17
17
"""Tests for Knit data structure"""
19
from cStringIO import StringIO
20
from io import BytesIO
33
from bzrlib.errors import (
34
RevisionAlreadyPresent,
39
from bzrlib.index import *
40
from bzrlib.knit import (
34
from ..bzr.index import *
35
from ..bzr.knit import (
41
36
AnnotatedKnitContent,
39
KnitDataStreamIncompatible,
40
KnitDataStreamUnknown,
42
KnitIndexUnknownMethod,
44
43
KnitVersionedFiles,
46
45
_VFContentMapGenerator,
53
from bzrlib.repofmt import pack_repo
54
from bzrlib.tests import (
51
from ..patiencediff import PatienceSequenceMatcher
58
58
TestCaseWithMemoryTransport,
59
59
TestCaseWithTransport,
62
from bzrlib.transport import get_transport
63
from bzrlib.transport.memory import MemoryTransport
64
from bzrlib.tuned_gzip import GzipFile
65
from bzrlib.versionedfile import (
62
from ..bzr.versionedfile import (
66
63
AbsentContentFactory,
68
65
network_bytes_to_kind_and_offset,
69
66
RecordingVersionedFilesDecorator,
73
compiled_knit_feature = tests.ModuleAvailableFeature(
74
'bzrlib._knit_load_data_pyx')
73
compiled_knit_feature = features.ModuleAvailableFeature(
74
'breezy.bzr._knit_load_data_pyx')
77
class ErrorTests(TestCase):
79
def test_knit_data_stream_incompatible(self):
80
error = KnitDataStreamIncompatible(
81
'stream format', 'target format')
82
self.assertEqual('Cannot insert knit data stream of format '
83
'"stream format" into knit of format '
84
'"target format".', str(error))
86
def test_knit_data_stream_unknown(self):
87
error = KnitDataStreamUnknown(
89
self.assertEqual('Cannot parse knit data stream of format '
90
'"stream format".', str(error))
92
def test_knit_header_error(self):
93
error = KnitHeaderError('line foo\n', 'path/to/file')
94
self.assertEqual("Knit header error: 'line foo\\n' unexpected"
95
" for file \"path/to/file\".", str(error))
97
def test_knit_index_unknown_method(self):
98
error = KnitIndexUnknownMethod('http://host/foo.kndx',
100
self.assertEqual("Knit index http://host/foo.kndx does not have a"
101
" known method in options: ['bad', 'no-eol']",
77
105
class KnitContentTestsMixin(object):
422
477
raise _TestException('foobar')
423
except _TestException, e:
478
except _TestException as e:
424
479
retry_exc = errors.RetryWithNewPacks(None, reload_occurred=False,
425
480
exc_info=sys.exc_info())
481
# GZ 2010-08-10: Cycle with exc_info affects 3 tests
428
484
def test_read_from_several_packs(self):
429
485
access, writer = self._get_access()
431
memos.extend(access.add_raw_records([('key', 10)], '1234567890'))
487
memos.extend(access.add_raw_records([(b'key', 10)], b'1234567890'))
433
489
access, writer = self._get_access('pack2', 'FOOBAR')
434
memos.extend(access.add_raw_records([('key', 5)], '12345'))
490
memos.extend(access.add_raw_records([(b'key', 5)], b'12345'))
436
492
access, writer = self._get_access('pack3', 'BAZ')
437
memos.extend(access.add_raw_records([('key', 5)], 'alpha'))
493
memos.extend(access.add_raw_records([(b'key', 5)], b'alpha'))
439
495
transport = self.get_transport()
440
access = _DirectPackAccess({"FOO":(transport, 'packfile'),
496
access = pack_repo._DirectPackAccess({"FOO":(transport, 'packfile'),
441
497
"FOOBAR":(transport, 'pack2'),
442
498
"BAZ":(transport, 'pack3')})
443
self.assertEqual(['1234567890', '12345', 'alpha'],
499
self.assertEqual([b'1234567890', b'12345', b'alpha'],
444
500
list(access.get_raw_records(memos)))
445
self.assertEqual(['1234567890'],
501
self.assertEqual([b'1234567890'],
446
502
list(access.get_raw_records(memos[0:1])))
447
self.assertEqual(['12345'],
503
self.assertEqual([b'12345'],
448
504
list(access.get_raw_records(memos[1:2])))
449
self.assertEqual(['alpha'],
505
self.assertEqual([b'alpha'],
450
506
list(access.get_raw_records(memos[2:3])))
451
self.assertEqual(['1234567890', 'alpha'],
507
self.assertEqual([b'1234567890', b'alpha'],
452
508
list(access.get_raw_records(memos[0:1] + memos[2:3])))
454
510
def test_set_writer(self):
455
511
"""The writer should be settable post construction."""
456
access = _DirectPackAccess({})
512
access = pack_repo._DirectPackAccess({})
457
513
transport = self.get_transport()
458
514
packname = 'packfile'
595
655
self.fail('Annotation was not identical with reloading.')
