75
44
# TODO: Parallel-extract that passes back each line along with a
76
45
# description of which revisions include it. Nice for checking all
46
# shas or calculating stats in parallel.
79
48
# TODO: Using a single _extract routine and then processing the output
80
49
# is probably inefficient. It's simple enough that we can afford to
81
50
# have slight specializations for different ways its used: annotate,
82
51
# basis for add, get, etc.
84
# TODO: Perhaps the API should work only in names to hide the integer
85
# indexes from the user?
53
# TODO: Probably the API should work only in names to hide the integer
54
# indexes from the user.
87
56
# TODO: Is there any potential performance win by having an add()
88
57
# variant that is passed a pre-cooked version of the single basis
94
from difflib import SequenceMatcher
99
class WeaveError(Exception):
100
"""Exception in processing weave"""
103
class WeaveFormatError(WeaveError):
104
"""Weave invariant violated"""
60
# TODO: Reweave can possibly be made faster by remembering diffs
61
# where the basis and destination are unchanged.
63
# FIXME: Sometimes we will be given a parents list for a revision
64
# that includes some redundant parents (i.e. already a parent of
65
# something in the list.) We should eliminate them. This can
66
# be done fairly efficiently because the sequence numbers constrain
67
# the possible relationships.
69
# FIXME: the conflict markers should be *7* characters
72
from cStringIO import StringIO
77
from bzrlib.lazy_import import lazy_import
78
lazy_import(globals(), """
79
from bzrlib import tsort
86
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
87
RevisionAlreadyPresent,
89
UnavailableRepresentation,
90
WeaveRevisionAlreadyPresent,
91
WeaveRevisionNotPresent,
93
from bzrlib.osutils import dirname, sha, sha_strings, split_lines
94
import bzrlib.patiencediff
95
from bzrlib.revision import NULL_REVISION
96
from bzrlib.symbol_versioning import *
97
from bzrlib.trace import mutter
98
from bzrlib.versionedfile import (
104
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
107
class WeaveContentFactory(ContentFactory):
108
"""Content factory for streaming from weaves.
110
:seealso ContentFactory:
113
def __init__(self, version, weave):
114
"""Create a WeaveContentFactory for version from weave."""
115
ContentFactory.__init__(self)
116
self.sha1 = weave.get_sha1s([version])[version]
117
self.key = (version,)
118
parents = weave.get_parent_map([version])[version]
119
self.parents = tuple((parent,) for parent in parents)
120
self.storage_kind = 'fulltext'
123
def get_bytes_as(self, storage_kind):
124
if storage_kind == 'fulltext':
125
return self._weave.get_text(self.key[-1])
126
elif storage_kind == 'chunked':
127
return self._weave.get_lines(self.key[-1])
129
raise UnavailableRepresentation(self.key, storage_kind, 'fulltext')
132
class Weave(VersionedFile):
108
133
"""weave - versioned text file storage.
110
135
A Weave manages versions of line-based text files, keeping track
211
276
return self._parents == other._parents \
212
277
and self._weave == other._weave \
213
278
and self._sha1s == other._sha1s
216
280
def __ne__(self, other):
217
281
return not self.__eq__(other)
220
def maybe_lookup(self, name_or_index):
221
"""Convert possible symbolic name to index, or pass through indexes."""
222
if isinstance(name_or_index, (int, long)):
225
return self.lookup(name_or_index)
228
def lookup(self, name):
283
def _idx_to_name(self, version):
284
return self._names[version]
286
def _lookup(self, name):
229
287
"""Convert symbolic version name to index."""
288
if not self._allow_reserved:
289
self.check_not_reserved_id(name)
231
291
return self._name_map[name]
233
raise WeaveError("name %r not present in weave %r" %
234
(name, self._weave_name))
237
def idx_to_name(self, version):
238
return self._names[version]
293
raise RevisionNotPresent(name, self._weave_name)
296
"""See VersionedFile.versions."""
297
return self._names[:]
299
def has_version(self, version_id):
300
"""See VersionedFile.has_version."""
