61
59
# where the basis and destination are unchanged.
63
61
# 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
62
# that includes some redundant parents (i.e. already a parent of
63
# something in the list.) We should eliminate them. This can
66
64
# be done fairly efficiently because the sequence numbers constrain
67
65
# the possible relationships.
71
from difflib import SequenceMatcher
67
# FIXME: the conflict markers should be *7* characters
70
from cStringIO import StringIO
75
from bzrlib.lazy_import import lazy_import
76
lazy_import(globals(), """
77
from bzrlib import tsort
84
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 *
73
95
from bzrlib.trace import mutter
74
from bzrlib.errors import WeaveError, WeaveFormatError, WeaveParentMismatch, \
75
WeaveRevisionNotPresent, WeaveRevisionAlreadyPresent
76
from bzrlib.tsort import topo_sort
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):
80
132
"""weave - versioned text file storage.
82
134
A Weave manages versions of line-based text files, keeping track
83
135
of the originating version for each line.
199
275
return self._parents == other._parents \
200
276
and self._weave == other._weave \
201
and self._sha1s == other._sha1s
277
and self._sha1s == other._sha1s
204
279
def __ne__(self, other):
205
280
return not self.__eq__(other)
207
def __contains__(self, name):
208
return self._name_map.has_key(name)
210
def maybe_lookup(self, name_or_index):
211
"""Convert possible symbolic name to index, or pass through indexes."""
212
if isinstance(name_or_index, (int, long)):
215
return self.lookup(name_or_index)
218
def lookup(self, name):
282
def _idx_to_name(self, version):
283
return self._names[version]
285
def _lookup(self, name):
219
286
"""Convert symbolic version name to index."""
287
if not self._allow_reserved:
288
self.check_not_reserved_id(name)
221
290
return self._name_map[name]
223
raise WeaveRevisionNotPresent(name, self)
292
raise RevisionNotPresent(name, self._weave_name)
295
"""See VersionedFile.versions."""
226
296
return self._names[:]
228
def iter_names(self):
229
"""Yield a list of all names in this weave."""
230
return iter(self._names)
232
def idx_to_name(self, version):
233
return self._names[version]
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:
235
385
def _check_repeated_add(self, name, parents, text, sha1):
236
386
"""Check that a duplicated add is OK.
238
388
If it is, return the (old) index; otherwise raise an exception.
240
idx = self.lookup(name)
390
idx = self._lookup(name)
241
391
if sorted(self._parents[idx]) != sorted(parents) \
242
392
or sha1 != self._sha1s[idx]:
243
raise WeaveRevisionAlreadyPresent(name, self)
393
raise RevisionAlreadyPresent(name, self._weave_name)
246
def add(self, name, parents, text, sha1=None):
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):
247
404
"""Add a single text on top of the weave.
249
406
Returns the index number of the newly added version.
252
409
Symbolic name for this version.
253
410
(Typically the revision-id of the revision that added it.)
411
If None, a name will be allocated based on the hash. (sha1:SHAHASH)
256
414
List or set of direct parent version numbers.
259
417
Sequence of lines to be added in the new version.
261
sha -- SHA-1 of the file, if known. This is trusted to be
419
:param nostore_sha: See VersionedFile.add_lines.
264
from bzrlib.osutils import sha_strings
266
assert isinstance(name, basestring)
268
sha1 = sha_strings(text)
269
if name in self._name_map:
270
return self._check_repeated_add(name, parents, text, sha1)
272
parents = map(self.maybe_lookup, parents)
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)
273
432
self._check_versions(parents)
274
## self._check_lines(text)
433
## self._check_lines(lines)
275
434
new_version = len(self._parents)
278
436
# if we abort after here the (in-memory) weave will be corrupt because only
279
437
# some fields are updated
438
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
439
# - Robert Collins 20060226
280
440
self._parents.append(parents[:])
281
441
self._sha1s.append(sha1)
282
self._names.append(name)
283
self._name_map[name] = new_version
442
self._names.append(version_id)
443
self._name_map[version_id] = new_version
287
447
# special case; adding with no parents revision; can do
288
448
# this more quickly by just appending unconditionally.
289
449
# even more specially, if we're adding an empty text we
290
450
# need do nothing at all.
292
452
self._weave.append(('{', new_version))
293
self._weave.extend(text)
453
self._weave.extend(lines)
294
454
self._weave.append(('}', None))
296
455
return new_version
298
457
if len(parents) == 1:
436
559
except IndexError:
437
560
raise IndexError("invalid version number %r" % i)
440
def annotate(self, name_or_index):
441
return list(self.annotate_iter(name_or_index))
444
def annotate_iter(self, name_or_index):
445
"""Yield list of (index-id, line) pairs for the specified version.