596
656
# Now delete the packs-in-use, which should trigger another reload, but
597
657
# this time we just raise an exception because we can't recover
598
for trans, name in vf._access._indices.itervalues():
658
for trans, name in vf._access._indices.values():
599
659
trans.delete(name)
600
660
self.assertRaises(errors.NoSuchFile, vf.annotate, key)
601
661
self.assertEqual([2, 1, 1], reload_counter)
603
663
def test__get_record_map_retries(self):
604
664
vf, reload_counter = self.make_vf_for_retrying()
605
keys = [('rev-1',), ('rev-2',), ('rev-3',)]
665
keys = [(b'rev-1',), (b'rev-2',), (b'rev-3',)]
606
666
records = vf._get_record_map(keys)
607
667
self.assertEqual(keys, sorted(records.keys()))
608
668
self.assertEqual([1, 1, 0], reload_counter)
609
669
# Now delete the packs-in-use, which should trigger another reload, but
610
670
# this time we just raise an exception because we can't recover
611
for trans, name in vf._access._indices.itervalues():
671
for trans, name in vf._access._indices.values():
612
672
trans.delete(name)
613
673
self.assertRaises(errors.NoSuchFile, vf._get_record_map, keys)
614
674
self.assertEqual([2, 1, 1], reload_counter)
616
676
def test_get_record_stream_retries(self):
617
677
vf, reload_counter = self.make_vf_for_retrying()
618
keys = [('rev-1',), ('rev-2',), ('rev-3',)]
678
keys = [(b'rev-1',), (b'rev-2',), (b'rev-3',)]
619
679
record_stream = vf.get_record_stream(keys, 'topological', False)
620
record = record_stream.next()
621
self.assertEqual(('rev-1',), record.key)
680
record = next(record_stream)
681
self.assertEqual((b'rev-1',), record.key)
622
682
self.assertEqual([0, 0, 0], reload_counter)
623
record = record_stream.next()
624
self.assertEqual(('rev-2',), record.key)
683
record = next(record_stream)
684
self.assertEqual((b'rev-2',), record.key)
625
685
self.assertEqual([1, 1, 0], reload_counter)
626
record = record_stream.next()
627
self.assertEqual(('rev-3',), record.key)
686
record = next(record_stream)
687
self.assertEqual((b'rev-3',), record.key)
628
688
self.assertEqual([1, 1, 0], reload_counter)
629
689
# Now delete all pack files, and see that we raise the right error
630
for trans, name in vf._access._indices.itervalues():
690
for trans, name in vf._access._indices.values():
631
691
trans.delete(name)
632
692
self.assertListRaises(errors.NoSuchFile,
633
693
vf.get_record_stream, keys, 'topological', False)
635
695
def test_iter_lines_added_or_present_in_keys_retries(self):
636
696
vf, reload_counter = self.make_vf_for_retrying()
637
keys = [('rev-1',), ('rev-2',), ('rev-3',)]
697
keys = [(b'rev-1',), (b'rev-2',), (b'rev-3',)]
638
698
# Unfortunately, iter_lines_added_or_present_in_keys iterates the
639
699
# result in random order (determined by the iteration order from a
640
700
# set()), so we don't have any solid way to trigger whether data is
708
768
def make_multiple_records(self):
709
769
"""Create the content for multiple records."""
710
sha1sum = osutils.sha('foo\nbar\n').hexdigest()
770
sha1sum = osutils.sha_string(b'foo\nbar\n')
712
gz_txt = self.create_gz_content('version rev-id-1 2 %s\n'
772
gz_txt = self.create_gz_content(b'version rev-id-1 2 %s\n'
717
777
record_1 = (0, len(gz_txt), sha1sum)
718
778
total_txt.append(gz_txt)
719
sha1sum = osutils.sha('baz\n').hexdigest()
720
gz_txt = self.create_gz_content('version rev-id-2 1 %s\n'
779
sha1sum = osutils.sha_string(b'baz\n')
780
gz_txt = self.create_gz_content(b'version rev-id-2 1 %s\n'
724
784
record_2 = (record_1[1], len(gz_txt), sha1sum)
725
785
total_txt.append(gz_txt)
726
786
return total_txt, record_1, record_2
728
788
def test_valid_knit_data(self):
729
sha1sum = osutils.sha('foo\nbar\n').hexdigest()
730
gz_txt = self.create_gz_content('version rev-id-1 2 %s\n'
789
sha1sum = osutils.sha_string(b'foo\nbar\n')
790
gz_txt = self.create_gz_content(b'version rev-id-1 2 %s\n'
735
795
transport = MockTransport([gz_txt])
736
796
access = _KnitKeyAccess(transport, ConstantMapper('filename'))
737
797
knit = KnitVersionedFiles(None, access)
738
records = [(('rev-id-1',), (('rev-id-1',), 0, len(gz_txt)))]
798
records = [((b'rev-id-1',), ((b'rev-id-1',), 0, len(gz_txt)))]
740
800
contents = list(knit._read_records_iter(records))
741
self.assertEqual([(('rev-id-1',), ['foo\n', 'bar\n'],
742
'4e48e2c9a3d2ca8a708cb0cc545700544efb5021')], contents)
801
self.assertEqual([((b'rev-id-1',), [b'foo\n', b'bar\n'],
802
b'4e48e2c9a3d2ca8a708cb0cc545700544efb5021')], contents)
744
804
raw_contents = list(knit._read_records_iter_raw(records))
745
self.assertEqual([(('rev-id-1',), gz_txt, sha1sum)], raw_contents)
805
self.assertEqual([((b'rev-id-1',), gz_txt, sha1sum)], raw_contents)
