~bzr-pqm/bzr/bzr.dev

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
# Copyright (C) 2006-2010 Canonical Ltd
#
# Authors:
#   Johan Rydberg <jrydberg@gnu.org>
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

"""Versioned text file storage api."""

from copy import copy
from cStringIO import StringIO
import os
import struct
from zlib import adler32

from bzrlib.lazy_import import lazy_import
lazy_import(globals(), """
import urllib

from bzrlib import (
    annotate,
    errors,
    graph as _mod_graph,
    groupcompress,
    index,
    knit,
    osutils,
    multiparent,
    tsort,
    revision,
    ui,
    )
from bzrlib.graph import DictParentsProvider, Graph, StackedParentsProvider
from bzrlib.transport.memory import MemoryTransport
""")
from bzrlib.registry import Registry
from bzrlib.symbol_versioning import *
from bzrlib.textmerge import TextMerge
from bzrlib import bencode


adapter_registry = Registry()
adapter_registry.register_lazy(('knit-delta-gz', 'fulltext'), 'bzrlib.knit',
    'DeltaPlainToFullText')
adapter_registry.register_lazy(('knit-ft-gz', 'fulltext'), 'bzrlib.knit',
    'FTPlainToFullText')
adapter_registry.register_lazy(('knit-annotated-delta-gz', 'knit-delta-gz'),
    'bzrlib.knit', 'DeltaAnnotatedToUnannotated')
adapter_registry.register_lazy(('knit-annotated-delta-gz', 'fulltext'),
    'bzrlib.knit', 'DeltaAnnotatedToFullText')
adapter_registry.register_lazy(('knit-annotated-ft-gz', 'knit-ft-gz'),
    'bzrlib.knit', 'FTAnnotatedToUnannotated')
adapter_registry.register_lazy(('knit-annotated-ft-gz', 'fulltext'),
    'bzrlib.knit', 'FTAnnotatedToFullText')
# adapter_registry.register_lazy(('knit-annotated-ft-gz', 'chunked'),
#     'bzrlib.knit', 'FTAnnotatedToChunked')


class ContentFactory(object):
    """Abstract interface for insertion and retrieval from a VersionedFile.

    :ivar sha1: None, or the sha1 of the content fulltext.
    :ivar storage_kind: The native storage kind of this factory. One of
        'mpdiff', 'knit-annotated-ft', 'knit-annotated-delta', 'knit-ft',
        'knit-delta', 'fulltext', 'knit-annotated-ft-gz',
        'knit-annotated-delta-gz', 'knit-ft-gz', 'knit-delta-gz'.
    :ivar key: The key of this content. Each key is a tuple with a single
        string in it.
    :ivar parents: A tuple of parent keys for self.key. If the object has
        no parent information, None (as opposed to () for an empty list of
        parents).
    """

    def __init__(self):
        """Create a ContentFactory."""
        self.sha1 = None
        self.storage_kind = None
        self.key = None
        self.parents = None


class ChunkedContentFactory(ContentFactory):
    """Static data content factory.

    This takes a 'chunked' list of strings. The only requirement on 'chunked' is
    that ''.join(lines) becomes a valid fulltext. A tuple of a single string
    satisfies this, as does a list of lines.

    :ivar sha1: None, or the sha1 of the content fulltext.
    :ivar storage_kind: The native storage kind of this factory. Always
        'chunked'
    :ivar key: The key of this content. Each key is a tuple with a single
        string in it.
    :ivar parents: A tuple of parent keys for self.key. If the object has
        no parent information, None (as opposed to () for an empty list of
        parents).
     """

    def __init__(self, key, parents, sha1, chunks):
        """Create a ContentFactory."""
        self.sha1 = sha1
        self.storage_kind = 'chunked'
        self.key = key
        self.parents = parents
        self._chunks = chunks

    def get_bytes_as(self, storage_kind):
        if storage_kind == 'chunked':
            return self._chunks
        elif storage_kind == 'fulltext':
            return ''.join(self._chunks)
        raise errors.UnavailableRepresentation(self.key, storage_kind,
            self.storage_kind)


class FulltextContentFactory(ContentFactory):
    """Static data content factory.

    This takes a fulltext when created and just returns that during
    get_bytes_as('fulltext').

    :ivar sha1: None, or the sha1 of the content fulltext.
    :ivar storage_kind: The native storage kind of this factory. Always
        'fulltext'.
    :ivar key: The key of this content. Each key is a tuple with a single
        string in it.
    :ivar parents: A tuple of parent keys for self.key. If the object has
        no parent information, None (as opposed to () for an empty list of
        parents).
     """

    def __init__(self, key, parents, sha1, text):
        """Create a ContentFactory."""
        self.sha1 = sha1
        self.storage_kind = 'fulltext'
        self.key = key
        self.parents = parents
        self._text = text

    def get_bytes_as(self, storage_kind):
        if storage_kind == self.storage_kind:
            return self._text
        elif storage_kind == 'chunked':
            return [self._text]
        raise errors.UnavailableRepresentation(self.key, storage_kind,
            self.storage_kind)


class AbsentContentFactory(ContentFactory):
    """A placeholder content factory for unavailable texts.

    :ivar sha1: None.
    :ivar storage_kind: 'absent'.
    :ivar key: The key of this content. Each key is a tuple with a single
        string in it.
    :ivar parents: None.
    """

    def __init__(self, key):
        """Create a ContentFactory."""
        self.sha1 = None
        self.storage_kind = 'absent'
        self.key = key
        self.parents = None

    def get_bytes_as(self, storage_kind):
        raise ValueError('A request was made for key: %s, but that'
                         ' content is not available, and the calling'
                         ' code does not handle if it is missing.'
                         % (self.key,))


class AdapterFactory(ContentFactory):
    """A content factory to adapt between key prefix's."""

    def __init__(self, key, parents, adapted):
        """Create an adapter factory instance."""
        self.key = key
        self.parents = parents
        self._adapted = adapted

    def __getattr__(self, attr):
        """Return a member from the adapted object."""
        if attr in ('key', 'parents'):
            return self.__dict__[attr]
        else:
            return getattr(self._adapted, attr)


def filter_absent(record_stream):
    """Adapt a record stream to remove absent records."""
    for record in record_stream:
        if record.storage_kind != 'absent':
            yield record


class VersionedFile(object):
    """Versioned text file storage.

    A versioned file manages versions of line-based text files,
    keeping track of the originating version for each line.

    To clients the "lines" of the file are represented as a list of
    strings. These strings will typically have terminal newline
    characters, but this is not required.  In particular files commonly
    do not have a newline at the end of the file.

    Texts are identified by a version-id string.
    """

    @staticmethod
    def check_not_reserved_id(version_id):
        revision.check_not_reserved_id(version_id)

    def copy_to(self, name, transport):
        """Copy this versioned file to name on transport."""
        raise NotImplementedError(self.copy_to)

    def get_record_stream(self, versions, ordering, include_delta_closure):
        """Get a stream of records for versions.

        :param versions: The versions to include. Each version is a tuple
            (version,).
        :param ordering: Either 'unordered' or 'topological'. A topologically
            sorted stream has compression parents strictly before their
            children.
        :param include_delta_closure: If True then the closure across any
            compression parents will be included (in the data content of the
            stream, not in the emitted records). This guarantees that
            'fulltext' can be used successfully on every record.
        :return: An iterator of ContentFactory objects, each of which is only
            valid until the iterator is advanced.
        """
        raise NotImplementedError(self.get_record_stream)

    def has_version(self, version_id):
        """Returns whether version is present."""
        raise NotImplementedError(self.has_version)

    def insert_record_stream(self, stream):
        """Insert a record stream into this versioned file.

