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
text_sha1="A" text_size="1"/>
86
<directory name="subdir"
87
file_id="foo-20050801201819-4139aa4a272f4250"
89
revision="mbp@foo-00"/>
90
<file executable="yes" file_id="bar-20050824000535-6bc48cfad47ed134"
91
name="bar" parent_id="foo-20050801201819-4139aa4a272f4250"
93
text_sha1="B" text_size="0"/>
97
_basis_inv_v5 = """<inventory revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92">
98
<file file_id="bar-20050901064931-73b4b1138abc9cd2"
99
name="bar" parent_id="TREE_ROOT"
100
revision="mbp@foo-123123"/>
101
<directory name="subdir"
102
file_id="foo-20050801201819-4139aa4a272f4250"
103
parent_id="TREE_ROOT"
104
revision="mbp@foo-00"/>
105
<file file_id="bar-20050824000535-6bc48cfad47ed134"
106
name="bar" parent_id="foo-20050801201819-4139aa4a272f4250"
107
revision="mbp@foo-00"/>
112
# DO NOT REFLOW THIS. Its the exact revision we want.
113
_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">
114
<message>- start splitting code for xml (de)serialization away from objects
115
preparatory to supporting multiple formats by a single library
118
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92" />
124
# DO NOT REFLOW THIS. Its the exact inventory we want.
125
_expected_inv_v5 = """<inventory format="5">
126
<file file_id="bar-20050901064931-73b4b1138abc9cd2" name="bar" revision="mbp@foo-123123" text_sha1="A" text_size="1" />
127
<directory file_id="foo-20050801201819-4139aa4a272f4250" name="subdir" revision="mbp@foo-00" />
128
<file executable="yes" file_id="bar-20050824000535-6bc48cfad47ed134" name="bar" parent_id="foo-20050801201819-4139aa4a272f4250" revision="mbp@foo-00" text_sha1="B" text_size="0" />
133
_expected_inv_v5_root = """<inventory file_id="f<" format="5" revision_id="mother!">
134
<file file_id="bar-20050901064931-73b4b1138abc9cd2" name="bar" parent_id="f<" revision="mbp@foo-123123" text_sha1="A" text_size="1" />
135
<directory file_id="foo-20050801201819-4139aa4a272f4250" name="subdir" parent_id="f<" revision="mbp@foo-00" />
136
<file executable="yes" file_id="bar-20050824000535-6bc48cfad47ed134" name="bar" parent_id="foo-20050801201819-4139aa4a272f4250" revision="mbp@foo-00" text_sha1="B" text_size="0" />
137
<symlink file_id="link-1" name="link" parent_id="foo-20050801201819-4139aa4a272f4250" revision="mbp@foo-00" symlink_target="a" />
141
_expected_inv_v7 = """<inventory format="7" revision_id="rev_outer">
142
<directory file_id="tree-root-321" name="" revision="rev_outer" />
143
<directory file_id="dir-id" name="dir" parent_id="tree-root-321" revision="rev_outer" />
144
<file file_id="file-id" name="file" parent_id="tree-root-321" revision="rev_outer" text_sha1="A" text_size="1" />
145
<symlink file_id="link-id" name="link" parent_id="tree-root-321" revision="rev_outer" symlink_target="a" />
146
<tree-reference file_id="nested-id" name="nested" parent_id="tree-root-321" revision="rev_outer" reference_revision="rev_inner" />
150
_revision_utf8_v5 = """<revision committer="Erik Bågfors <erik@foo.net>"
151
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
152
revision_id="erik@bågfors-02"
153
timestamp="1125907235.212"
155
<message>Include µnicode characters
158
<revision_ref revision_id="erik@bågfors-01"/>
163
_inventory_utf8_v5 = """<inventory file_id="TREé_ROOT" format="5"
164
revision_id="erik@bågfors-02">
165
<file file_id="bår-01"
166
name="bår" parent_id="TREé_ROOT"
167
revision="erik@bågfors-01"/>
168
<directory name="sµbdir"
169
file_id="sµbdir-01"
170
parent_id="TREé_ROOT"
171
revision="erik@bågfors-01"/>
172
<file executable="yes" file_id="bår-02"
173
name="bår" parent_id="sµbdir-01"
174
revision="erik@bågfors-02"/>
179
class TestSerializer(TestCase):
180
"""Test XML serialization"""
181
def test_canned_inventory(self):
182
"""Test unpacked a canned inventory v4 file."""
