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?
93
class WeaveError(Exception):
94
"""Exception in processing weave"""
97
class WeaveFormatError(WeaveError):
98
"""Weave invariant violated"""
53
# TODO: Probably the API should work only in names to hide the integer
54
# indexes from the user.
56
# TODO: Is there any potential performance win by having an add()
57
# variant that is passed a pre-cooked version of the single basis
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
84
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
85
RevisionAlreadyPresent,
87
UnavailableRepresentation,
88
WeaveRevisionAlreadyPresent,
89
WeaveRevisionNotPresent,
91
import bzrlib.errors as errors
92
from bzrlib.osutils import dirname, sha, sha_strings, split_lines
93
import bzrlib.patiencediff
94
from bzrlib.revision import NULL_REVISION
95
from bzrlib.symbol_versioning import *
96
from bzrlib.trace import mutter
97
from bzrlib.versionedfile import (
103
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
106
class WeaveContentFactory(ContentFactory):
107
"""Content factory for streaming from weaves.
109
:seealso ContentFactory:
112
def __init__(self, version, weave):
113
"""Create a WeaveContentFactory for version from weave."""
114
ContentFactory.__init__(self)
115
self.sha1 = weave.get_sha1s([version])[version]
116
self.key = (version,)
117
parents = weave.get_parent_map([version])[version]
118
self.parents = tuple((parent,) for parent in parents)
119
self.storage_kind = 'fulltext'
122
def get_bytes_as(self, storage_kind):
123
if storage_kind == 'fulltext':
124
return self._weave.get_text(self.key[-1])
126
raise UnavailableRepresentation(self.key, storage_kind, 'fulltext')
129
class Weave(VersionedFile):
102
130
"""weave - versioned text file storage.
104
132
A Weave manages versions of line-based text files, keeping track
199
273
return self._parents == other._parents \
200
274
and self._weave == other._weave \
201
275
and self._sha1s == other._sha1s
204
277
def __ne__(self, other):
205
278
return not self.__eq__(other)
280
def _idx_to_name(self, version):
281
return self._names[version]
208
def lookup(self, name):
283
def _lookup(self, name):
284
"""Convert symbolic version name to index."""
285
if not self._allow_reserved:
286
self.check_not_reserved_id(name)
210
288
return self._name_map[name]
212
raise WeaveError("name %s not present in weave" % name)
215
def add(self, name, parents, text):
290
raise RevisionNotPresent(name, self._weave_name)
293
"""See VersionedFile.versions."""
294
return self._names[:]
296
def has_version(self, version_id):
297
"""See VersionedFile.has_version."""
298
return (version_id in self._name_map)
300
__contains__ = has_version
302
def get_record_stream(self, versions, ordering, include_delta_closure):
303
"""Get a stream of records for versions.
305
:param versions: The versions to include. Each version is a tuple
307
:param ordering: Either 'unordered' or 'topological'. A topologically
308
sorted stream has compression parents strictly before their
310
:param include_delta_closure: If True then the closure across any
311
compression parents will be included (in the opaque data).
312
:return: An iterator of ContentFactory objects, each of which is only
313
valid until the iterator is advanced.
315
versions = [version[-1] for version in versions]
316
if ordering == 'topological':
317
parents = self.get_parent_map(versions)
318
new_versions = tsort.topo_sort(parents)
319
new_versions.extend(set(versions).difference(set(parents)))
320
versions = new_versions
321
for version in versions:
323
yield WeaveContentFactory(version, self)
325
yield AbsentContentFactory((version,))
327
def get_parent_map(self, version_ids):
328
"""See VersionedFile.get_parent_map."""
330
for version_id in version_ids:
331
if version_id == NULL_REVISION:
336
map(self._idx_to_name,
337
self._parents[self._lookup(version_id)]))
338
except RevisionNotPresent:
340
result[version_id] = parents
343
def get_parents_with_ghosts(self, version_id):
344
raise NotImplementedError(self.get_parents_with_ghosts)
346
def insert_record_stream(self, stream):
347
"""Insert a record stream into this versioned file.
349
:param stream: A stream of records to insert.
