~bzr-pqm/bzr/bzr.dev

5557.1.15 by John Arbash Meinel
Merge bzr.dev 5597 to resolve NEWS, aka bzr-2.3.txt
1
# Copyright (C) 2005-2011 Canonical Ltd
1887.1.1 by Adeodato Simó
Do not separate paragraphs in the copyright statement with blank lines,
2
#
1181 by Martin Pool
- add test for deserialization from a canned XML inventory
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.
1887.1.1 by Adeodato Simó
Do not separate paragraphs in the copyright statement with blank lines,
7
#
1181 by Martin Pool
- add test for deserialization from a canned XML inventory
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.
1887.1.1 by Adeodato Simó
Do not separate paragraphs in the copyright statement with blank lines,
12
#
1181 by Martin Pool
- add test for deserialization from a canned XML inventory
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
1181 by Martin Pool
- add test for deserialization from a canned XML inventory
16
17
from cStringIO import StringIO
18
2100.3.1 by Aaron Bentley
Start roundtripping tree-reference entries
19
from bzrlib import (
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
20
    errors,
4849.4.2 by John Arbash Meinel
Change from being a per-serializer attribute to being a per-repo attribute.
21
    fifo_cache,
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
22
    inventory,
4476.3.27 by Andrew Bennetts
Trivial improvement to test coverage in test_xml.
23
    xml6,
2100.3.1 by Aaron Bentley
Start roundtripping tree-reference entries
24
    xml7,
3311.3.2 by Aaron Bentley
Implement version numbers for format 8
25
    xml8,
2100.3.1 by Aaron Bentley
Start roundtripping tree-reference entries
26
    )
1185.31.25 by John Arbash Meinel
Renamed all of the tests from selftest/foo.py to tests/test_foo.py
27
from bzrlib.tests import TestCase
5579.3.1 by Jelmer Vernooij
Remove unused imports.
28
from bzrlib.inventory import Inventory
1732.1.26 by John Arbash Meinel
Switch to using bzrlib.xml5.serializer_v5 so that a plugin can override it if we want
29
import bzrlib.xml5
1181 by Martin Pool
- add test for deserialization from a canned XML inventory
30
1183 by Martin Pool
- implement version 5 xml storage, and tests
31
_revision_v5 = """<revision committer="Martin Pool &lt;mbp@sourcefrog.net&gt;"
32
    inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
33
    revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
2102.4.1 by John Arbash Meinel
Switch to using millisecond resolution in Revision XML
34
    timestamp="1125907235.212"
1183 by Martin Pool
- implement version 5 xml storage, and tests
35
    timezone="36000">
36
<message>- start splitting code for xml (de)serialization away from objects
37
  preparatory to supporting multiple formats by a single library
38
</message>
39
<parents>
40
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92"/>
41
</parents>
42
</revision>
43
"""
44
1913.1.2 by John Arbash Meinel
Add direct tests to xml serializer
45
_revision_v5_utc = """\
46
<revision committer="Martin Pool &lt;mbp@sourcefrog.net&gt;"
47
    inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
48
    revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
2102.4.1 by John Arbash Meinel
Switch to using millisecond resolution in Revision XML
49
    timestamp="1125907235.212"
1913.1.2 by John Arbash Meinel
Add direct tests to xml serializer
50
    timezone="0">
51
<message>- start splitting code for xml (de)serialization away from objects
52
  preparatory to supporting multiple formats by a single library
53
</message>
54
<parents>
55
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92"/>
56
</parents>
57
</revision>
58
"""
59
1183 by Martin Pool
- implement version 5 xml storage, and tests
60
_committed_inv_v5 = """<inventory>
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
61
<file file_id="bar-20050901064931-73b4b1138abc9cd2"
62
      name="bar" parent_id="TREE_ROOT"
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
63
      revision="mbp@foo-123123"
64
      text_sha1="A" text_size="1"/>
1183 by Martin Pool
- implement version 5 xml storage, and tests
65
<directory name="subdir"
66
           file_id="foo-20050801201819-4139aa4a272f4250"
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
67
           parent_id="TREE_ROOT"
1092.2.21 by Robert Collins
convert name_version to revision in inventory entries
68
           revision="mbp@foo-00"/>
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
69
<file executable="yes" file_id="bar-20050824000535-6bc48cfad47ed134"
70
      name="bar" parent_id="foo-20050801201819-4139aa4a272f4250"
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
71
      revision="mbp@foo-00"
72
      text_sha1="B" text_size="0"/>
1183 by Martin Pool
- implement version 5 xml storage, and tests
73
</inventory>
74
"""
75
1638.1.2 by Robert Collins
Change the basis-inventory file to not have the revision-id in the file name.
76
_basis_inv_v5 = """<inventory revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92">
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
77
<file file_id="bar-20050901064931-73b4b1138abc9cd2"
78
      name="bar" parent_id="TREE_ROOT"
1638.1.2 by Robert Collins
Change the basis-inventory file to not have the revision-id in the file name.
