1
# Copyright (C) 2005-2011 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
26
from ..sixish import (
29
from ..inventory import Inventory
30
from . import TestCase
33
_revision_v5 = """<revision committer="Martin Pool <mbp@sourcefrog.net>"
34
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
35
revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
36
timestamp="1125907235.212"
38
<message>- start splitting code for xml (de)serialization away from objects
39
preparatory to supporting multiple formats by a single library
42
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92"/>
47
_revision_v5_utc = """\
48
<revision committer="Martin Pool <mbp@sourcefrog.net>"
49
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
50
revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9"
51
timestamp="1125907235.212"
53
<message>- start splitting code for xml (de)serialization away from objects
54
preparatory to supporting multiple formats by a single library
57
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92"/>
62
_committed_inv_v5 = """<inventory>
63
<file file_id="bar-20050901064931-73b4b1138abc9cd2"
64
name="bar" parent_id="TREE_ROOT"
65
revision="mbp@foo-123123"
66
text_sha1="A" text_size="1"/>
67
<directory name="subdir"
68
file_id="foo-20050801201819-4139aa4a272f4250"
70
revision="mbp@foo-00"/>
71
<file executable="yes" file_id="bar-20050824000535-6bc48cfad47ed134"
72
name="bar" parent_id="foo-20050801201819-4139aa4a272f4250"
74
text_sha1="B" text_size="0"/>
78
_basis_inv_v5 = """<inventory revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92">
79
<file file_id="bar-20050901064931-73b4b1138abc9cd2"
80
name="bar" parent_id="TREE_ROOT"
81
revision="mbp@foo-123123"/>
82
<directory name="subdir"
83
file_id="foo-20050801201819-4139aa4a272f4250"
85
revision="mbp@foo-00"/>
86
<file file_id="bar-20050824000535-6bc48cfad47ed134"
87
name="bar" parent_id="foo-20050801201819-4139aa4a272f4250"
88
revision="mbp@foo-00"/>
93
# DO NOT REFLOW THIS. Its the exact revision we want.
94
_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">
95
<message>- start splitting code for xml (de)serialization away from objects
96
preparatory to supporting multiple formats by a single library
99
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92" />
105
# DO NOT REFLOW THIS. Its the exact inventory we want.
106
_expected_inv_v5 = """<inventory format="5">
107
<file file_id="bar-20050901064931-73b4b1138abc9cd2" name="bar" revision="mbp@foo-123123" text_sha1="A" text_size="1" />
108
<directory file_id="foo-20050801201819-4139aa4a272f4250" name="subdir" revision="mbp@foo-00" />
109
<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" />
114
_expected_inv_v5_root = """<inventory file_id="f<" format="5" revision_id="mother!">
115
<file file_id="bar-20050901064931-73b4b1138abc9cd2" name="bar" parent_id="f<" revision="mbp@foo-123123" text_sha1="A" text_size="1" />
116
<directory file_id="foo-20050801201819-4139aa4a272f4250" name="subdir" parent_id="f<" revision="mbp@foo-00" />
117
<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" />
118
<symlink file_id="link-1" name="link" parent_id="foo-20050801201819-4139aa4a272f4250" revision="mbp@foo-00" symlink_target="a" />
122
_expected_inv_v6 = """<inventory format="6" revision_id="rev_outer">