183
inp = StringIO(_working_inventory_v4)
184
inv = serializer_v4.read_inventory(inp)
185
self.assertEqual(len(inv), 4)
186
self.assert_('bar-20050901064931-73b4b1138abc9cd2' in inv)
188
def test_unpack_revision(self):
189
"""Test unpacking a canned revision v4"""
190
inp = StringIO(_revision_v4)
191
rev = serializer_v4.read_revision(inp)
192
eq = self.assertEqual
194
"Martin Pool <mbp@sourcefrog.net>")
196
"mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9")
197
eq(len(rev.parent_ids), 1)
198
eq(rev.parent_ids[0],
199
"mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
201
def test_unpack_revision_5(self):
202
"""Test unpacking a canned revision v5"""
203
inp = StringIO(_revision_v5)
204
rev = bzrlib.xml5.serializer_v5.read_revision(inp)
205
eq = self.assertEqual
207
"Martin Pool <mbp@sourcefrog.net>")
208
eq(len(rev.parent_ids), 1)
209
eq(rev.timezone, 36000)
210
eq(rev.parent_ids[0],
211
"mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
213
def test_unpack_revision_5_utc(self):
214
inp = StringIO(_revision_v5_utc)
215
rev = bzrlib.xml5.serializer_v5.read_revision(inp)
216
eq = self.assertEqual
218
"Martin Pool <mbp@sourcefrog.net>")
219
eq(len(rev.parent_ids), 1)
221
eq(rev.parent_ids[0],
222
"mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
224
def test_unpack_inventory_5(self):
225
"""Unpack canned new-style inventory"""
226
inp = StringIO(_committed_inv_v5)
227
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
228
eq = self.assertEqual
230
ie = inv['bar-20050824000535-6bc48cfad47ed134']
232
eq(ie.revision, 'mbp@foo-00')
234
eq(inv[ie.parent_id].kind, 'directory')
236
def test_unpack_basis_inventory_5(self):
237
"""Unpack canned new-style inventory"""
238
inp = StringIO(_basis_inv_v5)
239
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
240
eq = self.assertEqual
242
eq(inv.revision_id, 'mbp@sourcefrog.net-20050905063503-43948f59fa127d92')
243
ie = inv['bar-20050824000535-6bc48cfad47ed134']
245
eq(ie.revision, 'mbp@foo-00')
247
eq(inv[ie.parent_id].kind, 'directory')
249
def test_repack_inventory_5(self):
250
inp = StringIO(_committed_inv_v5)
251
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
253
bzrlib.xml5.serializer_v5.write_inventory(inv, outp)
254
self.assertEqualDiff(_expected_inv_v5, outp.getvalue())
255
inv2 = bzrlib.xml5.serializer_v5.read_inventory(StringIO(outp.getvalue()))
256
self.assertEqual(inv, inv2)
258
def assertRoundTrips(self, xml_string):
259
inp = StringIO(xml_string)
260
inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
262
bzrlib.xml5.serializer_v5.write_inventory(inv, outp)
263
self.assertEqualDiff(xml_string, outp.getvalue())
264
lines = bzrlib.xml5.serializer_v5.write_inventory_to_lines(inv)
266
self.assertEqual(outp.readlines(), lines)
267
inv2 = bzrlib.xml5.serializer_v5.read_inventory(StringIO(outp.getvalue()))
268
self.assertEqual(inv, inv2)
270
def tests_serialize_inventory_v5_with_root(self):
271
self.assertRoundTrips(_expected_inv_v5_root)
273
def check_repack_revision(self, txt):
274
"""Check that repacking a revision yields the same information"""
276
rev = bzrlib.xml5.serializer_v5.read_revision(inp)
278
bzrlib.xml5.serializer_v5.write_revision(rev, outp)
279
outfile_contents = outp.getvalue()
280
rev2 = bzrlib.xml5.serializer_v5.read_revision(StringIO(outfile_contents))
281
self.assertEqual(rev, rev2)
283
def test_repack_revision_5(self):
284
"""Round-trip revision to XML v5"""
285
self.check_repack_revision(_revision_v5)
287
def test_repack_revision_5_utc(self):
288
self.check_repack_revision(_revision_v5_utc)
290
def test_pack_revision_5(self):
291
"""Pack revision to XML v5"""
292