        :param stream: A stream of records to insert.
        :return: None
        :seealso VersionedFile.get_record_stream:
        """
        raise NotImplementedError

    def add_lines(self, version_id, parents, lines, parent_texts=None,
        left_matching_blocks=None, nostore_sha=None, random_id=False,
        check_content=True):
        """Add a single text on top of the versioned file.

        Must raise RevisionAlreadyPresent if the new version is
        already present in file history.

        Must raise RevisionNotPresent if any of the given parents are
        not present in file history.

        :param lines: A list of lines. Each line must be a bytestring. And all
            of them except the last must be terminated with \n and contain no
            other \n's. The last line may either contain no \n's or a single
            terminated \n. If the lines list does meet this constraint the add
            routine may error or may succeed - but you will be unable to read
            the data back accurately. (Checking the lines have been split
            correctly is expensive and extremely unlikely to catch bugs so it
            is not done at runtime unless check_content is True.)
        :param parent_texts: An optional dictionary containing the opaque
            representations of some or all of the parents of version_id to
            allow delta optimisations.  VERY IMPORTANT: the texts must be those
            returned by add_lines or data corruption can be caused.
        :param left_matching_blocks: a hint about which areas are common
            between the text and its left-hand-parent.  The format is
            the SequenceMatcher.get_matching_blocks format.
        :param nostore_sha: Raise ExistingContent and do not add the lines to
            the versioned file if the digest of the lines matches this.
        :param random_id: If True a random id has been selected rather than
            an id determined by some deterministic process such as a converter
            from a foreign VCS. When True the backend may choose not to check
            for uniqueness of the resulting key within the versioned file, so
            this should only be done when the result is expected to be unique
            anyway.
        :param check_content: If True, the lines supplied are verified to be
            bytestrings that are correctly formed lines.
        :return: The text sha1, the number of bytes in the text, and an opaque
                 representation of the inserted version which can be provided
                 back to future add_lines calls in the parent_texts dictionary.
        """
        self._check_write_ok()
        return self._add_lines(version_id, parents, lines, parent_texts,
            left_matching_blocks, nostore_sha, random_id, check_content)

    def _add_lines(self, version_id, parents, lines, parent_texts,
        left_matching_blocks, nostore_sha, random_id, check_content):
        """Helper to do the class specific add_lines."""
        raise NotImplementedError(self.add_lines)

    def add_lines_with_ghosts(self, version_id, parents, lines,
        parent_texts=None, nostore_sha=None, random_id=False,
        check_content=True, left_matching_blocks=None):
        """Add lines to the versioned file, allowing ghosts to be present.

        This takes the same parameters as add_lines and returns the same.
        """
        self._check_write_ok()
        return self._add_lines_with_ghosts(version_id, parents, lines,
            parent_texts, nostore_sha, random_id, check_content, left_matching_blocks)

    def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts,
        nostore_sha, random_id, check_content, left_matching_blocks):
        """Helper to do class specific add_lines_with_ghosts."""
        raise NotImplementedError(self.add_lines_with_ghosts)

    def check(self, progress_bar=None):
        """Check the versioned file for integrity."""
        raise NotImplementedError(self.check)

    def _check_lines_not_unicode(self, lines):
        """Check that lines being added to a versioned file are not unicode."""
        for line in lines:
            if line.__class__ is not str:
                raise errors.BzrBadParameterUnicode("lines")

    def _check_lines_are_lines(self, lines):
        """Check that the lines really are full lines without inline EOL."""
        for line in lines:
            if '\n' in line[:-1]:
                raise errors.BzrBadParameterContainsNewline("lines")

    def get_format_signature(self):
        """Get a text description of the data encoding in this file.

        :since: 0.90
        """
        raise NotImplementedError(self.get_format_signature)

    def make_mpdiffs(self, version_ids):
        """Create multiparent diffs for specified versions."""
        knit_versions = set()
        knit_versions.update(version_ids)
        parent_map = self.get_parent_map(version_ids)
        for version_id in version_ids:
            try:
                knit_versions.update(parent_map[version_id])
            except KeyError:
                raise errors.RevisionNotPresent(version_id, self)
        # We need to filter out ghosts, because we can't diff against them.
        knit_versions = set(self.get_parent_map(knit_versions).keys())
        lines = dict(zip(knit_versions,
            self._get_lf_split_line_list(knit_versions)))
        diffs = []
        for version_id in version_ids:
            target = lines[version_id]
            try:
                parents = [lines[p] for p in parent_map[version_id] if p in
                    knit_versions]
            except KeyError:
                # I don't know how this could ever trigger.
                # parent_map[version_id] was already triggered in the previous
                # for loop, and lines[p] has the 'if p in knit_versions' check,
                # so we again won't have a KeyError.
                raise errors.RevisionNotPresent(version_id, self)
            if len(parents) > 0:
                left_parent_blocks = self._extract_blocks(version_id,
                                                          parents[0], target)
            else:
                left_parent_blocks = None
            diffs.append(multiparent.MultiParent.from_lines(target, parents,
                         left_parent_blocks))
        return diffs

    def _extract_blocks(self, version_id, source, target):
        return None

    def add_mpdiffs(self, records):
        """Add mpdiffs to this VersionedFile.

        Records should be iterables of version, parents, expected_sha1,
        mpdiff. mpdiff should be a MultiParent instance.
        """
        # Does this need to call self._check_write_ok()? (IanC 20070919)
        vf_parents = {}
        mpvf = multiparent.MultiMemoryVersionedFile()
        versions = []
        for version, parent_ids, expected_sha1, mpdiff in records:
            versions.append(version)
            mpvf.add_diff(mpdiff, version, parent_ids)
        needed_parents = set()
        for version, parent_ids, expected_sha1, mpdiff in records:
            needed_parents.update(p for p in parent_ids
                                  if not mpvf.has_version(p))
        present_parents = set(self.get_parent_map(needed_parents).keys())
        for parent_id, lines in zip(present_parents,
                                 self._get_lf_split_line_list(present_parents)):
            mpvf.add_version(lines, parent_id, [])
        for (version, parent_ids, expected_sha1, mpdiff), lines in\
            zip(records, mpvf.get_line_list(versions)):
            if len(parent_ids) == 1:
                left_matching_blocks = list(mpdiff.get_matching_blocks(0,
                    mpvf.get_diff(parent_ids[0]).num_lines()))
            else:
                left_matching_blocks = None
            try:
                _, _, version_text = self.add_lines_with_ghosts(version,
                    parent_ids, lines, vf_parents,
                    left_matching_blocks=left_matching_blocks)
            except NotImplementedError:
                # The vf can't handle ghosts, so add lines normally, which will
                # (reasonably) fail if there are ghosts in the data.
                _, _, version_text = self.add_lines(version,
                    parent_ids, lines, vf_parents,
                    left_matching_blocks=left_matching_blocks)
            vf_parents[version] = version_text
        sha1s = self.get_sha1s(versions)
        for version, parent_ids, expected_sha1, mpdiff in records:
            if expected_sha1 != sha1s[version]:
                raise errors.VersionedFileInvalidChecksum(version)

    def get_text(self, version_id):
        """Return version contents as a text string.