123
<directory file_id="tree-root-321" name="" revision="rev_outer" />
124
<directory file_id="dir-id" name="dir" parent_id="tree-root-321" revision="rev_outer" />
125
<file file_id="file-id" name="file" parent_id="tree-root-321" revision="rev_outer" text_sha1="A" text_size="1" />
126
<symlink file_id="link-id" name="link" parent_id="tree-root-321" revision="rev_outer" symlink_target="a" />
130
_expected_inv_v7 = """<inventory format="7" revision_id="rev_outer">
131
<directory file_id="tree-root-321" name="" revision="rev_outer" />
132
<directory file_id="dir-id" name="dir" parent_id="tree-root-321" revision="rev_outer" />
133
<file file_id="file-id" name="file" parent_id="tree-root-321" revision="rev_outer" text_sha1="A" text_size="1" />
134
<symlink file_id="link-id" name="link" parent_id="tree-root-321" revision="rev_outer" symlink_target="a" />
135
<tree-reference file_id="nested-id" name="nested" parent_id="tree-root-321" revision="rev_outer" reference_revision="rev_inner" />
139
_expected_rev_v8 = """<revision committer="Martin Pool <mbp@sourcefrog.net>" format="8" inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41" revision_id="mbp@sourcefrog.net-20050905080035-e0439293f8b6b9f9" timestamp="1125907235.212" timezone="36000">
140
<message>- start splitting code for xml (de)serialization away from objects
141
preparatory to supporting multiple formats by a single library
144
<revision_ref revision_id="mbp@sourcefrog.net-20050905063503-43948f59fa127d92" />
149
_expected_inv_v8 = """<inventory format="8" revision_id="rev_outer">
150
<directory file_id="tree-root-321" name="" revision="rev_outer" />
151
<directory file_id="dir-id" name="dir" parent_id="tree-root-321" revision="rev_outer" />
152
<file file_id="file-id" name="file" parent_id="tree-root-321" revision="rev_outer" text_sha1="A" text_size="1" />
153
<symlink file_id="link-id" name="link" parent_id="tree-root-321" revision="rev_outer" symlink_target="a" />
157
_revision_utf8_v5 = """<revision committer="Erik Bågfors <erik@foo.net>"
158
inventory_sha1="e79c31c1deb64c163cf660fdedd476dd579ffd41"
159
revision_id="erik@bågfors-02"
160
timestamp="1125907235.212"
162
<message>Include µnicode characters
165
<revision_ref revision_id="erik@bågfors-01"/>
170
_inventory_utf8_v5 = """<inventory file_id="TREé_ROOT" format="5"
171
revision_id="erik@bågfors-02">
172
<file file_id="bår-01"
173
name="bår" parent_id="TREé_ROOT"
174
revision="erik@bågfors-01"/>
175
<directory name="sµbdir"
176
file_id="sµbdir-01"
177
parent_id="TREé_ROOT"
178
revision="erik@bågfors-01"/>
179
<file executable="yes" file_id="bår-02"
180
name="bår" parent_id="sµbdir-01"
181
revision="erik@bågfors-02"/>
185
# Before revision_id was always stored as an attribute
186
_inventory_v5a = """<inventory format="5">
190
# Before revision_id was always stored as an attribute
191
_inventory_v5b = """<inventory format="5" revision_id="a-rev-id">
196
class TestSerializer(TestCase):
197
"""Test XML serialization"""
199
def test_unpack_revision_5(self):
200
"""Test unpacking a canned revision v5"""
201
inp = BytesIO(_revision_v5)
202
rev = breezy.xml5.serializer_v5.read_revision(inp)
203
eq = self.assertEqual
205
"Martin Pool <mbp@sourcefrog.net>")
206
eq(len(rev.parent_ids), 1)
207
