48
50
# have slight specializations for different ways its used: annotate,
49
51
# basis for add, get, etc.
51
# TODO: Probably the API should work only in names to hide the integer
52
# indexes from the user.
53
# TODO: Perhaps the API should work only in names to hide the integer
54
# indexes from the user?
54
56
# TODO: Is there any potential performance win by having an add()
55
57
# variant that is passed a pre-cooked version of the single basis
58
# TODO: Reweave can possibly be made faster by remembering diffs
59
# where the basis and destination are unchanged.
61
# 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
# be done fairly efficiently because the sequence numbers constrain
65
# the possible relationships.
67
# FIXME: the conflict markers should be *7* characters
70
from cStringIO import StringIO
73
from bzrlib.lazy_import import lazy_import
74
lazy_import(globals(), """
75
from bzrlib import tsort
81
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
82
RevisionAlreadyPresent,
84
UnavailableRepresentation,
86
from bzrlib.osutils import dirname, sha, sha_strings, split_lines
87
import bzrlib.patiencediff
88
from bzrlib.revision import NULL_REVISION
89
from bzrlib.symbol_versioning import *
90
from bzrlib.trace import mutter
91
from bzrlib.versionedfile import (
98
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
101
class WeaveContentFactory(ContentFactory):
102
"""Content factory for streaming from weaves.
104
:seealso ContentFactory:
107
def __init__(self, version, weave):
108
"""Create a WeaveContentFactory for version from weave."""
109
ContentFactory.__init__(self)
110
self.sha1 = weave.get_sha1s([version])[version]
111
self.key = (version,)
112
parents = weave.get_parent_map([version])[version]
113
self.parents = tuple((parent,) for parent in parents)
114
self.storage_kind = 'fulltext'
117
def get_bytes_as(self, storage_kind):
118
if storage_kind == 'fulltext':
119
return self._weave.get_text(self.key[-1])
120
elif storage_kind == 'chunked':
121
return self._weave.get_lines(self.key[-1])
123
raise UnavailableRepresentation(self.key, storage_kind, 'fulltext')
126
class Weave(VersionedFile):
60
# TODO: Have a way to go back and insert a revision V1 that is a parent
61
# of an already-stored revision V2. This means some lines previously
62
# counted as new in V2 will be discovered to have actually come from V1.
63
# It is probably necessary to insert V1, then compute a whole new diff
64
# from the mashed ancestors to V2. This must be repeated for every
65
# direct child of V1. The deltas from V2 to its descendents won't change,
66
# although their location within the weave may change. It may be possible
67
# to just adjust the location of those instructions rather than
68
# re-weaving the whole thing. This is expected to be a fairly rare
69
# operation, only used when inserting data that was previously a ghost.
75
from difflib import SequenceMatcher
80
class WeaveError(Exception):
81
"""Exception in processing weave"""
84
class WeaveFormatError(WeaveError):
85
"""Weave invariant violated"""
127
89
"""weave - versioned text file storage.
129
91
A Weave manages versions of line-based text files, keeping track
130
92
of the originating version for each line.
270
205
return self._parents == other._parents \
271
206
and self._weave == other._weave \
272
and self._sha1s == other._sha1s
207
and self._sha1s == other._sha1s
274
210
def __ne__(self, other):
275
211
return not self.__eq__(other)
277
def _idx_to_name(self, version):
278
return self._names[version]
280
def _lookup(self, name):
214
def maybe_lookup(self, name_or_index):
215
"""Convert possible symbolic name to index, or pass through indexes."""
216
if isinstance(name_or_index, (int, long)):
219
return self.lookup(name_or_index)
222
def lookup(self, name):
281
223
"""Convert symbolic version name to index."""
282
if not self._allow_reserved:
283
self.check_not_reserved_id(name)
285
225
return self._name_map[name]
287
raise RevisionNotPresent(name, self._weave_name)
290
"""See VersionedFile.versions."""
291
return self._names[:]
293
def has_version(self, version_id):
294
"""See VersionedFile.has_version."""