        Raises RevisionNotPresent if version is not present in
        file history.
        """
        return ''.join(self.get_lines(version_id))
    get_string = get_text

    def get_texts(self, version_ids):
        """Return the texts of listed versions as a list of strings.

        Raises RevisionNotPresent if version is not present in
        file history.
        """
        return [''.join(self.get_lines(v)) for v in version_ids]

    def get_lines(self, version_id):
        """Return version contents as a sequence of lines.

        Raises RevisionNotPresent if version is not present in
        file history.
        """
        raise NotImplementedError(self.get_lines)

    def _get_lf_split_line_list(self, version_ids):
        return [StringIO(t).readlines() for t in self.get_texts(version_ids)]

    def get_ancestry(self, version_ids, topo_sorted=True):
        """Return a list of all ancestors of given version(s). This
        will not include the null revision.

        This list will not be topologically sorted if topo_sorted=False is
        passed.

        Must raise RevisionNotPresent if any of the given versions are
        not present in file history."""
        if isinstance(version_ids, basestring):
            version_ids = [version_ids]
        raise NotImplementedError(self.get_ancestry)

    def get_ancestry_with_ghosts(self, version_ids):
        """Return a list of all ancestors of given version(s). This
        will not include the null revision.

        Must raise RevisionNotPresent if any of the given versions are
        not present in file history.

        Ghosts that are known about will be included in ancestry list,
        but are not explicitly marked.
        """
        raise NotImplementedError(self.get_ancestry_with_ghosts)

    def get_parent_map(self, version_ids):
        """Get a map of the parents of version_ids.

        :param version_ids: The version ids to look up parents for.
        :return: A mapping from version id to parents.
        """
        raise NotImplementedError(self.get_parent_map)

    def get_parents_with_ghosts(self, version_id):
        """Return version names for parents of version_id.

        Will raise RevisionNotPresent if version_id is not present
        in the history.

        Ghosts that are known about will be included in the parent list,
        but are not explicitly marked.
        """
        try:
            return list(self.get_parent_map([version_id])[version_id])
        except KeyError:
            raise errors.RevisionNotPresent(version_id, self)

    def annotate(self, version_id):
        """Return a list of (version-id, line) tuples for version_id.

        :raise RevisionNotPresent: If the given version is
        not present in file history.
        """
        raise NotImplementedError(self.annotate)

    def iter_lines_added_or_present_in_versions(self, version_ids=None,
                                                pb=None):
        """Iterate over the lines in the versioned file from version_ids.

        This may return lines from other versions. Each item the returned
        iterator yields is a tuple of a line and a text version that that line
        is present in (not introduced in).

        Ordering of results is in whatever order is most suitable for the
        underlying storage format.

        If a progress bar is supplied, it may be used to indicate progress.
        The caller is responsible for cleaning up progress bars (because this
        is an iterator).

        NOTES: Lines are normalised: they will all have \n terminators.
               Lines are returned in arbitrary order.

        :return: An iterator over (line, version_id).
        """
        raise NotImplementedError(self.iter_lines_added_or_present_in_versions)

    def plan_merge(self, ver_a, ver_b):
        """Return pseudo-annotation indicating how the two versions merge.

        This is computed between versions a and b and their common
        base.

        Weave lines present in none of them are skipped entirely.

        Legend:
        killed-base Dead in base revision
        killed-both Killed in each revision
        killed-a    Killed in a
        killed-b    Killed in b
        unchanged   Alive in both a and b (possibly created in both)
        new-a       Created in a
        new-b       Created in b
        ghost-a     Killed in a, unborn in b
        ghost-b     Killed in b, unborn in a
        irrelevant  Not in either revision
        """
        raise NotImplementedError(VersionedFile.plan_merge)

    def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
                    b_marker=TextMerge.B_MARKER):
        return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]


class RecordingVersionedFilesDecorator(object):
    """A minimal versioned files that records calls made on it.

    Only enough methods have been added to support tests using it to date.

    :ivar calls: A list of the calls made; can be reset at any time by
        assigning [] to it.
    """

    def __init__(self, backing_vf):
        """Create a RecordingVersionedFilesDecorator decorating backing_vf.

        :param backing_vf: The versioned file to answer all methods.
        """
        self._backing_vf = backing_vf
        self.calls = []

    def add_lines(self, key, parents, lines, parent_texts=None,
        left_matching_blocks=None, nostore_sha=None, random_id=False,
        check_content=True):
        self.calls.append(("add_lines", key, parents, lines, parent_texts,
            left_matching_blocks, nostore_sha, random_id, check_content))
        return self._backing_vf.add_lines(key, parents, lines, parent_texts,
            left_matching_blocks, nostore_sha, random_id, check_content)

    def check(self):
        self._backing_vf.check()

    def get_parent_map(self, keys):
        self.calls.append(("get_parent_map", copy(keys)))
        return self._backing_vf.get_parent_map(keys)

    def get_record_stream(self, keys, sort_order, include_delta_closure):
        self.calls.append(("get_record_stream", list(keys), sort_order,
            include_delta_closure))
        return self._backing_vf.get_record_stream(keys, sort_order,
            include_delta_closure)

    def get_sha1s(self, keys):
        self.calls.append(("get_sha1s", copy(keys)))
        return self._backing_vf.get_sha1s(keys)

    def iter_lines_added_or_present_in_keys(self, keys, pb=None):
        self.calls.append(("iter_lines_added_or_present_in_keys", copy(keys)))
        return self._backing_vf.iter_lines_added_or_present_in_keys(keys, pb=pb)

    def keys(self):
        self.calls.append(("keys",))
        return self._backing_vf.keys()


class OrderingVersionedFilesDecorator(RecordingVersionedFilesDecorator):
    """A VF that records calls, and returns keys in specific order.

    :ivar calls: A list of the calls made; can be reset at any time by
        assigning [] to it.
    """

    def __init__(self, backing_vf, key_priority):
        """Create a RecordingVersionedFilesDecorator decorating backing_vf.

        :param backing_vf: The versioned file to answer all methods.
        :param key_priority: A dictionary defining what order keys should be
            returned from an 'unordered' get_record_stream request.
            Keys with lower priority are returned first, keys not present in
            the map get an implicit priority of 0, and are returned in
            lexicographical order.
        """
        RecordingVersionedFilesDecorator.__init__(self, backing_vf)
        self._key_priority = key_priority

    def get_record_stream(self, keys, sort_order, include_delta_closure):
        self.calls.append(("get_record_stream", list(keys), sort_order,
            include_delta_closure))
        if sort_order == 'unordered':
            def sort_key(key):
                return (self._key_priority.get(key, 0), key)
            # Use a defined order by asking for the keys one-by-one from the
            # backing_vf
            for key in sorted(keys, key=sort_key):
                for record in self._backing_vf.get_record_stream([key],
                                'unordered', include_delta_closure):
                    yield record
        else:
            for record in self._backing_vf.get_record_stream(keys, sort_order,
                            include_delta_closure):
                yield record


class KeyMapper(object):
    """KeyMappers map between keys and underlying partitioned storage."""

    def map(self, key):
        """Map key to an underlying storage identifier.

