44
74
# TODO: Parallel-extract that passes back each line along with a
45
75
# description of which revisions include it. Nice for checking all
46
# shas or calculating stats in parallel.
48
# TODO: Using a single _extract routine and then processing the output
49
# is probably inefficient. It's simple enough that we can afford to
50
# have slight specializations for different ways its used: annotate,
51
# basis for add, get, etc.
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
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):
81
class WeaveError(Exception):
82
"""Exception in processing weave"""
85
class WeaveFormatError(WeaveError):
86
"""Weave invariant violated"""
133
90
"""weave - versioned text file storage.
135
92
A Weave manages versions of line-based text files, keeping track
136
93
of the originating version for each line.
205
160
each version; the parent's parents are implied.
208
List of hex SHA-1 of each version.
211
List of symbolic names for each version. Each should be unique.
214
For each name, the version number.
217
Descriptive name of this weave; typically the filename if known.
163
List of hex SHA-1 of each version, or None if not recorded.
221
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
222
'_weave_name', '_matcher', '_allow_reserved']
224
def __init__(self, weave_name=None, access_mode='w', matcher=None,
225
get_scope=None, allow_reserved=False):
228
:param get_scope: A callable that returns an opaque object to be used
229
for detecting when this weave goes out of scope (should stop
230
answering requests or allowing mutation).
232
super(Weave, self).__init__()
166
__slots__ = ['_weave', '_parents', '_sha1s']
234
170
self._parents = []
238
self._weave_name = weave_name
240
self._matcher = bzrlib.patiencediff.PatienceSequenceMatcher
242
self._matcher = matcher
243
if get_scope is None:
244
get_scope = lambda:None
245
self._get_scope = get_scope
246
self._scope = get_scope()
247
self._access_mode = access_mode
248
self._allow_reserved = allow_reserved
251
return "Weave(%r)" % self._weave_name
253
def _check_write_ok(self):
254
"""Is the versioned file marked as 'finished' ? Raise if it is."""
255
if self._get_scope() != self._scope:
256
raise errors.OutSideTransaction()
257
if self._access_mode != 'w':
258
raise errors.ReadOnlyObjectDirtiedError(self)
261
"""Return a deep copy of self.
263
The copy can be modified without affecting the original weave."""
265
other._weave = self._weave[:]
266
other._parents = self._parents[:]
267
other._sha1s = self._sha1s[:]
268
other._names = self._names[:]
269
other._name_map = self._name_map.copy()
270
other._weave_name = self._weave_name
273
174
def __eq__(self, other):
274
175
if not isinstance(other, Weave):
276
177
return self._parents == other._parents \
277
and self._weave == other._weave \
278
and self._sha1s == other._sha1s
178
and self._weave == other._weave
280
181
def __ne__(self, other):
281
182
return not self.__eq__(other)
283
def _idx_to_name(self, version):
284
return self._names[version]
286
def _lookup(self, name):
287
"""Convert symbolic version name to index."""
288
if not self._allow_reserved:
289
self.check_not_reserved_id(name)
291
return self._name_map[name]
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(record))
377
self.add_lines(record.key[0], parents, lines)
378
except RevisionAlreadyPresent:
381
def _check_repeated_add(self, name, parents, text, sha1):
382
"""Check that a duplicated add is OK.
384
If it is, return the (old) index; otherwise raise an exception.
386
idx = self._lookup(name)
387
if sorted(self._parents[idx]) != sorted(parents) \
388
or sha1 != self._sha1s[idx]:
389
raise RevisionAlreadyPresent(name, self._weave_name)
392
def _add_lines(self, version_id, parents, lines, parent_texts,
393
left_matching_blocks, nostore_sha, random_id, check_content):
394
"""See VersionedFile.add_lines."""
395
idx = self._add(version_id, lines, map(self._lookup, parents),
396
nostore_sha=nostore_sha)
397
return sha_strings(lines), sum(map(len, lines)), idx
399
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
185
def add(self, parents, text):
400
186
"""Add a single text on top of the weave.