747
807
def test_multiple_records_valid(self):
748
808
total_txt, record_1, record_2 = self.make_multiple_records()
749
transport = MockTransport([''.join(total_txt)])
809
transport = MockTransport([b''.join(total_txt)])
750
810
access = _KnitKeyAccess(transport, ConstantMapper('filename'))
751
811
knit = KnitVersionedFiles(None, access)
752
records = [(('rev-id-1',), (('rev-id-1',), record_1[0], record_1[1])),
753
(('rev-id-2',), (('rev-id-2',), record_2[0], record_2[1]))]
812
records = [((b'rev-id-1',), ((b'rev-id-1',), record_1[0], record_1[1])),
813
((b'rev-id-2',), ((b'rev-id-2',), record_2[0], record_2[1]))]
755
815
contents = list(knit._read_records_iter(records))
756
self.assertEqual([(('rev-id-1',), ['foo\n', 'bar\n'], record_1[2]),
757
(('rev-id-2',), ['baz\n'], record_2[2])],
816
self.assertEqual([((b'rev-id-1',), [b'foo\n', b'bar\n'], record_1[2]),
817
((b'rev-id-2',), [b'baz\n'], record_2[2])],
760
820
raw_contents = list(knit._read_records_iter_raw(records))
761
self.assertEqual([(('rev-id-1',), total_txt[0], record_1[2]),
762
(('rev-id-2',), total_txt[1], record_2[2])],
821
self.assertEqual([((b'rev-id-1',), total_txt[0], record_1[2]),
822
((b'rev-id-2',), total_txt[1], record_2[2])],
765
825
def test_not_enough_lines(self):
766
sha1sum = osutils.sha('foo\n').hexdigest()
826
sha1sum = osutils.sha_string(b'foo\n')
767
827
# record says 2 lines data says 1
768
gz_txt = self.create_gz_content('version rev-id-1 2 %s\n'
828
gz_txt = self.create_gz_content(b'version rev-id-1 2 %s\n'
772
832
transport = MockTransport([gz_txt])
773
833
access = _KnitKeyAccess(transport, ConstantMapper('filename'))
774
834
knit = KnitVersionedFiles(None, access)
775
records = [(('rev-id-1',), (('rev-id-1',), 0, len(gz_txt)))]
776
self.assertRaises(errors.KnitCorrupt, list,
835
records = [((b'rev-id-1',), ((b'rev-id-1',), 0, len(gz_txt)))]
836
self.assertRaises(KnitCorrupt, list,
777
837
knit._read_records_iter(records))
779
839
# read_records_iter_raw won't detect that sort of mismatch/corruption
780
840
raw_contents = list(knit._read_records_iter_raw(records))
781
self.assertEqual([(('rev-id-1',), gz_txt, sha1sum)], raw_contents)
841
self.assertEqual([((b'rev-id-1',), gz_txt, sha1sum)], raw_contents)
783
843
def test_too_many_lines(self):
784
sha1sum = osutils.sha('foo\nbar\n').hexdigest()
844
sha1sum = osutils.sha_string(b'foo\nbar\n')
785
845
# record says 1 lines data says 2
786
gz_txt = self.create_gz_content('version rev-id-1 1 %s\n'
846
gz_txt = self.create_gz_content(b'version rev-id-1 1 %s\n'
791
851
transport = MockTransport([gz_txt])
792
852
access = _KnitKeyAccess(transport, ConstantMapper('filename'))
793
853
knit = KnitVersionedFiles(None, access)
794
records = [(('rev-id-1',), (('rev-id-1',), 0, len(gz_txt)))]
795
self.assertRaises(errors.KnitCorrupt, list,
854
records = [((b'rev-id-1',), ((b'rev-id-1',), 0, len(gz_txt)))]
855
self.assertRaises(KnitCorrupt, list,
796
856
knit._read_records_iter(records))
798
858
# read_records_iter_raw won't detect that sort of mismatch/corruption
799
859
raw_contents = list(knit._read_records_iter_raw(records))
800
self.assertEqual([(('rev-id-1',), gz_txt, sha1sum)], raw_contents)
860
self.assertEqual([((b'rev-id-1',), gz_txt, sha1sum)], raw_contents)
802
862
def test_mismatched_version_id(self):
803
sha1sum = osutils.sha('foo\nbar\n').hexdigest()
804
gz_txt = self.create_gz_content('version rev-id-1 2 %s\n'
863
sha1sum = osutils.sha_string(b'foo\nbar\n')
864
gz_txt = self.create_gz_content(b'version rev-id-1 2 %s\n'
809
869
transport = MockTransport([gz_txt])
810
870
access = _KnitKeyAccess(transport, ConstantMapper('filename'))
811
871
knit = KnitVersionedFiles(None, access)
812
872
# We are asking for rev-id-2, but the data is rev-id-1
813
records = [(('rev-id-2',), (('rev-id-2',), 0, len(gz_txt)))]
814
self.assertRaises(errors.KnitCorrupt, list,
873
records = [((b'rev-id-2',), ((b'rev-id-2',), 0, len(gz_txt)))]
874
self.assertRaises(KnitCorrupt, list,
815
875
knit._read_records_iter(records))
817
877
# read_records_iter_raw detects mismatches in the header
818
self.assertRaises(errors.KnitCorrupt, list,
878
self.assertRaises(KnitCorrupt, list,
819
879
knit._read_records_iter_raw(records))
821
881
def test_uncompressed_data(self):
822
sha1sum = osutils.sha('foo\nbar\n').hexdigest()
823
txt = ('version rev-id-1 2 %s\n'
882
sha1sum = osutils.sha_string(b'foo\nbar\n')
883
txt = (b'version rev-id-1 2 %s\n'
828
888
transport = MockTransport([txt])
829
889
access = _KnitKeyAccess(transport, ConstantMapper('filename'))
830
890
knit = KnitVersionedFiles(None, access)
831
records = [(('rev-id-1',), (('rev-id-1',), 0, len(txt)))]
