50
48
# have slight specializations for different ways its used: annotate,
51
49
# basis for add, get, etc.
53
# TODO: Perhaps the API should work only in names to hide the integer
54
# indexes from the user?
51
# TODO: Probably the API should work only in names to hide the integer
52
# indexes from the user.
56
54
# TODO: Is there any potential performance win by having an add()
57
55
# variant that is passed a pre-cooked version of the single basis
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
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 *
77
90
from bzrlib.trace import mutter
80
class WeaveError(Exception):
81
"""Exception in processing weave"""
84
class WeaveFormatError(WeaveError):
85
"""Weave invariant violated"""
88
class WeaveParentMismatch(WeaveError):
89
"""Parents are mismatched between two revisions."""
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):
93
127
"""weave - versioned text file storage.
95
129
A Weave manages versions of line-based text files, keeping track
96
130
of the originating version for each line.
209
270
return self._parents == other._parents \
210
271
and self._weave == other._weave \
211
and self._sha1s == other._sha1s
272
and self._sha1s == other._sha1s
214
274
def __ne__(self, other):
215
275
return not self.__eq__(other)
218
def maybe_lookup(self, name_or_index):
219
"""Convert possible symbolic name to index, or pass through indexes."""
220
if isinstance(name_or_index, (int, long)):
223
return self.lookup(name_or_index)
226
def lookup(self, name):
277
def _idx_to_name(self, version):
278
return self._names[version]
280
def _lookup(self, name):
227
281
"""Convert symbolic version name to index."""
282
if not self._allow_reserved:
283
self.check_not_reserved_id(name)
229
285
return self._name_map[name]
231
raise WeaveError("name %r not present in weave %r" %
232
(name, self._weave_name))
235
def iter_names(self):
236
"""Yield a list of all names in this weave."""
237
return iter(self._names)
239
def idx_to_name(self, version):
240
return self._names[version]
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:
243
380
def _check_repeated_add(self, name, parents, text, sha1):
244
381
"""Check that a duplicated add is OK.
246
383
If it is, return the (old) index; otherwise raise an exception.
248
idx = self.lookup(name)
249
if sorted(self._parents[idx]) != sorted(parents):
250
raise WeaveError("name \"%s\" already present in weave "
251
"with different parents" % name)
252
if sha1 != self._sha1s[idx]:
253
raise WeaveError("name \"%s\" already present in weave "
254
"with different text" % name)
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)
259
def add(self, name, parents, text, sha1=None):
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):
260
399
"""Add a single text on top of the weave.
262
401
Returns the index number of the newly added version.
265
404
Symbolic name for this version.
266
405
(Typically the revision-id of the revision that added it.)
406
If None, a name will be allocated based on the hash. (sha1:SHAHASH)
269
409
List or set of direct parent version numbers.
272
412
Sequence of lines to be added in the new version.
274
sha -- SHA-1 of the file, if known. This is trusted to be
414
:param nostore_sha: See VersionedFile.add_lines.
277
from bzrlib.osutils import sha_strings
279
assert isinstance(name, basestring)
281
sha1 = sha_strings(text)
282
if name in self._name_map:
283
return self._check_repeated_add(name, parents, text, sha1)
285
parents = map(self.maybe_lookup, parents)
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)
286
427
self._check_versions(parents)
287
## self._check_lines(text)
428
## self._check_lines(lines)
288
429
new_version = len(self._parents)
291
431
# if we abort after here the (in-memory) weave will be corrupt because only
292
432
# some fields are updated
433
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
434
# - Robert Collins 20060226
293
435
self._parents.append(parents[:])
294
436
self._sha1s.append(sha1)
295
self._names.append(name)
296
self._name_map[name] = new_version
437
self._names.append(version_id)
438
self._name_map[version_id] = new_version
300
442
# special case; adding with no parents revision; can do
301
443
# this more quickly by just appending unconditionally.
302
444
# even more specially, if we're adding an empty text we
303
445
# need do nothing at all.
305
447
self._weave.append(('{', new_version))
306
self._weave.extend(text)
448
self._weave.extend(lines)
307
449
self._weave.append(('}', None))
309
450
return new_version
311
452
if len(parents) == 1:
449
554
except IndexError:
450
555
raise IndexError("invalid version number %r" % i)
453
def annotate(self, name_or_index):
454
return list(self.annotate_iter(name_or_index))
457
def annotate_iter(self, name_or_index):
458
"""Yield list of (index-id, line) pairs for the specified version.
