59
61
# where the basis and destination are unchanged.
61
63
# FIXME: Sometimes we will be given a parents list for a revision
62
# that includes some redundant parents (i.e. already a parent of
63
# something in the list.) We should eliminate them. This can
64
# that includes some redundant parents (i.e. already a parent of
65
# something in the list.) We should eliminate them. This can
64
66
# be done fairly efficiently because the sequence numbers constrain
65
67
# the possible relationships.
67
# FIXME: the conflict markers should be *7* characters
70
from cStringIO import StringIO
72
from difflib import SequenceMatcher
75
from bzrlib.lazy_import import lazy_import
76
lazy_import(globals(), """
77
from bzrlib import tsort
74
from bzrlib.trace import mutter
84
75
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
85
RevisionAlreadyPresent,
87
UnavailableRepresentation,
88
WeaveRevisionAlreadyPresent,
89
WeaveRevisionNotPresent,
91
from bzrlib.osutils import dirname, sha, sha_strings, split_lines
92
import bzrlib.patiencediff
93
from bzrlib.revision import NULL_REVISION
94
from bzrlib.symbol_versioning import *
95
from bzrlib.trace import mutter
96
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])
125
elif storage_kind == 'chunked':
126
return self._weave.get_lines(self.key[-1])
128
raise UnavailableRepresentation(self.key, storage_kind, 'fulltext')
131
class Weave(VersionedFile):
76
WeaveRevisionNotPresent, WeaveRevisionAlreadyPresent)
77
import bzrlib.errors as errors
78
from bzrlib.tsort import topo_sort
132
82
"""weave - versioned text file storage.
134
84
A Weave manages versions of line-based text files, keeping track
135
85
of the originating version for each line.
220
170
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
221
'_weave_name', '_matcher', '_allow_reserved']
223
def __init__(self, weave_name=None, access_mode='w', matcher=None,
224
get_scope=None, allow_reserved=False):
227
:param get_scope: A callable that returns an opaque object to be used
228
for detecting when this weave goes out of scope (should stop
229
answering requests or allowing mutation).
231
super(Weave, self).__init__()
173
def __init__(self, weave_name=None):
233
175
self._parents = []
236
178
self._name_map = {}
237
179
self._weave_name = weave_name
239
self._matcher = bzrlib.patiencediff.PatienceSequenceMatcher
241
self._matcher = matcher
242
if get_scope is None:
243
get_scope = lambda:None
244
self._get_scope = get_scope
245
self._scope = get_scope()
246
self._access_mode = access_mode
247
self._allow_reserved = allow_reserved
249
181
def __repr__(self):
250
182
return "Weave(%r)" % self._weave_name
252
def _check_write_ok(self):
253
"""Is the versioned file marked as 'finished' ? Raise if it is."""
254
if self._get_scope() != self._scope:
255
raise errors.OutSideTransaction()
256
if self._access_mode != 'w':
257
raise errors.ReadOnlyObjectDirtiedError(self)
260
186
"""Return a deep copy of self.
262
188
The copy can be modified without affecting the original weave."""
264
190
other._weave = self._weave[:]
275
201
return self._parents == other._parents \
276
202
and self._weave == other._weave \
277
and self._sha1s == other._sha1s
203
and self._sha1s == other._sha1s
279
206
def __ne__(self, other):
280
207
return not self.__eq__(other)
282
def _idx_to_name(self, version):
283
return self._names[version]
285
def _lookup(self, name):
209
def __contains__(self, name):
210
return self._name_map.has_key(name)
212
def maybe_lookup(self, name_or_index):
213
"""Convert possible symbolic name to index, or pass through indexes."""
214
if isinstance(name_or_index, (int, long)):
217
return self.lookup(name_or_index)
220
def lookup(self, name):
286
221
"""Convert symbolic version name to index."""
287
if not self._allow_reserved:
288
self.check_not_reserved_id(name)
290
223
return self._name_map[name]
292
raise RevisionNotPresent(name, self._weave_name)
225
raise WeaveRevisionNotPresent(name, self)
295
"""See VersionedFile.versions."""
296
228
return self._names[:]
298
def has_version(self, version_id):
299
"""See VersionedFile.has_version."""
300
return (version_id in self._name_map)
302
__contains__ = has_version
304
def get_record_stream(self, versions, ordering, include_delta_closure):
305
"""Get a stream of records for versions.