891
records = [((b'rev-id-1',), ((b'rev-id-1',), 0, len(txt)))]
833
893
# We don't have valid gzip data ==> corrupt
834
self.assertRaises(errors.KnitCorrupt, list,
894
self.assertRaises(KnitCorrupt, list,
835
895
knit._read_records_iter(records))
837
897
# read_records_iter_raw will notice the bad data
838
self.assertRaises(errors.KnitCorrupt, list,
898
self.assertRaises(KnitCorrupt, list,
839
899
knit._read_records_iter_raw(records))
841
901
def test_corrupted_data(self):
842
sha1sum = osutils.sha('foo\nbar\n').hexdigest()
843
gz_txt = self.create_gz_content('version rev-id-1 2 %s\n'
902
sha1sum = osutils.sha_string(b'foo\nbar\n')
903
gz_txt = self.create_gz_content(b'version rev-id-1 2 %s\n'
848
908
# Change 2 bytes in the middle to \xff
849
gz_txt = gz_txt[:10] + '\xff\xff' + gz_txt[12:]
909
gz_txt = gz_txt[:10] + b'\xff\xff' + gz_txt[12:]
850
910
transport = MockTransport([gz_txt])
851
911
access = _KnitKeyAccess(transport, ConstantMapper('filename'))
852
912
knit = KnitVersionedFiles(None, access)
853
records = [(('rev-id-1',), (('rev-id-1',), 0, len(gz_txt)))]
854
self.assertRaises(errors.KnitCorrupt, list,
913
records = [((b'rev-id-1',), ((b'rev-id-1',), 0, len(gz_txt)))]
914
self.assertRaises(KnitCorrupt, list,
855
915
knit._read_records_iter(records))
856
916
# read_records_iter_raw will barf on bad gz data
857
self.assertRaises(errors.KnitCorrupt, list,
917
self.assertRaises(KnitCorrupt, list,
858
918
knit._read_records_iter_raw(records))
1143
1203
def test_get_parent_map(self):
1144
1204
transport = MockTransport([
1145
1205
_KndxIndex.HEADER,
1147
"b option 1 2 0 .c :",
1148
"c option 1 2 1 0 .e :"
1207
b"b option 1 2 0 .c :",
1208
b"c option 1 2 1 0 .e :"
1150
1210
index = self.get_knit_index(transport, "filename", "r")
1152
1212
self.assertEqual({
1154
("b",):(("a",), ("c",)),
1155
("c",):(("b",), ("a",), ("e",)),
1214
(b"b",): ((b"a",), (b"c",)),
1215
(b"c",): ((b"b",), (b"a",), (b"e",)),
1156
1216
}, index.get_parent_map(index.keys()))
1158
1218
def test_impossible_parent(self):
1159
1219
"""Test we get KnitCorrupt if the parent couldn't possibly exist."""
1160
1220
transport = MockTransport([
1161
1221
_KndxIndex.HEADER,
1163
"b option 0 1 4 :" # We don't have a 4th record
1223
b"b option 0 1 4 :" # We don't have a 4th record
1165
1225
index = self.get_knit_index(transport, 'filename', 'r')
1167
self.assertRaises(errors.KnitCorrupt, index.keys)
1168
except TypeError, e:
1169
if (str(e) == ('exceptions must be strings, classes, or instances,'
1170
' not exceptions.IndexError')
1171
and sys.version_info[0:2] >= (2,5)):
1172
self.knownFailure('Pyrex <0.9.5 fails with TypeError when'
1173
' raising new style exceptions with python'
1226
self.assertRaises(KnitCorrupt, index.keys)
1178
1228
def test_corrupted_parent(self):
1179
1229
transport = MockTransport([
1180
1230
_KndxIndex.HEADER,
1183
"c option 0 1 1v :", # Can't have a parent of '1v'
1233
b"c option 0 1 1v :", # Can't have a parent of '1v'
1185
1235
index = self.get_knit_index(transport, 'filename', 'r')
1187
self.assertRaises(errors.KnitCorrupt, index.keys)
1188
except TypeError, e:
1189
if (str(e) == ('exceptions must be strings, classes, or instances,'
1190
' not exceptions.ValueError')
1191
and sys.version_info[0:2] >= (2,5)):
1192
self.knownFailure('Pyrex <0.9.5 fails with TypeError when'
1193
' raising new style exceptions with python'
1236
self.assertRaises(KnitCorrupt, index.keys)
1198
1238
def test_corrupted_parent_in_list(self):
1199
1239
transport = MockTransport([
1200
1240
_KndxIndex.HEADER,
1203
"c option 0 1 1 v :", # Can't have a parent of 'v'
1243
b"c option 0 1 1 v :", # Can't have a parent of 'v'
1205
1245
index = self.get_knit_index(transport, 'filename', 'r')
1207
self.assertRaises(errors.KnitCorrupt, index.keys)
1208
except TypeError, e:
1209
if (str(e) == ('exceptions must be strings, classes, or instances,'
1210
' not exceptions.ValueError')
1211
and sys.version_info[0:2] >= (2,5)):
1212
self.knownFailure('Pyrex <0.9.5 fails with TypeError when'
1213
' raising new style exceptions with python'
1246
self.assertRaises(KnitCorrupt, index.keys)
1218
1248
def test_invalid_position(self):
1219
1249
transport = MockTransport([
1220
1250
_KndxIndex.HEADER,
1223
1253
index = self.get_knit_index(transport, 'filename', 'r')
1225
self.assertRaises(errors.KnitCorrupt, index.keys)
1226
except TypeError, e:
1227
if (str(e) == ('exceptions must be strings, classes, or instances,'
1228
' not exceptions.ValueError')
1229
and sys.version_info[0:2] >= (2,5)):
1230
self.knownFailure('Pyrex <0.9.5 fails with TypeError when'
1231
' raising new style exceptions with python'
