, InventoryEntry
27
from bzrlib.xml4 import serializer_v4
30
_working_inventory_v4 = """<inventory file_id="TREE_ROOT">
31
<entry file_id="bar-20050901064931-73b4b1138abc9cd2" kind="file" name="bar" parent_id="TREE_ROOT" />
32
<entry file_id="foo-20050801201819-4139aa4a272f4250" kind="directory" name="foo" parent_id="TREE_ROOT" />
33
<entry file_id="bar-20050824000535-6bc48cfad47ed134" kind="file" name="bar" parent_id="foo-20050801201819-4139aa4a272f4250" />
37
_revision_v4 = """<revision committer="Martin Pool <mbp@sourcefrog.net>"
38
inventory_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
39
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
40
revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
41
timestamp="1125907235.212"
43
<message>- start splitting code for xml (de)serialization away from objects
44
preparatory to supporting multiple formats by a single library
47
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92" revision_sha1="7bdf4cc8c5bdac739f8cf9b10b78cf4b68f915ff" />
52
_revision_v5 = """<revision committer="Martin Pool <mbp@sourcefrog.net>"
53
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
54
revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
55
timestamp="1125907235.212"
57
<message>- start splitting code for xml (de)serialization away from objects
58
preparatory to supporting multiple formats by a single library
61
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92"/>
66
_revision_v5_utc = """\
67
<revision committer="Martin Pool <mbp@sourcefrog.net>"
68
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
69
revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
70
timestamp="1125907235.212"
72
<message>- start splitting code for xml (de)serialization away from objects
73
preparatory to supporting multiple formats by a single library
76
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92"/>
81
_committed_inv_v5 = """<inventory>
82
<file file_id="bar-20050901064931-73b4b1138abc9cd2"
83
name="bar" parent_id="TREE_ROOT"
84
revision="mbp@foo-123123"/>
85
<directory name="subdir"
86
file_id="foo-20050801201819-4139aa4a272f4250"
88
revision="mbp@foo-00"/>
89
<file executable="yes" file_id="bar-20050824000535-6bc48cfad47ed134"
90
name="bar" parent_id="foo-20050801201819-4139aa4a272f4250"
91
revision="mbp@foo-00"/>
95
_basis_inv_v5 = """<inventory revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92">
96
<file file_id="bar-20050901064931-73b4b1138abc9cd2"
97
name="bar" parent_id="TREE_ROOT"
98
revision="mbp@foo-123123"/>
99
<directory name="subdir"
100
file_id="foo-20050801201819-4139aa4a272f4250"
101
parent_id="TREE_ROOT"
102
revision="mbp@foo-00"/>
103
<file file_id="bar-20050824000535-6bc48cfad47ed134"
104
name="bar" parent_id="foo-20050801201819-4139aa4a272f4250"
105
revision="mbp@foo-00"/>
110
# DO NOT REFLOW THIS. Its the exact revision we want.
111
_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">
112
<message>- start splitting code for xml (de)serialization away from objects
113
preparatory to supporting multiple formats by a single library
116
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92" />
122
# DO NOT REFLOW THIS. Its the exact inventory we want.
123
_expected_inv_v5 = """<inventory format="5">
124
<file file_id="bar-20050901064931-73b4b1138abc9cd2" name="bar" revision="mbp@foo-123123" />
125
<directory file_id="foo-20050801201819-4139aa4a272f4250" name="subdir" revision="mbp@foo-00" />
126
<file executable="yes" file_id="bar-20050824000535-6bc48cfad47ed134" name="bar" parent_id="foo-20050801201819-4139aa4a272f4250" revision="mbp@foo-00" />
131
_expected_inv_v5_root = """<inventory file_id="f<" format="5" revision_id="mother!">
132
<file file_id="bar-20050901064931-73b4b1138abc9cd2" name="bar" parent_id="f<" revision="mbp@foo-123123" />
133
<directory file_id="foo-20050801201819-4139aa4a272f4250" name="subdir" parent_id="f<" revision="mbp@foo-00" />
134
<file executable="yes" file_id="bar-20050824000535-6bc48cfad47ed134" name="bar" parent_id="foo-20050801201819-4139aa4a272f4250" revision="mbp@foo-00" />
138
_expected_inv_v7 = """<inventory format="7" revision_id="rev_outer">
139
<directory file_id="tree-root-321" name="" revision="rev_outer" />
140
<directory file_id="dir-id" name="dir" parent_id="tree-root-321" revision="rev_outer" />
141
<file file_id="file-id" name="file" parent_id="tree-root-321" revision="rev_outer" />
142
<symlink file_id="link-id" name="link" parent_id="tree-root-321" revision="rev_outer" />
143
<tree-reference file_id="nested-id" name="nested" parent_id="tree-root-321" revision="rev_outer" reference_revision="rev_inner" />
147
_revision_utf8_v5 = """<revision committer="Erik Bågfors <erik@foo.net>"
148
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
149
revision_id="erik@bågfors-02"
150
timestamp="1125907235.212"
152
<message>Include µnicode characters
155
<revision_ref revision_id="erik@bågfors-01"/>
160
_inventory_utf8_v5 = """<inventory file_id="TREé_ROOT" format="5"
161
revision_id="erik@bågfors-02">
162
<file file_id="bår-01"
163
name="bår" parent_id="TREé_ROOT"
164
revision="erik@bågfors-01"/>
165
<directory name="sµbdir"
166
file_id="sµbdir-01"
167
parent_id="TREé_ROOT"
168
revision="erik@bågfors-01"/>
169
<file executable="yes" file_id="bår-02"
170
name="bår" parent_id="sµbdir-01"
171
revision="erik@bågfors-02"/>
176
class TestSerializer(TestCase):
177
"""Test XML serialization"""
178
def test_canned_inventory(self):
179
"""Test unpacked a canned inventory v4 file."""