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.
447
573
The index indicates when the line originated in the weave."""
448
incls = [self.maybe_lookup(name_or_index)]
449
for origin, lineno, text in self._extract(incls):
456
(lineno, insert, deletes, text)
457
for each literal line.
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."""
463
597
lineno = 0 # line of weave, 0-based
465
599
for l in self._weave:
466
if isinstance(l, tuple):
600
if l.__class__ == tuple:
604
istack.append(self._names[v])
608
dset.add(self._names[v])
610
dset.remove(self._names[v])
479
612
raise WeaveFormatError('unexpected instruction %r' % v)
481
assert isinstance(l, basestring)
483
yield lineno, istack[-1], dset, l
614
yield lineno, istack[-1], frozenset(dset), l
618
raise WeaveFormatError("unclosed insertion blocks "
619
"at end of weave: %s" % istack)
621
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
624
def plan_merge(self, ver_a, ver_b):
625
"""Return pseudo-annotation indicating how the two versions merge.
627
This is computed between versions a and b and their common
630
Weave lines present in none of them are skipped entirely.
632
inc_a = set(self.get_ancestry([ver_a]))
633
inc_b = set(self.get_ancestry([ver_b]))
634
inc_c = inc_a & inc_b
636
for lineno, insert, deleteset, line in self._walk_internal([ver_a, ver_b]):
637
if deleteset & inc_c:
638
# killed in parent; can't be in either a or b
639
# not relevant to our work
640
yield 'killed-base', line
641
elif insert in inc_c:
642
# was inserted in base
643
killed_a = bool(deleteset & inc_a)
644
killed_b = bool(deleteset & inc_b)
645
if killed_a and killed_b:
646
yield 'killed-both', line
648
yield 'killed-a', line
650
yield 'killed-b', line
652
yield 'unchanged', line
653
elif insert in inc_a:
654
if deleteset & inc_a:
655
yield 'ghost-a', line
659
elif insert in inc_b:
660
if deleteset & inc_b:
661
yield 'ghost-b', line
665
# not in either revision
666
yield 'irrelevant', line
488
668
def _extract(self, versions):
489
669
"""Yield annotation of lines in included set.
548
def get_iter(self, name_or_index):
549
"""Yield lines for the specified version."""
550
incls = [self.maybe_lookup(name_or_index)]
551
for origin, lineno, line in self._extract(incls):
555
def get_text(self, name_or_index):
556
return ''.join(self.get_iter(name_or_index))
557
assert isinstance(version, int)
560
def get_lines(self, name_or_index):
561
return list(self.get_iter(name_or_index))
567
def mash_iter(self, included):
568
"""Return composed version of multiple included versions."""
569
included = map(self.maybe_lookup, included)
570
for origin, lineno, text in self._extract(included):
574
def dump(self, to_file):
575
from pprint import pprint
576
print >>to_file, "Weave._weave = ",
577
pprint(self._weave, to_file)
578
print >>to_file, "Weave._parents = ",
579
pprint(self._parents, to_file)
583
def numversions(self):
753
def _maybe_lookup(self, name_or_index):
754
"""Convert possible symbolic name to index, or pass through indexes.
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?"""
584
785
l = len(self._parents)
585
assert l == len(self._sha1s)
590
return self.numversions()
788
__len__ = num_versions
593
790
def check(self, progress_bar=None):
594
# check no circular inclusions
595
for version in range(self.numversions()):
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()):
596
795
inclusions = list(self._parents[version])
598
797
inclusions.sort()
600
799
raise WeaveFormatError("invalid included version %d for index %d"
601
800
% (inclusions[-1], version))
603
# try extracting all versions; this is a bit slow and parallel
604
# extraction could be used
605
nv = self.numversions()
606
for version in range(nv):
802
# try extracting all versions; parallel extraction is used
803
nv = self.num_versions()
808
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
809
# The problem is that set membership is much more expensive
810
name = self._idx_to_name(i)
813
new_inc = set([name])
814
for p in self._parents[i]:
815
new_inc.update(inclusions[self._idx_to_name(p)])
817
if set(new_inc) != set(self.get_ancestry(name)):
818
raise AssertionError(
820
% (set(new_inc), set(self.get_ancestry(name))))
821
inclusions[name] = new_inc
823
nlines = len(self._weave)
825
update_text = 'checking weave'
827
short_name = os.path.basename(self._weave_name)
828
update_text = 'checking %s' % (short_name,)
829
update_text = update_text[:25]
831
for lineno, insert, deleteset, line in self._walk_internal():
608
progress_bar.update('checking text', version, nv)
610
for l in self.get_iter(version):
613
expected = self._sha1s[version]
833
progress_bar.update(update_text, lineno, nlines)
835
for name, name_inclusions in inclusions.items():