295
return (version_id in self._name_map)
297
__contains__ = has_version
299
def get_record_stream(self, versions, ordering, include_delta_closure):
300
"""Get a stream of records for versions.
302
:param versions: The versions to include. Each version is a tuple
304
:param ordering: Either 'unordered' or 'topological'. A topologically
305
sorted stream has compression parents strictly before their
307
:param include_delta_closure: If True then the closure across any
308
compression parents will be included (in the opaque data).
309
:return: An iterator of ContentFactory objects, each of which is only
310
valid until the iterator is advanced.
312
versions = [version[-1] for version in versions]
313
if ordering == 'topological':
314
parents = self.get_parent_map(versions)
315
new_versions = tsort.topo_sort(parents)
316
new_versions.extend(set(versions).difference(set(parents)))
317
versions = new_versions
318
elif ordering == 'groupcompress':
319
parents = self.get_parent_map(versions)
320
new_versions = sort_groupcompress(parents)
321
new_versions.extend(set(versions).difference(set(parents)))
322
versions = new_versions
323
for version in versions:
325
yield WeaveContentFactory(version, self)
327
yield AbsentContentFactory((version,))
329
def get_parent_map(self, version_ids):
330
"""See VersionedFile.get_parent_map."""
332
for version_id in version_ids:
333
if version_id == NULL_REVISION:
338
map(self._idx_to_name,
339
self._parents[self._lookup(version_id)]))
340
except RevisionNotPresent:
342
result[version_id] = parents
345
def get_parents_with_ghosts(self, version_id):
346
raise NotImplementedError(self.get_parents_with_ghosts)
348
def insert_record_stream(self, stream):
349
"""Insert a record stream into this versioned file.
351
:param stream: A stream of records to insert.
353
:seealso VersionedFile.get_record_stream:
356
for record in stream:
357
# Raise an error when a record is missing.
358
if record.storage_kind == 'absent':
359
raise RevisionNotPresent([record.key[0]], self)
360
# adapt to non-tuple interface
361
parents = [parent[0] for parent in record.parents]
362
if (record.storage_kind == 'fulltext'
363
or record.storage_kind == 'chunked'):
364
self.add_lines(record.key[0], parents,
365
osutils.chunks_to_lines(record.get_bytes_as('chunked')))
367
adapter_key = record.storage_kind, 'fulltext'
369
adapter = adapters[adapter_key]
371
adapter_factory = adapter_registry.get(adapter_key)
372
adapter = adapter_factory(self)
373
adapters[adapter_key] = adapter
374
lines = split_lines(adapter.get_bytes(record))
376
self.add_lines(record.key[0], parents, lines)
377
except RevisionAlreadyPresent:
227
raise WeaveError("name %r not present in weave %r" %
228
(name, self._weave_name))
231
def iter_names(self):
232
"""Yield a list of all names in this weave."""
233
return iter(self._names)
235
def idx_to_name(self, version):
236
return self._names[version]
380
239
def _check_repeated_add(self, name, parents, text, sha1):
381
240
"""Check that a duplicated add is OK.
383
242
If it is, return the (old) index; otherwise raise an exception.
385
idx = self._lookup(name)
386
if sorted(self._parents[idx]) != sorted(parents) \
387
or sha1 != self._sha1s[idx]:
388
raise RevisionAlreadyPresent(name, self._weave_name)
244
idx = self.lookup(name)
245
if sorted(self._parents[idx]) != sorted(parents):
246
raise WeaveError("name \"%s\" already present in weave "
247
"with different parents" % name)
248
if sha1 != self._sha1s[idx]:
249
raise WeaveError("name \"%s\" already present in weave "
250
"with different text" % name)
391
def _add_lines(self, version_id, parents, lines, parent_texts,
392
left_matching_blocks, nostore_sha, random_id, check_content):
393
"""See VersionedFile.add_lines."""
394
idx = self._add(version_id, lines, map(self._lookup, parents),
395
nostore_sha=nostore_sha)
396
return sha_strings(lines), sum(map(len, lines)), idx
398
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
255
def add(self, name, parents, text, sha1=None):
399
256
"""Add a single text on top of the weave.