180
inp = StringIO(_working_inventory_v4)
181
inv = serializer_v4.read_inventory(inp)
182
self.assertEqual(len(inv), 4)
183
self.assert_('bar-20050901064931-73b4b1138abc9cd2' in inv)
185
def test_unpack_revision(self):
186
"""Test unpacking a canned revision v4"""
187
inp = StringIO(_revision_v4)
188
rev = serializer_v4.read_revision(inp)
189
eq = self.assertEqual
191
"Martin Pool <mbp@sourcefrog.net>")
193
"mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9")
194
eq(len(rev.parent_ids), 1)
195
eq(rev.parent_ids[0],
196
"mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
198
def test_unpack_revision_5(self):
199
"""Test unpacking a canned revision v5"""
200
inp = StringIO(_revision_v5)
201
rev = bzrlib.xml5.serializer_v5.read_revision(inp)
202
eq = self.assertEqual
204
"Martin Pool <mbp@sourcefrog.net>")
205
eq(len(rev.parent_ids), 1)
206
eq(rev.timezone, 36000)
207
eq(rev.parent_ids[0],
208
"mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
210
def test_unpack_revision_5_utc(self):
211
inp = StringIO(_revision_v5_utc)
212
rev = bzrlib.xml5.serializer_v5.read_revision(inp)
213
eq = self.assertEqual
215
"Martin Pool <mbp@sourcefrog.net>")
216
eq(len(rev.parent_ids), 1)
218
eq(rev.parent_ids[0],
219
"mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
221
def test_unpack_inventory_5(self):
222
"""Unpack canned new-style inventory"""
223
inp = StringIO(_committed_inv_v5)
224
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
225
eq = self.assertEqual
227
ie = inv['bar-20050824000535-6bc48cfad47ed134']
229
eq(ie.revision, 'mbp@foo-00')
231
eq(inv[ie.parent_id].kind, 'directory')
233
def test_unpack_basis_inventory_5(self):
234
"""Unpack canned new-style inventory"""
235
inp = StringIO(_basis_inv_v5)
236
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
237
eq = self.assertEqual
239
eq(inv.revision_id, 'mbp@sourcefrog.net-20050905063503-43948f59fa127d92')
240
ie = inv['bar-20050824000535-6bc48cfad47ed134']
242
eq(ie.revision, 'mbp@foo-00')
244
eq(inv[ie.parent_id].kind, 'directory')
246
def test_repack_inventory_5(self):
247
inp = StringIO(_committed_inv_v5)
248
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
250
bzrlib.xml5.serializer_v5.write_inventory(inv, outp)
251
self.assertEqualDiff(_expected_inv_v5, outp.getvalue())
252
inv2 = bzrlib.xml5.serializer_v5.read_inventory(StringIO(outp.getvalue()))
253
self.assertEqual(inv, inv2)
255
def assertRoundTrips(self, xml_string):
256
inp = StringIO(xml_string)
257
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
259
bzrlib.xml5.serializer_v5.write_inventory(inv, outp)
260
self.assertEqualDiff(xml_string, outp.getvalue())
261
inv2 = bzrlib.xml5.serializer_v5.read_inventory(StringIO(outp.getvalue()))
262
self.assertEqual(inv, inv2)
264
def tests_serialize_inventory_v5_with_root(self):
265
self.assertRoundTrips(_expected_inv_v5_root)
267
def check_repack_revision(self, txt):
268
"""Check that repacking a revision yields the same information"""
270
rev = bzrlib.xml5.serializer_v5.read_revision(inp)
272
bzrlib.xml5.serializer_v5.write_revision(rev, outp)
273
outfile_contents = outp.getvalue()
274
rev2 = bzrlib.xml5.serializer_v5.read_revision(StringIO(outfile_contents))
275
self.assertEqual(rev, rev2)
277
def test_repack_revision_5(self):
278
"""Round-trip revision to XML v5"""
279
self.check_repack_revision(_revision_v5)
281
def test_repack_revision_5_utc(self):