351
:seealso VersionedFile.get_record_stream:
354
for record in stream:
355
# Raise an error when a record is missing.
356
if record.storage_kind == 'absent':
357
raise RevisionNotPresent([record.key[0]], self)
358
# adapt to non-tuple interface
359
parents = [parent[0] for parent in record.parents]
360
if record.storage_kind == 'fulltext':
361
self.add_lines(record.key[0], parents,
362
split_lines(record.get_bytes_as('fulltext')))
364
adapter_key = record.storage_kind, 'fulltext'
366
adapter = adapters[adapter_key]
368
adapter_factory = adapter_registry.get(adapter_key)
369
adapter = adapter_factory(self)
370
adapters[adapter_key] = adapter
371
lines = split_lines(adapter.get_bytes(
372
record, record.get_bytes_as(record.storage_kind)))
374
self.add_lines(record.key[0], parents, lines)
375
except RevisionAlreadyPresent:
378
def _check_repeated_add(self, name, parents, text, sha1):
379
"""Check that a duplicated add is OK.
381
If it is, return the (old) index; otherwise raise an exception.
383
idx = self._lookup(name)
384
if sorted(self._parents[idx]) != sorted(parents) \
385
or sha1 != self._sha1s[idx]:
386
raise RevisionAlreadyPresent(name, self._weave_name)
389
def _add_lines(self, version_id, parents, lines, parent_texts,
390
left_matching_blocks, nostore_sha, random_id, check_content):
391
"""See VersionedFile.add_lines."""
392
idx = self._add(version_id, lines, map(self._lookup, parents),
393
nostore_sha=nostore_sha)
394
return sha_strings(lines), sum(map(len, lines)), idx
396
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
216
397
"""Add a single text on top of the weave.
218
399
Returns the index number of the newly added version.
221
402
Symbolic name for this version.
222
403
(Typically the revision-id of the revision that added it.)
225
406
List or set of direct parent version numbers.
228
Sequence of lines to be added in the new version."""
230
assert isinstance(name, basestring)
231
if name in self._name_map:
232
raise WeaveError("name %r already present in weave" % name)
409
Sequence of lines to be added in the new version.
411
:param nostore_sha: See VersionedFile.add_lines.
413
self._check_lines_not_unicode(lines)
414
self._check_lines_are_lines(lines)
416
sha1 = sha_strings(lines)
417
if sha1 == nostore_sha:
418
raise errors.ExistingContent
419
if version_id in self._name_map:
420
return self._check_repeated_add(version_id, parents, lines, sha1)
234
422
self._check_versions(parents)
235
## self._check_lines(text)
423
## self._check_lines(lines)
236
424
new_version = len(self._parents)
243
426
# if we abort after here the (in-memory) weave will be corrupt because only
244
427
# some fields are updated
428
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
429
# - Robert Collins 20060226
245
430
self._parents.append(parents[:])
246
431
self._sha1s.append(sha1)
247
self._names.append(name)
248
self._name_map[name] = new_version
432
self._names.append(version_id)
433
self._name_map[version_id] = new_version
419
590
lineno = 0 # line of weave, 0-based
421
592
for l in self._weave:
422
if isinstance(l, tuple):
593
if l.__class__ == tuple:
597
istack.append(self._names[v])
601
dset.add(self._names[v])
435
raise WeaveFormatError('unexpected instruction %r'
438
assert isinstance(l, basestring)
440
yield lineno, istack[-1], dset, l
445
def _extract(self, versions):
446
"""Yield annotation of lines in included set.
448
Yields a sequence of tuples (origin, lineno, text), where
449
origin is the origin version, lineno the index in the weave,
450
and text the text of the line.
452
The set typically but not necessarily corresponds to a version.
454
included = self.inclusions(versions)
459
lineno = 0 # line of weave, 0-based
465
WFE = WeaveFormatError
467
for l in self._weave:
468
if isinstance(l, tuple):
472
assert v not in istack
486
assert isinstance(l, basestring)
488
isactive = (not dset) and istack and (istack[-1] in included)
490
result.append((istack[-1], lineno, l))
603
dset.remove(self._names[v])
605
raise WeaveFormatError('unexpected instruction %r' % v)
607
yield lineno, istack[-1], frozenset(dset), l
494
raise WFE("unclosed insertion blocks at end of weave",
611
raise WeaveFormatError("unclosed insertion blocks "
612
"at end of weave: %s" % istack)
497
raise WFE("unclosed deletion blocks at end of weave",
504
def get_iter(self, version):
505
"""Yield lines for the specified version."""
506
for origin, lineno, line in self._extract([version]):
510
def get(self, index):
511
return list(self.get_iter(index))
514
def mash_iter(self, included):
515
"""Return composed version of multiple included versions."""