eq(rev.timezone, 36000)
208
eq(rev.parent_ids[0],
209
"mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
211
def test_unpack_revision_5_utc(self):
212
inp = BytesIO(_revision_v5_utc)
213
rev = breezy.xml5.serializer_v5.read_revision(inp)
214
eq = self.assertEqual
216
"Martin Pool <mbp@sourcefrog.net>")
217
eq(len(rev.parent_ids), 1)
219
eq(rev.parent_ids[0],
220
"mbp@sourcefrog.net-20050905063503-43948f59fa127d92")
222
def test_unpack_inventory_5(self):
223
"""Unpack canned new-style inventory"""
224
inp = BytesIO(_committed_inv_v5)
225
inv = breezy.xml5.serializer_v5.read_inventory(inp)
226
eq = self.assertEqual
228
ie = inv['bar-20050824000535-6bc48cfad47ed134']
230
eq(ie.revision, 'mbp@foo-00')
232
eq(inv[ie.parent_id].kind, 'directory')
234
def test_unpack_basis_inventory_5(self):
235
"""Unpack canned new-style inventory"""
236
inp = BytesIO(_basis_inv_v5)
237
inv = breezy.xml5.serializer_v5.read_inventory(inp)
238
eq = self.assertEqual
240
eq(inv.revision_id, 'mbp@sourcefrog.net-20050905063503-43948f59fa127d92')
241
ie = inv['bar-20050824000535-6bc48cfad47ed134']
243
eq(ie.revision, 'mbp@foo-00')
245
eq(inv[ie.parent_id].kind, 'directory')
247
def test_unpack_inventory_5a(self):
248
inv = breezy.xml5.serializer_v5.read_inventory_from_string(
249
_inventory_v5a, revision_id='test-rev-id')
250
self.assertEqual('test-rev-id', inv.root.revision)
252
def test_unpack_inventory_5a_cache_and_copy(self):
253
# Passing an entry_cache should get populated with the objects
254
# But the returned objects should be copies if return_from_cache is
256
entry_cache = fifo_cache.FIFOCache()
257
inv = breezy.xml5.serializer_v5.read_inventory_from_string(
258
_inventory_v5a, revision_id='test-rev-id',
259
entry_cache=entry_cache, return_from_cache=False)
260
for entry in inv.iter_just_entries():
261
key = (entry.file_id, entry.revision)
262
if entry.file_id is inv.root.file_id:
263
# The root id is inferred for xml v5
264
self.assertFalse(key in entry_cache)
266
self.assertIsNot(entry, entry_cache[key])
268
def test_unpack_inventory_5a_cache_no_copy(self):
269
# Passing an entry_cache should get populated with the objects
270
# The returned objects should be exact if return_from_cache is
272
entry_cache = fifo_cache.FIFOCache()
273
inv = breezy.xml5.serializer_v5.read_inventory_from_string(
274
_inventory_v5a, revision_id='test-rev-id',
275
entry_cache=entry_cache, return_from_cache=True)
276
for entry in inv.iter_just_entries():
277
key = (entry.file_id, entry.revision)
278
if entry.file_id is inv.root.file_id:
279
# The root id is inferred for xml v5
280
self.assertFalse(key in entry_cache)
282
self.assertIs(entry, entry_cache[key])
284
def test_unpack_inventory_5b(self):
285
inv = breezy.xml5.serializer_v5.read_inventory_from_string(
286
_inventory_v5b, revision_id='test-rev-id')
287
self.assertEqual('a-rev-id', inv.root.revision)
289
def test_repack_inventory_5(self):
290
inp = BytesIO(_committed_inv_v5)
291
inv = breezy.xml5.serializer_v5.read_inventory(inp)
293
breezy.xml5.serializer_v5.write_inventory(inv, outp)
294
self.assertEqualDiff(_expected_inv_v5, outp.getvalue())
295
inv2 = breezy.xml5.serializer_v5.read_inventory(BytesIO(outp.getvalue()))
296
self.assertEqual(inv, inv2)
298
def assertRoundTrips(self, xml_string):
299