282
self.check_repack_revision(_revision_v5_utc)
284
def test_pack_revision_5(self):
285
"""Pack revision to XML v5"""
286
# fixed 20051025, revisions should have final newline
287
rev = bzrlib.xml5.serializer_v5.read_revision_from_string(_revision_v5)
289
bzrlib.xml5.serializer_v5.write_revision(rev, outp)
290
outfile_contents = outp.getvalue()
291
self.assertEqual(outfile_contents[-1], '\n')
292
self.assertEqualDiff(outfile_contents, bzrlib.xml5.serializer_v5.write_revision_to_string(rev))
293
self.assertEqualDiff(outfile_contents, _expected_rev_v5)
295
def test_empty_property_value(self):
296
"""Create an empty property value check that it serializes correctly"""
297
s_v5 = bzrlib.xml5.serializer_v5
298
rev = s_v5.read_revision_from_string(_revision_v5)
300
props = {'empty':'', 'one':'one'}
301
rev.properties = props
302
txt = s_v5.write_revision_to_string(rev)
303
new_rev = s_v5.read_revision_from_string(txt)
304
self.assertEqual(props, new_rev.properties)
306
def test_roundtrip_inventory_v7(self):
307
inv = Inventory('tree-root-321', revision_id='rev_outer')
308
inv.add(inventory.TreeReference('nested-id', 'nested', 'tree-root-321',
309
'rev_outer', 'rev_inner'))
310
inv.add(inventory.InventoryFile('file-id', 'file', 'tree-root-321'))
311
inv.add(inventory.InventoryDirectory('dir-id', 'dir',
313
inv.add(inventory.InventoryLink('link-id', 'link', 'tree-root-321'))
314
inv['tree-root-321'].revision = 'rev_outer'
315
inv['dir-id'].revision = 'rev_outer'
316
inv['file-id'].revision = 'rev_outer'
317
inv['link-id'].revision = 'rev_outer'
318
txt = xml7.serializer_v7.write_inventory_to_string(inv)
319
self.assertEqualDiff(_expected_inv_v7, txt)
320
inv2 = xml7.serializer_v7.read_inventory_from_string(txt)
321
self.assertEqual(5, len(inv2))
322
for path, ie in inv.iter_entries():
323
self.assertEqual(ie, inv2[ie.file_id])
325
def test_wrong_format_v7(self):
326
"""Can't accidentally open a file with wrong serializer"""
327
s_v6 = bzrlib.xml6.serializer_v6
328
s_v7 = xml7.serializer_v7
329
self.assertRaises(errors.UnexpectedInventoryFormat,
330
s_v7.read_inventory_from_string, _expected_inv_v5)
331
self.assertRaises(errors.UnexpectedInventoryFormat,
332
s_v6.read_inventory_from_string, _expected_inv_v7)
334
def test_tree_reference(self):
335
s_v5 = bzrlib.xml5.serializer_v5
336
s_v6 = bzrlib.xml6.serializer_v6
337
s_v7 = xml7.serializer_v7
338
inv = Inventory('tree-root-321')
339
inv.add(inventory.TreeReference('nested-id', 'nested', 'tree-root-321',
340
'rev-outer', 'rev-inner'))
341
self.assertRaises(errors.UnsupportedInventoryKind,
342
s_v5.write_inventory_to_string, inv)
343
self.assertRaises(errors.UnsupportedInventoryKind,
344
s_v6.write_inventory_to_string, inv)
345
txt = s_v7.write_inventory_to_string(inv)
346
inv2 = s_v7.read_inventory_from_string(txt)
347
self.assertEqual('tree-root-321', inv2['nested-id'].parent_id)
348
self.assertEqual('rev-outer', inv2['nested-id'].revision)
349
self.assertEqual('rev-inner', inv2['nested-id'].reference_revision)
350
self.assertRaises(errors.UnsupportedInventoryKind,
351
s_v6.read_inventory_from_string,
352
txt.replace('format="7"', 'format="6"'))
353
self.assertRaises(errors.UnsupportedInventoryKind,
354
s_v5.read_inventory_from_string,
355
txt.replace('format="7"', 'format="5"'))
357
def test_revision_ids_are_utf8(self):
358
"""Parsed revision_ids should all be utf-8 strings, not unicode."""