402
188
Returns the index number of the newly added version.
405
Symbolic name for this version.
406
(Typically the revision-id of the revision that added it.)
409
191
List or set of direct parent version numbers.
412
Sequence of lines to be added in the new version.
414
:param nostore_sha: See VersionedFile.add_lines.
416
self._check_lines_not_unicode(lines)
417
self._check_lines_are_lines(lines)
419
sha1 = sha_strings(lines)
420
if sha1 == nostore_sha:
421
raise errors.ExistingContent
422
if version_id in self._name_map:
423
return self._check_repeated_add(version_id, parents, lines, sha1)
194
Sequence of lines to be added in the new version."""
425
196
self._check_versions(parents)
426
## self._check_lines(lines)
197
## self._check_lines(text)
427
198
new_version = len(self._parents)
429
# if we abort after here the (in-memory) weave will be corrupt because only
430
# some fields are updated
431
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
432
# - Robert Collins 20060226
433
self._parents.append(parents[:])
206
# if we abort after here the weave will be corrupt
207
self._parents.append(frozenset(parents))
434
208
self._sha1s.append(sha1)
435
self._names.append(version_id)
436
self._name_map[version_id] = new_version
440
212
# special case; adding with no parents revision; can do
441
213
# this more quickly by just appending unconditionally.
442
214
# even more specially, if we're adding an empty text we
443
215
# need do nothing at all.
445
217
self._weave.append(('{', new_version))
446
self._weave.extend(lines)
447
self._weave.append(('}', None))
218
self._weave.extend(text)
219
self._weave.append(('}', new_version))
448
221
return new_version
450
223
if len(parents) == 1:
552
351
except IndexError:
553
352
raise IndexError("invalid version number %r" % i)
555
def _compatible_parents(self, my_parents, other_parents):
556
"""During join check that other_parents are joinable with my_parents.
558
Joinable is defined as 'is a subset of' - supersets may require
559
regeneration of diffs, but subsets do not.
561
return len(other_parents.difference(my_parents)) == 0
563
def annotate(self, version_id):
564
"""Return a list of (version-id, line) tuples for version_id.
355
def annotate(self, index):
356
return list(self.annotate_iter(index))
359
def annotate_iter(self, version):
360
"""Yield list of (index-id, line) pairs for the specified version.
566
362
The index indicates when the line originated in the weave."""
567
incls = [self._lookup(version_id)]
568
return [(self._idx_to_name(origin), text) for origin, lineno, text in
569
self._extract(incls)]
571
def iter_lines_added_or_present_in_versions(self, version_ids=None,
573
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
574
if version_ids is None:
575
version_ids = self.versions()
576
version_ids = set(version_ids)
577
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
578
# if inserted not in version_ids then it was inserted before the
579
# versions we care about, but because weaves cannot represent ghosts
580
# properly, we do not filter down to that
581
# if inserted not in version_ids: continue
583
yield line + '\n', inserted
587
def _walk_internal(self, version_ids=None):
588
"""Helper method for weave actions."""
363
for origin, lineno, text in self._extract([version]):
371
(lineno, insert, deletes, text)
372
for each literal line.
593
378
lineno = 0 # line of weave, 0-based
595
380
for l in self._weave:
596
if l.__class__ == tuple:
381
if isinstance(l, tuple):
600
istack.append(self._names[v])
604
dset.add(self._names[v])
606
dset.remove(self._names[v])
608
raise WeaveFormatError('unexpected instruction %r' % v)
610
yield lineno, istack[-1], frozenset(dset), l
394
raise WeaveFormatError('unexpected instruction %r'
397
assert isinstance(l, basestring)
399
yield lineno, istack[-1], dset, l
404
def _extract(self, versions):
405
"""Yield annotation of lines in included set.