79
      revision="mbp@foo-123123"/>
80
<directory name="subdir"
81
           file_id="foo-20050801201819-4139aa4a272f4250"
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
82
           parent_id="TREE_ROOT"
1638.1.2 by Robert Collins
Change the basis-inventory file to not have the revision-id in the file name.
83
           revision="mbp@foo-00"/>
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
84
<file file_id="bar-20050824000535-6bc48cfad47ed134"
85
      name="bar" parent_id="foo-20050801201819-4139aa4a272f4250"
1638.1.2 by Robert Collins
Change the basis-inventory file to not have the revision-id in the file name.
86
      revision="mbp@foo-00"/>
87
</inventory>
88
"""
89
1886.1.1 by John Arbash Meinel
Fix bug #47782,
90
1934.1.3 by John Arbash Meinel
[merge] robert's custom XML serializer, and cleanup for benchmarks and iter_entries() differences
91
# DO NOT REFLOW THIS. Its the exact revision we want.
2102.4.1 by John Arbash Meinel
Switch to using millisecond resolution in Revision XML
92
_expected_rev_v5 = """<revision committer="Martin Pool &lt;mbp@sourcefrog.net&gt;" format="5" inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41" revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9" timestamp="1125907235.212" timezone="36000">
1934.1.3 by John Arbash Meinel
[merge] robert's custom XML serializer, and cleanup for benchmarks and iter_entries() differences
93
<message>- start splitting code for xml (de)serialization away from objects
94
  preparatory to supporting multiple formats by a single library
95
</message>
96
<parents>
97
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92" />
98
</parents>
99
</revision>
100
"""
101
102
103
# DO NOT REFLOW THIS. Its the exact inventory we want.
104
_expected_inv_v5 = """<inventory format="5">
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
105
<file file_id="bar-20050901064931-73b4b1138abc9cd2" name="bar" revision="mbp@foo-123123" text_sha1="A" text_size="1" />
1934.1.3 by John Arbash Meinel
[merge] robert's custom XML serializer, and cleanup for benchmarks and iter_entries() differences
106
<directory file_id="foo-20050801201819-4139aa4a272f4250" name="subdir" revision="mbp@foo-00" />
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
107
<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" />
1934.1.3 by John Arbash Meinel
[merge] robert's custom XML serializer, and cleanup for benchmarks and iter_entries() differences
108
</inventory>
109
"""
110
111
112
_expected_inv_v5_root = """<inventory file_id="f&lt;" format="5" revision_id="mother!">
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
113
<file file_id="bar-20050901064931-73b4b1138abc9cd2" name="bar" parent_id="f&lt;" revision="mbp@foo-123123" text_sha1="A" text_size="1" />
1934.1.3 by John Arbash Meinel
[merge] robert's custom XML serializer, and cleanup for benchmarks and iter_entries() differences
114
<directory file_id="foo-20050801201819-4139aa4a272f4250" name="subdir" parent_id="f&lt;" revision="mbp@foo-00" />
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
115
<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" />
116
<symlink file_id="link-1" name="link" parent_id="foo-20050801201819-4139aa4a272f4250" revision="mbp@foo-00" symlink_target="a" />
1934.1.3 by John Arbash Meinel
[merge] robert's custom XML serializer, and cleanup for benchmarks and iter_entries() differences
117
</inventory>
118
"""
119
4476.3.27 by Andrew Bennetts
Trivial improvement to test coverage in test_xml.
120
_expected_inv_v6 = """<inventory format="6" revision_id="rev_outer">
121
<directory file_id="tree-root-321" name="" revision="rev_outer" />
122
<directory file_id="dir-id" name="dir" parent_id="tree-root-321" revision="rev_outer" />
123
<file file_id="file-id" name="file" parent_id="tree-root-321" revision="rev_outer" text_sha1="A" text_size="1" />
124
<symlink file_id="link-id" name="link" parent_id="tree-root-321" revision="rev_outer" symlink_target="a" />
125
</inventory>
126
"""
127
2100.3.2 by Aaron Bentley
Add tests for format 7, enforce number
128
_expected_inv_v7 = """<inventory format="7" revision_id="rev_outer">
129
<directory file_id="tree-root-321" name="" revision="rev_outer" />
130
<directory file_id="dir-id" name="dir" parent_id="tree-root-321" revision="rev_outer" />
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
131
<file file_id="file-id" name="file" parent_id="tree-root-321" revision="rev_outer" text_sha1="A" text_size="1" />
132
<symlink file_id="link-id" name="link" parent_id="tree-root-321" revision="rev_outer" symlink_target="a" />
2100.3.2 by Aaron Bentley
Add tests for format 7, enforce number
133
<tree-reference file_id="nested-id" name="nested" parent_id="tree-root-321" revision="rev_outer" reference_revision="rev_inner" />
134
</inventory>
135
"""
1934.1.3 by John Arbash Meinel
[merge] robert's custom XML serializer, and cleanup for benchmarks and iter_entries() differences
136
3311.3.2 by Aaron Bentley
Implement version numbers for format 8
137
_expected_rev_v8 = """<revision committer="Martin Pool &lt;mbp@sourcefrog.net&gt;" format="8" inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41" revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9" timestamp="1125907235.212" timezone="36000">
138
<message>- start splitting code for xml (de)serialization away from objects
139
  preparatory to supporting multiple formats by a single library