1254
self.assertRaises(KnitCorrupt, index.keys)
1236
1256
def test_invalid_size(self):
1237
1257
transport = MockTransport([
1238
1258
_KndxIndex.HEADER,
1241
1261
index = self.get_knit_index(transport, 'filename', 'r')
1243
self.assertRaises(errors.KnitCorrupt, index.keys)
1244
except TypeError, e:
1245
if (str(e) == ('exceptions must be strings, classes, or instances,'
1246
' not exceptions.ValueError')
1247
and sys.version_info[0:2] >= (2,5)):
1248
self.knownFailure('Pyrex <0.9.5 fails with TypeError when'
1249
' raising new style exceptions with python'
1262
self.assertRaises(KnitCorrupt, index.keys)
1254
1264
def test_scan_unvalidated_index_not_implemented(self):
1255
1265
transport = MockTransport()
1263
1273
def test_short_line(self):
1264
1274
transport = MockTransport([
1265
1275
_KndxIndex.HEADER,
1267
"b option 10 10 0", # This line isn't terminated, ignored
1277
b"b option 10 10 0", # This line isn't terminated, ignored
1269
1279
index = self.get_knit_index(transport, "filename", "r")
1270
self.assertEqual(set([('a',)]), index.keys())
1280
self.assertEqual({(b'a',)}, index.keys())
1272
1282
def test_skip_incomplete_record(self):
1273
1283
# A line with bogus data should just be skipped
1274
1284
transport = MockTransport([
1275
1285
_KndxIndex.HEADER,
1277
"b option 10 10 0", # This line isn't terminated, ignored
1278
"c option 20 10 0 :", # Properly terminated, and starts with '\n'
1287
b"b option 10 10 0", # This line isn't terminated, ignored
1288
b"c option 20 10 0 :", # Properly terminated, and starts with '\n'
1280
1290
index = self.get_knit_index(transport, "filename", "r")
1281
self.assertEqual(set([('a',), ('c',)]), index.keys())
1291
self.assertEqual({(b'a',), (b'c',)}, index.keys())
1283
1293
def test_trailing_characters(self):
1284
1294
# A line with bogus data should just be skipped
1285
1295
transport = MockTransport([
1286
1296
_KndxIndex.HEADER,
1288
"b option 10 10 0 :a", # This line has extra trailing characters
1289
"c option 20 10 0 :", # Properly terminated, and starts with '\n'
1298
b"b option 10 10 0 :a", # This line has extra trailing characters
1299
b"c option 20 10 0 :", # Properly terminated, and starts with '\n'
1291
1301
index = self.get_knit_index(transport, "filename", "r")
1292
self.assertEqual(set([('a',), ('c',)]), index.keys())
1302
self.assertEqual({(b'a',), (b'c',)}, index.keys())
1295
1305
class LowLevelKnitIndexTests_c(LowLevelKnitIndexTests):
1395
1405
# We should not try to access any parent_lines content, because we know
1396
1406
# we already have the matching blocks
1397
1407
par_ann, blocks = ann._get_parent_annotations_and_matches(rev_key,
1398
['1\n', '2\n', '3\n'], parent_key)
1408
[b'1\n', b'2\n', b'3\n'], parent_key)
1399
1409
self.assertEqual(parent_ann, par_ann)
1400
1410
self.assertEqual([(0, 1, 1), (3, 3, 0)], blocks)
1401
1411
self.assertEqual({}, ann._matching_blocks)
1403
1413
def test__process_pending(self):
1404
1414
ann = self.make_annotator()
1405
rev_key = ('rev-id',)
1408
record = ['0,1,1\n', 'new-line\n']
1415
rev_key = (b'rev-id',)
1416
p1_key = (b'p1-id',)
1417
p2_key = (b'p2-id',)
1418
record = [b'0,1,1\n', b'new-line\n']
1409
1419
details = ('line-delta', False)
1410
p1_record = ['line1\n', 'line2\n']
1420
p1_record = [b'line1\n', b'line2\n']
1411
1421
ann._num_compression_children[p1_key] = 1
1412
res = ann._expand_record(rev_key, (p1_key,p2_key), p1_key,
1422
res = ann._expand_record(rev_key, (p1_key, p2_key), p1_key,
1413
1423
record, details)
1414
1424
self.assertEqual(None, res)
1415
1425
# self.assertTrue(p1_key in ann._pending_deltas)
1416
1426
self.assertEqual({}, ann._pending_annotation)
1417
1427
# Now insert p1, and we should be able to expand the delta
1418
1428
res = ann._expand_record(p1_key, (), None, p1_record,
1419
('fulltext', False))
1429
(b'fulltext', False))
1420
1430
self.assertEqual(p1_record, res)
1421
1431
ann._annotations_cache[p1_key] = [(p1_key,)]*2
1422
1432
res = ann._process_pending(p1_key)
1510
1520
idx = knit._index
1511
1521
idx.add_records([(('a-1',), ['fulltext'], (('a-1',), 0, 0), [])])
1512
1522
self.check_file_contents('test.kndx',
1513
'# bzr knit index 8\n'
1515
'a-1 fulltext 0 0 :'
1523
b'# bzr knit index 8\n'
1525
b'a-1 fulltext 0 0 :'
1517
1527
idx.add_records([
1518
1528
(('a-2',), ['fulltext'], (('a-2',), 0, 0), [('a-1',)]),
1519
1529
(('a-3',), ['fulltext'], (('a-3',), 0, 0), [('a-2',)]),
1521
1531
self.check_file_contents('test.kndx',
1522
'# bzr knit index 8\n'
1524
'a-1 fulltext 0 0 :\n'
1525
'a-2 fulltext 0 0 0 :\n'
1526
'a-3 fulltext 0 0 1 :'
1532
b'# bzr knit index 8\n'
1534
b'a-1 fulltext 0 0 :\n'
1535