401
258
Returns the index number of the newly added version.
404
261
Symbolic name for this version.
405
262
(Typically the revision-id of the revision that added it.)
406
If None, a name will be allocated based on the hash. (sha1:SHAHASH)
409
265
List or set of direct parent version numbers.
412
268
Sequence of lines to be added in the new version.
414
:param nostore_sha: See VersionedFile.add_lines.
270
sha -- SHA-1 of the file, if known. This is trusted to be
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 is None:
423
version_id = "sha1:" + sha1
424
if version_id in self._name_map:
425
return self._check_repeated_add(version_id, parents, lines, sha1)
273
from bzrlib.osutils import sha_strings
275
assert isinstance(name, basestring)
277
sha1 = sha_strings(text)
278
if name in self._name_map:
279
return self._check_repeated_add(name, parents, text, sha1)
281
parents = map(self.maybe_lookup, parents)
427
282
self._check_versions(parents)
428
## self._check_lines(lines)
283
## self._check_lines(text)
429
284
new_version = len(self._parents)
431
287
# if we abort after here the (in-memory) weave will be corrupt because only
432
288
# some fields are updated
433
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
434
# - Robert Collins 20060226
435
289
self._parents.append(parents[:])
436
290
self._sha1s.append(sha1)
437
self._names.append(version_id)
438
self._name_map[version_id] = new_version
291
self._names.append(name)
292
self._name_map[name] = new_version
442
296
# special case; adding with no parents revision; can do
443
297
# this more quickly by just appending unconditionally.
444
298
# even more specially, if we're adding an empty text we
445
299
# need do nothing at all.
447
301
self._weave.append(('{', new_version))
448
self._weave.extend(lines)
302
self._weave.extend(text)
449
303
self._weave.append(('}', None))
450
305
return new_version
452
307
if len(parents) == 1:
554
440
except IndexError:
555
441
raise IndexError("invalid version number %r" % i)
557
def _compatible_parents(self, my_parents, other_parents):
558
"""During join check that other_parents are joinable with my_parents.
560
Joinable is defined as 'is a subset of' - supersets may require
561
regeneration of diffs, but subsets do not.
563
return len(other_parents.difference(my_parents)) == 0
565
def annotate(self, version_id):
566
"""Return a list of (version-id, line) tuples for version_id.
444
def annotate(self, name_or_index):
445
return list(self.annotate_iter(name_or_index))
448
def annotate_iter(self, name_or_index):
449
"""Yield list of (index-id, line) pairs for the specified version.
568
451
The index indicates when the line originated in the weave."""
569
incls = [self._lookup(version_id)]
570
return [(self._idx_to_name(origin), text) for origin, lineno, text in
571
self._extract(incls)]
573
def iter_lines_added_or_present_in_versions(self, version_ids=None,
575
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
576
if version_ids is None:
577
version_ids = self.versions()
578
version_ids = set(version_ids)
579
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
580
if inserted not in version_ids: continue
582
yield line + '\n', inserted
586
def _walk_internal(self, version_ids=None):
587
"""Helper method for weave actions."""
452
incls = [self.maybe_lookup(name_or_index)]
453
for origin, lineno, text in self._extract(incls):
461
(lineno, insert, deletes, text)
462
for each literal line.
592
468
lineno = 0 # line of weave, 0-based
594
470
for l in self._weave:
595
if l.__class__ == tuple:
471
if isinstance(l, tuple):
599
istack.append(self._names[v])
603
dset.add(self._names[v])
605
dset.remove(self._names[v])
607
raise WeaveFormatError('unexpected instruction %r' % v)
609
yield lineno, istack[-1], frozenset(dset), l
484
raise WeaveFormatError('unexpected instruction %r'
487
assert isinstance(l, basestring)
489
yield lineno, istack[-1], dset, l
494
def _extract(self, versions):
495
"""Yield annotation of lines in included set.
497
Yields a sequence of tuples (origin, lineno, text), where
498
origin is the origin version, lineno the index in the weave,
499
and text the text of the line.