301
return (version_id in self._name_map)
303
__contains__ = has_version
305
def get_record_stream(self, versions, ordering, include_delta_closure):
306
"""Get a stream of records for versions.
308
:param versions: The versions to include. Each version is a tuple
310
:param ordering: Either 'unordered' or 'topological'. A topologically
311
sorted stream has compression parents strictly before their
313
:param include_delta_closure: If True then the closure across any
314
compression parents will be included (in the opaque data).
315
:return: An iterator of ContentFactory objects, each of which is only
316
valid until the iterator is advanced.
318
versions = [version[-1] for version in versions]
319
if ordering == 'topological':
320
parents = self.get_parent_map(versions)
321
new_versions = tsort.topo_sort(parents)
322
new_versions.extend(set(versions).difference(set(parents)))
323
versions = new_versions
324
for version in versions:
326
yield WeaveContentFactory(version, self)
328
yield AbsentContentFactory((version,))
330
def get_parent_map(self, version_ids):
331
"""See VersionedFile.get_parent_map."""
333
for version_id in version_ids:
334
if version_id == NULL_REVISION:
339
map(self._idx_to_name,
340
self._parents[self._lookup(version_id)]))
341
except RevisionNotPresent:
343
result[version_id] = parents
346
def get_parents_with_ghosts(self, version_id):
347
raise NotImplementedError(self.get_parents_with_ghosts)
349
def insert_record_stream(self, stream):
350
"""Insert a record stream into this versioned file.
352
:param stream: A stream of records to insert.
354
:seealso VersionedFile.get_record_stream:
357
for record in stream:
358
# Raise an error when a record is missing.
359
if record.storage_kind == 'absent':
360
raise RevisionNotPresent([record.key[0]], self)
361
# adapt to non-tuple interface
362
parents = [parent[0] for parent in record.parents]
363
if (record.storage_kind == 'fulltext'
364
or record.storage_kind == 'chunked'):
365
self.add_lines(record.key[0], parents,
366
osutils.chunks_to_lines(record.get_bytes_as('chunked')))
368
adapter_key = record.storage_kind, 'fulltext'
370
adapter = adapters[adapter_key]
372
adapter_factory = adapter_registry.get(adapter_key)
373
adapter = adapter_factory(self)
374
adapters[adapter_key] = adapter
375
lines = split_lines(adapter.get_bytes(
376
record, record.get_bytes_as(record.storage_kind)))
378
self.add_lines(record.key[0], parents, lines)
379
except RevisionAlreadyPresent:
241
382
def _check_repeated_add(self, name, parents, text, sha1):
242
383
"""Check that a duplicated add is OK.
244
385
If it is, return the (old) index; otherwise raise an exception.
246
idx = self.lookup(name)
247
if sorted(self._parents[idx]) != sorted(parents):
248
raise WeaveError("name \"%s\" already present in weave "
249
"with different parents" % name)
250
if sha1 != self._sha1s[idx]:
251
raise WeaveError("name \"%s\" already present in weave "
252
"with different text" % name)
387
idx = self._lookup(name)
388
if sorted(self._parents[idx]) != sorted(parents) \
389
or sha1 != self._sha1s[idx]:
390
raise RevisionAlreadyPresent(name, self._weave_name)
257
def add(self, name, parents, text, sha1=None):
393
def _add_lines(self, version_id, parents, lines, parent_texts,
394
left_matching_blocks, nostore_sha, random_id, check_content):
395
"""See VersionedFile.add_lines."""
396
idx = self._add(version_id, lines, map(self._lookup, parents),
397
nostore_sha=nostore_sha)
398
return sha_strings(lines), sum(map(len, lines)), idx
400
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
258
401
"""Add a single text on top of the weave.
260
403
Returns the index number of the newly added version.
263
406
Symbolic name for this version.
264
407
(Typically the revision-id of the revision that added it.)
267
410
List or set of direct parent version numbers.
270
413
Sequence of lines to be added in the new version.