307
:param versions: The versions to include. Each version is a tuple
309
:param ordering: Either 'unordered' or 'topological'. A topologically
310
sorted stream has compression parents strictly before their
312
:param include_delta_closure: If True then the closure across any
313
compression parents will be included (in the opaque data).
314
:return: An iterator of ContentFactory objects, each of which is only
315
valid until the iterator is advanced.
317
versions = [version[-1] for version in versions]
318
if ordering == 'topological':
319
parents = self.get_parent_map(versions)
320
new_versions = tsort.topo_sort(parents)
321
new_versions.extend(set(versions).difference(set(parents)))
322
versions = new_versions
323
elif ordering == 'groupcompress':
324
parents = self.get_parent_map(versions)
325
new_versions = sort_groupcompress(parents)
326
new_versions.extend(set(versions).difference(set(parents)))
327
versions = new_versions
328
for version in versions:
330
yield WeaveContentFactory(version, self)
332
yield AbsentContentFactory((version,))
334
def get_parent_map(self, version_ids):
335
"""See VersionedFile.get_parent_map."""
337
for version_id in version_ids:
338
if version_id == NULL_REVISION:
343
map(self._idx_to_name,
344
self._parents[self._lookup(version_id)]))
345
except RevisionNotPresent:
347
result[version_id] = parents
350
def get_parents_with_ghosts(self, version_id):
351
raise NotImplementedError(self.get_parents_with_ghosts)
353
def insert_record_stream(self, stream):
354
"""Insert a record stream into this versioned file.
356
:param stream: A stream of records to insert.
358
:seealso VersionedFile.get_record_stream:
361
for record in stream:
362
# Raise an error when a record is missing.
363
if record.storage_kind == 'absent':
364
raise RevisionNotPresent([record.key[0]], self)
365
# adapt to non-tuple interface
366
parents = [parent[0] for parent in record.parents]
367
if (record.storage_kind == 'fulltext'
368
or record.storage_kind == 'chunked'):
369
self.add_lines(record.key[0], parents,
370
osutils.chunks_to_lines(record.get_bytes_as('chunked')))
372
adapter_key = record.storage_kind, 'fulltext'
374
adapter = adapters[adapter_key]
376
adapter_factory = adapter_registry.get(adapter_key)
377
adapter = adapter_factory(self)
378
adapters[adapter_key] = adapter
379
lines = split_lines(adapter.get_bytes(record))
381
self.add_lines(record.key[0], parents, lines)
382
except RevisionAlreadyPresent:
230
def iter_names(self):
231
"""Yield a list of all names in this weave."""
232
return iter(self._names)
234
def idx_to_name(self, version):
235
return self._names[version]
385
237
def _check_repeated_add(self, name, parents, text, sha1):
386
238
"""Check that a duplicated add is OK.
388
240
If it is, return the (old) index; otherwise raise an exception.
390
idx = self._lookup(name)
242
idx = self.lookup(name)
391
243
if sorted(self._parents[idx]) != sorted(parents) \
392
244
or sha1 != self._sha1s[idx]:
393
raise RevisionAlreadyPresent(name, self._weave_name)
245
raise WeaveRevisionAlreadyPresent(name, self)
396
def _add_lines(self, version_id, parents, lines, parent_texts,
397
left_matching_blocks, nostore_sha, random_id, check_content):
398
"""See VersionedFile.add_lines."""
399
idx = self._add(version_id, lines, map(self._lookup, parents),
400
nostore_sha=nostore_sha)
401
return sha_strings(lines), sum(map(len, lines)), idx
403
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
248
def add(self, name, parents, text, sha1=None):
404
249
"""Add a single text on top of the weave.
406
251
Returns the index number of the newly added version.
409
254
Symbolic name for this version.
410
255
(Typically the revision-id of the revision that added it.)
411
If None, a name will be allocated based on the hash. (sha1:SHAHASH)
414
258
List or set of direct parent version numbers.
417
261
Sequence of lines to be added in the new version.
419
:param nostore_sha: See VersionedFile.add_lines.