836
# The active inclusion must be an ancestor,
837
# and no ancestors must have deleted this line,
838
# because we don't support resurrection.
839
if (insert in name_inclusions) and not (deleteset & name_inclusions):
840
sha1s[name].update(line)
843
version = self._idx_to_name(i)
844
hd = sha1s[version].hexdigest()
845
expected = self._sha1s[i]
614
846
if hd != expected:
615
raise WeaveError("mismatched sha1 for version %d; "
616
"got %s, expected %s"
617
% (version, hd, expected))
847
raise errors.WeaveInvalidChecksum(
848
"mismatched sha1 for version %s: "
849
"got %s, expected %s"
850
% (version, hd, expected))
619
852
# TODO: check insertions are properly nested, that there are
620
853
# no lines outside of insertion blocks, that deletions are
621
854
# properly paired, etc.
625
def merge(self, merge_versions):
626
"""Automerge and mark conflicts between versions.
628
This returns a sequence, each entry describing alternatives
629
for a chunk of the file. Each of the alternatives is given as
632
If there is a chunk of the file where there's no diagreement,
633
only one alternative is given.
635
# approach: find the included versions common to all the
637
raise NotImplementedError()
641
def _delta(self, included, lines):
642
"""Return changes from basis to new revision.
644
The old text for comparison is the union of included revisions.
646
This is used in inserting a new text.
648
Delta is returned as a sequence of
649
(weave1, weave2, newlines).
651
This indicates that weave1:weave2 of the old weave should be
652
replaced by the sequence of lines in newlines. Note that
653
these line numbers are positions in the total weave and don't
654
correspond to the lines in any extracted version, or even the
655
extracted union of included versions.
657
If line1=line2, this is a pure insert; if newlines=[] this is a
658
pure delete. (Similar to difflib.)
660
raise NotImplementedError()
663
def plan_merge(self, ver_a, ver_b):
664
"""Return pseudo-annotation indicating how the two versions merge.
666
This is computed between versions a and b and their common
669
Weave lines present in none of them are skipped entirely.
671
inc_a = self.inclusions([ver_a])
672
inc_b = self.inclusions([ver_b])
673
inc_c = inc_a & inc_b
675
for lineno, insert, deleteset, line in self._walk():
676
if deleteset & inc_c:
677
# killed in parent; can't be in either a or b
678
# not relevant to our work
679
yield 'killed-base', line
680
elif insert in inc_c:
681
# was inserted in base
682
killed_a = bool(deleteset & inc_a)
683
killed_b = bool(deleteset & inc_b)
684
if killed_a and killed_b:
685
yield 'killed-both', line
687
yield 'killed-a', line
689
yield 'killed-b', line
691
yield 'unchanged', line
692
elif insert in inc_a:
693
if deleteset & inc_a:
694
yield 'ghost-a', line
698
elif insert in inc_b:
699
if deleteset & inc_b:
700
yield 'ghost-b', line
704
# not in either revision
705
yield 'irrelevant', line
707
yield 'unchanged', '' # terminator
711
def weave_merge(self, plan):
716
for state, line in plan:
717
if state == 'unchanged' or state == 'killed-both':
718
# resync and flush queued conflicts changes if any
719
if not lines_a and not lines_b:
721
elif ch_a and not ch_b:
723
for l in lines_a: yield l
724
elif ch_b and not ch_a:
725
for l in lines_b: yield l
726
elif lines_a == lines_b:
727
for l in lines_a: yield l
730
for l in lines_a: yield l
732
for l in lines_b: yield l
739
if state == 'unchanged':
742
elif state == 'killed-a':
745
elif state == 'killed-b':
748
elif state == 'new-a':
751
elif state == 'new-b':
755
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
760
def join(self, other):
761
"""Integrate versions from other into this weave.
763
The resulting weave contains all the history of both weaves;
764
any version you could retrieve from either self or other can be
765
retrieved from self after this call.
767
It is illegal for the two weaves to contain different values
768
or different parents for any version. See also reweave().