        :param key: A key tuple e.g. ('file-id', 'revision-id').
        :return: An underlying storage identifier, specific to the partitioning
            mechanism.
        """
        raise NotImplementedError(self.map)

    def unmap(self, partition_id):
        """Map a partitioned storage id back to a key prefix.

        :param partition_id: The underlying partition id.
        :return: As much of a key (or prefix) as is derivable from the partition
            id.
        """
        raise NotImplementedError(self.unmap)


class ConstantMapper(KeyMapper):
    """A key mapper that maps to a constant result."""

    def __init__(self, result):
        """Create a ConstantMapper which will return result for all maps."""
        self._result = result

    def map(self, key):
        """See KeyMapper.map()."""
        return self._result


class URLEscapeMapper(KeyMapper):
    """Base class for use with transport backed storage.

    This provides a map and unmap wrapper that respectively url escape and
    unescape their outputs and inputs.
    """

    def map(self, key):
        """See KeyMapper.map()."""
        return urllib.quote(self._map(key))

    def unmap(self, partition_id):
        """See KeyMapper.unmap()."""
        return self._unmap(urllib.unquote(partition_id))


class PrefixMapper(URLEscapeMapper):
    """A key mapper that extracts the first component of a key.

    This mapper is for use with a transport based backend.
    """

    def _map(self, key):
        """See KeyMapper.map()."""
        return key[0]

    def _unmap(self, partition_id):
        """See KeyMapper.unmap()."""
        return (partition_id,)


class HashPrefixMapper(URLEscapeMapper):
    """A key mapper that combines the first component of a key with a hash.

    This mapper is for use with a transport based backend.
    """

    def _map(self, key):
        """See KeyMapper.map()."""
        prefix = self._escape(key[0])
        return "%02x/%s" % (adler32(prefix) & 0xff, prefix)

    def _escape(self, prefix):
        """No escaping needed here."""
        return prefix

    def _unmap(self, partition_id):
        """See KeyMapper.unmap()."""
        return (self._unescape(osutils.basename(partition_id)),)

    def _unescape(self, basename):
        """No unescaping needed for HashPrefixMapper."""
        return basename


class HashEscapedPrefixMapper(HashPrefixMapper):
    """Combines the escaped first component of a key with a hash.

    This mapper is for use with a transport based backend.
    """

    _safe = "abcdefghijklmnopqrstuvwxyz0123456789-_@,."

    def _escape(self, prefix):
        """Turn a key element into a filesystem safe string.

        This is similar to a plain urllib.quote, except
        it uses specific safe characters, so that it doesn't
        have to translate a lot of valid file ids.
        """
        # @ does not get escaped. This is because it is a valid
        # filesystem character we use all the time, and it looks
        # a lot better than seeing %40 all the time.
        r = [((c in self._safe) and c or ('%%%02x' % ord(c)))
             for c in prefix]
        return ''.join(r)

    def _unescape(self, basename):
        """Escaped names are easily unescaped by urlutils."""
        return urllib.unquote(basename)


def make_versioned_files_factory(versioned_file_factory, mapper):
    """Create a ThunkedVersionedFiles factory.

    This will create a callable which when called creates a
    ThunkedVersionedFiles on a transport, using mapper to access individual
    versioned files, and versioned_file_factory to create each individual file.
    """
    def factory(transport):
        return ThunkedVersionedFiles(transport, versioned_file_factory, mapper,
            lambda:True)
    return factory


class VersionedFiles(object):
    """Storage for many versioned files.

    This object allows a single keyspace for accessing the history graph and
    contents of named bytestrings.

    Currently no implementation allows the graph of different key prefixes to
    intersect, but the API does allow such implementations in the future.

    The keyspace is expressed via simple tuples. Any instance of VersionedFiles
    may have a different length key-size, but that size will be constant for
    all texts added to or retrieved from it. For instance, bzrlib uses
    instances with a key-size of 2 for storing user files in a repository, with
    the first element the fileid, and the second the version of that file.

    The use of tuples allows a single code base to support several different
    uses with only the mapping logic changing from instance to instance.
    """

    def add_lines(self, key, parents, lines, parent_texts=None,
        left_matching_blocks=None, nostore_sha=None, random_id=False,
        check_content=True):
        """Add a text to the store.

        :param key: The key tuple of the text to add. If the last element is
            None, a CHK string will be generated during the addition.
        :param parents: The parents key tuples of the text to add.
        :param lines: A list of lines. Each line must be a bytestring. And all
            of them except the last must be terminated with \n and contain no
            other \n's. The last line may either contain no \n's or a single
            terminating \n. If the lines list does meet this constraint the add
            routine may error or may succeed - but you will be unable to read
            the data back accurately. (Checking the lines have been split
            correctly is expensive and extremely unlikely to catch bugs so it
            is not done at runtime unless check_content is True.)
        :param parent_texts: An optional dictionary containing the opaque
            representations of some or all of the parents of version_id to
            allow delta optimisations.  VERY IMPORTANT: the texts must be those
            returned by add_lines or data corruption can be caused.
        :param left_matching_blocks: a hint about which areas are common
            between the text and its left-hand-parent.  The format is
            the SequenceMatcher.get_matching_blocks format.
        :param nostore_sha: Raise ExistingContent and do not add the lines to
            the versioned file if the digest of the lines matches this.
        :param random_id: If True a random id has been selected rather than
            an id determined by some deterministic process such as a converter
            from a foreign VCS. When True the backend may choose not to check
            for uniqueness of the resulting key within the versioned file, so
            this should only be done when the result is expected to be unique
            anyway.
        :param check_content: If True, the lines supplied are verified to be
            bytestrings that are correctly formed lines.
        :return: The text sha1, the number of bytes in the text, and an opaque
                 representation of the inserted version which can be provided
                 back to future add_lines calls in the parent_texts dictionary.
        """
        raise NotImplementedError(self.add_lines)

    def _add_text(self, key, parents, text, nostore_sha=None, random_id=False):
        """Add a text to the store.

        This is a private function for use by CommitBuilder.

        :param key: The key tuple of the text to add. If the last element is
            None, a CHK string will be generated during the addition.
        :param parents: The parents key tuples of the text to add.
        :param text: A string containing the text to be committed.
        :param nostore_sha: Raise ExistingContent and do not add the lines to
            the versioned file if the digest of the lines matches this.
        :param random_id: If True a random id has been selected rather than
            an id determined by some deterministic process such as a converter
            from a foreign VCS. When True the backend may choose not to check
            for uniqueness of the resulting key within the versioned file, so
            this should only be done when the result is expected to be unique
            anyway.
        :param check_content: If True, the lines supplied are verified to be
            bytestrings that are correctly formed lines.
        :return: The text sha1, the number of bytes in the text, and an opaque
                 representation of the inserted version which can be provided
                 back to future _add_text calls in the parent_texts dictionary.
        """
        # The default implementation just thunks over to .add_lines(),
        # inefficient, but it works.
        return self.add_lines(key, parents, osutils.split_lines(text),
                              nostore_sha=nostore_sha,
                              random_id=random_id,
                              check_content=True)

    def add_mpdiffs(self, records):
        """Add mpdiffs to this VersionedFile.