inp = BytesIO(xml_string)
300
inv = breezy.xml5.serializer_v5.read_inventory(inp)
302
breezy.xml5.serializer_v5.write_inventory(inv, outp)
303
self.assertEqualDiff(xml_string, outp.getvalue())
304
lines = breezy.xml5.serializer_v5.write_inventory_to_lines(inv)
306
self.assertEqual(outp.readlines(), lines)
307
inv2 = breezy.xml5.serializer_v5.read_inventory(BytesIO(outp.getvalue()))
308
self.assertEqual(inv, inv2)
310
def tests_serialize_inventory_v5_with_root(self):
311
self.assertRoundTrips(_expected_inv_v5_root)
313
def check_repack_revision(self, txt):
314
"""Check that repacking a revision yields the same information"""
316
rev = breezy.xml5.serializer_v5.read_revision(inp)
318
breezy.xml5.serializer_v5.write_revision(rev, outp)
319
outfile_contents = outp.getvalue()
320
rev2 = breezy.xml5.serializer_v5.read_revision(BytesIO(outfile_contents))
321
self.assertEqual(rev, rev2)
323
def test_repack_revision_5(self):
324
"""Round-trip revision to XML v5"""
325
self.check_repack_revision(_revision_v5)
327
def test_repack_revision_5_utc(self):
328
self.check_repack_revision(_revision_v5_utc)
330
def test_pack_revision_5(self):
331
"""Pack revision to XML v5"""
332
# fixed 20051025, revisions should have final newline
333
rev = breezy.xml5.serializer_v5.read_revision_from_string(_revision_v5)
335
breezy.xml5.serializer_v5.write_revision(rev, outp)
336
outfile_contents = outp.getvalue()
337
self.assertEqual(outfile_contents[-1], '\n')
338
self.assertEqualDiff(outfile_contents, breezy.xml5.serializer_v5.write_revision_to_string(rev))
339
self.assertEqualDiff(outfile_contents, _expected_rev_v5)
341
def test_empty_property_value(self):
342
"""Create an empty property value check that it serializes correctly"""
343
s_v5 = breezy.xml5.serializer_v5
344
rev = s_v5.read_revision_from_string(_revision_v5)
346
props = {'empty':'', 'one':'one'}
347
rev.properties = props
348
txt = s_v5.write_revision_to_string(rev)
349
new_rev = s_v5.read_revision_from_string(txt)
350
self.assertEqual(props, new_rev.properties)
352
def get_sample_inventory(self):
353
inv = Inventory('tree-root-321', revision_id='rev_outer')
354
inv.add(inventory.InventoryFile('file-id', 'file', 'tree-root-321'))
355
inv.add(inventory.InventoryDirectory('dir-id', 'dir',
357
inv.add(inventory.InventoryLink('link-id', 'link', 'tree-root-321'))
358
inv['tree-root-321'].revision = 'rev_outer'
359
inv['dir-id'].revision = 'rev_outer'
360
inv['file-id'].revision = 'rev_outer'
361
inv['file-id'].text_sha1 = 'A'
362
inv['file-id'].text_size = 1
363
inv['link-id'].revision = 'rev_outer'
364
inv['link-id'].symlink_target = 'a'
367
def test_roundtrip_inventory_v7(self):
368
inv = self.get_sample_inventory()
369
inv.add(inventory.TreeReference('nested-id', 'nested', 'tree-root-321',
370
'rev_outer', 'rev_inner'))
371
txt = xml7.serializer_v7.write_inventory_to_string(inv)
372
lines = xml7.serializer_v7.write_inventory_to_lines(inv)
373
self.assertEqual(breezy.osutils.split_lines(txt), lines)
374
self.assertEqualDiff(_expected_inv_v7, txt)
375
inv2 = xml7.serializer_v7.read_inventory_from_string(txt)
376
self.assertEqual(5, len(inv2))
377
for path, ie in inv.iter_entries():
378
self.assertEqual(ie, inv2[ie.file_id])
380
def test_roundtrip_inventory_v6(self):
381
inv = self.get_sample_inventory()