140
</message>
141
<parents>
142
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92" />
143
</parents>
144
</revision>
145
"""
146
147
_expected_inv_v8 = """<inventory format="8" revision_id="rev_outer">
148
<directory file_id="tree-root-321" name="" revision="rev_outer" />
149
<directory file_id="dir-id" name="dir" parent_id="tree-root-321" revision="rev_outer" />
150
<file file_id="file-id" name="file" parent_id="tree-root-321" revision="rev_outer" text_sha1="A" text_size="1" />
151
<symlink file_id="link-id" name="link" parent_id="tree-root-321" revision="rev_outer" symlink_target="a" />
152
</inventory>
153
"""
154
2294.1.5 by John Arbash Meinel
Fix Inventory.iter_entries_by_dir to return Unicode paths,
155
_revision_utf8_v5 = """<revision committer="Erik B&#229;gfors &lt;erik@foo.net&gt;"
156
    inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
157
    revision_id="erik@b&#229;gfors-02"
158
    timestamp="1125907235.212"
159
    timezone="36000">
160
<message>Include &#181;nicode characters
161
</message>
162
<parents>
163
<revision_ref revision_id="erik@b&#229;gfors-01"/>
164
</parents>
165
</revision>
166
"""
167
168
_inventory_utf8_v5 = """<inventory file_id="TRE&#233;_ROOT" format="5"
169
                                   revision_id="erik@b&#229;gfors-02">
170
<file file_id="b&#229;r-01"
2294.1.9 by John Arbash Meinel
Minor performance improvement, use None as signal rather than ROOT_ID
171
      name="b&#229;r" parent_id="TRE&#233;_ROOT"
2294.1.5 by John Arbash Meinel
Fix Inventory.iter_entries_by_dir to return Unicode paths,
172
      revision="erik@b&#229;gfors-01"/>
173
<directory name="s&#181;bdir"
174
           file_id="s&#181;bdir-01"
2294.1.9 by John Arbash Meinel
Minor performance improvement, use None as signal rather than ROOT_ID
175
           parent_id="TRE&#233;_ROOT"
2294.1.5 by John Arbash Meinel
Fix Inventory.iter_entries_by_dir to return Unicode paths,
176
           revision="erik@b&#229;gfors-01"/>
177
<file executable="yes" file_id="b&#229;r-02"
178
      name="b&#229;r" parent_id="s&#181;bdir-01"
179
      revision="erik@b&#229;gfors-02"/>
180
</inventory>
181
"""
182
2917.2.1 by John Arbash Meinel
Fix bug #152360. The xml5 serializer should be using
183
# Before revision_id was always stored as an attribute
184
_inventory_v5a = """<inventory format="5">
185
</inventory>
186
"""
187
188
# Before revision_id was always stored as an attribute
189
_inventory_v5b = """<inventory format="5" revision_id="a-rev-id">
190
</inventory>
191
"""
192
2294.1.5 by John Arbash Meinel
Fix Inventory.iter_entries_by_dir to return Unicode paths,
193
1181 by Martin Pool
- add test for deserialization from a canned XML inventory
194
class TestSerializer(TestCase):
195
    """Test XML serialization"""
2889.1.1 by Robert Collins
* The class ``bzrlib.repofmt.knitrepo.KnitRepository3`` has been folded into
196
1183 by Martin Pool
- implement version 5 xml storage, and tests
197
    def test_unpack_revision_5(self):
198
        """Test unpacking a canned revision v5"""
199
        inp = StringIO(_revision_v5)
1732.1.26 by John Arbash Meinel
Switch to using bzrlib.xml5.serializer_v5 so that a plugin can override it if we want
200
        rev = bzrlib.xml5.serializer_v5.read_revision(inp)
1183 by Martin Pool
- implement version 5 xml storage, and tests
201
        eq = self.assertEqual
202
        eq(rev.committer,
203
           "Martin Pool <mbp@sourcefrog.net>")
1313 by Martin Pool
- rename to Revision.parent_ids to avoid confusion with old usage
204
        eq(len(rev.parent_ids), 1)
1183 by Martin Pool
- implement version 5 xml storage, and tests
205
        eq(rev.timezone, 36000)
1313 by Martin Pool
- rename to Revision.parent_ids to avoid confusion with old usage
206
        eq(rev.parent_ids[0],
1183 by Martin Pool
- implement version 5 xml storage, and tests
207
           "mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
208
1913.1.2 by John Arbash Meinel
Add direct tests to xml serializer
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
213
        eq(rev.committer,
214
           "Martin Pool <mbp@sourcefrog.net>")
215
        eq(len(rev.parent_ids), 1)
216
        eq(rev.timezone, 0)
217
        eq(rev.parent_ids[0],
218
           "mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
219
1183 by Martin Pool
- implement version 5 xml storage, and tests
220
    def test_unpack_inventory_5(self):
221
        """Unpack canned new-style inventory"""
222
        inp = StringIO(_committed_inv_v5)
1732.1.26 by John Arbash Meinel
Switch to using bzrlib.xml5.serializer_v5 so that a plugin can override it if we want
223
        inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
1183 by Martin Pool
- implement version 5 xml storage, and tests
224
        eq = self.assertEqual
225
        eq(len(inv), 4)
226
        ie = inv['bar-20050824000535-6bc48cfad47ed134']
227
        eq(ie.kind, 'file')
1092.2.21 by Robert Collins
convert name_version to revision in inventory entries
228
        eq(ie.revision, 'mbp@foo-00')
1183 by Martin Pool
- implement version 5 xml storage, and tests
229
        eq(ie.name, 'bar')
230
        eq(inv[ie.parent_id].kind, 'directory')
1184 by Martin Pool
- fix v5 packing of inventory entries
231
1638.1.2 by Robert Collins
Change the basis-inventory file to not have the revision-id in the file name.