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.
460
568
The index indicates when the line originated in the weave."""
461
incls = [self.maybe_lookup(name_or_index)]
462
for origin, lineno, text in self._extract(incls):
470
(lineno, insert, deletes, text)
471
for each literal line.
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."""
477
592
lineno = 0 # line of weave, 0-based
479
594
for l in self._weave:
480
if isinstance(l, tuple):
595
if l.__class__ == tuple:
599
istack.append(self._names[v])
603
dset.add(self._names[v])
605
dset.remove(self._names[v])
493
raise WeaveFormatError('unexpected instruction %r'
607
raise WeaveFormatError('unexpected instruction %r' % v)
496
assert isinstance(l, basestring)
498
yield lineno, istack[-1], dset, l
609
yield lineno, istack[-1], frozenset(dset), l
613
raise WeaveFormatError("unclosed insertion blocks "
614
"at end of weave: %s" % istack)
616
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
619
def plan_merge(self, ver_a, ver_b):
620
"""Return pseudo-annotation indicating how the two versions merge.
622
This is computed between versions a and b and their common
625
Weave lines present in none of them are skipped entirely.
627
inc_a = set(self.get_ancestry([ver_a]))
628
inc_b = set(self.get_ancestry([ver_b]))
629
inc_c = inc_a & inc_b
631
for lineno, insert, deleteset, line in self._walk_internal([ver_a, ver_b]):
632
if deleteset & inc_c:
633
# killed in parent; can't be in either a or b
634
# not relevant to our work
635
yield 'killed-base', line
636
elif insert in inc_c:
637
# was inserted in base
638
killed_a = bool(deleteset & inc_a)
639
killed_b = bool(deleteset & inc_b)
640
if killed_a and killed_b:
641
yield 'killed-both', line
643
yield 'killed-a', line
645
yield 'killed-b', line
647
yield 'unchanged', line
648
elif insert in inc_a:
649
if deleteset & inc_a:
650
yield 'ghost-a', line
654
elif insert in inc_b:
655
if deleteset & inc_b:
656
yield 'ghost-b', line
660
# not in either revision
661
yield 'irrelevant', line
503
663
def _extract(self, versions):
504
664
"""Yield annotation of lines in included set.
527
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.
529
717
for l in self._weave:
530
if isinstance(l, tuple):
718
if l.__class__ == tuple:
534
assert v not in istack
725
iset.remove(istack.pop())
539
727
if v in included:
544
730
if v in included:
733
raise AssertionError()
548
assert isinstance(l, basestring)
549
735
if isactive is None:
550
736
isactive = (not dset) and istack and (istack[-1] in included)
552
738
result.append((istack[-1], lineno, l))
556
raise WFE("unclosed insertion blocks at end of weave",
741
raise WeaveFormatError("unclosed insertion blocks "
742
"at end of weave: %s" % istack)
559
raise WFE("unclosed deletion blocks at end of weave",
566
def get_iter(self, name_or_index):
567
"""Yield lines for the specified version."""
568
incls = [self.maybe_lookup(name_or_index)]
569
for origin, lineno, line in self._extract(incls):
573
def get_text(self, name_or_index):
574
return ''.join(self.get_iter(name_or_index))
575
assert isinstance(version, int)
578
def get_lines(self, name_or_index):
579
return list(self.get_iter(name_or_index))
585
def mash_iter(self, included):
586
"""Return composed version of multiple included versions."""