        Records should be iterables of version, parents, expected_sha1,
        mpdiff. mpdiff should be a MultiParent instance.
        """
        vf_parents = {}
        mpvf = multiparent.MultiMemoryVersionedFile()
        versions = []
        for version, parent_ids, expected_sha1, mpdiff in records:
            versions.append(version)
            mpvf.add_diff(mpdiff, version, parent_ids)
        needed_parents = set()
        for version, parent_ids, expected_sha1, mpdiff in records:
            needed_parents.update(p for p in parent_ids
                                  if not mpvf.has_version(p))
        # It seems likely that adding all the present parents as fulltexts can
        # easily exhaust memory.
        chunks_to_lines = osutils.chunks_to_lines
        for record in self.get_record_stream(needed_parents, 'unordered',
            True):
            if record.storage_kind == 'absent':
                continue
            mpvf.add_version(chunks_to_lines(record.get_bytes_as('chunked')),
                record.key, [])
        for (key, parent_keys, expected_sha1, mpdiff), lines in\
            zip(records, mpvf.get_line_list(versions)):
            if len(parent_keys) == 1:
                left_matching_blocks = list(mpdiff.get_matching_blocks(0,
                    mpvf.get_diff(parent_keys[0]).num_lines()))
            else:
                left_matching_blocks = None
            version_sha1, _, version_text = self.add_lines(key,
                parent_keys, lines, vf_parents,
                left_matching_blocks=left_matching_blocks)
            if version_sha1 != expected_sha1:
                raise errors.VersionedFileInvalidChecksum(version)
            vf_parents[key] = version_text

    def annotate(self, key):
        """Return a list of (version-key, line) tuples for the text of key.

        :raise RevisionNotPresent: If the key is not present.
        """
        raise NotImplementedError(self.annotate)

    def check(self, progress_bar=None):
        """Check this object for integrity.
        
        :param progress_bar: A progress bar to output as the check progresses.
        :param keys: Specific keys within the VersionedFiles to check. When
            this parameter is not None, check() becomes a generator as per
            get_record_stream. The difference to get_record_stream is that
            more or deeper checks will be performed.
        :return: None, or if keys was supplied a generator as per
            get_record_stream.
        """
        raise NotImplementedError(self.check)

    @staticmethod
    def check_not_reserved_id(version_id):
        revision.check_not_reserved_id(version_id)

    def clear_cache(self):
        """Clear whatever caches this VersionedFile holds.

        This is generally called after an operation has been performed, when we
        don't expect to be using this versioned file again soon.
        """

    def _check_lines_not_unicode(self, lines):
        """Check that lines being added to a versioned file are not unicode."""
        for line in lines:
            if line.__class__ is not str:
                raise errors.BzrBadParameterUnicode("lines")

    def _check_lines_are_lines(self, lines):
        """Check that the lines really are full lines without inline EOL."""
        for line in lines:
            if '\n' in line[:-1]:
                raise errors.BzrBadParameterContainsNewline("lines")

    def get_known_graph_ancestry(self, keys):
        """Get a KnownGraph instance with the ancestry of keys."""
        # most basic implementation is a loop around get_parent_map
        pending = set(keys)
        parent_map = {}
        while pending:
            this_parent_map = self.get_parent_map(pending)
            parent_map.update(this_parent_map)
            pending = set()
            map(pending.update, this_parent_map.itervalues())
            pending = pending.difference(parent_map)
        kg = _mod_graph.KnownGraph(parent_map)
        return kg

    def get_parent_map(self, keys):
        """Get a map of the parents of keys.

        :param keys: The keys to look up parents for.
        :return: A mapping from keys to parents. Absent keys are absent from
            the mapping.
        """
        raise NotImplementedError(self.get_parent_map)

    def get_record_stream(self, keys, ordering, include_delta_closure):
        """Get a stream of records for keys.

        :param keys: The keys to include.
        :param ordering: Either 'unordered' or 'topological'. A topologically
            sorted stream has compression parents strictly before their
            children.
        :param include_delta_closure: If True then the closure across any
            compression parents will be included (in the opaque data).
        :return: An iterator of ContentFactory objects, each of which is only
            valid until the iterator is advanced.
        """
        raise NotImplementedError(self.get_record_stream)

    def get_sha1s(self, keys):
        """Get the sha1's of the texts for the given keys.

        :param keys: The names of the keys to lookup
        :return: a dict from key to sha1 digest. Keys of texts which are not
            present in the store are not present in the returned
            dictionary.
        """
        raise NotImplementedError(self.get_sha1s)

    has_key = index._has_key_from_parent_map

    def get_missing_compression_parent_keys(self):
        """Return an iterable of keys of missing compression parents.

        Check this after calling insert_record_stream to find out if there are
        any missing compression parents.  If there are, the records that
        depend on them are not able to be inserted safely. The precise
        behaviour depends on the concrete VersionedFiles class in use.

        Classes that do not support this will raise NotImplementedError.
        """
        raise NotImplementedError(self.get_missing_compression_parent_keys)

    def insert_record_stream(self, stream):
        """Insert a record stream into this container.

        :param stream: A stream of records to insert.
        :return: None
        :seealso VersionedFile.get_record_stream:
        """
        raise NotImplementedError

    def iter_lines_added_or_present_in_keys(self, keys, pb=None):
        """Iterate over the lines in the versioned files from keys.

        This may return lines from other keys. Each item the returned
        iterator yields is a tuple of a line and a text version that that line
        is present in (not introduced in).

        Ordering of results is in whatever order is most suitable for the
        underlying storage format.

        If a progress bar is supplied, it may be used to indicate progress.
        The caller is responsible for cleaning up progress bars (because this
        is an iterator).

        NOTES:
         * Lines are normalised by the underlying store: they will all have \n
           terminators.
         * Lines are returned in arbitrary order.

        :return: An iterator over (line, key).
        """
        raise NotImplementedError(self.iter_lines_added_or_present_in_keys)

    def keys(self):
        """Return a iterable of the keys for all the contained texts."""
        raise NotImplementedError(self.keys)

    def make_mpdiffs(self, keys):
        """Create multiparent diffs for specified keys."""
        keys_order = tuple(keys)
        keys = frozenset(keys)
        knit_keys = set(keys)
        parent_map = self.get_parent_map(keys)
        for parent_keys in parent_map.itervalues():
            if parent_keys:
                knit_keys.update(parent_keys)
        missing_keys = keys - set(parent_map)
        if missing_keys:
            raise errors.RevisionNotPresent(list(missing_keys)[0], self)
        # We need to filter out ghosts, because we can't diff against them.
        maybe_ghosts = knit_keys - keys
        ghosts = maybe_ghosts - set(self.get_parent_map(maybe_ghosts))
        knit_keys.difference_update(ghosts)
        lines = {}
        chunks_to_lines = osutils.chunks_to_lines
        for record in self.get_record_stream(knit_keys, 'topological', True):
            lines[record.key] = chunks_to_lines(record.get_bytes_as('chunked'))
            # line_block_dict = {}
            # for parent, blocks in record.extract_line_blocks():
            #   line_blocks[parent] = blocks
            # line_blocks[record.key] = line_block_dict
        diffs = []
        for key in keys_order:
            target = lines[key]
            parents = parent_map[key] or []
            # Note that filtering knit_keys can lead to a parent difference
            # between the creation and the application of the mpdiff.
            parent_lines = [lines[p] for p in parents if p in knit_keys]
            if len(parent_lines) > 0:
                left_parent_blocks = self._extract_blocks(key, parent_lines[0],
                    target)
            else:
                left_parent_blocks = None
            diffs.append(multiparent.MultiParent.from_lines(target,
                parent_lines, left_parent_blocks))
        return diffs

    missing_keys = index._missing_keys_from_parent_map

    def _extract_blocks(self, version_id, source, target):
        return None


class ThunkedVersionedFiles(VersionedFiles):
    """Storage for many versioned files thunked onto a 'VersionedFile' class.