# fixed 20051025, revisions should have final newline
293
rev = bzrlib.xml5.serializer_v5.read_revision_from_string(_revision_v5)
295
bzrlib.xml5.serializer_v5.write_revision(rev, outp)
296
outfile_contents = outp.getvalue()
297
self.assertEqual(outfile_contents[-1], '\n')
298
self.assertEqualDiff(outfile_contents, bzrlib.xml5.serializer_v5.write_revision_to_string(rev))
299
self.assertEqualDiff(outfile_contents, _expected_rev_v5)
301
def test_empty_property_value(self):
302
"""Create an empty property value check that it serializes correctly"""
303
s_v5 = bzrlib.xml5.serializer_v5
304
rev = s_v5.read_revision_from_string(_revision_v5)
306
props = {'empty':'', 'one':'one'}
307
rev.properties = props
308
txt = s_v5.write_revision_to_string(rev)
309
new_rev = s_v5.read_revision_from_string(txt)
310
self.assertEqual(props, new_rev.properties)
312
def test_roundtrip_inventory_v7(self):
313
inv = Inventory('tree-root-321', revision_id='rev_outer')
314
inv.add(inventory.TreeReference('nested-id', 'nested', 'tree-root-321',
315
'rev_outer', 'rev_inner'))
316
inv.add(inventory.InventoryFile('file-id', 'file', 'tree-root-321'))
317
inv.add(inventory.InventoryDirectory('dir-id', 'dir',
319
inv.add(inventory.InventoryLink('link-id', 'link', 'tree-root-321'))
320
inv['tree-root-321'].revision = 'rev_outer'
321
inv['dir-id'].revision = 'rev_outer'
322
inv['file-id'].revision = 'rev_outer'
323
inv['file-id'].text_sha1 = 'A'
324
inv['file-id'].text_size = 1
325
inv['link-id'].revision = 'rev_outer'
326
inv['link-id'].symlink_target = 'a'
327
txt = xml7.serializer_v7.write_inventory_to_string(inv)
328
lines = xml7.serializer_v7.write_inventory_to_lines(inv)
329
self.assertEqual(bzrlib.osutils.split_lines(txt), lines)
330
self.assertEqualDiff(_expected_inv_v7, txt)
331
inv2 = xml7.serializer_v7.read_inventory_from_string(txt)
332
self.assertEqual(5, len(inv2))
333
for path, ie in inv.iter_entries():
334
self.assertEqual(ie, inv2[ie.file_id])
336
def test_wrong_format_v7(self):
337
"""Can't accidentally open a file with wrong serializer"""
338
s_v6 = bzrlib.xml6.serializer_v6
339
s_v7 = xml7.serializer_v7
340
self.assertRaises(errors.UnexpectedInventoryFormat,
341
s_v7.read_inventory_from_string, _expected_inv_v5)
342
self.assertRaises(errors.UnexpectedInventoryFormat,
343
s_v6.read_inventory_from_string, _expected_inv_v7)
345
def test_tree_reference(self):
346
s_v5 = bzrlib.xml5.serializer_v5
347
s_v6 = bzrlib.xml6.serializer_v6
348
s_v7 = xml7.serializer_v7
349
inv = Inventory('tree-root-321', revision_id='rev-outer')
350
inv.root.revision = 'root-rev'
351
inv.add(inventory.TreeReference('nested-id', 'nested', 'tree-root-321',
352
'rev-outer', 'rev-inner'))
353
self.assertRaises(errors.UnsupportedInventoryKind,
354
s_v5.write_inventory_to_string, inv)
355
self.assertRaises(errors.UnsupportedInventoryKind,
356
s_v6.write_inventory_to_string, inv)
357
txt = s_v7.write_inventory_to_string(inv)
358
lines = s_v7.write_inventory_to_lines(inv)
359
self.assertEqual(bzrlib.osutils.split_lines(txt), lines)
360
inv2 = s_v7.read_inventory_from_string(txt)
361
self.assertEqual('tree-root-321', inv2['nested-id'].parent_id)
362
self.assertEqual('rev-outer', inv2['nested-id'].revision)
363
self.assertEqual('rev-inner', inv2['nested-id'].reference_revision)
364
self.assertRaises(errors.UnsupportedInventoryKind,
365
s_v6.read_inventory_from_string,
366
txt.replace('format="7"', 'format="6"'))
367
self.assertRaises(errors.UnsupportedInventoryKind,
368
s_v5.read_inventory_from_string,
369
txt.replace('format="7"', 'format="5"'))
371
def test_revision_ids_are_utf8(self):
372
"""Parsed revision_ids should all be utf-8 strings, not unicode."""