272
sha -- SHA-1 of the file, if known. This is trusted to be
415
:param nostore_sha: See VersionedFile.add_lines.
275
from bzrlib.osutils import sha_strings
277
assert isinstance(name, basestring)
279
sha1 = sha_strings(text)
280
if name in self._name_map:
281
return self._check_repeated_add(name, parents, text, sha1)
283
parents = map(self.maybe_lookup, parents)
417
self._check_lines_not_unicode(lines)
418
self._check_lines_are_lines(lines)
420
sha1 = sha_strings(lines)
421
if sha1 == nostore_sha:
422
raise errors.ExistingContent
423
if version_id in self._name_map:
424
return self._check_repeated_add(version_id, parents, lines, sha1)
284
426
self._check_versions(parents)
285
## self._check_lines(text)
427
## self._check_lines(lines)
286
428
new_version = len(self._parents)
289
430
# if we abort after here the (in-memory) weave will be corrupt because only
290
431
# some fields are updated
432
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
433
# - Robert Collins 20060226
291
434
self._parents.append(parents[:])
292
435
self._sha1s.append(sha1)
293
self._names.append(name)
294
self._name_map[name] = new_version
436
self._names.append(version_id)
437
self._name_map[version_id] = new_version
520
690
WFE = WeaveFormatError
693
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
694
# +285282 0 1676.8040 1676.8040 +<isinstance>
695
# 1.6 seconds in 'isinstance'.
696
# changing the first isinstance:
697
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
698
# +140414 0 762.8050 762.8050 +<isinstance>
699
# note that the inline time actually dropped (less function calls)
700
# and total processing time was halved.
701
# we're still spending ~1/4 of the method in isinstance though.
702
# so lets hard code the acceptable string classes we expect:
703
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
704
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
706
# yay, down to ~1/4 the initial extract time, and our inline time
707
# has shrunk again, with isinstance no longer dominating.
708
# tweaking the stack inclusion test to use a set gives:
709
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
710
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
712
# - a 5% win, or possibly just noise. However with large istacks that
713
# 'in' test could dominate, so I'm leaving this change in place -
714
# when its fast enough to consider profiling big datasets we can review.
522
719
for l in self._weave:
523
if isinstance(l, tuple):
720
if l.__class__ == tuple:
527
assert v not in istack
727
iset.remove(istack.pop())
532
729
if v in included:
537
732
if v in included:
735
raise AssertionError()
541
assert isinstance(l, basestring)
542
737
if isactive is None:
543
738
isactive = (not dset) and istack and (istack[-1] in included)
545
740
result.append((istack[-1], lineno, l))
549
raise WFE("unclosed insertion blocks at end of weave",
743
raise WeaveFormatError("unclosed insertion blocks "
744
"at end of weave: %s" % istack)
552
raise WFE("unclosed deletion blocks at end of weave",
559
def get_iter(self, name_or_index):
560
"""Yield lines for the specified version."""
561
incls = [self.maybe_lookup(name_or_index)]
562
for origin, lineno, line in self._extract(incls):
566
def get_text(self, name_or_index):
567
return ''.join(self.get_iter(name_or_index))
568
assert isinstance(version, int)
571
def get_lines(self, name_or_index):
572
return list(self.get_iter(name_or_index))
578
def mash_iter(self, included):
579
"""Return composed version of multiple included versions."""
580
included = map(self.maybe_lookup, included)
581
for origin, lineno, text in self._extract(included):
585
def dump(self, to_file):
586
from pprint import pprint
587
print >>to_file, "Weave._weave = ",
588
pprint(self._weave, to_file)
589
print >>to_file, "Weave._parents = ",
590
pprint(self._parents, to_file)
594
def numversions(self):
746
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
750
def _maybe_lookup(self, name_or_index):
751
"""Convert possible symbolic name to index, or pass through indexes.
755
if isinstance(name_or_index, (int, long)):
758
return self._lookup(name_or_index)
760
def get_lines(self, version_id):
761
"""See VersionedFile.get_lines()."""