359
s_v5 = bzrlib.xml5.serializer_v5
360
rev = s_v5.read_revision_from_string(_revision_utf8_v5)
361
self.assertEqual('erik@b\xc3\xa5gfors-02', rev.revision_id)
362
self.assertIsInstance(rev.revision_id, str)
363
self.assertEqual(['erik@b\xc3\xa5gfors-01'], rev.parent_ids)
364
for parent_id in rev.parent_ids:
365
self.assertIsInstance(parent_id, str)
366
self.assertEqual(u'Include \xb5nicode characters\n', rev.message)
367
self.assertIsInstance(rev.message, unicode)
369
# ie.revision should either be None or a utf-8 revision id
370
inv = s_v5.read_inventory_from_string(_inventory_utf8_v5)
371
rev_id_1 = u'erik@b\xe5gfors-01'.encode('utf8')
372
rev_id_2 = u'erik@b\xe5gfors-02'.encode('utf8')
373
fid_root = u'TRE\xe9_ROOT'.encode('utf8')
374
fid_bar1 = u'b\xe5r-01'.encode('utf8')
375
fid_sub = u's\xb5bdir-01'.encode('utf8')
376
fid_bar2 = u'b\xe5r-02'.encode('utf8')
377
expected = [(u'', fid_root, None, None),
378
(u'b\xe5r', fid_bar1, fid_root, rev_id_1),
379
(u's\xb5bdir', fid_sub, fid_root, rev_id_1),
380
(u's\xb5bdir/b\xe5r', fid_bar2, fid_sub, rev_id_2),
382
self.assertEqual(rev_id_2, inv.revision_id)
383
self.assertIsInstance(inv.revision_id, str)
385
actual = list(inv.iter_entries_by_dir())
386
for ((exp_path, exp_file_id, exp_parent_id, exp_rev_id),
387
(act_path, act_ie)) in zip(expected, actual):
388
self.assertEqual(exp_path, act_path)
389
self.assertIsInstance(act_path, unicode)
390
self.assertEqual(exp_file_id, act_ie.file_id)
391
self.assertIsInstance(act_ie.file_id, str)
392
self.assertEqual(exp_parent_id, act_ie.parent_id)
393
if exp_parent_id is not None:
394
self.assertIsInstance(act_ie.parent_id, str)
395
self.assertEqual(exp_rev_id, act_ie.revision)
396
if exp_rev_id is not None:
397
self.assertIsInstance(act_ie.revision, str)
399
self.assertEqual(len(expected), len(actual))
401
def test_registry(self):
402
self.assertIs(serializer_v4,
403
xml_serializer.format_registry.get('4'))
404
self.assertIs(bzrlib.xml5.serializer_v5,
405
xml_serializer.format_registry.get('5'))
406
self.assertIs(bzrlib.xml6.serializer_v6,
407
xml_serializer.format_registry.get('6'))
408
self.assertIs(bzrlib.xml7.serializer_v7,
409
xml_serializer.format_registry.get('7'))
412
class TestEncodeAndEscape(TestCase):
413
"""Whitebox testing of the _encode_and_escape function."""
416
# Keep the cache clear before and after the test
417
bzrlib.xml5._ensure_utf8_re()
418
bzrlib.xml5._clear_cache()
419
self.addCleanup(bzrlib.xml5._clear_cache)
421
def test_simple_ascii(self):
422
# _encode_and_escape always appends a final ", because these parameters
423
# are being used in xml attributes, and by returning it now, we have to
424
# do fewer string operations later.
425
val = bzrlib.xml5._encode_and_escape('foo bar')
426
self.assertEqual('foo bar"', val)
427
# The second time should be cached
428
val2 = bzrlib.xml5._encode_and_escape('foo bar')
429
self.assertIs(val2, val)
431
def test_ascii_with_xml(self):
432
self.assertEqual('&'"<>"',
433
bzrlib.xml5._encode_and_escape('&\'"<>'))
435
def test_utf8_with_xml(self):
437
utf8_str = '\xc2\xb5\xc3\xa5&\xd8\xac'
438
self.assertEqual('µå&ج"',
439
bzrlib.xml5._encode_and_escape(utf8_str))
441
def test_unicode(self):
442
uni_str = u'\xb5\xe5&\u062c'
443
self.assertEqual('µå&ج"',
444
bzrlib.xml5._encode_and_escape(uni_str))
added
removed
bzrlib/tests/test_xml.py