    This object allows a single keyspace for accessing the history graph and
    contents of named bytestrings.

    Currently no implementation allows the graph of different key prefixes to
    intersect, but the API does allow such implementations in the future.
    """

    def __init__(self, transport, file_factory, mapper, is_locked):
        """Create a ThunkedVersionedFiles."""
        self._transport = transport
        self._file_factory = file_factory
        self._mapper = mapper
        self._is_locked = is_locked

    def add_lines(self, key, parents, lines, parent_texts=None,
        left_matching_blocks=None, nostore_sha=None, random_id=False,
        check_content=True):
        """See VersionedFiles.add_lines()."""
        path = self._mapper.map(key)
        version_id = key[-1]
        parents = [parent[-1] for parent in parents]
        vf = self._get_vf(path)
        try:
            try:
                return vf.add_lines_with_ghosts(version_id, parents, lines,
                    parent_texts=parent_texts,
                    left_matching_blocks=left_matching_blocks,
                    nostore_sha=nostore_sha, random_id=random_id,
                    check_content=check_content)
            except NotImplementedError:
                return vf.add_lines(version_id, parents, lines,
                    parent_texts=parent_texts,
                    left_matching_blocks=left_matching_blocks,
                    nostore_sha=nostore_sha, random_id=random_id,
                    check_content=check_content)
        except errors.NoSuchFile:
            # parent directory may be missing, try again.
            self._transport.mkdir(osutils.dirname(path))
            try:
                return vf.add_lines_with_ghosts(version_id, parents, lines,
                    parent_texts=parent_texts,
                    left_matching_blocks=left_matching_blocks,
                    nostore_sha=nostore_sha, random_id=random_id,
                    check_content=check_content)
            except NotImplementedError:
                return vf.add_lines(version_id, parents, lines,
                    parent_texts=parent_texts,
                    left_matching_blocks=left_matching_blocks,
                    nostore_sha=nostore_sha, random_id=random_id,
                    check_content=check_content)

    def annotate(self, key):
        """Return a list of (version-key, line) tuples for the text of key.

        :raise RevisionNotPresent: If the key is not present.
        """
        prefix = key[:-1]
        path = self._mapper.map(prefix)
        vf = self._get_vf(path)
        origins = vf.annotate(key[-1])
        result = []
        for origin, line in origins:
            result.append((prefix + (origin,), line))
        return result

    def get_annotator(self):
        return annotate.Annotator(self)

    def check(self, progress_bar=None, keys=None):
        """See VersionedFiles.check()."""
        # XXX: This is over-enthusiastic but as we only thunk for Weaves today
        # this is tolerable. Ideally we'd pass keys down to check() and 
        # have the older VersiondFile interface updated too.
        for prefix, vf in self._iter_all_components():
            vf.check()
        if keys is not None:
            return self.get_record_stream(keys, 'unordered', True)

    def get_parent_map(self, keys):
        """Get a map of the parents of keys.

        :param keys: The keys to look up parents for.
        :return: A mapping from keys to parents. Absent keys are absent from
            the mapping.
        """
        prefixes = self._partition_keys(keys)
        result = {}
        for prefix, suffixes in prefixes.items():
            path = self._mapper.map(prefix)
            vf = self._get_vf(path)
            parent_map = vf.get_parent_map(suffixes)
            for key, parents in parent_map.items():
                result[prefix + (key,)] = tuple(
                    prefix + (parent,) for parent in parents)
        return result

    def _get_vf(self, path):
        if not self._is_locked():
            raise errors.ObjectNotLocked(self)
        return self._file_factory(path, self._transport, create=True,
            get_scope=lambda:None)

    def _partition_keys(self, keys):
        """Turn keys into a dict of prefix:suffix_list."""
        result = {}
        for key in keys:
            prefix_keys = result.setdefault(key[:-1], [])
            prefix_keys.append(key[-1])
        return result

    def _get_all_prefixes(self):
        # Identify all key prefixes.
        # XXX: A bit hacky, needs polish.
        if type(self._mapper) == ConstantMapper:
            paths = [self._mapper.map(())]
            prefixes = [()]
        else:
            relpaths = set()
            for quoted_relpath in self._transport.iter_files_recursive():
                path, ext = os.path.splitext(quoted_relpath)
                relpaths.add(path)
            paths = list(relpaths)
            prefixes = [self._mapper.unmap(path) for path in paths]
        return zip(paths, prefixes)

    def get_record_stream(self, keys, ordering, include_delta_closure):
        """See VersionedFiles.get_record_stream()."""
        # Ordering will be taken care of by each partitioned store; group keys
        # by partition.
        keys = sorted(keys)
        for prefix, suffixes, vf in self._iter_keys_vf(keys):
            suffixes = [(suffix,) for suffix in suffixes]
            for record in vf.get_record_stream(suffixes, ordering,
                include_delta_closure):
                if record.parents is not None:
                    record.parents = tuple(
                        prefix + parent for parent in record.parents)
                record.key = prefix + record.key
                yield record

    def _iter_keys_vf(self, keys):
        prefixes = self._partition_keys(keys)
        sha1s = {}
        for prefix, suffixes in prefixes.items():
            path = self._mapper.map(prefix)
            vf = self._get_vf(path)
            yield prefix, suffixes, vf

    def get_sha1s(self, keys):
        """See VersionedFiles.get_sha1s()."""
        sha1s = {}
        for prefix,suffixes, vf in self._iter_keys_vf(keys):
            vf_sha1s = vf.get_sha1s(suffixes)
            for suffix, sha1 in vf_sha1s.iteritems():
                sha1s[prefix + (suffix,)] = sha1
        return sha1s

    def insert_record_stream(self, stream):
        """Insert a record stream into this container.

        :param stream: A stream of records to insert.
        :return: None
        :seealso VersionedFile.get_record_stream:
        """
        for record in stream:
            prefix = record.key[:-1]
            key = record.key[-1:]
            if record.parents is not None:
                parents = [parent[-1:] for parent in record.parents]
            else:
                parents = None
            thunk_record = AdapterFactory(key, parents, record)
            path = self._mapper.map(prefix)
            # Note that this parses the file many times; we can do better but
            # as this only impacts weaves in terms of performance, it is
            # tolerable.
            vf = self._get_vf(path)
            vf.insert_record_stream([thunk_record])

    def iter_lines_added_or_present_in_keys(self, keys, pb=None):
        """Iterate over the lines in the versioned files from keys.

        This may return lines from other keys. Each item the returned
        iterator yields is a tuple of a line and a text version that that line
        is present in (not introduced in).

        Ordering of results is in whatever order is most suitable for the
        underlying storage format.

        If a progress bar is supplied, it may be used to indicate progress.
        The caller is responsible for cleaning up progress bars (because this
        is an iterator).

        NOTES:
         * Lines are normalised by the underlying store: they will all have \n
           terminators.
         * Lines are returned in arbitrary order.