762
int_index = self._maybe_lookup(version_id)
763
result = [line for (origin, lineno, line) in self._extract([int_index])]
764
expected_sha1 = self._sha1s[int_index]
765
measured_sha1 = sha_strings(result)
766
if measured_sha1 != expected_sha1:
767
raise errors.WeaveInvalidChecksum(
768
'file %s, revision %s, expected: %s, measured %s'
769
% (self._weave_name, version_id,
770
expected_sha1, measured_sha1))
773
def get_sha1s(self, version_ids):
774
"""See VersionedFile.get_sha1s()."""
776
for v in version_ids:
777
result[v] = self._sha1s[self._lookup(v)]
780
def num_versions(self):
781
"""How many versions are in this weave?"""
595
782
l = len(self._parents)
596
assert l == len(self._sha1s)
601
return self.numversions()
785
__len__ = num_versions
604
787
def check(self, progress_bar=None):
605
# check no circular inclusions
606
for version in range(self.numversions()):
788
# TODO evaluate performance hit of using string sets in this routine.
789
# TODO: check no circular inclusions
790
# TODO: create a nested progress bar
791
for version in range(self.num_versions()):
607
792
inclusions = list(self._parents[version])
609
794
inclusions.sort()
611
796
raise WeaveFormatError("invalid included version %d for index %d"
612
797
% (inclusions[-1], version))
614
# try extracting all versions; this is a bit slow and parallel
615
# extraction could be used
616
nv = self.numversions()
617
for version in range(nv):
799
# try extracting all versions; parallel extraction is used
800
nv = self.num_versions()
805
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
806
# The problem is that set membership is much more expensive
807
name = self._idx_to_name(i)
810
new_inc = set([name])
811
for p in self._parents[i]:
812
new_inc.update(inclusions[self._idx_to_name(p)])
814
if set(new_inc) != set(self.get_ancestry(name)):
815
raise AssertionError(
817
% (set(new_inc), set(self.get_ancestry(name))))
818
inclusions[name] = new_inc
820
nlines = len(self._weave)
822
update_text = 'checking weave'
824
short_name = os.path.basename(self._weave_name)
825
update_text = 'checking %s' % (short_name,)
826
update_text = update_text[:25]
828
for lineno, insert, deleteset, line in self._walk_internal():
619
progress_bar.update('checking text', version, nv)
621
for l in self.get_iter(version):
624
expected = self._sha1s[version]
830
progress_bar.update(update_text, lineno, nlines)
832
for name, name_inclusions in inclusions.items():
833
# The active inclusion must be an ancestor,
834
# and no ancestors must have deleted this line,
835
# because we don't support resurrection.
836
if (insert in name_inclusions) and not (deleteset & name_inclusions):
837
sha1s[name].update(line)
840
version = self._idx_to_name(i)
841
hd = sha1s[version].hexdigest()
842
expected = self._sha1s[i]
625
843
if hd != expected:
626
raise WeaveError("mismatched sha1 for version %d; "
627
"got %s, expected %s"
628
% (version, hd, expected))
844
raise errors.WeaveInvalidChecksum(
845
"mismatched sha1 for version %s: "
846
"got %s, expected %s"
847
% (version, hd, expected))
630
849
# TODO: check insertions are properly nested, that there are
631
850
# no lines outside of insertion blocks, that deletions are
632
851
# properly paired, etc.
636
def merge(self, merge_versions):
637
"""Automerge and mark conflicts between versions.
639
This returns a sequence, each entry describing alternatives
640
for a chunk of the file. Each of the alternatives is given as
643
If there is a chunk of the file where there's no diagreement,
644
only one alternative is given.
647
# approach: find the included versions common to all the
649
raise NotImplementedError()
653
def _delta(self, included, lines):
654
"""Return changes from basis to new revision.
656
The old text for comparison is the union of included revisions.
658
This is used in inserting a new text.
660
Delta is returned as a sequence of
661
(weave1, weave2, newlines).