501
The set typically but not necessarily corresponds to a version.
504
if not isinstance(i, int):
507
included = self.inclusions(versions)
512
lineno = 0 # line of weave, 0-based
518
WFE = WeaveFormatError
520
for l in self._weave:
521
if isinstance(l, tuple):
525
assert v not in istack
539
assert isinstance(l, basestring)
541
isactive = (not dset) and istack and (istack[-1] in included)
543
result.append((istack[-1], lineno, l))
613
raise WeaveFormatError("unclosed insertion blocks "
614
"at end of weave: %s" % istack)
547
raise WFE("unclosed insertion blocks at end of weave",
616
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
550
raise WFE("unclosed deletion blocks at end of weave",
557
def get_iter(self, name_or_index):
558
"""Yield lines for the specified version."""
559
incls = [self.maybe_lookup(name_or_index)]
560
for origin, lineno, line in self._extract(incls):
564
def get_text(self, name_or_index):
565
return ''.join(self.get_iter(name_or_index))
566
assert isinstance(version, int)
569
def get_lines(self, name_or_index):
570
return list(self.get_iter(name_or_index))
576
def mash_iter(self, included):
577
"""Return composed version of multiple included versions."""
578
included = map(self.maybe_lookup, included)
579
for origin, lineno, text in self._extract(included):
583
def dump(self, to_file):
584
from pprint import pprint
585
print >>to_file, "Weave._weave = ",
586
pprint(self._weave, to_file)
587
print >>to_file, "Weave._parents = ",
588
pprint(self._parents, to_file)
592
def numversions(self):
593
l = len(self._parents)
594
assert l == len(self._sha1s)
599
return self.numversions()
602
def check(self, progress_bar=None):
603
# check no circular inclusions
604
for version in range(self.numversions()):
605
inclusions = list(self._parents[version])
608
if inclusions[-1] >= version:
609
raise WeaveFormatError("invalid included version %d for index %d"
610
% (inclusions[-1], version))
612
# try extracting all versions; this is a bit slow and parallel
613
# extraction could be used
614
nv = self.numversions()
615
for version in range(nv):
617
progress_bar.update('checking text', version, nv)
619
for l in self.get_iter(version):
622
expected = self._sha1s[version]
624
raise WeaveError("mismatched sha1 for version %d; "
625
"got %s, expected %s"
626
% (version, hd, expected))
628
# TODO: check insertions are properly nested, that there are
629
# no lines outside of insertion blocks, that deletions are
630
# properly paired, etc.
634
def merge(self, merge_versions):
635
"""Automerge and mark conflicts between versions.
637
This returns a sequence, each entry describing alternatives
638
for a chunk of the file. Each of the alternatives is given as
641
If there is a chunk of the file where there's no diagreement,
642
only one alternative is given.
644
# approach: find the included versions common to all the
646
raise NotImplementedError()
650
def _delta(self, included, lines):
651
"""Return changes from basis to new revision.
653
The old text for comparison is the union of included revisions.
655
This is used in inserting a new text.
657
Delta is returned as a sequence of
658
(weave1, weave2, newlines).
660
This indicates that weave1:weave2 of the old weave should be
661
replaced by the sequence of lines in newlines. Note that
662
these line numbers are positions in the total weave and don't
663
correspond to the lines in any extracted version, or even the
664
extracted union of included versions.
666
If line1=line2, this is a pure insert; if newlines=[] this is a
667
pure delete. (Similar to difflib.)
669
raise NotImplementedError()
619
672
def plan_merge(self, ver_a, ver_b):
620
673
"""Return pseudo-annotation indicating how the two versions merge.
660
713
# not in either revision
661
714
yield 'irrelevant', line
663
def _extract(self, versions):
664
"""Yield annotation of lines in included set.
666
Yields a sequence of tuples (origin, lineno, text), where
667
origin is the origin version, lineno the index in the weave,
668
and text the text of the line.
670
The set typically but not necessarily corresponds to a version.