587
included = map(self.maybe_lookup, included)
588
for origin, lineno, text in self._extract(included):
592
def dump(self, to_file):
593
from pprint import pprint
594
print >>to_file, "Weave._weave = ",
595
pprint(self._weave, to_file)
596
print >>to_file, "Weave._parents = ",
597
pprint(self._parents, to_file)
601
def numversions(self):
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.
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?"""
602
780
l = len(self._parents)
603
assert l == len(self._sha1s)
608
return self.numversions()
783
__len__ = num_versions
611
785
def check(self, progress_bar=None):
612
# check no circular inclusions
613
for version in range(self.numversions()):
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()):
614
790
inclusions = list(self._parents[version])
616
792
inclusions.sort()
618
794
raise WeaveFormatError("invalid included version %d for index %d"
619
795
% (inclusions[-1], version))
621
# try extracting all versions; this is a bit slow and parallel
622
# extraction could be used
623
nv = self.numversions()
624
for version in range(nv):
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():
626
progress_bar.update('checking text', version, nv)
628
for l in self.get_iter(version):
631
expected = self._sha1s[version]
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]
632
841
if hd != expected:
633
raise WeaveError("mismatched sha1 for version %d; "
634
"got %s, expected %s"
635
% (version, hd, expected))
842
raise errors.WeaveInvalidChecksum(
843
"mismatched sha1 for version %s: "
844
"got %s, expected %s"
845
% (version, hd, expected))
637
847
# TODO: check insertions are properly nested, that there are
638
848
# no lines outside of insertion blocks, that deletions are
639
849
# properly paired, etc.
643
def merge(self, merge_versions):
644
"""Automerge and mark conflicts between versions.
646
This returns a sequence, each entry describing alternatives
647
for a chunk of the file. Each of the alternatives is given as
650
If there is a chunk of the file where there's no diagreement,
651
only one alternative is given.
653
# approach: find the included versions common to all the
655
raise NotImplementedError()
659
def _delta(self, included, lines):
660
"""Return changes from basis to new revision.
662
The old text for comparison is the union of included revisions.
664
This is used in inserting a new text.
666
Delta is returned as a sequence of
667
(weave1, weave2, newlines).
669
This indicates that weave1:weave2 of the old weave should be
670
replaced by the sequence of lines in newlines. Note that
671
these line numbers are positions in the total weave and don't
672
correspond to the lines in any extracted version, or even the
673
extracted union of included versions.
675
If line1=line2, this is a pure insert; if newlines=[] this is a
676
pure delete. (Similar to difflib.)
678
raise NotImplementedError()
681
def plan_merge(self, ver_a, ver_b):
682
"""Return pseudo-annotation indicating how the two versions merge.
684
This is computed between versions a and b and their common
687
Weave lines present in none of them are skipped entirely.
689
inc_a = self.inclusions([ver_a])
690
inc_b = self.inclusions([ver_b])
691
inc_c = inc_a & inc_b
693
for lineno, insert, deleteset, line in self._walk():
694
if deleteset & inc_c:
695
# killed in parent; can't be in either a or b
696
# not relevant to our work
697
yield 'killed-base', line
698
elif insert in inc_c:
699
# was inserted in base
700
killed_a = bool(deleteset & inc_a)
701
killed_b = bool(deleteset & inc_b)
702
if killed_a and killed_b:
703
yield 'killed-both', line
705
yield 'killed-a', line
707
yield 'killed-b', line
709
yield 'unchanged', line
710
elif insert in inc_a:
711
if deleteset & inc_a:
712
yield 'ghost-a', line
716
elif insert in inc_b:
717
if deleteset & inc_b:
718
yield 'ghost-b', line
722
# not in either revision
723
yield 'irrelevant', line
725
yield 'unchanged', '' # terminator
729
def weave_merge(self, plan):
734
for state, line in plan:
735
if state == 'unchanged' or state == 'killed-both':
736
# resync and flush queued conflicts changes if any
737
if not lines_a and not lines_b:
739
elif ch_a and not ch_b:
741
for l in lines_a: yield l
742
elif ch_b and not ch_a:
743
for l in lines_b: yield l
744
elif lines_a == lines_b:
745
for l in lines_a: yield l
748
for l in lines_a: yield l
750
for l in lines_b: yield l
757
if state == 'unchanged':
760
elif state == 'killed-a':
763
elif state == 'killed-b':
766
elif state == 'new-a':
769
elif state == 'new-b':
773
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
778
def join(self, other):
779
"""Integrate versions from other into this weave.