263
sha -- SHA-1 of the file, if known. This is trusted to be
421
self._check_lines_not_unicode(lines)
422
self._check_lines_are_lines(lines)
424
sha1 = sha_strings(lines)
425
if sha1 == nostore_sha:
426
raise errors.ExistingContent
427
if version_id is None:
428
version_id = "sha1:" + sha1
429
if version_id in self._name_map:
430
return self._check_repeated_add(version_id, parents, lines, sha1)
266
from bzrlib.osutils import sha_strings
268
assert isinstance(name, basestring)
270
sha1 = sha_strings(text)
271
if name in self._name_map:
272
return self._check_repeated_add(name, parents, text, sha1)
274
parents = map(self.maybe_lookup, parents)
432
275
self._check_versions(parents)
433
## self._check_lines(lines)
276
## self._check_lines(text)
434
277
new_version = len(self._parents)
436
280
# if we abort after here the (in-memory) weave will be corrupt because only
437
281
# some fields are updated
438
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
439
# - Robert Collins 20060226
440
282
self._parents.append(parents[:])
441
283
self._sha1s.append(sha1)
442
self._names.append(version_id)
443
self._name_map[version_id] = new_version
284
self._names.append(name)
285
self._name_map[name] = new_version
447
289
# special case; adding with no parents revision; can do
448
290
# this more quickly by just appending unconditionally.
449
291
# even more specially, if we're adding an empty text we
450
292
# need do nothing at all.
452
294
self._weave.append(('{', new_version))
453
self._weave.extend(lines)
295
self._weave.extend(text)
454
296
self._weave.append(('}', None))
455
298
return new_version
457
300
if len(parents) == 1:
559
438
except IndexError:
560
439
raise IndexError("invalid version number %r" % i)
562
def _compatible_parents(self, my_parents, other_parents):
563
"""During join check that other_parents are joinable with my_parents.
565
Joinable is defined as 'is a subset of' - supersets may require
566
regeneration of diffs, but subsets do not.
568
return len(other_parents.difference(my_parents)) == 0
570
def annotate(self, version_id):
571
"""Return a list of (version-id, line) tuples for version_id.
442
def annotate(self, name_or_index):
443
return list(self.annotate_iter(name_or_index))
446
def annotate_iter(self, name_or_index):
447
"""Yield list of (index-id, line) pairs for the specified version.
573
449
The index indicates when the line originated in the weave."""
574
incls = [self._lookup(version_id)]
575
return [(self._idx_to_name(origin), text) for origin, lineno, text in
576
self._extract(incls)]
578
def iter_lines_added_or_present_in_versions(self, version_ids=None,
580
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
581
if version_ids is None:
582
version_ids = self.versions()
583
version_ids = set(version_ids)
584
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
585
if inserted not in version_ids: continue
587
yield line + '\n', inserted
591
def _walk_internal(self, version_ids=None):
592
"""Helper method for weave actions."""
450
incls = [self.maybe_lookup(name_or_index)]
451
for origin, lineno, text in self._extract(incls):
458
(lineno, insert, deletes, text)
459
for each literal line.
597
465
lineno = 0 # line of weave, 0-based
599
467
for l in self._weave:
600
if l.__class__ == tuple:
468
if isinstance(l, tuple):
604
istack.append(self._names[v])
608
dset.add(self._names[v])
610
dset.remove(self._names[v])
612
481
raise WeaveFormatError('unexpected instruction %r' % v)
614
yield lineno, istack[-1], frozenset(dset), l
483
assert isinstance(l, basestring)
485
yield lineno, istack[-1], dset, l
753
def _maybe_lookup(self, name_or_index):
754
"""Convert possible symbolic name to index, or pass through indexes.
555
def get_iter(self, name_or_index):
556
"""Yield lines for the specified version."""
557
incls = [self.maybe_lookup(name_or_index)]
562
# We don't have sha1 sums for multiple entries
564
for origin, lineno, line in self._extract(incls):
569
expected_sha1 = self._sha1s[index]
570
measured_sha1 = cur_sha.hexdigest()
571
if measured_sha1 != expected_sha1:
572
raise errors.WeaveInvalidChecksum(
573
'file %s, revision %s, expected: %s, measured %s'
574
% (self._weave_name, self._names[index],
575
expected_sha1, measured_sha1))
578
def get_text(self, name_or_index):
579
return ''.join(self.get_iter(name_or_index))
580
assert isinstance(version, int)
583
def get_lines(self, name_or_index):
584
return list(self.get_iter(name_or_index))
590
def get_sha1(self, name):
591
"""Get the stored sha1 sum for the given revision.
593
:param name: The name of the version to lookup
758
if isinstance(name_or_index, (int, long)):
761
return self._lookup(name_or_index)
763
def get_lines(self, version_id):
764
"""See VersionedFile.get_lines()."""