232
    def test_unpack_basis_inventory_5(self):
233
        """Unpack canned new-style inventory"""
234
        inp = StringIO(_basis_inv_v5)
1732.1.26 by John Arbash Meinel
Switch to using bzrlib.xml5.serializer_v5 so that a plugin can override it if we want
235
        inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
1638.1.2 by Robert Collins
Change the basis-inventory file to not have the revision-id in the file name.
236
        eq = self.assertEqual
237
        eq(len(inv), 4)
238
        eq(inv.revision_id, 'mbp@sourcefrog.net-20050905063503-43948f59fa127d92')
239
        ie = inv['bar-20050824000535-6bc48cfad47ed134']
240
        eq(ie.kind, 'file')
241
        eq(ie.revision, 'mbp@foo-00')
242
        eq(ie.name, 'bar')
243
        eq(inv[ie.parent_id].kind, 'directory')
244
2917.2.1 by John Arbash Meinel
Fix bug #152360. The xml5 serializer should be using
245
    def test_unpack_inventory_5a(self):
246
        inv = bzrlib.xml5.serializer_v5.read_inventory_from_string(
247
                _inventory_v5a, revision_id='test-rev-id')
248
        self.assertEqual('test-rev-id', inv.root.revision)
249
4849.4.2 by John Arbash Meinel
Change from being a per-serializer attribute to being a per-repo attribute.
250
    def test_unpack_inventory_5a_cache_and_copy(self):
251
        # Passing an entry_cache should get populated with the objects
252
        # But the returned objects should be copies if return_from_cache is
253
        # False
254
        entry_cache = fifo_cache.FIFOCache()
255
        inv = bzrlib.xml5.serializer_v5.read_inventory_from_string(
256
            _inventory_v5a, revision_id='test-rev-id',
257
            entry_cache=entry_cache, return_from_cache=False)
258
        for entry in inv.iter_just_entries():
259
            key = (entry.file_id, entry.revision)
260
            if entry.file_id is inv.root.file_id:
261
                # The root id is inferred for xml v5
262
                self.assertFalse(key in entry_cache)
263
            else:
264
                self.assertIsNot(entry, entry_cache[key])
265
266
    def test_unpack_inventory_5a_cache_no_copy(self):
267
        # Passing an entry_cache should get populated with the objects
268
        # The returned objects should be exact if return_from_cache is
269
        # True
270
        entry_cache = fifo_cache.FIFOCache()
271
        inv = bzrlib.xml5.serializer_v5.read_inventory_from_string(
272
            _inventory_v5a, revision_id='test-rev-id',
273
            entry_cache=entry_cache, return_from_cache=True)
274
        for entry in inv.iter_just_entries():
275
            key = (entry.file_id, entry.revision)
276
            if entry.file_id is inv.root.file_id:
277
                # The root id is inferred for xml v5
278
                self.assertFalse(key in entry_cache)
279
            else:
280
                self.assertIs(entry, entry_cache[key])
281
2917.2.1 by John Arbash Meinel
Fix bug #152360. The xml5 serializer should be using
282
    def test_unpack_inventory_5b(self):
283
        inv = bzrlib.xml5.serializer_v5.read_inventory_from_string(
284
                _inventory_v5b, revision_id='test-rev-id')
285
        self.assertEqual('a-rev-id', inv.root.revision)
286
1184 by Martin Pool
- fix v5 packing of inventory entries
287
    def test_repack_inventory_5(self):
288
        inp = StringIO(_committed_inv_v5)
1732.1.26 by John Arbash Meinel
Switch to using bzrlib.xml5.serializer_v5 so that a plugin can override it if we want
289
        inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
1184 by Martin Pool
- fix v5 packing of inventory entries
290
        outp = StringIO()
1732.1.26 by John Arbash Meinel
Switch to using bzrlib.xml5.serializer_v5 so that a plugin can override it if we want
291
        bzrlib.xml5.serializer_v5.write_inventory(inv, outp)
1934.1.3 by John Arbash Meinel
[merge] robert's custom XML serializer, and cleanup for benchmarks and iter_entries() differences
292
        self.assertEqualDiff(_expected_inv_v5, outp.getvalue())
293
        inv2 = bzrlib.xml5.serializer_v5.read_inventory(StringIO(outp.getvalue()))
294
        self.assertEqual(inv, inv2)
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
295
1934.1.3 by John Arbash Meinel
[merge] robert's custom XML serializer, and cleanup for benchmarks and iter_entries() differences
296
    def assertRoundTrips(self, xml_string):
297