673
if not isinstance(i, int):
676
included = self._inclusions(versions)
682
lineno = 0 # line of weave, 0-based
688
WFE = WeaveFormatError
691
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
692
# +285282 0 1676.8040 1676.8040 +<isinstance>
693
# 1.6 seconds in 'isinstance'.
694
# changing the first isinstance:
695
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
696
# +140414 0 762.8050 762.8050 +<isinstance>
697
# note that the inline time actually dropped (less function calls)
698
# and total processing time was halved.
699
# we're still spending ~1/4 of the method in isinstance though.
700
# so lets hard code the acceptable string classes we expect:
701
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
702
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
704
# yay, down to ~1/4 the initial extract time, and our inline time
705
# has shrunk again, with isinstance no longer dominating.
706
# tweaking the stack inclusion test to use a set gives:
707
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
708
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
710
# - a 5% win, or possibly just noise. However with large istacks that
711
# 'in' test could dominate, so I'm leaving this change in place -
712
# when its fast enough to consider profiling big datasets we can review.
717
for l in self._weave:
718
if l.__class__ == tuple:
725
iset.remove(istack.pop())
716
yield 'unchanged', '' # terminator
720
def weave_merge(self, plan):
725
for state, line in plan:
726
if state == 'unchanged' or state == 'killed-both':
727
# resync and flush queued conflicts changes if any
728
if not lines_a and not lines_b:
730
elif ch_a and not ch_b:
732
for l in lines_a: yield l
733
elif ch_b and not ch_a:
734
for l in lines_b: yield l
735
elif lines_a == lines_b:
736
for l in lines_a: yield l
733
raise AssertionError()
739
for l in lines_a: yield l
741
for l in lines_b: yield l
748
if state == 'unchanged':
751
elif state == 'killed-a':
754
elif state == 'killed-b':
757
elif state == 'new-a':
760
elif state == 'new-b':
736
isactive = (not dset) and istack and (istack[-1] in included)
738
result.append((istack[-1], lineno, l))
741
raise WeaveFormatError("unclosed insertion blocks "
742
"at end of weave: %s" % istack)
744
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
748
def _maybe_lookup(self, name_or_index):
749
"""Convert possible symbolic name to index, or pass through indexes.
764
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
769
def join(self, other):
770
"""Integrate versions from other into this weave.
772
The resulting weave contains all the history of both weaves;
773
any version you could retrieve from either self or other can be
774
retrieved from self after this call.
776
It is illegal for the two weaves to contain different values
753
if isinstance(name_or_index, (int, long)):
756
return self._lookup(name_or_index)
758
def get_lines(self, version_id):
759
"""See VersionedFile.get_lines()."""
760
int_index = self._maybe_lookup(version_id)
761
result = [line for (origin, lineno, line) in self._extract([int_index])]
762
expected_sha1 = self._sha1s[int_index]
763
measured_sha1 = sha_strings(result)
764
if measured_sha1 != expected_sha1:
765
raise errors.WeaveInvalidChecksum(
766
'file %s, revision %s, expected: %s, measured %s'
767
% (self._weave_name, version_id,
768
expected_sha1, measured_sha1))
771
def get_sha1s(self, version_ids):
772
"""See VersionedFile.get_sha1s()."""
774
for v in version_ids:
775
result[v] = self._sha1s[self._lookup(v)]
778
def num_versions(self):
779
"""How many versions are in this weave?"""
780
l = len(self._parents)
783
__len__ = num_versions
785
def check(self, progress_bar=None):