        :return: An iterator over (line, key).
        """
        for prefix, suffixes, vf in self._iter_keys_vf(keys):
            for line, version in vf.iter_lines_added_or_present_in_versions(suffixes):
                yield line, prefix + (version,)

    def _iter_all_components(self):
        for path, prefix in self._get_all_prefixes():
            yield prefix, self._get_vf(path)

    def keys(self):
        """See VersionedFiles.keys()."""
        result = set()
        for prefix, vf in self._iter_all_components():
            for suffix in vf.versions():
                result.add(prefix + (suffix,))
        return result


class _PlanMergeVersionedFile(VersionedFiles):
    """A VersionedFile for uncommitted and committed texts.

    It is intended to allow merges to be planned with working tree texts.
    It implements only the small part of the VersionedFiles interface used by
    PlanMerge.  It falls back to multiple versionedfiles for data not stored in
    _PlanMergeVersionedFile itself.

    :ivar: fallback_versionedfiles a list of VersionedFiles objects that can be
        queried for missing texts.
    """

    def __init__(self, file_id):
        """Create a _PlanMergeVersionedFile.

        :param file_id: Used with _PlanMerge code which is not yet fully
            tuple-keyspace aware.
        """
        self._file_id = file_id
        # fallback locations
        self.fallback_versionedfiles = []
        # Parents for locally held keys.
        self._parents = {}
        # line data for locally held keys.
        self._lines = {}
        # key lookup providers
        self._providers = [DictParentsProvider(self._parents)]

    def plan_merge(self, ver_a, ver_b, base=None):
        """See VersionedFile.plan_merge"""
        from bzrlib.merge import _PlanMerge
        if base is None:
            return _PlanMerge(ver_a, ver_b, self, (self._file_id,)).plan_merge()
        old_plan = list(_PlanMerge(ver_a, base, self, (self._file_id,)).plan_merge())
        new_plan = list(_PlanMerge(ver_a, ver_b, self, (self._file_id,)).plan_merge())
        return _PlanMerge._subtract_plans(old_plan, new_plan)

    def plan_lca_merge(self, ver_a, ver_b, base=None):
        from bzrlib.merge import _PlanLCAMerge
        graph = Graph(self)
        new_plan = _PlanLCAMerge(ver_a, ver_b, self, (self._file_id,), graph).plan_merge()
        if base is None:
            return new_plan
        old_plan = _PlanLCAMerge(ver_a, base, self, (self._file_id,), graph).plan_merge()
        return _PlanLCAMerge._subtract_plans(list(old_plan), list(new_plan))

    def add_lines(self, key, parents, lines):
        """See VersionedFiles.add_lines

        Lines are added locally, not to fallback versionedfiles.  Also, ghosts
        are permitted.  Only reserved ids are permitted.
        """
        if type(key) is not tuple:
            raise TypeError(key)
        if not revision.is_reserved_id(key[-1]):
            raise ValueError('Only reserved ids may be used')
        if parents is None:
            raise ValueError('Parents may not be None')
        if lines is None:
            raise ValueError('Lines may not be None')
        self._parents[key] = tuple(parents)
        self._lines[key] = lines

    def get_record_stream(self, keys, ordering, include_delta_closure):
        pending = set(keys)
        for key in keys:
            if key in self._lines:
                lines = self._lines[key]
                parents = self._parents[key]
                pending.remove(key)
                yield ChunkedContentFactory(key, parents, None, lines)
        for versionedfile in self.fallback_versionedfiles:
            for record in versionedfile.get_record_stream(
                pending, 'unordered', True):
                if record.storage_kind == 'absent':
                    continue
                else:
                    pending.remove(record.key)
                    yield record
            if not pending:
                return
        # report absent entries
        for key in pending:
            yield AbsentContentFactory(key)

    def get_parent_map(self, keys):
        """See VersionedFiles.get_parent_map"""
        # We create a new provider because a fallback may have been added.
        # If we make fallbacks private we can update a stack list and avoid
        # object creation thrashing.
        keys = set(keys)
        result = {}
        if revision.NULL_REVISION in keys:
            keys.remove(revision.NULL_REVISION)
            result[revision.NULL_REVISION] = ()
        self._providers = self._providers[:1] + self.fallback_versionedfiles
        result.update(
            StackedParentsProvider(self._providers).get_parent_map(keys))
        for key, parents in result.iteritems():
            if parents == ():
                result[key] = (revision.NULL_REVISION,)
        return result


class PlanWeaveMerge(TextMerge):
    """Weave merge that takes a plan as its input.

    This exists so that VersionedFile.plan_merge is implementable.
    Most callers will want to use WeaveMerge instead.
    """

    def __init__(self, plan, a_marker=TextMerge.A_MARKER,
                 b_marker=TextMerge.B_MARKER):
        TextMerge.__init__(self, a_marker, b_marker)
        self.plan = list(plan)

    def _merge_struct(self):
        lines_a = []
        lines_b = []
        ch_a = ch_b = False

        def outstanding_struct():
            if not lines_a and not lines_b:
                return
            elif ch_a and not ch_b:
                # one-sided change:
                yield(lines_a,)
            elif ch_b and not ch_a:
                yield (lines_b,)
            elif lines_a == lines_b:
                yield(lines_a,)
            else:
                yield (lines_a, lines_b)

        # We previously considered either 'unchanged' or 'killed-both' lines
        # to be possible places to resynchronize.  However, assuming agreement
        # on killed-both lines may be too aggressive. -- mbp 20060324
        for state, line in self.plan:
            if state == 'unchanged':
                # resync and flush queued conflicts changes if any
                for struct in outstanding_struct():
                    yield struct
                lines_a = []
                lines_b = []
                ch_a = ch_b = False

            if state == 'unchanged':
                if line:
                    yield ([line],)
            elif state == 'killed-a':
                ch_a = True
                lines_b.append(line)
            elif state == 'killed-b':
                ch_b = True
                lines_a.append(line)
            elif state == 'new-a':
                ch_a = True
                lines_a.append(line)
            elif state == 'new-b':
                ch_b = True
                lines_b.append(line)
            elif state == 'conflicted-a':
                ch_b = ch_a = True
                lines_a.append(line)
            elif state == 'conflicted-b':
                ch_b = ch_a = True
                lines_b.append(line)
            elif state == 'killed-both':
                # This counts as a change, even though there is no associated
                # line
                ch_b = ch_a = True
            else:
                if state not in ('irrelevant', 'ghost-a', 'ghost-b',
                        'killed-base'):
                    raise AssertionError(state)
        for struct in outstanding_struct():
            yield struct

    def base_from_plan(self):
        """Construct a BASE file from the plan text."""
        base_lines = []
        for state, line in self.plan:
            if state in ('killed-a', 'killed-b', 'killed-both', 'unchanged'):
                # If unchanged, then this line is straight from base. If a or b
                # or both killed the line, then it *used* to be in base.
                base_lines.append(line)
            else:
                if state not in ('killed-base', 'irrelevant',
                                 'ghost-a', 'ghost-b',
                                 'new-a', 'new-b',
                                 'conflicted-a', 'conflicted-b'):
                    # killed-base, irrelevant means it doesn't apply
                    # ghost-a/ghost-b are harder to say for sure, but they
                    # aren't in the 'inc_c' which means they aren't in the
                    # shared base of a & b. So we don't include them.  And
                    # obviously if the line is newly inserted, it isn't in base