765
int_index = self._maybe_lookup(version_id)
766
result = [line for (origin, lineno, line) in self._extract([int_index])]
767
expected_sha1 = self._sha1s[int_index]
768
measured_sha1 = sha_strings(result)
769
if measured_sha1 != expected_sha1:
770
raise errors.WeaveInvalidChecksum(
771
'file %s, revision %s, expected: %s, measured %s'
772
% (self._weave_name, version_id,
773
expected_sha1, measured_sha1))
776
def get_sha1s(self, version_ids):
777
"""See VersionedFile.get_sha1s()."""
779
for v in version_ids:
780
result[v] = self._sha1s[self._lookup(v)]
783
def num_versions(self):
784
"""How many versions are in this weave?"""
595
return self._sha1s[self.lookup(name)]
597
def mash_iter(self, included):
598
"""Return composed version of multiple included versions."""
599
included = map(self.maybe_lookup, included)
600
for origin, lineno, text in self._extract(included):
604
def dump(self, to_file):
605
from pprint import pprint
606
print >>to_file, "Weave._weave = ",
607
pprint(self._weave, to_file)
608
print >>to_file, "Weave._parents = ",
609
pprint(self._parents, to_file)
613
def numversions(self):
785
614
l = len(self._parents)
615
assert l == len(self._sha1s)
788
__len__ = num_versions
620
return self.numversions()
790
622
def check(self, progress_bar=None):
791
# TODO evaluate performance hit of using string sets in this routine.
792
# TODO: check no circular inclusions
793
# TODO: create a nested progress bar
794
for version in range(self.num_versions()):
623
# check no circular inclusions
624
for version in range(self.numversions()):
795
625
inclusions = list(self._parents[version])
797
627
inclusions.sort()
828
652
update_text = 'checking %s' % (short_name,)
829
653
update_text = update_text[:25]
831
for lineno, insert, deleteset, line in self._walk_internal():
655
for lineno, insert, deleteset, line in self._walk():
833
657
progress_bar.update(update_text, lineno, nlines)
835
for name, name_inclusions in inclusions.items():
659
for j, j_inc in enumerate(inclusions):
836
660
# The active inclusion must be an ancestor,
837
661
# and no ancestors must have deleted this line,
838
662
# because we don't support resurrection.
839
if (insert in name_inclusions) and not (deleteset & name_inclusions):
840
sha1s[name].update(line)
663
if (insert in j_inc) and not (deleteset & j_inc):
664
sha1s[j].update(line)
843
version = self._idx_to_name(i)
666
for version in range(nv):
844
667
hd = sha1s[version].hexdigest()
845
expected = self._sha1s[i]
668
expected = self._sha1s[version]
846
669
if hd != expected:
847
670
raise errors.WeaveInvalidChecksum(
848
671
"mismatched sha1 for version %s: "
849
672
"got %s, expected %s"
850
% (version, hd, expected))
673
% (self._names[version], hd, expected))
852
675
# TODO: check insertions are properly nested, that there are
853
676
# no lines outside of insertion blocks, that deletions are
854
677
# properly paired, etc.
679
def _delta(self, included, lines):
680
"""Return changes from basis to new revision.
682
The old text for comparison is the union of included revisions.
684
This is used in inserting a new text.
686
Delta is returned as a sequence of
687
(weave1, weave2, newlines).
689
This indicates that weave1:weave2 of the old weave should be
690
replaced by the sequence of lines in newlines. Note that
691
these line numbers are positions in the total weave and don't
692
correspond to the lines in any extracted version, or even the
693
extracted union of included versions.
695
If line1=line2, this is a pure insert; if newlines=[] this is a
696
pure delete. (Similar to difflib.)
698
raise NotImplementedError()
701
def plan_merge(self, ver_a, ver_b):
702
"""Return pseudo-annotation indicating how the two versions merge.
704
This is computed between versions a and b and their common
707
Weave lines present in none of them are skipped entirely.