407
Yields a sequence of tuples (origin, lineno, text), where
408
origin is the origin version, lineno the index in the weave,
409
and text the text of the line.
411
The set typically but not necessarily corresponds to a version.
413
included = self.inclusions(versions)
418
lineno = 0 # line of weave, 0-based
424
WFE = WeaveFormatError
426
for l in self._weave:
427
if isinstance(l, tuple):
431
assert v not in istack
446
assert isinstance(l, basestring)
448
isactive = (not dset) and istack and (istack[-1] in included)
450
result.append((istack[-1], lineno, l))
614
raise WeaveFormatError("unclosed insertion blocks "
615
"at end of weave: %s" % istack)
454
raise WFE("unclosed insertion blocks at end of weave",
617
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
457
raise WFE("unclosed deletion blocks at end of weave",
464
def get_iter(self, version):
465
"""Yield lines for the specified version."""
466
for origin, lineno, line in self._extract([version]):
470
def get(self, index):
471
return list(self.get_iter(index))
474
def mash_iter(self, included):
475
"""Return composed version of multiple included versions."""
476
for origin, lineno, text in self._extract(included):
480
def dump(self, to_file):
481
from pprint import pprint
482
print >>to_file, "Weave._weave = ",
483
pprint(self._weave, to_file)
484
print >>to_file, "Weave._parents = ",
485
pprint(self._parents, to_file)
489
def numversions(self):
490
l = len(self._parents)
491
assert l == len(self._sha1s)
496
return self.numversions()
499
def check(self, progress_bar=None):
500
# check no circular inclusions
501
for version in range(self.numversions()):
502
inclusions = list(self._parents[version])
505
if inclusions[-1] >= version:
506
raise WeaveFormatError("invalid included version %d for index %d"
507
% (inclusions[-1], version))
509
# try extracting all versions; this is a bit slow and parallel
510
# extraction could be used
512
nv = self.numversions()
513
for version in range(nv):
515
progress_bar.update('checking text', version, nv)
517
for l in self.get_iter(version):
520
expected = self._sha1s[version]
522
raise WeaveError("mismatched sha1 for version %d; "
523
"got %s, expected %s"
524
% (version, hd, expected))
526
# TODO: check insertions are properly nested, that there are
527
# no lines outside of insertion blocks, that deletions are
528
# properly paired, etc.
532
def merge(self, merge_versions):
533
"""Automerge and mark conflicts between versions.
535
This returns a sequence, each entry describing alternatives
536
for a chunk of the file. Each of the alternatives is given as
539
If there is a chunk of the file where there's no diagreement,
540
only one alternative is given.
543
# approach: find the included versions common to all the
545
raise NotImplementedError()
549
def _delta(self, included, lines):
550
"""Return changes from basis to new revision.
552
The old text for comparison is the union of included revisions.
554
This is used in inserting a new text.
556
Delta is returned as a sequence of
557
(weave1, weave2, newlines).
559
This indicates that weave1:weave2 of the old weave should be
560
replaced by the sequence of lines in newlines. Note that
561
these line numbers are positions in the total weave and don't
562
correspond to the lines in any extracted version, or even the
563
extracted union of included versions.
565
If line1=line2, this is a pure insert; if newlines=[] this is a
566
pure delete. (Similar to difflib.)
620
571
def plan_merge(self, ver_a, ver_b):
621
572
"""Return pseudo-annotation indicating how the two versions merge.
661
612
# not in either revision
662
613
yield 'irrelevant', line
664
def _extract(self, versions):
665
"""Yield annotation of lines in included set.
667
Yields a sequence of tuples (origin, lineno, text), where
668
origin is the origin version, lineno the index in the weave,
669
and text the text of the line.
671
The set typically but not necessarily corresponds to a version.