                    # If 'conflicted-a' or b, then it is new vs one base, but
                    # old versus another base. However, if we make it present
                    # in the base, it will be deleted from the target, and it
                    # seems better to get a line doubled in the merge result,
                    # rather than have it deleted entirely.
                    # Example, each node is the 'text' at that point:
                    #           MN
                    #          /   \
                    #        MaN   MbN
                    #         |  X  |
                    #        MabN MbaN
                    #          \   /
                    #           ???
                    # There was a criss-cross conflict merge. Both sides
                    # include the other, but put themselves first.
                    # Weave marks this as a 'clean' merge, picking OTHER over
                    # THIS. (Though the details depend on order inserted into
                    # weave, etc.)
                    # LCA generates a plan:
                    # [('unchanged', M),
                    #  ('conflicted-b', b),
                    #  ('unchanged', a),
                    #  ('conflicted-a', b),
                    #  ('unchanged', N)]
                    # If you mark 'conflicted-*' as part of BASE, then a 3-way
                    # merge tool will cleanly generate "MaN" (as BASE vs THIS
                    # removes one 'b', and BASE vs OTHER removes the other)
                    # If you include neither, 3-way creates a clean "MbabN" as
                    # THIS adds one 'b', and OTHER does too.
                    # It seems that having the line 2 times is better than
                    # having it omitted. (Easier to manually delete than notice
                    # it needs to be added.)
                    raise AssertionError('Unknown state: %s' % (state,))
        return base_lines


class WeaveMerge(PlanWeaveMerge):
    """Weave merge that takes a VersionedFile and two versions as its input."""

    def __init__(self, versionedfile, ver_a, ver_b,
        a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
        plan = versionedfile.plan_merge(ver_a, ver_b)
        PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)


class VirtualVersionedFiles(VersionedFiles):
    """Dummy implementation for VersionedFiles that uses other functions for
    obtaining fulltexts and parent maps.

    This is always on the bottom of the stack and uses string keys
    (rather than tuples) internally.
    """

    def __init__(self, get_parent_map, get_lines):
        """Create a VirtualVersionedFiles.

        :param get_parent_map: Same signature as Repository.get_parent_map.
        :param get_lines: Should return lines for specified key or None if
                          not available.
        """
        super(VirtualVersionedFiles, self).__init__()
        self._get_parent_map = get_parent_map
        self._get_lines = get_lines

    def check(self, progressbar=None):
        """See VersionedFiles.check.

        :note: Always returns True for VirtualVersionedFiles.
        """
        return True

    def add_mpdiffs(self, records):
        """See VersionedFiles.mpdiffs.

        :note: Not implemented for VirtualVersionedFiles.
        """
        raise NotImplementedError(self.add_mpdiffs)

    def get_parent_map(self, keys):
        """See VersionedFiles.get_parent_map."""
        return dict([((k,), tuple([(p,) for p in v]))
            for k,v in self._get_parent_map([k for (k,) in keys]).iteritems()])

    def get_sha1s(self, keys):
        """See VersionedFiles.get_sha1s."""
        ret = {}
        for (k,) in keys:
            lines = self._get_lines(k)
            if lines is not None:
                if not isinstance(lines, list):
                    raise AssertionError
                ret[(k,)] = osutils.sha_strings(lines)
        return ret

    def get_record_stream(self, keys, ordering, include_delta_closure):
        """See VersionedFiles.get_record_stream."""
        for (k,) in list(keys):
            lines = self._get_lines(k)
            if lines is not None:
                if not isinstance(lines, list):
                    raise AssertionError
                yield ChunkedContentFactory((k,), None,
                        sha1=osutils.sha_strings(lines),
                        chunks=lines)
            else:
                yield AbsentContentFactory((k,))

    def iter_lines_added_or_present_in_keys(self, keys, pb=None):
        """See VersionedFile.iter_lines_added_or_present_in_versions()."""
        for i, (key,) in enumerate(keys):
            if pb is not None:
                pb.update("Finding changed lines", i, len(keys))
            for l in self._get_lines(key):
                yield (l, key)


def network_bytes_to_kind_and_offset(network_bytes):
    """Strip of a record kind from the front of network_bytes.

    :param network_bytes: The bytes of a record.
    :return: A tuple (storage_kind, offset_of_remaining_bytes)
    """
    line_end = network_bytes.find('\n')
    storage_kind = network_bytes[:line_end]
    return storage_kind, line_end + 1


class NetworkRecordStream(object):
    """A record_stream which reconstitures a serialised stream."""

    def __init__(self, bytes_iterator):
        """Create a NetworkRecordStream.

        :param bytes_iterator: An iterator of bytes. Each item in this
            iterator should have been obtained from a record_streams'
            record.get_bytes_as(record.storage_kind) call.
        """
        self._bytes_iterator = bytes_iterator
        self._kind_factory = {
            'fulltext': fulltext_network_to_record,
            'groupcompress-block': groupcompress.network_block_to_records,
            'knit-ft-gz': knit.knit_network_to_record,
            'knit-delta-gz': knit.knit_network_to_record,
            'knit-annotated-ft-gz': knit.knit_network_to_record,
            'knit-annotated-delta-gz': knit.knit_network_to_record,
            'knit-delta-closure': knit.knit_delta_closure_to_records,
            }

    def read(self):
        """Read the stream.

        :return: An iterator as per VersionedFiles.get_record_stream().
        """
        for bytes in self._bytes_iterator:
            storage_kind, line_end = network_bytes_to_kind_and_offset(bytes)
            for record in self._kind_factory[storage_kind](
                storage_kind, bytes, line_end):
                yield record


def fulltext_network_to_record(kind, bytes, line_end):
    """Convert a network fulltext record to record."""
    meta_len, = struct.unpack('!L', bytes[line_end:line_end+4])
    record_meta = bytes[line_end+4:line_end+4+meta_len]
    key, parents = bencode.bdecode_as_tuple(record_meta)
    if parents == 'nil':
        parents = None
    fulltext = bytes[line_end+4+meta_len:]
    return [FulltextContentFactory(key, parents, None, fulltext)]


def _length_prefix(bytes):
    return struct.pack('!L', len(bytes))


def record_to_fulltext_bytes(record):
    if record.parents is None:
        parents = 'nil'
    else:
        parents = record.parents
    record_meta = bencode.bencode((record.key, parents))
    record_content = record.get_bytes_as('fulltext')
    return "fulltext\n%s%s%s" % (
        _length_prefix(record_meta), record_meta, record_content)


def sort_groupcompress(parent_map):
    """Sort and group the keys in parent_map into groupcompress order.

    groupcompress is defined (currently) as reverse-topological order, grouped
    by the key prefix.

    :return: A sorted-list of keys
    """
    # gc-optimal ordering is approximately reverse topological,
    # properly grouped by file-id.
    per_prefix_map = {}
    for item in parent_map.iteritems():
        key = item[0]
        if isinstance(key, str) or len(key) == 1:
            prefix = ''
        else:
            prefix = key[0]
        try:
            per_prefix_map[prefix].append(item)
        except KeyError:
            per_prefix_map[prefix] = [item]

    present_keys = []
    for prefix in sorted(per_prefix_map):
        present_keys.extend(reversed(tsort.topo_sort(per_prefix_map[prefix])))
    return present_keys