        inp = StringIO(xml_string)
298
        inv = bzrlib.xml5.serializer_v5.read_inventory(inp)
299
        outp = StringIO()
300
        bzrlib.xml5.serializer_v5.write_inventory(inv, outp)
301
        self.assertEqualDiff(xml_string, outp.getvalue())
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
302
        lines = bzrlib.xml5.serializer_v5.write_inventory_to_lines(inv)
303
        outp.seek(0)
304
        self.assertEqual(outp.readlines(), lines)
1934.1.3 by John Arbash Meinel
[merge] robert's custom XML serializer, and cleanup for benchmarks and iter_entries() differences
305
        inv2 = bzrlib.xml5.serializer_v5.read_inventory(StringIO(outp.getvalue()))
306
        self.assertEqual(inv, inv2)
307
308
    def tests_serialize_inventory_v5_with_root(self):
309
        self.assertRoundTrips(_expected_inv_v5_root)
1185 by Martin Pool
- add xml round-trip test for revisions
310
1913.1.2 by John Arbash Meinel
Add direct tests to xml serializer
311
    def check_repack_revision(self, txt):
312
        """Check that repacking a revision yields the same information"""
313
        inp = StringIO(txt)
314
        rev = bzrlib.xml5.serializer_v5.read_revision(inp)
315
        outp = StringIO()
316
        bzrlib.xml5.serializer_v5.write_revision(rev, outp)
317
        outfile_contents = outp.getvalue()
318
        rev2 = bzrlib.xml5.serializer_v5.read_revision(StringIO(outfile_contents))
319
        self.assertEqual(rev, rev2)
320
1185 by Martin Pool
- add xml round-trip test for revisions
321
    def test_repack_revision_5(self):
1185.16.123 by Martin Pool
Fix syntax of serializer_v5.pack_revision_to_string
322
        """Round-trip revision to XML v5"""
1913.1.2 by John Arbash Meinel
Add direct tests to xml serializer
323
        self.check_repack_revision(_revision_v5)
324
325
    def test_repack_revision_5_utc(self):
326
        self.check_repack_revision(_revision_v5_utc)
1185 by Martin Pool
- add xml round-trip test for revisions
327
1185.16.123 by Martin Pool
Fix syntax of serializer_v5.pack_revision_to_string
328
    def test_pack_revision_5(self):
329
        """Pack revision to XML v5"""
330
        # fixed 20051025, revisions should have final newline
1732.1.26 by John Arbash Meinel
Switch to using bzrlib.xml5.serializer_v5 so that a plugin can override it if we want
331
        rev = bzrlib.xml5.serializer_v5.read_revision_from_string(_revision_v5)
1185.16.123 by Martin Pool
Fix syntax of serializer_v5.pack_revision_to_string
332
        outp = StringIO()
1732.1.26 by John Arbash Meinel
Switch to using bzrlib.xml5.serializer_v5 so that a plugin can override it if we want
333
        bzrlib.xml5.serializer_v5.write_revision(rev, outp)
1185.16.123 by Martin Pool
Fix syntax of serializer_v5.pack_revision_to_string
334
        outfile_contents = outp.getvalue()
335
        self.assertEqual(outfile_contents[-1], '\n')
1732.1.26 by John Arbash Meinel
Switch to using bzrlib.xml5.serializer_v5 so that a plugin can override it if we want
336
        self.assertEqualDiff(outfile_contents, bzrlib.xml5.serializer_v5.write_revision_to_string(rev))
1934.1.3 by John Arbash Meinel
[merge] robert's custom XML serializer, and cleanup for benchmarks and iter_entries() differences
337
        self.assertEqualDiff(outfile_contents, _expected_rev_v5)
1886.1.1 by John Arbash Meinel
Fix bug #47782,
338
339
    def test_empty_property_value(self):
340
        """Create an empty property value check that it serializes correctly"""
341
        s_v5 = bzrlib.xml5.serializer_v5
342
        rev = s_v5.read_revision_from_string(_revision_v5)
343
        outp = StringIO()
344
        props = {'empty':'', 'one':'one'}
345
        rev.properties = props
346
        txt = s_v5.write_revision_to_string(rev)
347
        new_rev = s_v5.read_revision_from_string(txt)
348
        self.assertEqual(props, new_rev.properties)
2100.3.1 by Aaron Bentley
Start roundtripping tree-reference entries
349
3311.3.2 by Aaron Bentley
Implement version numbers for format 8
350
    def get_sample_inventory(self):
2100.3.2 by Aaron Bentley
Add tests for format 7, enforce number
351
        inv = Inventory('tree-root-321', revision_id='rev_outer')
352
        inv.add(inventory.InventoryFile('file-id', 'file', 'tree-root-321'))
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
353
        inv.add(inventory.InventoryDirectory('dir-id', 'dir',
2100.3.2 by Aaron Bentley