674
if not isinstance(i, int):
677
included = self._inclusions(versions)
683
lineno = 0 # line of weave, 0-based
689
WFE = WeaveFormatError
692
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
693
# +285282 0 1676.8040 1676.8040 +<isinstance>
694
# 1.6 seconds in 'isinstance'.
695
# changing the first isinstance:
696
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
697
# +140414 0 762.8050 762.8050 +<isinstance>
698
# note that the inline time actually dropped (less function calls)
699
# and total processing time was halved.
700
# we're still spending ~1/4 of the method in isinstance though.
701
# so lets hard code the acceptable string classes we expect:
702
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
703
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
705
# yay, down to ~1/4 the initial extract time, and our inline time
706
# has shrunk again, with isinstance no longer dominating.
707
# tweaking the stack inclusion test to use a set gives:
708
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
709
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
711
# - a 5% win, or possibly just noise. However with large istacks that
712
# 'in' test could dominate, so I'm leaving this change in place -
713
# when its fast enough to consider profiling big datasets we can review.
718
for l in self._weave:
719
if l.__class__ == tuple:
726
iset.remove(istack.pop())
615
yield 'unchanged', '' # terminator
619
def weave_merge(self, plan):
624
for state, line in plan:
625
if state == 'unchanged' or state == 'killed-both':
626
# resync and flush queued conflicts changes if any
627
if not lines_a and not lines_b:
629
elif ch_a and not ch_b:
631
for l in lines_a: yield l
632
elif ch_b and not ch_a:
633
for l in lines_b: yield l
634
elif lines_a == lines_b:
635
for l in lines_a: yield l
734
raise AssertionError()
737
isactive = (not dset) and istack and (istack[-1] in included)
739
result.append((istack[-1], lineno, l))
742
raise WeaveFormatError("unclosed insertion blocks "
743
"at end of weave: %s" % istack)
745
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
749
def _maybe_lookup(self, name_or_index):
750
"""Convert possible symbolic name to index, or pass through indexes.
754
if isinstance(name_or_index, (int, long)):
757
return self._lookup(name_or_index)
759
def get_lines(self, version_id):
760
"""See VersionedFile.get_lines()."""
761
int_index = self._maybe_lookup(version_id)
762
result = [line for (origin, lineno, line) in self._extract([int_index])]
763
expected_sha1 = self._sha1s[int_index]
764
measured_sha1 = sha_strings(result)
765
if measured_sha1 != expected_sha1:
766
raise errors.WeaveInvalidChecksum(
767
'file %s, revision %s, expected: %s, measured %s'
768
% (self._weave_name, version_id,
769
expected_sha1, measured_sha1))
772
def get_sha1s(self, version_ids):
773
"""See VersionedFile.get_sha1s()."""
775
for v in version_ids:
776
result[v] = self._sha1s[self._lookup(v)]
779
def num_versions(self):
780
"""How many versions are in this weave?"""
781
l = len(self._parents)
784
__len__ = num_versions
786
def check(self, progress_bar=None):
787
# TODO evaluate performance hit of using string sets in this routine.
788
# TODO: check no circular inclusions
789
# TODO: create a nested progress bar
790
for version in range(self.num_versions()):
791
inclusions = list(self._parents[version])
794
if inclusions[-1] >= version:
795
raise WeaveFormatError("invalid included version %d for index %d"
796
% (inclusions[-1], version))
798
# try extracting all versions; parallel extraction is used
799
nv = self.num_versions()
804
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
805
# The problem is that set membership is much more expensive
806
name = self._idx_to_name(i)
809
new_inc = set([name])
810
for p in self._parents[i]:
811
new_inc.update(inclusions[self._idx_to_name(p)])
813
if set(new_inc) != set(self.get_ancestry(name)):
814
raise AssertionError(
816
% (set(new_inc), set(self.get_ancestry(name))))
817
inclusions[name] = new_inc
819
nlines = len(self._weave)
821
update_text = 'checking weave'
823
short_name = os.path.basename(self._weave_name)
824
update_text = 'checking %s' % (short_name,)
825
update_text = update_text[:25]
827
for lineno, insert, deleteset, line in self._walk_internal():
829
progress_bar.update(update_text, lineno, nlines)
831
for name, name_inclusions in inclusions.items():