516
for origin, lineno, text in self._extract(included):
520
def dump(self, to_file):
521
from pprint import pprint
522
print >>to_file, "Weave._weave = ",
523
pprint(self._weave, to_file)
524
print >>to_file, "Weave._parents = ",
525
pprint(self._parents, to_file)
529
def numversions(self):
530
l = len(self._parents)
531
assert l == len(self._sha1s)
536
return self.numversions()
539
def check(self, progress_bar=None):
540
# check no circular inclusions
541
for version in range(self.numversions()):
542
inclusions = list(self._parents[version])
545
if inclusions[-1] >= version:
546
raise WeaveFormatError("invalid included version %d for index %d"
547
% (inclusions[-1], version))
549
# try extracting all versions; this is a bit slow and parallel
550
# extraction could be used
551
nv = self.numversions()
552
for version in range(nv):
554
progress_bar.update('checking text', version, nv)
556
for l in self.get_iter(version):
559
expected = self._sha1s[version]
561
raise WeaveError("mismatched sha1 for version %d; "
562
"got %s, expected %s"
563
% (version, hd, expected))
565
# TODO: check insertions are properly nested, that there are
566
# no lines outside of insertion blocks, that deletions are
567
# properly paired, etc.
571
def merge(self, merge_versions):
572
"""Automerge and mark conflicts between versions.
574
This returns a sequence, each entry describing alternatives
575
for a chunk of the file. Each of the alternatives is given as
578
If there is a chunk of the file where there's no diagreement,
579
only one alternative is given.
582
# approach: find the included versions common to all the
584
raise NotImplementedError()
588
def _delta(self, included, lines):
589
"""Return changes from basis to new revision.
591
The old text for comparison is the union of included revisions.
593
This is used in inserting a new text.
595
Delta is returned as a sequence of
596
(weave1, weave2, newlines).
598
This indicates that weave1:weave2 of the old weave should be
599
replaced by the sequence of lines in newlines. Note that
600
these line numbers are positions in the total weave and don't
601
correspond to the lines in any extracted version, or even the
602
extracted union of included versions.
604
If line1=line2, this is a pure insert; if newlines=[] this is a
605
pure delete. (Similar to difflib.)
614
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
610
617
def plan_merge(self, ver_a, ver_b):
611
618
"""Return pseudo-annotation indicating how the two versions merge.
651
658
# not in either revision
652
659
yield 'irrelevant', line
654
yield 'unchanged', '' # terminator
658
def weave_merge(self, plan):
663
for state, line in plan:
664
if state == 'unchanged' or state == 'killed-both':
665
# resync and flush queued conflicts changes if any
666
if not lines_a and not lines_b:
668
elif ch_a and not ch_b:
670
for l in lines_a: yield l
671
elif ch_b and not ch_a:
672
for l in lines_b: yield l
673
elif lines_a == lines_b:
674
for l in lines_a: yield l
661
def _extract(self, versions):
662
"""Yield annotation of lines in included set.
664
Yields a sequence of tuples (origin, lineno, text), where
665
origin is the origin version, lineno the index in the weave,
666
and text the text of the line.
668
The set typically but not necessarily corresponds to a version.
671
if not isinstance(i, int):
674
included = self._inclusions(versions)
680
lineno = 0 # line of weave, 0-based
686
WFE = WeaveFormatError
689
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
690
# +285282 0 1676.8040 1676.8040 +<isinstance>
691
# 1.6 seconds in 'isinstance'.
692
# changing the first isinstance:
693
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
694
# +140414 0 762.8050 762.8050 +<isinstance>
695
# note that the inline time actually dropped (less function calls)
696
# and total processing time was halved.
697
# we're still spending ~1/4 of the method in isinstance though.
698
# so lets hard code the acceptable string classes we expect:
699
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
700
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
702
# yay, down to ~1/4 the initial extract time, and our inline time
703
# has shrunk again, with isinstance no longer dominating.
704
# tweaking the stack inclusion test to use a set gives:
705
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
706
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
708
# - a 5% win, or possibly just noise. However with large istacks that
709
# 'in' test could dominate, so I'm leaving this change in place -
710
# when its fast enough to consider profiling big datasets we can review.
715
for l in self._weave:
716
if l.__class__ == tuple:
723
iset.remove(istack.pop())
677
for l in lines_a: yield l
679
for l in lines_b: yield l
686
if state == 'unchanged':
689
elif state == 'killed-a':
692
elif state == 'killed-b':
695
elif state == 'new-a':
698
elif state == 'new-b':
702
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
731
raise AssertionError()
734
isactive = (not dset) and istack and (istack[-1] in included)
736
result.append((istack[-1], lineno, l))
739
raise WeaveFormatError("unclosed insertion blocks "
740
"at end of weave: %s" % istack)
742
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
746
def _maybe_lookup(self, name_or_index):
747
"""Convert possible symbolic name to index, or pass through indexes.