786
# TODO evaluate performance hit of using string sets in this routine.
787
# TODO: check no circular inclusions
788
# TODO: create a nested progress bar
789
for version in range(self.num_versions()):
790
inclusions = list(self._parents[version])
793
if inclusions[-1] >= version:
794
raise WeaveFormatError("invalid included version %d for index %d"
795
% (inclusions[-1], version))
797
# try extracting all versions; parallel extraction is used
798
nv = self.num_versions()
803
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
804
# The problem is that set membership is much more expensive
805
name = self._idx_to_name(i)
808
new_inc = set([name])
809
for p in self._parents[i]:
810
new_inc.update(inclusions[self._idx_to_name(p)])
812
if set(new_inc) != set(self.get_ancestry(name)):
813
raise AssertionError(
815
% (set(new_inc), set(self.get_ancestry(name))))
816
inclusions[name] = new_inc
818
nlines = len(self._weave)
820
update_text = 'checking weave'
822
short_name = os.path.basename(self._weave_name)
823
update_text = 'checking %s' % (short_name,)
824
update_text = update_text[:25]
826
for lineno, insert, deleteset, line in self._walk_internal():
828
progress_bar.update(update_text, lineno, nlines)
830
for name, name_inclusions in inclusions.items():
831
# The active inclusion must be an ancestor,
832
# and no ancestors must have deleted this line,
833
# because we don't support resurrection.
834
if (insert in name_inclusions) and not (deleteset & name_inclusions):
835
sha1s[name].update(line)
838
version = self._idx_to_name(i)
839
hd = sha1s[version].hexdigest()
840
expected = self._sha1s[i]
842
raise errors.WeaveInvalidChecksum(
843
"mismatched sha1 for version %s: "
844
"got %s, expected %s"
845
% (version, hd, expected))
847
# TODO: check insertions are properly nested, that there are
848
# no lines outside of insertion blocks, that deletions are
849
# properly paired, etc.
779
if other.numversions() == 0:
780
return # nothing to update, easy
781
# work through in index order to make sure we get all dependencies
782
for other_idx, name in enumerate(other._names):
783
# TODO: If all the parents of the other version are already
784
# present then we can avoid some work by just taking the delta
785
# and adjusting the offsets.
786
if self._check_version_consistent(other, other_idx, name):
788
new_parents = self._imported_parents(other, other_idx)
789
lines = other.get_lines(other_idx)
790
sha1 = other._sha1s[other_idx]
791
self.add(name, new_parents, lines, sha1)
851
794
def _imported_parents(self, other, other_idx):
852
795
"""Return list of parents in self corresponding to indexes in other."""
854
797
for parent_idx in other._parents[other_idx]:
855
798
parent_name = other._names[parent_idx]
856
if parent_name not in self._name_map:
799
if parent_name not in self._names:
857
800
# should not be possible
858
raise WeaveError("missing parent {%s} of {%s} in %r"
801
raise WeaveError("missing parent {%s} of {%s} in %r"
859
802
% (parent_name, other._name_map[other_idx], self))
860
803
new_parents.append(self._name_map[parent_name])
861
804
return new_parents
869
812
* the same direct parents (by name, not index, and disregarding
872
815
If present & correct return True;
873
if not present in self return False;
816
if not present in self return False;
874
817
if inconsistent raise error."""
875
818
this_idx = self._name_map.get(name, -1)
876
819
if this_idx != -1:
877
820
if self._sha1s[this_idx] != other._sha1s[other_idx]:
878
raise errors.WeaveTextDiffers(name, self, other)
821
raise WeaveError("inconsistent texts for version {%s} "
822
"when joining weaves"
879
824
self_parents = self._parents[this_idx]
880
825
other_parents = other._parents[other_idx]
881
n1 = set([self._names[i] for i in self_parents])
882
n2 = set([other._names[i] for i in other_parents])
883
if not self._compatible_parents(n1, n2):
884
raise WeaveParentMismatch("inconsistent parents "
885
"for version {%s}: %s vs %s" % (name, n1, n2))
826
n1 = [self._names[i] for i in self_parents]
827
n2 = [other._names[i] for i in other_parents]
831
raise WeaveError("inconsistent parents for version {%s}: "
887
835
return True # ok!
891
def _reweave(self, other, pb, msg):
892
"""Reweave self with other - internal helper for join().
894
:param other: The other weave to merge
895
:param pb: An optional progress bar, indicating how far done we are
896
:param msg: An optional message for the progress
898
new_weave = _reweave(self, other, pb=pb, msg=msg)
899
self._copy_weave_content(new_weave)
901
def _copy_weave_content(self, otherweave):
902
"""adsorb the content from otherweave."""