373
s_v5 = bzrlib.xml5.serializer_v5
374
rev = s_v5.read_revision_from_string(_revision_utf8_v5)
375
self.assertEqual('erik@b\xc3\xa5gfors-02', rev.revision_id)
376
self.assertIsInstance(rev.revision_id, str)
377
self.assertEqual(['erik@b\xc3\xa5gfors-01'], rev.parent_ids)
378
for parent_id in rev.parent_ids:
379
self.assertIsInstance(parent_id, str)
380
self.assertEqual(u'Include \xb5nicode characters\n', rev.message)
381
self.assertIsInstance(rev.message, unicode)
383
# ie.revision should either be None or a utf-8 revision id
384
inv = s_v5.read_inventory_from_string(_inventory_utf8_v5)
385
rev_id_1 = u'erik@b\xe5gfors-01'.encode('utf8')
386
rev_id_2 = u'erik@b\xe5gfors-02'.encode('utf8')
387
fid_root = u'TRE\xe9_ROOT'.encode('utf8')
388
fid_bar1 = u'b\xe5r-01'.encode('utf8')
389
fid_sub = u's\xb5bdir-01'.encode('utf8')
390
fid_bar2 = u'b\xe5r-02'.encode('utf8')
391
expected = [(u'', fid_root, None, None),
392
(u'b\xe5r', fid_bar1, fid_root, rev_id_1),
393
(u's\xb5bdir', fid_sub, fid_root, rev_id_1),
394
(u's\xb5bdir/b\xe5r', fid_bar2, fid_sub, rev_id_2),
396
self.assertEqual(rev_id_2, inv.revision_id)
397
self.assertIsInstance(inv.revision_id, str)
399
actual = list(inv.iter_entries_by_dir())
400
for ((exp_path, exp_file_id, exp_parent_id, exp_rev_id),
401
(act_path, act_ie)) in zip(expected, actual):
402
self.assertEqual(exp_path, act_path)
403
self.assertIsInstance(act_path, unicode)
404
self.assertEqual(exp_file_id, act_ie.file_id)
405
self.assertIsInstance(act_ie.file_id, str)
406
self.assertEqual(exp_parent_id, act_ie.parent_id)
407
if exp_parent_id is not None:
408
self.assertIsInstance(act_ie.parent_id, str)
409
self.assertEqual(exp_rev_id, act_ie.revision)
410
if exp_rev_id is not None:
411
self.assertIsInstance(act_ie.revision, str)
413
self.assertEqual(len(expected), len(actual))
415
def test_registry(self):
416
self.assertIs(serializer_v4,
417
xml_serializer.format_registry.get('4'))
418
self.assertIs(bzrlib.xml5.serializer_v5,
419
xml_serializer.format_registry.get('5'))
420
self.assertIs(bzrlib.xml6.serializer_v6,
421
xml_serializer.format_registry.get('6'))
422
self.assertIs(bzrlib.xml7.serializer_v7,
423
xml_serializer.format_registry.get('7'))
426
class TestEncodeAndEscape(TestCase):
427
"""Whitebox testing of the _encode_and_escape function."""
430
# Keep the cache clear before and after the test
431
bzrlib.xml5._ensure_utf8_re()
432
bzrlib.xml5._clear_cache()
433
self.addCleanup(bzrlib.xml5._clear_cache)
435
def test_simple_ascii(self):
436
# _encode_and_escape always appends a final ", because these parameters
437
# are being used in xml attributes, and by returning it now, we have to
438
# do fewer string operations later.
439
val = bzrlib.xml5._encode_and_escape('foo bar')
440
self.assertEqual('foo bar"', val)
441
# The second time should be cached
442
val2 = bzrlib.xml5._encode_and_escape('foo bar')
443
self.assertIs(val2, val)
445
def test_ascii_with_xml(self):
446
self.assertEqual('&'"<>"',
447
bzrlib.xml5._encode_and_escape('&\'"<>'))
449
def test_utf8_with_xml(self):
451
utf8_str = '\xc2\xb5\xc3\xa5&\xd8\xac'
452
self.assertEqual('µå&ج"',
453
bzrlib.xml5._encode_and_escape(utf8_str))
455
def test_unicode(self):
456
uni_str = u'\xb5\xe5&\u062c'
457
self.assertEqual('µå&ج"',
458
bzrlib.xml5._encode_and_escape(uni_str))