b'a-2 fulltext 0 0 0 :\n'
1536
b'a-3 fulltext 0 0 1 :'
1528
self.assertEqual(set([('a-3',), ('a-1',), ('a-2',)]), idx.keys())
1538
self.assertEqual({('a-3',), ('a-1',), ('a-2',)}, idx.keys())
1529
1539
self.assertEqual({
1530
1540
('a-1',): ((('a-1',), 0, 0), None, (), ('fulltext', False)),
1531
1541
('a-2',): ((('a-2',), 0, 0), None, (('a-1',),), ('fulltext', False)),
1532
1542
('a-3',): ((('a-3',), 0, 0), None, (('a-2',),), ('fulltext', False)),
1533
1543
}, idx.get_build_details(idx.keys()))
1534
self.assertEqual({('a-1',):(),
1535
('a-2',):(('a-1',),),
1536
('a-3',):(('a-2',),),},
1544
self.assertEqual({('a-1',): (),
1545
('a-2',): (('a-1',),),
1546
('a-3',): (('a-2',),),},
1537
1547
idx.get_parent_map(idx.keys()))
1539
1549
def test_add_versions_fails_clean(self):
1639
1649
def test_keys(self):
1640
1650
index = self.two_graph_index()
1641
self.assertEqual(set([('tail',), ('tip',), ('parent',), ('separate',)]),
1651
self.assertEqual({(b'tail',), (b'tip',), (b'parent',), (b'separate',)},
1642
1652
set(index.keys()))
1644
1654
def test_get_position(self):
1645
1655
index = self.two_graph_index()
1646
self.assertEqual((index._graph_index._indices[0], 0, 100), index.get_position(('tip',)))
1647
self.assertEqual((index._graph_index._indices[1], 100, 78), index.get_position(('parent',)))
1656
self.assertEqual((index._graph_index._indices[0], 0, 100), index.get_position((b'tip',)))
1657
self.assertEqual((index._graph_index._indices[1], 100, 78), index.get_position((b'parent',)))
1649
1659
def test_get_method_deltas(self):
1650
1660
index = self.two_graph_index(deltas=True)
1651
self.assertEqual('fulltext', index.get_method(('tip',)))
1652
self.assertEqual('line-delta', index.get_method(('parent',)))
1661
self.assertEqual('fulltext', index.get_method((b'tip',)))
1662
self.assertEqual('line-delta', index.get_method((b'parent',)))
1654
1664
def test_get_method_no_deltas(self):
1655
1665
# check that the parent-history lookup is ignored with deltas=False.
1656
1666
index = self.two_graph_index(deltas=False)
1657
self.assertEqual('fulltext', index.get_method(('tip',)))
1658
self.assertEqual('fulltext', index.get_method(('parent',)))
1667
self.assertEqual('fulltext', index.get_method((b'tip',)))
1668
self.assertEqual('fulltext', index.get_method((b'parent',)))
1660
1670
def test_get_options_deltas(self):
1661
1671
index = self.two_graph_index(deltas=True)
1870
1880
index.scan_unvalidated_index(g_index_1)
1871
1881
index.scan_unvalidated_index(g_index_2)
1872
1882
self.assertEqual(
1873
frozenset([('one-missing-parent',), ('two-missing-parent',)]),
1883
frozenset([(b'one-missing-parent',), (b'two-missing-parent',)]),
1874
1884
index.get_missing_compression_parents())
1876
1886
def test_add_mulitiple_unvalidated_indices_with_mutual_dependencies(self):
1877
1887
graph_index_a = self.make_g_index('one', 2,
1878
[(('parent-one', ), ' 100 78', ([('non-compression-parent',)], [])),
1879
(('child-of-two', ), ' 100 78',
1880
([('parent-two',)], [('parent-two',)]))])
1888
[((b'parent-one', ), b' 100 78', ([(b'non-compression-parent',)], [])),
1889
((b'child-of-two', ), b' 100 78',
1890
([(b'parent-two',)], [(b'parent-two',)]))])
1881
1891
graph_index_b = self.make_g_index('two', 2,
1882
[(('parent-two', ), ' 100 78', ([('non-compression-parent',)], [])),
1883
(('child-of-one', ), ' 100 78',
1884
([('parent-one',)], [('parent-one',)]))])
1892
[((b'parent-two', ), b' 100 78', ([(b'non-compression-parent',)], [])),
1893
((b'child-of-one', ), b' 100 78',
1894
([(b'parent-one',)], [(b'parent-one',)]))])
1885
1895
combined = CombinedGraphIndex([graph_index_a, graph_index_b])
1886
1896
index = _KnitGraphIndex(combined, lambda: True, deltas=True)
1887
1897
index.scan_unvalidated_index(graph_index_a)
2040
2050
def test_add_versions_same_dup(self):
2041
2051
index = self.two_graph_index(catch_adds=True)
2042
2052
# options can be spelt two different ways
2043
index.add_records([(('tip',), 'fulltext,no-eol', (None, 0, 100), [])])
2044
index.add_records([(('tip',), 'no-eol,fulltext', (None, 0, 100), [])])
2053
index.add_records([((b'tip',), 'fulltext,no-eol', (None, 0, 100), [])])
2054
index.add_records([((b'tip',), 'no-eol,fulltext', (None, 0, 100), [])])
2045
2055
# position/length are ignored (because each pack could have fulltext or
2046
2056
# delta, and be at a different position.
2047
index.add_records([(('tip',), 'fulltext,no-eol', (None, 50, 100), [])])
2048
index.add_records([(('tip',), 'fulltext,no-eol', (None, 0, 1000), [])])
2057
index.add_records([((b'tip',), 'fulltext,no-eol', (None, 50, 100), [])])
2058
index.add_records([((b'tip',), 'fulltext,no-eol', (None, 0, 1000), [])])