903
for attr in self.__slots__:
904
if attr != '_weave_name':
905
setattr(self, attr, copy(getattr(otherweave, attr)))
908
class WeaveFile(Weave):
909
"""A WeaveFile represents a Weave on disk and writes on change."""
911
WEAVE_SUFFIX = '.weave'
913
def __init__(self, name, transport, filemode=None, create=False, access_mode='w', get_scope=None):
914
"""Create a WeaveFile.
916
:param create: If not True, only open an existing knit.
918
super(WeaveFile, self).__init__(name, access_mode, get_scope=get_scope,
919
allow_reserved=False)
920
self._transport = transport
921
self._filemode = filemode
923
f = self._transport.get(name + WeaveFile.WEAVE_SUFFIX)
924
_read_weave_v5(StringIO(f.read()), 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)
973
def _reweave(wa, wb, pb=None, msg=None):
974
"""Combine two weaves and return the result.
976
This works even if a revision R has different parents in
977
wa and wb. In the resulting weave all the parents are given.
979
This is done by just building up a new weave, maintaining ordering
980
of the versions in the two inputs. More efficient approaches
981
might be possible but it should only be necessary to do
982
this operation rarely, when a new previously ghost version is
985
:param pb: An optional progress bar, indicating how far done we are
986
:param msg: An optional message for the progress
990
queue_a = range(wa.num_versions())
991
queue_b = range(wb.num_versions())
992
# first determine combined parents of all versions
993
# map from version name -> all parent names
994
combined_parents = _reweave_parent_graphs(wa, wb)
995
mutter("combined parents: %r", combined_parents)
996
order = tsort.topo_sort(combined_parents.iteritems())
997
mutter("order to reweave: %r", order)
1002
for idx, name in enumerate(order):
1004
pb.update(msg, idx, len(order))
1005
if name in wa._name_map:
1006
lines = wa.get_lines(name)
1007
if name in wb._name_map:
1008
lines_b = wb.get_lines(name)
1009
if lines != lines_b:
1010
mutter('Weaves differ on content. rev_id {%s}', name)
1011
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1013
lines = list(difflib.unified_diff(lines, lines_b,
1014
wa._weave_name, wb._weave_name))
1015
mutter('lines:\n%s', ''.join(lines))
1016
raise errors.WeaveTextDiffers(name, wa, wb)
1018
lines = wb.get_lines(name)
1019
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1023
def _reweave_parent_graphs(wa, wb):
1024
"""Return combined parent ancestry for two weaves.
1026
Returned as a list of (version_name, set(parent_names))"""
1028
for weave in [wa, wb]:
1029
for idx, name in enumerate(weave._names):
1030
p = combined.setdefault(name, set())
1031
p.update(map(weave._idx_to_name, weave._parents[idx]))
841
"""Show the weave's table-of-contents"""
842
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
843
for i in (6, 50, 10, 10):
846
for i in range(w.numversions()):
849
parent_str = ' '.join(map(str, w._parents[i]))
850
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
854
def weave_stats(weave_file, pb):
855
from bzrlib.weavefile import read_weave
857
wf = file(weave_file, 'rb')
859
# FIXME: doesn't work on pipes
860
weave_size = wf.tell()
864
for i in range(vers):
865
pb.update('checking sizes', i, vers)
866
for origin, lineno, line in w._extract([i]):
871
print 'versions %9d' % vers
872
print 'weave file %9d bytes' % weave_size
873
print 'total contents %9d bytes' % total
874
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
877
print 'average size %9d bytes' % avg
878
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
882
print """bzr weave tool
884
Experimental tool for weave algorithm.
888
Create an empty weave file
889
weave get WEAVEFILE VERSION
890
Write out specified version.
891
weave check WEAVEFILE
892
Check consistency of all versions.
894
Display table of contents.
895
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
896
Add NEWTEXT, with specified parent versions.
897
weave annotate WEAVEFILE VERSION
898
Display origin of each line.
899
weave mash WEAVEFILE VERSION...
900
Display composite of all selected versions.
901
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
902
Auto-merge two versions and display conflicts.