Add tests for format 7, enforce number
354
                                             'tree-root-321'))
355
        inv.add(inventory.InventoryLink('link-id', 'link', 'tree-root-321'))
356
        inv['tree-root-321'].revision = 'rev_outer'
357
        inv['dir-id'].revision = 'rev_outer'
358
        inv['file-id'].revision = 'rev_outer'
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
359
        inv['file-id'].text_sha1 = 'A'
360
        inv['file-id'].text_size = 1
2100.3.2 by Aaron Bentley
Add tests for format 7, enforce number
361
        inv['link-id'].revision = 'rev_outer'
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
362
        inv['link-id'].symlink_target = 'a'
3311.3.2 by Aaron Bentley
Implement version numbers for format 8
363
        return inv
364
365
    def test_roundtrip_inventory_v7(self):
366
        inv = self.get_sample_inventory()
367
        inv.add(inventory.TreeReference('nested-id', 'nested', 'tree-root-321',
368
                                        'rev_outer', 'rev_inner'))
2100.3.2 by Aaron Bentley
Add tests for format 7, enforce number
369
        txt = xml7.serializer_v7.write_inventory_to_string(inv)
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
370
        lines = xml7.serializer_v7.write_inventory_to_lines(inv)
371
        self.assertEqual(bzrlib.osutils.split_lines(txt), lines)
2100.3.2 by Aaron Bentley
Add tests for format 7, enforce number
372
        self.assertEqualDiff(_expected_inv_v7, txt)
373
        inv2 = xml7.serializer_v7.read_inventory_from_string(txt)
374
        self.assertEqual(5, len(inv2))
375
        for path, ie in inv.iter_entries():
376
            self.assertEqual(ie, inv2[ie.file_id])
377
4476.3.27 by Andrew Bennetts
Trivial improvement to test coverage in test_xml.
378
    def test_roundtrip_inventory_v6(self):
379
        inv = self.get_sample_inventory()
380
        txt = xml6.serializer_v6.write_inventory_to_string(inv)
381
        lines = xml6.serializer_v6.write_inventory_to_lines(inv)
382
        self.assertEqual(bzrlib.osutils.split_lines(txt), lines)
383
        self.assertEqualDiff(_expected_inv_v6, txt)
384
        inv2 = xml6.serializer_v6.read_inventory_from_string(txt)
385
        self.assertEqual(4, len(inv2))
386
        for path, ie in inv.iter_entries():
387
            self.assertEqual(ie, inv2[ie.file_id])
388
2100.3.2 by Aaron Bentley
Add tests for format 7, enforce number
389
    def test_wrong_format_v7(self):
390
        """Can't accidentally open a file with wrong serializer"""
391
        s_v6 = bzrlib.xml6.serializer_v6
392
        s_v7 = xml7.serializer_v7
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
393
        self.assertRaises(errors.UnexpectedInventoryFormat,
2100.3.2 by Aaron Bentley
Add tests for format 7, enforce number
394
                          s_v7.read_inventory_from_string, _expected_inv_v5)
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
395
        self.assertRaises(errors.UnexpectedInventoryFormat,
2100.3.2 by Aaron Bentley
Add tests for format 7, enforce number
396
                          s_v6.read_inventory_from_string, _expected_inv_v7)
397
2100.3.1 by Aaron Bentley
Start roundtripping tree-reference entries
398
    def test_tree_reference(self):
399
        s_v5 = bzrlib.xml5.serializer_v5
400
        s_v6 = bzrlib.xml6.serializer_v6
401
        s_v7 = xml7.serializer_v7
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
402
        inv = Inventory('tree-root-321', revision_id='rev-outer')
403
        inv.root.revision = 'root-rev'
2100.3.1 by Aaron Bentley
Start roundtripping tree-reference entries
404
        inv.add(inventory.TreeReference('nested-id', 'nested', 'tree-root-321',
405
                                        'rev-outer', 'rev-inner'))
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
406
        self.assertRaises(errors.UnsupportedInventoryKind,
2100.3.1 by Aaron Bentley
Start roundtripping tree-reference entries
407
                          s_v5.write_inventory_to_string, inv)
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
408
        self.assertRaises(errors.UnsupportedInventoryKind,
2100.3.1 by Aaron Bentley
Start roundtripping tree-reference entries
409
                          s_v6.write_inventory_to_string, inv)
410
        txt = s_v7.write_inventory_to_string(inv)
2817.2.1 by Robert Collins
* Inventory serialisation no longer double-sha's the content.