709
inc_a = self.inclusions([ver_a])
710
inc_b = self.inclusions([ver_b])
711
inc_c = inc_a & inc_b
713
for lineno, insert, deleteset, line in self._walk():
714
if deleteset & inc_c:
715
# killed in parent; can't be in either a or b
716
# not relevant to our work
717
yield 'killed-base', line
718
elif insert in inc_c:
719
# was inserted in base
720
killed_a = bool(deleteset & inc_a)
721
killed_b = bool(deleteset & inc_b)
722
if killed_a and killed_b:
723
yield 'killed-both', line
725
yield 'killed-a', line
727
yield 'killed-b', line
729
yield 'unchanged', line
730
elif insert in inc_a:
731
if deleteset & inc_a:
732
yield 'ghost-a', line
736
elif insert in inc_b:
737
if deleteset & inc_b:
738
yield 'ghost-b', line
742
# not in either revision
743
yield 'irrelevant', line
745
yield 'unchanged', '' # terminator
749
def weave_merge(self, plan):
753
# TODO: Return a structured form of the conflicts (e.g. 2-tuples for
754
# conflicted regions), rather than just inserting the markers.
756
# TODO: Show some version information (e.g. author, date) on
757
# conflicted regions.
758
for state, line in plan:
759
if state == 'unchanged' or state == 'killed-both':
760
# resync and flush queued conflicts changes if any
761
if not lines_a and not lines_b:
763
elif ch_a and not ch_b:
765
for l in lines_a: yield l
766
elif ch_b and not ch_a:
767
for l in lines_b: yield l
768
elif lines_a == lines_b:
769
for l in lines_a: yield l
772
for l in lines_a: yield l
774
for l in lines_b: yield l
781
if state == 'unchanged':
784
elif state == 'killed-a':
787
elif state == 'killed-b':
790
elif state == 'new-a':
793
elif state == 'new-b':
797
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
802
def join(self, other, pb=None, msg=None):
804
"""Integrate versions from other into this weave.
806
The resulting weave contains all the history of both weaves;
807
any version you could retrieve from either self or other can be
808
retrieved from self after this call.
810
It is illegal for the two weaves to contain different values
811
or different parents for any version. See also reweave().
813
:param other: The other weave to pull into this one
814
:param pb: An optional progress bar
815
:param msg: An optional message to display for progress
817
if other.numversions() == 0:
818
return # nothing to update, easy
819
# two loops so that we do not change ourselves before verifying it
821
# work through in index order to make sure we get all dependencies
824
for other_idx, name in enumerate(other._names):
825
self._check_version_consistent(other, other_idx, name)
826
sha1 = other._sha1s[other_idx]
830
if name in self._name_map:
831
idx = self.lookup(name)
832
n1 = set(map(other.idx_to_name, other._parents[other_idx]))
833
n2 = set(map(self.idx_to_name, self._parents[idx]))
834
if sha1 == self._sha1s[idx] and n1 == n2:
837
names_to_join.append((other_idx, name))
844
for other_idx, name in names_to_join:
845
# TODO: If all the parents of the other version are already
846
# present then we can avoid some work by just taking the delta
847
# and adjusting the offsets.
848
new_parents = self._imported_parents(other, other_idx)
849
sha1 = other._sha1s[other_idx]
854
pb.update(msg, merged, len(names_to_join))
856
lines = other.get_lines(other_idx)
857
self.add(name, new_parents, lines, sha1)
859
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
860
merged, processed, self._weave_name, time.time( )-time0))
856
862
def _imported_parents(self, other, other_idx):
857
863
"""Return list of parents in self corresponding to indexes in other."""
859
865
for parent_idx in other._parents[other_idx]:
860
866
parent_name = other._names[parent_idx]
861
if parent_name not in self._name_map:
867
if parent_name not in self._names:
862
868
# should not be possible
863
raise WeaveError("missing parent {%s} of {%s} in %r"
869
raise WeaveError("missing parent {%s} of {%s} in %r"
864
870
% (parent_name, other._name_map[other_idx], self))
865
871
new_parents.append(self._name_map[parent_name])
866
872
return new_parents
896
def _reweave(self, other, pb, msg):
897
"""Reweave self with other - internal helper for join().
904
def reweave(self, other, pb=None, msg=None):
905
"""Reweave self with other.
899
907
:param other: The other weave to merge
900
908
:param pb: An optional progress bar, indicating how far done we are
901
909
:param msg: An optional message for the progress
903
new_weave = _reweave(self, other, pb=pb, msg=msg)
904
self._copy_weave_content(new_weave)
906
def _copy_weave_content(self, otherweave):
907
"""adsorb the content from otherweave."""
911
new_weave = reweave(self, other, pb=pb, msg=msg)
908
912
for attr in self.__slots__:
909
if attr != '_weave_name':
910
setattr(self, attr, copy(getattr(otherweave, attr)))
913
class WeaveFile(Weave):
914
"""A WeaveFile represents a Weave on disk and writes on change."""