903
weave diff WEAVEFILE VERSION1 VERSION2
904
Show differences between two versions.
908
% weave init foo.weave
910
% weave add foo.weave ver0 < foo.txt
913
(create updated version)
915
% weave get foo.weave 0 | diff -u - foo.txt
916
% weave add foo.weave ver1 0 < foo.txt
919
% weave get foo.weave 0 > foo.txt (create forked version)
921
% weave add foo.weave ver2 0 < foo.txt
924
% weave merge foo.weave 1 2 > foo.txt (merge them)
925
% vi foo.txt (resolve conflicts)
926
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
938
# in case we're run directly from the subdirectory
939
sys.path.append('..')
941
from bzrlib.weavefile import write_weave, read_weave
942
from bzrlib.progress import ProgressBar
957
return read_weave(file(argv[2], 'rb'))
963
# at the moment, based on everything in the file
965
parents = map(int, argv[4:])
966
lines = sys.stdin.readlines()
967
ver = w.add(name, parents, lines)
968
write_weave(w, file(argv[2], 'wb'))
969
print 'added version %r %d' % (name, ver)
972
if os.path.exists(fn):
973
raise IOError("file exists")
975
write_weave(w, file(fn, 'wb'))
976
elif cmd == 'get': # get one version
978
sys.stdout.writelines(w.get_iter(int(argv[3])))
980
elif cmd == 'mash': # get composite
982
sys.stdout.writelines(w.mash_iter(map(int, argv[3:])))
985
from difflib import unified_diff
988
v1, v2 = map(int, argv[3:5])
991
diff_gen = unified_diff(lines1, lines2,
992
'%s version %d' % (fn, v1),
993
'%s version %d' % (fn, v2))
994
sys.stdout.writelines(diff_gen)
996
elif cmd == 'annotate':
998
# newline is added to all lines regardless; too hard to get
999
# reasonable formatting otherwise
1001
for origin, text in w.annotate(int(argv[3])):
1002
text = text.rstrip('\r\n')
1004
print ' | %s' % (text)
1006
print '%5d | %s' % (origin, text)
1012
elif cmd == 'stats':
1013
weave_stats(argv[2], ProgressBar())
1015
elif cmd == 'check':
1020
print '%d versions ok' % w.numversions()
1022
elif cmd == 'inclusions':
1024
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1026
elif cmd == 'parents':
1028
print ' '.join(map(str, w._parents[int(argv[3])]))
1030
elif cmd == 'plan-merge':
1032
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1034
print '%14s | %s' % (state, line),
1036
elif cmd == 'merge':
1038
p = w.plan_merge(int(argv[3]), int(argv[4]))
1039
sys.stdout.writelines(w.weave_merge(p))
1041
elif cmd == 'mash-merge':
1047
v1, v2 = map(int, argv[3:5])
1049
basis = w.inclusions([v1]).intersection(w.inclusions([v2]))
1051
base_lines = list(w.mash_iter(basis))
1052
a_lines = list(w.get(v1))
1053
b_lines = list(w.get(v2))
1055
from bzrlib.merge3 import Merge3
1056
m3 = Merge3(base_lines, a_lines, b_lines)
1058
name_a = 'version %d' % v1
1059
name_b = 'version %d' % v2
1060
sys.stdout.writelines(m3.merge_lines(name_a=name_a, name_b=name_b))
1062
raise ValueError('unknown command %r' % cmd)
1066
def profile_main(argv):
1067
import tempfile, hotshot, hotshot.stats
1069
prof_f = tempfile.NamedTemporaryFile()
1071
prof = hotshot.Profile(prof_f.name)
1073
ret = prof.runcall(main, argv)
1076
stats = hotshot.stats.load(prof_f.name)
1078
stats.sort_stats('cumulative')
1079
## XXX: Might like to write to stderr or the trace file instead but
1080
## print_stats seems hardcoded to stdout
1081
stats.print_stats(20)
1086
if __name__ == '__main__':
1088
if '--profile' in sys.argv:
1090
args.remove('--profile')
1091
sys.exit(profile_main(args))
1093
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