411
        lines = s_v7.write_inventory_to_lines(inv)
412
        self.assertEqual(bzrlib.osutils.split_lines(txt), lines)
2100.3.1 by Aaron Bentley
Start roundtripping tree-reference entries
413
        inv2 = s_v7.read_inventory_from_string(txt)
414
        self.assertEqual('tree-root-321', inv2['nested-id'].parent_id)
415
        self.assertEqual('rev-outer', inv2['nested-id'].revision)
416
        self.assertEqual('rev-inner', inv2['nested-id'].reference_revision)
2255.6.1 by Aaron Bentley
Merge from by-reference-trees
417
3311.3.2 by Aaron Bentley
Implement version numbers for format 8
418
    def test_roundtrip_inventory_v8(self):
419
        inv = self.get_sample_inventory()
420
        txt = xml8.serializer_v8.write_inventory_to_string(inv)
421
        inv2 = xml8.serializer_v8.read_inventory_from_string(txt)
422
        self.assertEqual(4, len(inv2))
423
        for path, ie in inv.iter_entries():
424
            self.assertEqual(ie, inv2[ie.file_id])
425
426
    def test_inventory_text_v8(self):
427
        inv = self.get_sample_inventory()
428
        txt = xml8.serializer_v8.write_inventory_to_string(inv)
429
        lines = xml8.serializer_v8.write_inventory_to_lines(inv)
430
        self.assertEqual(bzrlib.osutils.split_lines(txt), lines)
431
        self.assertEqualDiff(_expected_inv_v8, txt)
432
3311.3.3 by Aaron Bentley
Handle format 5 revision
433
    def test_revision_text_v6(self):
434
        """Pack revision to XML v6"""
435
        rev = bzrlib.xml6.serializer_v6.read_revision_from_string(
436
            _expected_rev_v5)
437
        serialized = bzrlib.xml6.serializer_v6.write_revision_to_string(rev)
438
        self.assertEqualDiff(serialized, _expected_rev_v5)
439
440
    def test_revision_text_v7(self):
441
        """Pack revision to XML v7"""
442
        rev = bzrlib.xml7.serializer_v7.read_revision_from_string(
443
            _expected_rev_v5)
444
        serialized = bzrlib.xml7.serializer_v7.write_revision_to_string(rev)
445
        self.assertEqualDiff(serialized, _expected_rev_v5)
446
3311.3.2 by Aaron Bentley
Implement version numbers for format 8
447
    def test_revision_text_v8(self):
448
        """Pack revision to XML v8"""
449
        rev = bzrlib.xml8.serializer_v8.read_revision_from_string(
450
            _expected_rev_v8)
451
        serialized = bzrlib.xml8.serializer_v8.write_revision_to_string(rev)
452
        self.assertEqualDiff(serialized, _expected_rev_v8)
453
2249.5.4 by John Arbash Meinel
When reading XML, always return utf-8 revision ids.
454
    def test_revision_ids_are_utf8(self):
455
        """Parsed revision_ids should all be utf-8 strings, not unicode."""
456
        s_v5 = bzrlib.xml5.serializer_v5
2294.1.5 by John Arbash Meinel
Fix Inventory.iter_entries_by_dir to return Unicode paths,
457
        rev = s_v5.read_revision_from_string(_revision_utf8_v5)
458
        self.assertEqual('erik@b\xc3\xa5gfors-02', rev.revision_id)
2249.5.4 by John Arbash Meinel
When reading XML, always return utf-8 revision ids.
459
        self.assertIsInstance(rev.revision_id, str)
2294.1.5 by John Arbash Meinel
Fix Inventory.iter_entries_by_dir to return Unicode paths,
460
        self.assertEqual(['erik@b\xc3\xa5gfors-01'], rev.parent_ids)
2249.5.4 by John Arbash Meinel
When reading XML, always return utf-8 revision ids.
461
        for parent_id in rev.parent_ids:
462
            self.assertIsInstance(parent_id, str)
2294.1.5 by John Arbash Meinel
Fix Inventory.iter_entries_by_dir to return Unicode paths,
463
        self.assertEqual(u'Include \xb5nicode characters\n', rev.message)
464
        self.assertIsInstance(rev.message, unicode)
2249.5.4 by John Arbash Meinel
When reading XML, always return utf-8 revision ids.
465
466
        # ie.revision should either be None or a utf-8 revision id
2294.1.5 by John Arbash Meinel
Fix Inventory.iter_entries_by_dir to return Unicode paths,
467
        inv = s_v5.read_inventory_from_string(_inventory_utf8_v5)
468
        rev_id_1 = u'erik@b\xe5gfors-01'.encode('utf8')
469
        rev_id_2 = u'erik@b\xe5gfors-02'.encode('utf8')
2294.1.10 by John Arbash Meinel
Switch all apis over to utf8 file ids. All tests pass
470
        fid_root = u'TRE\xe9_ROOT'.encode('utf8')
471
        fid_bar1 = u'b\xe5r-01'.encode('utf8')
472
        fid_sub = u's\xb5bdir-01'.encode('utf8')
473
        fid_bar2 = u'b\xe5r-02'.encode('utf8')
2889.1.1 by Robert Collins
* The class ``bzrlib.repofmt.knitrepo.KnitRepository3`` has been folded into
474
        expected = [(u'', fid_root, None, rev_id_2),
2294.1.6 by John Arbash Meinel
Include parent_id checks in the xml serializer tests.