663
This indicates that weave1:weave2 of the old weave should be
664
replaced by the sequence of lines in newlines. Note that
665
these line numbers are positions in the total weave and don't
666
correspond to the lines in any extracted version, or even the
667
extracted union of included versions.
669
If line1=line2, this is a pure insert; if newlines=[] this is a
670
pure delete. (Similar to difflib.)
675
def plan_merge(self, ver_a, ver_b):
676
"""Return pseudo-annotation indicating how the two versions merge.
678
This is computed between versions a and b and their common
681
Weave lines present in none of them are skipped entirely.
683
inc_a = self.inclusions([ver_a])
684
inc_b = self.inclusions([ver_b])
685
inc_c = inc_a & inc_b
687
for lineno, insert, deleteset, line in self._walk():
688
if deleteset & inc_c:
689
# killed in parent; can't be in either a or b
690
# not relevant to our work
691
yield 'killed-base', line
692
elif insert in inc_c:
693
# was inserted in base
694
killed_a = bool(deleteset & inc_a)
695
killed_b = bool(deleteset & inc_b)
696
if killed_a and killed_b:
697
yield 'killed-both', line
699
yield 'killed-a', line
701
yield 'killed-b', line
703
yield 'unchanged', line
704
elif insert in inc_a:
705
if deleteset & inc_a:
706
yield 'ghost-a', line
710
elif insert in inc_b:
711
if deleteset & inc_b:
712
yield 'ghost-b', line
716
# not in either revision
717
yield 'irrelevant', line
719
yield 'unchanged', '' # terminator
723
def weave_merge(self, plan):
728
for state, line in plan:
729
if state == 'unchanged' or state == 'killed-both':
730
# resync and flush queued conflicts changes if any
731
if not lines_a and not lines_b:
733
elif ch_a and not ch_b:
735
for l in lines_a: yield l
736
elif ch_b and not ch_a:
737
for l in lines_b: yield l
738
elif lines_a == lines_b:
739
for l in lines_a: yield l
742
for l in lines_a: yield l
744
for l in lines_b: yield l
751
if state == 'unchanged':
754
elif state == 'killed-a':
757
elif state == 'killed-b':
760
elif state == 'new-a':
763
elif state == 'new-b':
767
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
853
def _imported_parents(self, other, other_idx):
854
"""Return list of parents in self corresponding to indexes in other."""
856
for parent_idx in other._parents[other_idx]:
857
parent_name = other._names[parent_idx]
858
if parent_name not in self._name_map:
859
# should not be possible
860
raise WeaveError("missing parent {%s} of {%s} in %r"
861
% (parent_name, other._name_map[other_idx], self))
862
new_parents.append(self._name_map[parent_name])
865
def _check_version_consistent(self, other, other_idx, name):
866
"""Check if a version in consistent in this and other.
868
To be consistent it must have:
871
* the same direct parents (by name, not index, and disregarding
874
If present & correct return True;
875
if not present in self return False;
876
if inconsistent raise error."""
877
this_idx = self._name_map.get(name, -1)
879
if self._sha1s[this_idx] != other._sha1s[other_idx]:
880
raise errors.WeaveTextDiffers(name, self, other)
881
self_parents = self._parents[this_idx]
882
other_parents = other._parents[other_idx]
883
n1 = set([self._names[i] for i in self_parents])
884
n2 = set([other._names[i] for i in other_parents])
885
if not self._compatible_parents(n1, n2):
886
raise WeaveParentMismatch("inconsistent parents "
887
"for version {%s}: %s vs %s" % (name, n1, n2))
893
def _reweave(self, other, pb, msg):
894
"""Reweave self with other - internal helper for join().
896
:param other: The other weave to merge
897
:param pb: An optional progress bar, indicating how far done we are
898
:param msg: An optional message for the progress
900
new_weave = _reweave(self, other, pb=pb, msg=msg)
901
self._copy_weave_content(new_weave)
903
def _copy_weave_content(self, otherweave):
904
"""adsorb the content from otherweave."""