770
if other.numversions() == 0:
771
return # nothing to update, easy
772
# two loops so that we do not change ourselves before verifying it
774
# work through in index order to make sure we get all dependencies
775
for other_idx, name in enumerate(other._names):
776
if self._check_version_consistent(other, other_idx, name):
778
for other_idx, name in enumerate(other._names):
779
# TODO: If all the parents of the other version are already
780
# present then we can avoid some work by just taking the delta
781
# and adjusting the offsets.
782
new_parents = self._imported_parents(other, other_idx)
783
lines = other.get_lines(other_idx)
784
sha1 = other._sha1s[other_idx]
785
self.add(name, new_parents, lines, sha1)
788
856
def _imported_parents(self, other, other_idx):
789
857
"""Return list of parents in self corresponding to indexes in other."""
791
859
for parent_idx in other._parents[other_idx]:
792
860
parent_name = other._names[parent_idx]
793
if parent_name not in self._names:
861
if parent_name not in self._name_map:
794
862
# should not be possible
795
raise WeaveError("missing parent {%s} of {%s} in %r"
863
raise WeaveError("missing parent {%s} of {%s} in %r"
796
864
% (parent_name, other._name_map[other_idx], self))
797
865
new_parents.append(self._name_map[parent_name])
798
866
return new_parents
832
def reweave(self, other):
833
"""Reweave self with other."""
834
new_weave = reweave(self, other)
896
def _reweave(self, other, pb, msg):
897
"""Reweave self with other - internal helper for join().
899
:param other: The other weave to merge
900
:param pb: An optional progress bar, indicating how far done we are
901
: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."""
835
908
for attr in self.__slots__:
836
setattr(self, attr, getattr(new_weave, attr))
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):
840
978
"""Combine two weaves and return the result.
842
This works even if a revision R has different parents in
980
This works even if a revision R has different parents in
843
981
wa and wb. In the resulting weave all the parents are given.
845
This is done by just building up a new weave, maintaining ordering
983
This is done by just building up a new weave, maintaining ordering
846
984
of the versions in the two inputs. More efficient approaches
847
might be possible but it should only be necessary to do
848
this operation rarely, when a new previously ghost version is
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
853
queue_a = range(wa.numversions())
854
queue_b = range(wb.numversions())
994
queue_a = range(wa.num_versions())
995
queue_b = range(wb.num_versions())
855
996
# first determine combined parents of all versions
856
997
# map from version name -> all parent names
857
998
combined_parents = _reweave_parent_graphs(wa, wb)
858
999
mutter("combined parents: %r", combined_parents)
859
order = topo_sort(combined_parents.iteritems())
1000
order = tsort.topo_sort(combined_parents.iteritems())
860
1001
mutter("order to reweave: %r", order)
1006
for idx, name in enumerate(order):
1008
pb.update(msg, idx, len(order))
862
1009
if name in wa._name_map:
863
1010
lines = wa.get_lines(name)
864
1011
if name in wb._name_map:
865
assert lines == wb.get_lines(name)
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)
867
1022
lines = wb.get_lines(name)
868
wr.add(name, combined_parents[name], lines)
1023
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
872
1027
def _reweave_parent_graphs(wa, wb):
873
1028
"""Return combined parent ancestry for two weaves.
875
1030
Returned as a list of (version_name, set(parent_names))"""
877
1032
for weave in [wa, wb]:
878
1033
for idx, name in enumerate(weave._names):
879
1034
p = combined.setdefault(name, set())
880
p.update(map(weave.idx_to_name, weave._parents[idx]))
1035
p.update(map(weave._idx_to_name, weave._parents[idx]))
885
"""Show the weave's table-of-contents"""
886
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
887
for i in (6, 50, 10, 10):
890
for i in range(w.numversions()):
893
parent_str = ' '.join(map(str, w._parents[i]))
894
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
898
def weave_stats(weave_file, pb):
899
from bzrlib.weavefile import read_weave
901
wf = file(weave_file, 'rb')
903
# FIXME: doesn't work on pipes
904
weave_size = wf.tell()
908
for i in range(vers):
909
pb.update('checking sizes', i, vers)
910
for origin, lineno, line in w._extract([i]):
915
print 'versions %9d' % vers
916
print 'weave file %9d bytes' % weave_size
917
print 'total contents %9d bytes' % total
918
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
921
print 'average size %9d bytes' % avg
922
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
926
print """bzr weave tool
928
Experimental tool for weave algorithm.
932
Create an empty weave file
933
weave get WEAVEFILE VERSION
934
Write out specified version.