832
# The active inclusion must be an ancestor,
833
# and no ancestors must have deleted this line,
834
# because we don't support resurrection.
835
if (insert in name_inclusions) and not (deleteset & name_inclusions):
836
sha1s[name].update(line)
839
version = self._idx_to_name(i)
840
hd = sha1s[version].hexdigest()
841
expected = self._sha1s[i]
843
raise errors.WeaveInvalidChecksum(
844
"mismatched sha1 for version %s: "
845
"got %s, expected %s"
846
% (version, hd, expected))
848
# TODO: check insertions are properly nested, that there are
849
# no lines outside of insertion blocks, that deletions are
850
# properly paired, etc.
852
def _imported_parents(self, other, other_idx):
853
"""Return list of parents in self corresponding to indexes in other."""
855
for parent_idx in other._parents[other_idx]:
856
parent_name = other._names[parent_idx]
857
if parent_name not in self._name_map:
858
# should not be possible
859
raise WeaveError("missing parent {%s} of {%s} in %r"
860
% (parent_name, other._name_map[other_idx], self))
861
new_parents.append(self._name_map[parent_name])
864
def _check_version_consistent(self, other, other_idx, name):
865
"""Check if a version in consistent in this and other.
867
To be consistent it must have:
870
* the same direct parents (by name, not index, and disregarding
873
If present & correct return True;
874
if not present in self return False;
875
if inconsistent raise error."""
876
this_idx = self._name_map.get(name, -1)
878
if self._sha1s[this_idx] != other._sha1s[other_idx]:
879
raise errors.WeaveTextDiffers(name, self, other)
880
self_parents = self._parents[this_idx]
881
other_parents = other._parents[other_idx]
882
n1 = set([self._names[i] for i in self_parents])
883
n2 = set([other._names[i] for i in other_parents])
884
if not self._compatible_parents(n1, n2):
885
raise WeaveParentMismatch("inconsistent parents "
886
"for version {%s}: %s vs %s" % (name, n1, n2))
892
def _reweave(self, other, pb, msg):
893
"""Reweave self with other - internal helper for join().
895
:param other: The other weave to merge
896
:param pb: An optional progress bar, indicating how far done we are
897
:param msg: An optional message for the progress
899
new_weave = _reweave(self, other, pb=pb, msg=msg)
900
self._copy_weave_content(new_weave)
902
def _copy_weave_content(self, otherweave):
903
"""adsorb the content from otherweave."""
904
for attr in self.__slots__:
905
if attr != '_weave_name':
906
setattr(self, attr, copy(getattr(otherweave, attr)))
909
class WeaveFile(Weave):
910
"""A WeaveFile represents a Weave on disk and writes on change."""
912
WEAVE_SUFFIX = '.weave'
914
def __init__(self, name, transport, filemode=None, create=False, access_mode='w', get_scope=None):
915
"""Create a WeaveFile.
917
:param create: If not True, only open an existing knit.
919
super(WeaveFile, self).__init__(name, access_mode, get_scope=get_scope,
920
allow_reserved=False)
921
self._transport = transport
922
self._filemode = filemode
924
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
925
except errors.NoSuchFile:
931
def _add_lines(self, version_id, parents, lines, parent_texts,
932
left_matching_blocks, nostore_sha, random_id, check_content):
933
"""Add a version and save the weave."""
934
self.check_not_reserved_id(version_id)
935
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
936
parent_texts, left_matching_blocks, nostore_sha, random_id,
941
def copy_to(self, name, transport):
942
"""See VersionedFile.copy_to()."""
943
# as we are all in memory always, just serialise to the new place.
945
write_weave_v5(self, sio)
947
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
950
"""Save the weave."""