781
The resulting weave contains all the history of both weaves;
782
any version you could retrieve from either self or other can be
783
retrieved from self after this call.
785
It is illegal for the two weaves to contain different values
786
or different parents for any version. See also reweave().
788
if other.numversions() == 0:
789
return # nothing to update, easy
790
# two loops so that we do not change ourselves before verifying it
792
# work through in index order to make sure we get all dependencies
793
for other_idx, name in enumerate(other._names):
794
if self._check_version_consistent(other, other_idx, name):
796
for other_idx, name in enumerate(other._names):
797
# TODO: If all the parents of the other version are already
798
# present then we can avoid some work by just taking the delta
799
# and adjusting the offsets.
800
new_parents = self._imported_parents(other, other_idx)
801
lines = other.get_lines(other_idx)
802
sha1 = other._sha1s[other_idx]
803
self.add(name, new_parents, lines, sha1)
806
851
def _imported_parents(self, other, other_idx):
807
852
"""Return list of parents in self corresponding to indexes in other."""
809
854
for parent_idx in other._parents[other_idx]:
810
855
parent_name = other._names[parent_idx]
811
if parent_name not in self._names:
856
if parent_name not in self._name_map:
812
857
# should not be possible
813
raise WeaveError("missing parent {%s} of {%s} in %r"
858
raise WeaveError("missing parent {%s} of {%s} in %r"
814
859
% (parent_name, other._name_map[other_idx], self))
815
860
new_parents.append(self._name_map[parent_name])
816
861
return new_parents
850
def reweave(self, other):
851
"""Reweave self with other."""
852
new_weave = reweave(self, other)
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."""
853
903
for attr in self.__slots__:
854
setattr(self, attr, getattr(new_weave, attr))
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
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
924
except errors.NoSuchFile:
930
def _add_lines(self, version_id, parents, lines, parent_texts,
931
left_matching_blocks, nostore_sha, random_id, check_content):
932
"""Add a version and save the weave."""
933
self.check_not_reserved_id(version_id)
934
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
935
parent_texts, left_matching_blocks, nostore_sha, random_id,
940
def copy_to(self, name, transport):
941
"""See VersionedFile.copy_to()."""
942
# as we are all in memory always, just serialise to the new place.
944
write_weave_v5(self, sio)
946
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
949
"""Save the weave."""
950
self._check_write_ok()
952
write_weave_v5(self, sio)
954
bytes = sio.getvalue()
955
path = self._weave_name + WeaveFile.WEAVE_SUFFIX
957
self._transport.put_bytes(path, bytes, self._filemode)
958
except errors.NoSuchFile:
959
self._transport.mkdir(dirname(path))
960
self._transport.put_bytes(path, bytes, self._filemode)
964
"""See VersionedFile.get_suffixes()."""
965
return [WeaveFile.WEAVE_SUFFIX]
967
def insert_record_stream(self, stream):
968
super(WeaveFile, self).insert_record_stream(stream)
972
def _reweave(wa, wb, pb=None, msg=None):
858
973
"""Combine two weaves and return the result.
860
This works even if a revision R has different parents in
975
This works even if a revision R has different parents in
861
976
wa and wb. In the resulting weave all the parents are given.