916
WEAVE_SUFFIX = '.weave'
918
def __init__(self, name, transport, filemode=None, create=False, access_mode='w', get_scope=None):
919
"""Create a WeaveFile.
921
:param create: If not True, only open an existing knit.
923
super(WeaveFile, self).__init__(name, access_mode, get_scope=get_scope,
924
allow_reserved=False)
925
self._transport = transport
926
self._filemode = filemode
928
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
929
except errors.NoSuchFile:
935
def _add_lines(self, version_id, parents, lines, parent_texts,
936
left_matching_blocks, nostore_sha, random_id, check_content):
937
"""Add a version and save the weave."""
938
self.check_not_reserved_id(version_id)
939
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
940
parent_texts, left_matching_blocks, nostore_sha, random_id,
945
def copy_to(self, name, transport):
946
"""See VersionedFile.copy_to()."""
947
# as we are all in memory always, just serialise to the new place.
949
write_weave_v5(self, sio)
951
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
954
"""Save the weave."""
955
self._check_write_ok()
957
write_weave_v5(self, sio)
959
bytes = sio.getvalue()
960
path = self._weave_name + WeaveFile.WEAVE_SUFFIX
962
self._transport.put_bytes(path, bytes, self._filemode)
963
except errors.NoSuchFile:
964
self._transport.mkdir(dirname(path))
965
self._transport.put_bytes(path, bytes, self._filemode)
969
"""See VersionedFile.get_suffixes()."""
970
return [WeaveFile.WEAVE_SUFFIX]
972
def insert_record_stream(self, stream):
973
super(WeaveFile, self).insert_record_stream(stream)
977
def _reweave(wa, wb, pb=None, msg=None):
913
setattr(self, attr, getattr(new_weave, attr))
916
def reweave(wa, wb, pb=None, msg=None):
978
917
"""Combine two weaves and return the result.
980
This works even if a revision R has different parents in
919
This works even if a revision R has different parents in
981
920
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
922
This is done by just building up a new weave, maintaining ordering
984
923
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
924
might be possible but it should only be necessary to do
925
this operation rarely, when a new previously ghost version is
989
928
:param pb: An optional progress bar, indicating how far done we are
1020
959
raise errors.WeaveTextDiffers(name, wa, wb)
1022
961
lines = wb.get_lines(name)
1023
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
962
wr.add(name, combined_parents[name], lines)
1027
966
def _reweave_parent_graphs(wa, wb):
1028
967
"""Return combined parent ancestry for two weaves.
1030
969
Returned as a list of (version_name, set(parent_names))"""
1032
971
for weave in [wa, wb]:
1033
972
for idx, name in enumerate(weave._names):
1034
973
p = combined.setdefault(name, set())
1035
p.update(map(weave._idx_to_name, weave._parents[idx]))
974
p.update(map(weave.idx_to_name, weave._parents[idx]))
979
"""Show the weave's table-of-contents"""
980
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
981
for i in (6, 50, 10, 10):
984
for i in range(w.numversions()):
987
parent_str = ' '.join(map(str, w._parents[i]))
988
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
992
def weave_stats(weave_file, pb):
993
from bzrlib.weavefile import read_weave
995
wf = file(weave_file, 'rb')
997
# FIXME: doesn't work on pipes
998
weave_size = wf.tell()
1002
for i in range(vers):
1003
pb.update('checking sizes', i, vers)
1004
for origin, lineno, line in w._extract([i]):
1009
print 'versions %9d' % vers
1010
print 'weave file %9d bytes' % weave_size
1011
print 'total contents %9d bytes' % total
1012
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1015
print 'average size %9d bytes' % avg
1016
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1020
print """bzr weave tool
1022
Experimental tool for weave algorithm.
1025
weave init WEAVEFILE
1026
Create an empty weave file
1027
weave get WEAVEFILE VERSION
1028
Write out specified version.
1029
weave check WEAVEFILE
1030
Check consistency of all versions.
1032
Display table of contents.
1033
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1034
Add NEWTEXT, with specified parent versions.
1035
weave annotate WEAVEFILE VERSION
1036
Display origin of each line.
1037
weave mash WEAVEFILE VERSION...
1038
Display composite of all selected versions.
1039
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1040
Auto-merge two versions and display conflicts.