2049
2059
# but neither should have added data.
2050
2060
self.assertEqual([[], [], [], []], self.caught_entries)
2052
2062
def test_add_versions_different_dup(self):
2053
2063
index = self.two_graph_index(catch_adds=True)
2054
2064
# change options
2055
self.assertRaises(errors.KnitCorrupt, index.add_records,
2056
[(('tip',), 'no-eol,line-delta', (None, 0, 100), [])])
2057
self.assertRaises(errors.KnitCorrupt, index.add_records,
2058
[(('tip',), 'line-delta,no-eol', (None, 0, 100), [])])
2059
self.assertRaises(errors.KnitCorrupt, index.add_records,
2060
[(('tip',), 'fulltext', (None, 0, 100), [])])
2065
self.assertRaises(KnitCorrupt, index.add_records,
2066
[((b'tip',), 'no-eol,line-delta', (None, 0, 100), [])])
2067
self.assertRaises(KnitCorrupt, index.add_records,
2068
[((b'tip',), 'line-delta,no-eol', (None, 0, 100), [])])
2069
self.assertRaises(KnitCorrupt, index.add_records,
2070
[((b'tip',), 'fulltext', (None, 0, 100), [])])
2062
self.assertRaises(errors.KnitCorrupt, index.add_records,
2063
[(('tip',), 'fulltext,no-eol', (None, 0, 100), [('parent',)])])
2072
self.assertRaises(KnitCorrupt, index.add_records,
2073
[((b'tip',), 'fulltext,no-eol', (None, 0, 100), [('parent',)])])
2064
2074
# change options in the second record
2065
self.assertRaises(errors.KnitCorrupt, index.add_records,
2066
[(('tip',), 'fulltext,no-eol', (None, 0, 100), []),
2067
(('tip',), 'no-eol,line-delta', (None, 0, 100), [])])
2075
self.assertRaises(KnitCorrupt, index.add_records,
2076
[((b'tip',), 'fulltext,no-eol', (None, 0, 100), []),
2077
((b'tip',), 'no-eol,line-delta', (None, 0, 100), [])])
2068
2078
self.assertEqual([], self.caught_entries)
2126
2136
_min_buffer_size=125)
2128
2138
def test__split_by_prefix(self):
2129
self.assertSplitByPrefix({'f': [('f', 'a'), ('f', 'b')],
2130
'g': [('g', 'b'), ('g', 'a')],
2132
[('f', 'a'), ('g', 'b'),
2133
('g', 'a'), ('f', 'b')])
2135
self.assertSplitByPrefix({'f': [('f', 'a'), ('f', 'b')],
2136
'g': [('g', 'b'), ('g', 'a')],
2138
[('f', 'a'), ('f', 'b'),
2139
('g', 'b'), ('g', 'a')])
2141
self.assertSplitByPrefix({'f': [('f', 'a'), ('f', 'b')],
2142
'g': [('g', 'b'), ('g', 'a')],
2144
[('f', 'a'), ('f', 'b'),
2145
('g', 'b'), ('g', 'a')])
2147
self.assertSplitByPrefix({'f': [('f', 'a'), ('f', 'b')],
2148
'g': [('g', 'b'), ('g', 'a')],
2149
'': [('a',), ('b',)]
2151
[('f', 'a'), ('g', 'b'),
2153
('g', 'a'), ('f', 'b')])
2139
self.assertSplitByPrefix({b'f': [(b'f', b'a'), (b'f', b'b')],
2140
b'g': [(b'g', b'b'), (b'g', b'a')],
2142
[(b'f', b'a'), (b'g', b'b'),
2143
(b'g', b'a'), (b'f', b'b')])
2145
self.assertSplitByPrefix({b'f': [(b'f', b'a'), (b'f', b'b')],
2146
b'g': [(b'g', b'b'), (b'g', b'a')],
2148
[(b'f', b'a'), (b'f', b'b'),
2149
(b'g', b'b'), (b'g', b'a')])
2151
self.assertSplitByPrefix({b'f': [(b'f', b'a'), (b'f', b'b')],
2152
b'g': [(b'g', b'b'), (b'g', b'a')],
2154
[(b'f', b'a'), (b'f', b'b'),
2155
(b'g', b'b'), (b'g', b'a')])
2157
self.assertSplitByPrefix({b'f': [(b'f', b'a'), (b'f', b'b')],
2158
b'g': [(b'g', b'b'), (b'g', b'a')],
2159
b'': [(b'a',), (b'b',)]
2160
}, [b'f', b'g', b''],
2161
[(b'f', b'a'), (b'g', b'b'),
2163
(b'g', b'a'), (b'f', b'b')])
2156
2166
class TestStacking(KnitTests):
2172
2182
def test_add_lines(self):
2173
2183
# lines added to the test are not added to the basis
2174
2184
basis, test = self.get_basis_and_test_knit()
2176
key_basis = ('bar',)
2177
key_cross_border = ('quux',)
2178
key_delta = ('zaphod',)
2179
test.add_lines(key, (), ['foo\n'])
2186
key_basis = (b'bar',)
2187
key_cross_border = (b'quux',)
2188
key_delta = (b'zaphod',)
2189
test.add_lines(key, (), [b'foo\n'])
2180
2190
self.assertEqual({}, basis.get_parent_map([key]))
2181
2191
# lines added to the test that reference across the stack do a
2183
basis.add_lines(key_basis, (), ['foo\n'])
2193
basis.add_lines(key_basis, (), [b'foo\n'])
2184
2194
basis.calls = []
2185
test.add_lines(key_cross_border, (key_basis,), ['foo\n'])
2195
test.add_lines(key_cross_border, (key_basis,), [b'foo\n'])
2186
2196
self.assertEqual('fulltext', test._index.get_method(key_cross_border))
2187
2197
# we don't even need to look at the basis to see that this should be
2188
2198
# stored as a fulltext
2189
2199
self.assertEqual([], basis.calls)