905
for attr in self.__slots__:
906
if attr != '_weave_name':
907
setattr(self, attr, copy(getattr(otherweave, attr)))
910
class WeaveFile(Weave):
911
"""A WeaveFile represents a Weave on disk and writes on change."""
913
WEAVE_SUFFIX = '.weave'
915
def __init__(self, name, transport, filemode=None, create=False, access_mode='w', get_scope=None):
916
"""Create a WeaveFile.
918
:param create: If not True, only open an existing knit.
920
super(WeaveFile, self).__init__(name, access_mode, get_scope=get_scope,
921
allow_reserved=False)
922
self._transport = transport
923
self._filemode = filemode
925
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
926
except errors.NoSuchFile:
932
def _add_lines(self, version_id, parents, lines, parent_texts,
933
left_matching_blocks, nostore_sha, random_id, check_content):
934
"""Add a version and save the weave."""
935
self.check_not_reserved_id(version_id)
936
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
937
parent_texts, left_matching_blocks, nostore_sha, random_id,
942
def copy_to(self, name, transport):
943
"""See VersionedFile.copy_to()."""
944
# as we are all in memory always, just serialise to the new place.
946
write_weave_v5(self, sio)
948
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
951
"""Save the weave."""
952
self._check_write_ok()
954
write_weave_v5(self, sio)
956
bytes = sio.getvalue()
957
path = self._weave_name + WeaveFile.WEAVE_SUFFIX
959
self._transport.put_bytes(path, bytes, self._filemode)
960
except errors.NoSuchFile:
961
self._transport.mkdir(dirname(path))
962
self._transport.put_bytes(path, bytes, self._filemode)
966
"""See VersionedFile.get_suffixes()."""
967
return [WeaveFile.WEAVE_SUFFIX]
969
def insert_record_stream(self, stream):
970
super(WeaveFile, self).insert_record_stream(stream)
973
@deprecated_method(one_five)
974
def join(self, other, pb=None, msg=None, version_ids=None,
975
ignore_missing=False):
976
"""Join other into self and save."""
977
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
981
def _reweave(wa, wb, pb=None, msg=None):
982
"""Combine two weaves and return the result.
984
This works even if a revision R has different parents in
985
wa and wb. In the resulting weave all the parents are given.
987
This is done by just building up a new weave, maintaining ordering
988
of the versions in the two inputs. More efficient approaches
989
might be possible but it should only be necessary to do
990
this operation rarely, when a new previously ghost version is
993
:param pb: An optional progress bar, indicating how far done we are
994
:param msg: An optional message for the progress
998
queue_a = range(wa.num_versions())
999
queue_b = range(wb.num_versions())
1000
# first determine combined parents of all versions
1001
# map from version name -> all parent names
1002
combined_parents = _reweave_parent_graphs(wa, wb)
1003
mutter("combined parents: %r", combined_parents)
1004
order = tsort.topo_sort(combined_parents.iteritems())
1005
mutter("order to reweave: %r", order)
1010
for idx, name in enumerate(order):
1012
pb.update(msg, idx, len(order))
1013
if name in wa._name_map:
1014
lines = wa.get_lines(name)
1015
if name in wb._name_map:
1016
lines_b = wb.get_lines(name)
1017
if lines != lines_b:
1018
mutter('Weaves differ on content. rev_id {%s}', name)
1019
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1021
lines = list(difflib.unified_diff(lines, lines_b,
1022
wa._weave_name, wb._weave_name))
1023
mutter('lines:\n%s', ''.join(lines))
1024
raise errors.WeaveTextDiffers(name, wa, wb)
1026
lines = wb.get_lines(name)
1027
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1030
def _reweave_parent_graphs(wa, wb):
1031
"""Return combined parent ancestry for two weaves.
1033
Returned as a list of (version_name, set(parent_names))"""
1035
for weave in [wa, wb]:
1036
for idx, name in enumerate(weave._names):
1037
p = combined.setdefault(name, set())
1038
p.update(map(weave._idx_to_name, weave._parents[idx]))
777
1042
def weave_toc(w):
added
removed
bzrlib/weave.py