863
This is done by just building up a new weave, maintaining ordering
978
This is done by just building up a new weave, maintaining ordering
864
979
of the versions in the two inputs. More efficient approaches
865
might be possible but it should only be necessary to do
866
this operation rarely, when a new previously ghost version is
980
might be possible but it should only be necessary to do
981
this operation rarely, when a new previously ghost version is
984
:param pb: An optional progress bar, indicating how far done we are
985
:param msg: An optional message for the progress
871
queue_a = range(wa.numversions())
872
queue_b = range(wb.numversions())
989
queue_a = range(wa.num_versions())
990
queue_b = range(wb.num_versions())
873
991
# first determine combined parents of all versions
874
992
# map from version name -> all parent names
875
993
combined_parents = _reweave_parent_graphs(wa, wb)
876
994
mutter("combined parents: %r", combined_parents)
877
order = _make_reweave_order(wa._names, wb._names, combined_parents)
995
order = tsort.topo_sort(combined_parents.iteritems())
878
996
mutter("order to reweave: %r", order)
1001
for idx, name in enumerate(order):
1003
pb.update(msg, idx, len(order))
880
1004
if name in wa._name_map:
881
1005
lines = wa.get_lines(name)
882
1006
if name in wb._name_map:
883
assert lines == wb.get_lines(name)
1007
lines_b = wb.get_lines(name)
1008
if lines != lines_b:
1009
mutter('Weaves differ on content. rev_id {%s}', name)
1010
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1012
lines = list(difflib.unified_diff(lines, lines_b,
1013
wa._weave_name, wb._weave_name))
1014
mutter('lines:\n%s', ''.join(lines))
1015
raise errors.WeaveTextDiffers(name, wa, wb)
885
1017
lines = wb.get_lines(name)
886
wr.add(name, combined_parents[name], lines)
1018
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
890
1022
def _reweave_parent_graphs(wa, wb):
891
1023
"""Return combined parent ancestry for two weaves.
893
1025
Returned as a list of (version_name, set(parent_names))"""
895
1027
for weave in [wa, wb]:
896
1028
for idx, name in enumerate(weave._names):
897
1029
p = combined.setdefault(name, set())
898
p.update(map(weave.idx_to_name, weave._parents[idx]))
1030
p.update(map(weave._idx_to_name, weave._parents[idx]))
902
def _make_reweave_order(wa_order, wb_order, combined_parents):
903
"""Return an order to reweave versions respecting parents."""
907
next_a = next_b = None
908
len_a = len(wa_order)
909
len_b = len(wb_order)
910
while ia < len(wa_order) or ib < len(wb_order):
912
next_a = wa_order[ia]
916
if combined_parents[next_a].issubset(done):
918
result.append(next_a)
922
next_b = wb_order[ib]
926
elif combined_parents[next_b].issubset(done):
928
result.append(next_b)
931
raise WeaveError("don't know how to reweave at {%s} and {%s}"
937
"""Show the weave's table-of-contents"""
938
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
939
for i in (6, 50, 10, 10):
942
for i in range(w.numversions()):
945
parent_str = ' '.join(map(str, w._parents[i]))
946
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
950
def weave_stats(weave_file, pb):
951
from bzrlib.weavefile import read_weave
953
wf = file(weave_file, 'rb')
955
# FIXME: doesn't work on pipes
956
weave_size = wf.tell()
960
for i in range(vers):
961
pb.update('checking sizes', i, vers)
962
for origin, lineno, line in w._extract([i]):
967
print 'versions %9d' % vers
968
print 'weave file %9d bytes' % weave_size
969
print 'total contents %9d bytes' % total
970
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
973
print 'average size %9d bytes' % avg
974
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
978
print """bzr weave tool
980
Experimental tool for weave algorithm.
984
Create an empty weave file
985
weave get WEAVEFILE VERSION
986
Write out specified version.
987
weave check WEAVEFILE
988
Check consistency of all versions.
990
Display table of contents.
991
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
992
Add NEWTEXT, with specified parent versions.
993
weave annotate WEAVEFILE VERSION
994
Display origin of each line.
995
weave mash WEAVEFILE VERSION...