1041
weave diff WEAVEFILE VERSION1 VERSION2
1042
Show differences between two versions.
1046
% weave init foo.weave
1048
% weave add foo.weave ver0 < foo.txt
1051
(create updated version)
1053
% weave get foo.weave 0 | diff -u - foo.txt
1054
% weave add foo.weave ver1 0 < foo.txt
1057
% weave get foo.weave 0 > foo.txt (create forked version)
1059
% weave add foo.weave ver2 0 < foo.txt
1062
% weave merge foo.weave 1 2 > foo.txt (merge them)
1063
% vi foo.txt (resolve conflicts)
1064
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1076
# in case we're run directly from the subdirectory
1077
sys.path.append('..')
1079
from bzrlib.weavefile import write_weave, read_weave
1080
from bzrlib.progress import ProgressBar
1095
return read_weave(file(argv[2], 'rb'))
1101
# at the moment, based on everything in the file
1103
parents = map(int, argv[4:])
1104
lines = sys.stdin.readlines()
1105
ver = w.add(name, parents, lines)
1106
write_weave(w, file(argv[2], 'wb'))
1107
print 'added version %r %d' % (name, ver)
1110
if os.path.exists(fn):
1111
raise IOError("file exists")
1113
write_weave(w, file(fn, 'wb'))
1114
elif cmd == 'get': # get one version
1116
sys.stdout.writelines(w.get_iter(int(argv[3])))
1118
elif cmd == 'mash': # get composite
1120
sys.stdout.writelines(w.mash_iter(map(int, argv[3:])))
1123
from difflib import unified_diff
1126
v1, v2 = map(int, argv[3:5])
1129
diff_gen = unified_diff(lines1, lines2,
1130
'%s version %d' % (fn, v1),
1131
'%s version %d' % (fn, v2))
1132
sys.stdout.writelines(diff_gen)
1134
elif cmd == 'annotate':
1136
# newline is added to all lines regardless; too hard to get
1137
# reasonable formatting otherwise
1139
for origin, text in w.annotate(int(argv[3])):
1140
text = text.rstrip('\r\n')
1142
print ' | %s' % (text)
1144
print '%5d | %s' % (origin, text)
1150
elif cmd == 'stats':
1151
weave_stats(argv[2], ProgressBar())
1153
elif cmd == 'check':
1158
print '%d versions ok' % w.numversions()
1160
elif cmd == 'inclusions':
1162
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1164
elif cmd == 'parents':
1166
print ' '.join(map(str, w._parents[int(argv[3])]))
1168
elif cmd == 'plan-merge':
1170
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1172
print '%14s | %s' % (state, line),
1174
elif cmd == 'merge':
1176
p = w.plan_merge(int(argv[3]), int(argv[4]))
1177
sys.stdout.writelines(w.weave_merge(p))
1179
elif cmd == 'mash-merge':
1185
v1, v2 = map(int, argv[3:5])
1187
basis = w.inclusions([v1]).intersection(w.inclusions([v2]))
1189
base_lines = list(w.mash_iter(basis))
1190
a_lines = list(w.get(v1))
1191
b_lines = list(w.get(v2))
1193
from bzrlib.merge3 import Merge3
1194
m3 = Merge3(base_lines, a_lines, b_lines)
1196
name_a = 'version %d' % v1
1197
name_b = 'version %d' % v2
1198
sys.stdout.writelines(m3.merge_lines(name_a=name_a, name_b=name_b))
1200
raise ValueError('unknown command %r' % cmd)
1204
def profile_main(argv):
1205
import tempfile, hotshot, hotshot.stats
1207
prof_f = tempfile.NamedTemporaryFile()
1209
prof = hotshot.Profile(prof_f.name)
1211
ret = prof.runcall(main, argv)
1214
stats = hotshot.stats.load(prof_f.name)
1216
stats.sort_stats('cumulative')
1217
## XXX: Might like to write to stderr or the trace file instead but
1218
## print_stats seems hardcoded to stdout
1219
stats.print_stats(20)
1224
def lsprofile_main(argv):
1225
from bzrlib.lsprof import profile
1226
ret,stats = profile(main, argv)
1232
if __name__ == '__main__':
1234
if '--profile' in sys.argv:
1236
args.remove('--profile')
1237
sys.exit(profile_main(args))
1238
elif '--lsprof' in sys.argv:
1240
args.remove('--lsprof')
1241
sys.exit(lsprofile_main(args))
1243
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