751
if isinstance(name_or_index, (int, long)):
754
return self._lookup(name_or_index)
756
def get_lines(self, version_id):
757
"""See VersionedFile.get_lines()."""
758
int_index = self._maybe_lookup(version_id)
759
result = [line for (origin, lineno, line) in self._extract([int_index])]
760
expected_sha1 = self._sha1s[int_index]
761
measured_sha1 = sha_strings(result)
762
if measured_sha1 != expected_sha1:
763
raise errors.WeaveInvalidChecksum(
764
'file %s, revision %s, expected: %s, measured %s'
765
% (self._weave_name, version_id,
766
expected_sha1, measured_sha1))
769
def get_sha1s(self, version_ids):
770
"""See VersionedFile.get_sha1s()."""
772
for v in version_ids:
773
result[v] = self._sha1s[self._lookup(v)]
776
def num_versions(self):
777
"""How many versions are in this weave?"""
778
l = len(self._parents)
781
__len__ = num_versions
783
def check(self, progress_bar=None):
784
# TODO evaluate performance hit of using string sets in this routine.
785
# TODO: check no circular inclusions
786
# TODO: create a nested progress bar
787
for version in range(self.num_versions()):
788
inclusions = list(self._parents[version])
791
if inclusions[-1] >= version:
792
raise WeaveFormatError("invalid included version %d for index %d"
793
% (inclusions[-1], version))
795
# try extracting all versions; parallel extraction is used
796
nv = self.num_versions()
801
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
802
# The problem is that set membership is much more expensive
803
name = self._idx_to_name(i)
806
new_inc = set([name])
807
for p in self._parents[i]:
808
new_inc.update(inclusions[self._idx_to_name(p)])
810
if set(new_inc) != set(self.get_ancestry(name)):
811
raise AssertionError(
813
% (set(new_inc), set(self.get_ancestry(name))))
814
inclusions[name] = new_inc
816
nlines = len(self._weave)
818
update_text = 'checking weave'
820
short_name = os.path.basename(self._weave_name)
821
update_text = 'checking %s' % (short_name,)
822
update_text = update_text[:25]
824
for lineno, insert, deleteset, line in self._walk_internal():
826
progress_bar.update(update_text, lineno, nlines)
828
for name, name_inclusions in inclusions.items():
829
# The active inclusion must be an ancestor,
830
# and no ancestors must have deleted this line,
831
# because we don't support resurrection.
832
if (insert in name_inclusions) and not (deleteset & name_inclusions):
833
sha1s[name].update(line)
836
version = self._idx_to_name(i)
837
hd = sha1s[version].hexdigest()
838
expected = self._sha1s[i]
840
raise errors.WeaveInvalidChecksum(
841
"mismatched sha1 for version %s: "
842
"got %s, expected %s"
843
% (version, hd, expected))
845
# TODO: check insertions are properly nested, that there are
846
# no lines outside of insertion blocks, that deletions are
847
# properly paired, etc.
849
def _imported_parents(self, other, other_idx):
850
"""Return list of parents in self corresponding to indexes in other."""
852
for parent_idx in other._parents[other_idx]:
853
parent_name = other._names[parent_idx]
854
if parent_name not in self._name_map:
855
# should not be possible
856
raise WeaveError("missing parent {%s} of {%s} in %r"
857
% (parent_name, other._name_map[other_idx], self))
858
new_parents.append(self._name_map[parent_name])
861
def _check_version_consistent(self, other, other_idx, name):
862
"""Check if a version in consistent in this and other.
864
To be consistent it must have:
867
* the same direct parents (by name, not index, and disregarding
870
If present & correct return True;
871
if not present in self return False;
872
if inconsistent raise error."""
873
this_idx = self._name_map.get(name, -1)
875
if self._sha1s[this_idx] != other._sha1s[other_idx]:
876
raise errors.WeaveTextDiffers(name, self, other)
877
self_parents = self._parents[this_idx]
878
other_parents = other._parents[other_idx]
879
n1 = set([self._names[i] for i in self_parents])
880
n2 = set([other._names[i] for i in other_parents])
881
if not self._compatible_parents(n1, n2):
882
raise WeaveParentMismatch("inconsistent parents "
883
"for version {%s}: %s vs %s" % (name, n1, n2))
889
def _reweave(self, other, pb, msg):
890
"""Reweave self with other - internal helper for join().