382
txt = xml6.serializer_v6.write_inventory_to_string(inv)
383
lines = xml6.serializer_v6.write_inventory_to_lines(inv)
384
self.assertEqual(breezy.osutils.split_lines(txt), lines)
385
self.assertEqualDiff(_expected_inv_v6, txt)
386
inv2 = xml6.serializer_v6.read_inventory_from_string(txt)
387
self.assertEqual(4, len(inv2))
388
for path, ie in inv.iter_entries():
389
self.assertEqual(ie, inv2[ie.file_id])
391
def test_wrong_format_v7(self):
392
"""Can't accidentally open a file with wrong serializer"""
393
s_v6 = breezy.xml6.serializer_v6
394
s_v7 = xml7.serializer_v7
395
self.assertRaises(errors.UnexpectedInventoryFormat,
396
s_v7.read_inventory_from_string, _expected_inv_v5)
397
self.assertRaises(errors.UnexpectedInventoryFormat,
398
s_v6.read_inventory_from_string, _expected_inv_v7)
400
def test_tree_reference(self):
401
s_v5 = breezy.xml5.serializer_v5
402
s_v6 = breezy.xml6.serializer_v6
403
s_v7 = xml7.serializer_v7
404
inv = Inventory('tree-root-321', revision_id='rev-outer')
405
inv.root.revision = 'root-rev'
406
inv.add(inventory.TreeReference('nested-id', 'nested', 'tree-root-321',
407
'rev-outer', 'rev-inner'))
408
self.assertRaises(errors.UnsupportedInventoryKind,
409
s_v5.write_inventory_to_string, inv)
410
self.assertRaises(errors.UnsupportedInventoryKind,
411
s_v6.write_inventory_to_string, inv)
412
txt = s_v7.write_inventory_to_string(inv)
413
lines = s_v7.write_inventory_to_lines(inv)
414
self.assertEqual(breezy.osutils.split_lines(txt), lines)
415
inv2 = s_v7.read_inventory_from_string(txt)
416
self.assertEqual('tree-root-321', inv2['nested-id'].parent_id)
417
self.assertEqual('rev-outer', inv2['nested-id'].revision)
418
self.assertEqual('rev-inner', inv2['nested-id'].reference_revision)
420
def test_roundtrip_inventory_v8(self):
421
inv = self.get_sample_inventory()
422
txt = xml8.serializer_v8.write_inventory_to_string(inv)
423
inv2 = xml8.serializer_v8.read_inventory_from_string(txt)
424
self.assertEqual(4, len(inv2))
425
for path, ie in inv.iter_entries():
426
self.assertEqual(ie, inv2[ie.file_id])
428
def test_inventory_text_v8(self):
429
inv = self.get_sample_inventory()
430
txt = xml8.serializer_v8.write_inventory_to_string(inv)
431
lines = xml8.serializer_v8.write_inventory_to_lines(inv)
432
self.assertEqual(breezy.osutils.split_lines(txt), lines)
433
self.assertEqualDiff(_expected_inv_v8, txt)
435
def test_revision_text_v6(self):
436
"""Pack revision to XML v6"""
437
rev = breezy.xml6.serializer_v6.read_revision_from_string(
439
serialized = breezy.xml6.serializer_v6.write_revision_to_string(rev)
440
self.assertEqualDiff(serialized, _expected_rev_v5)
442
def test_revision_text_v7(self):
443
"""Pack revision to XML v7"""
444
rev = breezy.xml7.serializer_v7.read_revision_from_string(
446
serialized = breezy.xml7.serializer_v7.write_revision_to_string(rev)
447
self.assertEqualDiff(serialized, _expected_rev_v5)
449
def test_revision_text_v8(self):
450
"""Pack revision to XML v8"""
451
rev = breezy.xml8.serializer_v8.read_revision_from_string(
453
serialized = breezy.xml8.serializer_v8.write_revision_to_string(rev)
454
self.assertEqualDiff(serialized, _expected_rev_v8)
456
def test_revision_ids_are_utf8(self):
457
"""Parsed revision_ids should all be utf-8 strings, not unicode."""