2190
2200
# Subsequent adds do delta.
2191
2201
basis.calls = []
2192
test.add_lines(key_delta, (key_cross_border,), ['foo\n'])
2202
test.add_lines(key_delta, (key_cross_border,), [b'foo\n'])
2193
2203
self.assertEqual('line-delta', test._index.get_method(key_delta))
2194
2204
self.assertEqual([], basis.calls)
2196
2206
def test_annotate(self):
2197
2207
# annotations from the test knit are answered without asking the basis
2198
2208
basis, test = self.get_basis_and_test_knit()
2200
key_basis = ('bar',)
2201
key_missing = ('missing',)
2202
test.add_lines(key, (), ['foo\n'])
2210
key_basis = (b'bar',)
2211
key_missing = (b'missing',)
2212
test.add_lines(key, (), [b'foo\n'])
2203
2213
details = test.annotate(key)
2204
self.assertEqual([(key, 'foo\n')], details)
2214
self.assertEqual([(key, b'foo\n')], details)
2205
2215
self.assertEqual([], basis.calls)
2206
2216
# But texts that are not in the test knit are looked for in the basis
2208
basis.add_lines(key_basis, (), ['foo\n', 'bar\n'])
2218
basis.add_lines(key_basis, (), [b'foo\n', b'bar\n'])
2209
2219
basis.calls = []
2210
2220
details = test.annotate(key_basis)
2211
self.assertEqual([(key_basis, 'foo\n'), (key_basis, 'bar\n')], details)
2221
self.assertEqual([(key_basis, b'foo\n'), (key_basis, b'bar\n')], details)
2212
2222
# Not optimised to date:
2213
2223
# self.assertEqual([("annotate", key_basis)], basis.calls)
2214
self.assertEqual([('get_parent_map', set([key_basis])),
2215
('get_parent_map', set([key_basis])),
2224
self.assertEqual([('get_parent_map', {key_basis}),
2225
('get_parent_map', {key_basis}),
2216
2226
('get_record_stream', [key_basis], 'topological', True)],
2323
2333
self.assertEqual(record.get_bytes_as('fulltext'), result[3])
2324
2334
# It's not strictly minimal, but it seems reasonable for now for it to
2325
2335
# ask which fallbacks have which parents.
2327
("get_parent_map", set([key_basis, key_basis_2, key_missing])),
2328
# topological is requested from the fallback, because that is what
2329
# was requested at the top level.
2330
("get_record_stream", [key_basis_2, key_basis], 'topological', True)],
2336
self.assertEqual(2, len(calls))
2338
("get_parent_map", {key_basis, key_basis_2, key_missing}),
2340
# topological is requested from the fallback, because that is what
2341
# was requested at the top level.
2344
("get_record_stream", [key_basis_2, key_basis], 'topological', True),
2345
("get_record_stream", [key_basis, key_basis_2], 'topological', True)])
2333
2347
def test_get_record_stream_unordered_deltas(self):
2334
2348
# records from the test knit are answered without asking the basis:
2335
2349
basis, test = self.get_basis_and_test_knit()
2337
key_basis = ('bar',)
2338
key_missing = ('missing',)
2339
test.add_lines(key, (), ['foo\n'])
2351
key_basis = (b'bar',)
2352
key_missing = (b'missing',)
2353
test.add_lines(key, (), [b'foo\n'])
2340
2354
records = list(test.get_record_stream([key], 'unordered', False))
2341
2355
self.assertEqual(1, len(records))
2342
2356
self.assertEqual([], basis.calls)
2343
2357
# Missing (from test knit) objects are retrieved from the basis:
2344
basis.add_lines(key_basis, (), ['foo\n', 'bar\n'])
2358
basis.add_lines(key_basis, (), [b'foo\n', b'bar\n'])
2345
2359
basis.calls = []
2346
2360
records = list(test.get_record_stream([key_basis, key_missing],
2347
2361
'unordered', False))
2530
2544
def test_make_mpdiffs(self):
2531
2545
# Generating an mpdiff across a stacking boundary should detect parent
2532
2546
# texts regions.
2535
key_right = ('zaphod',)
2548
key_left = (b'bar',)
2549
key_right = (b'zaphod',)
2536
2550
basis, test = self.get_basis_and_test_knit()
2537
basis.add_lines(key_left, (), ['bar\n'])
2538
basis.add_lines(key_right, (), ['zaphod\n'])
2551
basis.add_lines(key_left, (), [b'bar\n'])
2552
basis.add_lines(key_right, (), [b'zaphod\n'])
2539
2553
basis.calls = []
2540
2554
test.add_lines(key, (key_left, key_right),
2541
['bar\n', 'foo\n', 'zaphod\n'])
2555
[b'bar\n', b'foo\n', b'zaphod\n'])
2542
2556
diffs = test.make_mpdiffs([key])
2543
2557
self.assertEqual([
2544
2558
multiparent.MultiParent([multiparent.ParentText(0, 0, 0, 1),
2545
multiparent.NewText(['foo\n']),
2559
multiparent.NewText([b'foo\n']),
2546
2560
multiparent.ParentText(1, 0, 2, 1)])],
2548
2562
self.assertEqual(3, len(basis.calls))
2549
2563
self.assertEqual([
2550
("get_parent_map", set([key_left, key_right])),
2551
("get_parent_map", set([key_left, key_right])),
2564
("get_parent_map", {key_left, key_right}),
2565
("get_parent_map", {key_left, key_right}),
2553
2567
basis.calls[:-1])
2554
2568
last_call = basis.calls[-1]
2555
2569
self.assertEqual('get_record_stream', last_call[0])
2556
self.assertEqual(set([key_left, key_right]), set(last_call[1]))
2570
self.assertEqual({key_left, key_right}, set(last_call[1]))
2557
2571
self.assertEqual('topological', last_call[2])
2558
2572
self.assertEqual(True, last_call[3])
added
removed
breezy/tests/test_knit.py