892
:param other: The other weave to merge
893
:param pb: An optional progress bar, indicating how far done we are
894
:param msg: An optional message for the progress
896
new_weave = _reweave(self, other, pb=pb, msg=msg)
897
self._copy_weave_content(new_weave)
899
def _copy_weave_content(self, otherweave):
900
"""adsorb the content from otherweave."""
901
for attr in self.__slots__:
902
if attr != '_weave_name':
903
setattr(self, attr, copy(getattr(otherweave, attr)))
906
class WeaveFile(Weave):
907
"""A WeaveFile represents a Weave on disk and writes on change."""
909
WEAVE_SUFFIX = '.weave'
911
def __init__(self, name, transport, filemode=None, create=False, access_mode='w', get_scope=None):
912
"""Create a WeaveFile.
914
:param create: If not True, only open an existing knit.
916
super(WeaveFile, self).__init__(name, access_mode, get_scope=get_scope,
917
allow_reserved=False)
918
self._transport = transport
919
self._filemode = filemode
921
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
922
except errors.NoSuchFile:
928
def _add_lines(self, version_id, parents, lines, parent_texts,
929
left_matching_blocks, nostore_sha, random_id, check_content):
930
"""Add a version and save the weave."""
931
self.check_not_reserved_id(version_id)
932
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
933
parent_texts, left_matching_blocks, nostore_sha, random_id,
938
def copy_to(self, name, transport):
939
"""See VersionedFile.copy_to()."""
940
# as we are all in memory always, just serialise to the new place.
942
write_weave_v5(self, sio)
944
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
947
"""Save the weave."""
948
self._check_write_ok()
950
write_weave_v5(self, sio)
952
bytes = sio.getvalue()
953
path = self._weave_name + WeaveFile.WEAVE_SUFFIX
955
self._transport.put_bytes(path, bytes, self._filemode)
956
except errors.NoSuchFile:
957
self._transport.mkdir(dirname(path))
958
self._transport.put_bytes(path, bytes, self._filemode)
962
"""See VersionedFile.get_suffixes()."""
963
return [WeaveFile.WEAVE_SUFFIX]
965
def insert_record_stream(self, stream):
966
super(WeaveFile, self).insert_record_stream(stream)
969
@deprecated_method(one_five)
970
def join(self, other, pb=None, msg=None, version_ids=None,
971
ignore_missing=False):
972
"""Join other into self and save."""
973
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
977
def _reweave(wa, wb, pb=None, msg=None):
978
"""Combine two weaves and return the result.
980
This works even if a revision R has different parents in
981
wa and wb. In the resulting weave all the parents are given.
983
This is done by just building up a new weave, maintaining ordering
984
of the versions in the two inputs. More efficient approaches
985
might be possible but it should only be necessary to do
986
this operation rarely, when a new previously ghost version is
989
:param pb: An optional progress bar, indicating how far done we are
990
:param msg: An optional message for the progress
994
queue_a = range(wa.num_versions())
995
queue_b = range(wb.num_versions())
996
# first determine combined parents of all versions
997
# map from version name -> all parent names
998
combined_parents = _reweave_parent_graphs(wa, wb)
999
mutter("combined parents: %r", combined_parents)
1000
order = tsort.topo_sort(combined_parents.iteritems())
1001
mutter("order to reweave: %r", order)
1006
for idx, name in enumerate(order):
1008
pb.update(msg, idx, len(order))
1009
if name in wa._name_map:
1010
lines = wa.get_lines(name)
1011
if name in wb._name_map:
1012
lines_b = wb.get_lines(name)
1013
if lines != lines_b:
1014
mutter('Weaves differ on content. rev_id {%s}', name)
1015
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1017
lines = list(difflib.unified_diff(lines, lines_b,
1018
wa._weave_name, wb._weave_name))
1019
mutter('lines:\n%s', ''.join(lines))
1020
raise errors.WeaveTextDiffers(name, wa, wb)
1022
lines = wb.get_lines(name)
1023
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1026
def _reweave_parent_graphs(wa, wb):
1027
"""Return combined parent ancestry for two weaves.
1029
Returned as a list of (version_name, set(parent_names))"""
1031
for weave in [wa, wb]:
1032
for idx, name in enumerate(weave._names):
1033
p = combined.setdefault(name, set())
1034
p.update(map(weave._idx_to_name, weave._parents[idx]))
712
1038
def weave_toc(w):
added
removed
bzrlib/weave.py