475
                    (u'b\xe5r', fid_bar1, fid_root, rev_id_1),
476
                    (u's\xb5bdir', fid_sub, fid_root, rev_id_1),
477
                    (u's\xb5bdir/b\xe5r', fid_bar2, fid_sub, rev_id_2),
2294.1.5 by John Arbash Meinel
Fix Inventory.iter_entries_by_dir to return Unicode paths,
478
                   ]
479
        self.assertEqual(rev_id_2, inv.revision_id)
480
        self.assertIsInstance(inv.revision_id, str)
481
482
        actual = list(inv.iter_entries_by_dir())
2294.1.6 by John Arbash Meinel
Include parent_id checks in the xml serializer tests.
483
        for ((exp_path, exp_file_id, exp_parent_id, exp_rev_id),
484
             (act_path, act_ie)) in zip(expected, actual):
2294.1.5 by John Arbash Meinel
Fix Inventory.iter_entries_by_dir to return Unicode paths,
485
            self.assertEqual(exp_path, act_path)
486
            self.assertIsInstance(act_path, unicode)
487
            self.assertEqual(exp_file_id, act_ie.file_id)
2294.1.10 by John Arbash Meinel
Switch all apis over to utf8 file ids. All tests pass
488
            self.assertIsInstance(act_ie.file_id, str)
2294.1.6 by John Arbash Meinel
Include parent_id checks in the xml serializer tests.
489
            self.assertEqual(exp_parent_id, act_ie.parent_id)
490
            if exp_parent_id is not None:
2294.1.10 by John Arbash Meinel
Switch all apis over to utf8 file ids. All tests pass
491
                self.assertIsInstance(act_ie.parent_id, str)
2294.1.5 by John Arbash Meinel
Fix Inventory.iter_entries_by_dir to return Unicode paths,
492
            self.assertEqual(exp_rev_id, act_ie.revision)
493
            if exp_rev_id is not None:
494
                self.assertIsInstance(act_ie.revision, str)
495
496
        self.assertEqual(len(expected), len(actual))
2249.5.10 by John Arbash Meinel
Make sure xml5 can handle unicode or utf8 strings
497
498
499
class TestEncodeAndEscape(TestCase):
500
    """Whitebox testing of the _encode_and_escape function."""
501
502
    def setUp(self):
6552.1.4 by Vincent Ladeuil
Remaining tests matching setup(self) that can be rewritten with super().
503
        super(TestEncodeAndEscape, self).setUp()
2249.5.10 by John Arbash Meinel
Make sure xml5 can handle unicode or utf8 strings
504
        # Keep the cache clear before and after the test
6355.1.8 by Jelmer Vernooij
Fix tests.
505
        bzrlib.xml_serializer._clear_cache()
506
        self.addCleanup(bzrlib.xml_serializer._clear_cache)
2249.5.10 by John Arbash Meinel
Make sure xml5 can handle unicode or utf8 strings
507
508
    def test_simple_ascii(self):
509
        # _encode_and_escape always appends a final ", because these parameters
510
        # are being used in xml attributes, and by returning it now, we have to
511
        # do fewer string operations later.
6355.1.8 by Jelmer Vernooij
Fix tests.
512
        val = bzrlib.xml_serializer.encode_and_escape('foo bar')
2249.5.10 by John Arbash Meinel
Make sure xml5 can handle unicode or utf8 strings
513
        self.assertEqual('foo bar"', val)
514
        # The second time should be cached
6355.1.8 by Jelmer Vernooij
Fix tests.
515
        val2 = bzrlib.xml_serializer.encode_and_escape('foo bar')
2249.5.10 by John Arbash Meinel
Make sure xml5 can handle unicode or utf8 strings
516
        self.assertIs(val2, val)
517
518
    def test_ascii_with_xml(self):
519
        self.assertEqual('&amp;&apos;&quot;&lt;&gt;"',
6355.1.8 by Jelmer Vernooij
Fix tests.
520
                         bzrlib.xml_serializer.encode_and_escape('&\'"<>'))
2249.5.10 by John Arbash Meinel
Make sure xml5 can handle unicode or utf8 strings
521
522
    def test_utf8_with_xml(self):
523
        # u'\xb5\xe5&\u062c'
524
        utf8_str = '\xc2\xb5\xc3\xa5&\xd8\xac'
525
        self.assertEqual('&#181;&#229;&amp;&#1580;"',
6355.1.8 by Jelmer Vernooij
Fix tests.
526
                         bzrlib.xml_serializer.encode_and_escape(utf8_str))
2249.5.10 by John Arbash Meinel
Make sure xml5 can handle unicode or utf8 strings
527
528
    def test_unicode(self):
529
        uni_str = u'\xb5\xe5&\u062c'
530
        self.assertEqual('&#181;&#229;&amp;&#1580;"',
6355.1.8 by Jelmer Vernooij
Fix tests.
531
                         bzrlib.xml_serializer.encode_and_escape(uni_str))
5671.2.4 by Jelmer Vernooij
Fix unescape_xml test.
532
533
534
class TestMisc(TestCase):
535
536
    def test_unescape_xml(self):
537
        """We get some kind of error when malformed entities are passed"""
538
        self.assertRaises(KeyError, bzrlib.xml8._unescape_xml, 'foo&bar;')