951
self._check_write_ok()
953
write_weave_v5(self, sio)
955
bytes = sio.getvalue()
956
path = self._weave_name + WeaveFile.WEAVE_SUFFIX
958
self._transport.put_bytes(path, bytes, self._filemode)
959
except errors.NoSuchFile:
960
self._transport.mkdir(dirname(path))
961
self._transport.put_bytes(path, bytes, self._filemode)
965
"""See VersionedFile.get_suffixes()."""
966
return [WeaveFile.WEAVE_SUFFIX]
968
def insert_record_stream(self, stream):
969
super(WeaveFile, self).insert_record_stream(stream)
972
@deprecated_method(one_five)
973
def join(self, other, pb=None, msg=None, version_ids=None,
974
ignore_missing=False):
975
"""Join other into self and save."""
976
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
980
def _reweave(wa, wb, pb=None, msg=None):
981
"""Combine two weaves and return the result.
983
This works even if a revision R has different parents in
984
wa and wb. In the resulting weave all the parents are given.
986
This is done by just building up a new weave, maintaining ordering
987
of the versions in the two inputs. More efficient approaches
988
might be possible but it should only be necessary to do
989
this operation rarely, when a new previously ghost version is
992
:param pb: An optional progress bar, indicating how far done we are
993
:param msg: An optional message for the progress
997
queue_a = range(wa.num_versions())
998
queue_b = range(wb.num_versions())
999
# first determine combined parents of all versions
1000
# map from version name -> all parent names
1001
combined_parents = _reweave_parent_graphs(wa, wb)
1002
mutter("combined parents: %r", combined_parents)
1003
order = tsort.topo_sort(combined_parents.iteritems())
1004
mutter("order to reweave: %r", order)
1009
for idx, name in enumerate(order):
1011
pb.update(msg, idx, len(order))
1012
if name in wa._name_map:
1013
lines = wa.get_lines(name)
1014
if name in wb._name_map:
1015
lines_b = wb.get_lines(name)
1016
if lines != lines_b:
1017
mutter('Weaves differ on content. rev_id {%s}', name)
1018
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1020
lines = list(difflib.unified_diff(lines, lines_b,
1021
wa._weave_name, wb._weave_name))
1022
mutter('lines:\n%s', ''.join(lines))
1023
raise errors.WeaveTextDiffers(name, wa, wb)
1025
lines = wb.get_lines(name)
1026
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1029
def _reweave_parent_graphs(wa, wb):
1030
"""Return combined parent ancestry for two weaves.
1032
Returned as a list of (version_name, set(parent_names))"""
1034
for weave in [wa, wb]:
1035
for idx, name in enumerate(weave._names):
1036
p = combined.setdefault(name, set())
1037
p.update(map(weave._idx_to_name, weave._parents[idx]))
1042
"""Show the weave's table-of-contents"""
1043
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1044
for i in (6, 50, 10, 10):
638
for l in lines_a: yield l
640
for l in lines_b: yield l
647
if state == 'unchanged':
650
elif state == 'killed-a':
653
elif state == 'killed-b':
656
elif state == 'new-a':
659
elif state == 'new-b':
663
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
674
"""Show some text information about the weave."""
675
print '%6s %40s %20s' % ('ver', 'sha1', 'parents')
676
for i in (6, 40, 20):
1047
for i in range(w.num_versions()):
679
for i in range(w.numversions()):
1048
680
sha1 = w._sha1s[i]
1050
parent_str = ' '.join(map(str, w._parents[i]))
1051
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1055
def weave_stats(weave_file, pb):
681
print '%6d %40s %s' % (i, sha1, ' '.join(map(str, w._parents[i])))
685
def weave_stats(weave_file):
686
from bzrlib.progress import ProgressBar
1056
687
from bzrlib.weavefile import read_weave
1058
691
wf = file(weave_file, 'rb')
1059
692
w = read_weave(wf)
1060
693
# FIXME: doesn't work on pipes
added
removed
bzrlib/weave.py