935
weave check WEAVEFILE
936
Check consistency of all versions.
938
Display table of contents.
939
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
940
Add NEWTEXT, with specified parent versions.
941
weave annotate WEAVEFILE VERSION
942
Display origin of each line.
943
weave mash WEAVEFILE VERSION...
944
Display composite of all selected versions.
945
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
946
Auto-merge two versions and display conflicts.
947
weave diff WEAVEFILE VERSION1 VERSION2
948
Show differences between two versions.
952
% weave init foo.weave
954
% weave add foo.weave ver0 < foo.txt
957
(create updated version)
959
% weave get foo.weave 0 | diff -u - foo.txt
960
% weave add foo.weave ver1 0 < foo.txt
963
% weave get foo.weave 0 > foo.txt (create forked version)
965
% weave add foo.weave ver2 0 < foo.txt
968
% weave merge foo.weave 1 2 > foo.txt (merge them)
969
% vi foo.txt (resolve conflicts)
970
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
982
# in case we're run directly from the subdirectory
983
sys.path.append('..')
985
from bzrlib.weavefile import write_weave, read_weave
986
from bzrlib.progress import ProgressBar
1001
return read_weave(file(argv[2], 'rb'))
1007
# at the moment, based on everything in the file
1009
parents = map(int, argv[4:])
1010
lines = sys.stdin.readlines()
1011
ver = w.add(name, parents, lines)
1012
write_weave(w, file(argv[2], 'wb'))
1013
print 'added version %r %d' % (name, ver)
1016
if os.path.exists(fn):
1017
raise IOError("file exists")
1019
write_weave(w, file(fn, 'wb'))
1020
elif cmd == 'get': # get one version
1022
sys.stdout.writelines(w.get_iter(int(argv[3])))
1024
elif cmd == 'mash': # get composite
1026
sys.stdout.writelines(w.mash_iter(map(int, argv[3:])))
1029
from difflib import unified_diff
1032
v1, v2 = map(int, argv[3:5])
1035
diff_gen = unified_diff(lines1, lines2,
1036
'%s version %d' % (fn, v1),
1037
'%s version %d' % (fn, v2))
1038
sys.stdout.writelines(diff_gen)
1040
elif cmd == 'annotate':
1042
# newline is added to all lines regardless; too hard to get
1043
# reasonable formatting otherwise
1045
for origin, text in w.annotate(int(argv[3])):
1046
text = text.rstrip('\r\n')
1048
print ' | %s' % (text)
1050
print '%5d | %s' % (origin, text)
1056
elif cmd == 'stats':
1057
weave_stats(argv[2], ProgressBar())
1059
elif cmd == 'check':
1064
print '%d versions ok' % w.numversions()
1066
elif cmd == 'inclusions':
1068
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1070
elif cmd == 'parents':
1072
print ' '.join(map(str, w._parents[int(argv[3])]))
1074
elif cmd == 'plan-merge':
1076
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1078
print '%14s | %s' % (state, line),
1080
elif cmd == 'merge':
1082
p = w.plan_merge(int(argv[3]), int(argv[4]))
1083
sys.stdout.writelines(w.weave_merge(p))
1085
elif cmd == 'mash-merge':
1091
v1, v2 = map(int, argv[3:5])
1093
basis = w.inclusions([v1]).intersection(w.inclusions([v2]))
1095
base_lines = list(w.mash_iter(basis))
1096
a_lines = list(w.get(v1))
1097
b_lines = list(w.get(v2))
1099
from bzrlib.merge3 import Merge3
1100
m3 = Merge3(base_lines, a_lines, b_lines)
1102
name_a = 'version %d' % v1
1103
name_b = 'version %d' % v2
1104
sys.stdout.writelines(m3.merge_lines(name_a=name_a, name_b=name_b))
1106
raise ValueError('unknown command %r' % cmd)
1110
def profile_main(argv):
1111
import tempfile, hotshot, hotshot.stats
1113
prof_f = tempfile.NamedTemporaryFile()
1115
prof = hotshot.Profile(prof_f.name)
1117
ret = prof.runcall(main, argv)
1120
stats = hotshot.stats.load(prof_f.name)
1122
stats.sort_stats('cumulative')
1123
## XXX: Might like to write to stderr or the trace file instead but
1124
## print_stats seems hardcoded to stdout
1125
stats.print_stats(20)
1130
if __name__ == '__main__':
1132
if '--profile' in sys.argv:
1134
args.remove('--profile')
1135
sys.exit(profile_main(args))
1137
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