458
s_v5 = breezy.xml5.serializer_v5
459
rev = s_v5.read_revision_from_string(_revision_utf8_v5)
460
self.assertEqual('erik@b\xc3\xa5gfors-02', rev.revision_id)
461
self.assertIsInstance(rev.revision_id, str)
462
self.assertEqual(['erik@b\xc3\xa5gfors-01'], rev.parent_ids)
463
for parent_id in rev.parent_ids:
464
self.assertIsInstance(parent_id, str)
465
self.assertEqual(u'Include \xb5nicode characters\n', rev.message)
466
self.assertIsInstance(rev.message, unicode)
468
# ie.revision should either be None or a utf-8 revision id
469
inv = s_v5.read_inventory_from_string(_inventory_utf8_v5)
470
rev_id_1 = u'erik@b\xe5gfors-01'.encode('utf8')
471
rev_id_2 = u'erik@b\xe5gfors-02'.encode('utf8')
472
fid_root = u'TRE\xe9_ROOT'.encode('utf8')
473
fid_bar1 = u'b\xe5r-01'.encode('utf8')
474
fid_sub = u's\xb5bdir-01'.encode('utf8')
475
fid_bar2 = u'b\xe5r-02'.encode('utf8')
476
expected = [(u'', fid_root, None, rev_id_2),
477
(u'b\xe5r', fid_bar1, fid_root, rev_id_1),
478
(u's\xb5bdir', fid_sub, fid_root, rev_id_1),
479
(u's\xb5bdir/b\xe5r', fid_bar2, fid_sub, rev_id_2),
481
self.assertEqual(rev_id_2, inv.revision_id)
482
self.assertIsInstance(inv.revision_id, str)
484
actual = list(inv.iter_entries_by_dir())
485
for ((exp_path, exp_file_id, exp_parent_id, exp_rev_id),
486
(act_path, act_ie)) in zip(expected, actual):
487
self.assertEqual(exp_path, act_path)
488
self.assertIsInstance(act_path, unicode)
489
self.assertEqual(exp_file_id, act_ie.file_id)
490
self.assertIsInstance(act_ie.file_id, str)
491
self.assertEqual(exp_parent_id, act_ie.parent_id)
492
if exp_parent_id is not None:
493
self.assertIsInstance(act_ie.parent_id, str)
494
self.assertEqual(exp_rev_id, act_ie.revision)
495
if exp_rev_id is not None:
496
self.assertIsInstance(act_ie.revision, str)
498
self.assertEqual(len(expected), len(actual))
501
class TestEncodeAndEscape(TestCase):
502
"""Whitebox testing of the _encode_and_escape function."""
505
super(TestEncodeAndEscape, self).setUp()
506
# Keep the cache clear before and after the test
507
breezy.xml_serializer._clear_cache()
508
self.addCleanup(breezy.xml_serializer._clear_cache)
510
def test_simple_ascii(self):
511
# _encode_and_escape always appends a final ", because these parameters
512
# are being used in xml attributes, and by returning it now, we have to
513
# do fewer string operations later.
514
val = breezy.xml_serializer.encode_and_escape('foo bar')
515
self.assertEqual('foo bar"', val)
516
# The second time should be cached
517
val2 = breezy.xml_serializer.encode_and_escape('foo bar')
518
self.assertIs(val2, val)
520
def test_ascii_with_xml(self):
521
self.assertEqual('&'"<>"',
522
breezy.xml_serializer.encode_and_escape('&\'"<>'))
524
def test_utf8_with_xml(self):
526
utf8_str = '\xc2\xb5\xc3\xa5&\xd8\xac'
527
self.assertEqual('µå&ج"',
528
breezy.xml_serializer.encode_and_escape(utf8_str))
530
def test_unicode(self):
531
uni_str = u'\xb5\xe5&\u062c'
532
self.assertEqual('µå&ج"',
533
breezy.xml_serializer.encode_and_escape(uni_str))
536
class TestMisc(TestCase):
538
def test_unescape_xml(self):
539
"""We get some kind of error when malformed entities are passed"""
540
self.assertRaises(KeyError, breezy.xml8._unescape_xml, 'foo&bar;')