996
Display composite of all selected versions.
997
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
998
Auto-merge two versions and display conflicts.
999
weave diff WEAVEFILE VERSION1 VERSION2
1000
Show differences between two versions.
1004
% weave init foo.weave
1006
% weave add foo.weave ver0 < foo.txt
1009
(create updated version)
1011
% weave get foo.weave 0 | diff -u - foo.txt
1012
% weave add foo.weave ver1 0 < foo.txt
1015
% weave get foo.weave 0 > foo.txt (create forked version)
1017
% weave add foo.weave ver2 0 < foo.txt
1020
% weave merge foo.weave 1 2 > foo.txt (merge them)
1021
% vi foo.txt (resolve conflicts)
1022
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1034
# in case we're run directly from the subdirectory
1035
sys.path.append('..')
1037
from bzrlib.weavefile import write_weave, read_weave
1038
from bzrlib.progress import ProgressBar
1053
return read_weave(file(argv[2], 'rb'))
1059
# at the moment, based on everything in the file
1061
parents = map(int, argv[4:])
1062
lines = sys.stdin.readlines()
1063
ver = w.add(name, parents, lines)
1064
write_weave(w, file(argv[2], 'wb'))
1065
print 'added version %r %d' % (name, ver)
1068
if os.path.exists(fn):
1069
raise IOError("file exists")
1071
write_weave(w, file(fn, 'wb'))
1072
elif cmd == 'get': # get one version
1074
sys.stdout.writelines(w.get_iter(int(argv[3])))
1076
elif cmd == 'mash': # get composite
1078
sys.stdout.writelines(w.mash_iter(map(int, argv[3:])))
1081
from difflib import unified_diff
1084
v1, v2 = map(int, argv[3:5])
1087
diff_gen = unified_diff(lines1, lines2,
1088
'%s version %d' % (fn, v1),
1089
'%s version %d' % (fn, v2))
1090
sys.stdout.writelines(diff_gen)
1092
elif cmd == 'annotate':
1094
# newline is added to all lines regardless; too hard to get
1095
# reasonable formatting otherwise
1097
for origin, text in w.annotate(int(argv[3])):
1098
text = text.rstrip('\r\n')
1100
print ' | %s' % (text)
1102
print '%5d | %s' % (origin, text)
1108
elif cmd == 'stats':
1109
weave_stats(argv[2], ProgressBar())
1111
elif cmd == 'check':
1116
print '%d versions ok' % w.numversions()
1118
elif cmd == 'inclusions':
1120
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1122
elif cmd == 'parents':
1124
print ' '.join(map(str, w._parents[int(argv[3])]))
1126
elif cmd == 'plan-merge':
1128
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1130
print '%14s | %s' % (state, line),
1132
elif cmd == 'merge':
1134
p = w.plan_merge(int(argv[3]), int(argv[4]))
1135
sys.stdout.writelines(w.weave_merge(p))
1137
elif cmd == 'mash-merge':
1143
v1, v2 = map(int, argv[3:5])
1145
basis = w.inclusions([v1]).intersection(w.inclusions([v2]))
1147
base_lines = list(w.mash_iter(basis))
1148
a_lines = list(w.get(v1))
1149
b_lines = list(w.get(v2))
1151
from bzrlib.merge3 import Merge3
1152
m3 = Merge3(base_lines, a_lines, b_lines)
1154
name_a = 'version %d' % v1
1155
name_b = 'version %d' % v2
1156
sys.stdout.writelines(m3.merge_lines(name_a=name_a, name_b=name_b))
1158
raise ValueError('unknown command %r' % cmd)
1162
def profile_main(argv):
1163
import tempfile, hotshot, hotshot.stats
1165
prof_f = tempfile.NamedTemporaryFile()
1167
prof = hotshot.Profile(prof_f.name)
1169
ret = prof.runcall(main, argv)
1172
stats = hotshot.stats.load(prof_f.name)
1174
stats.sort_stats('cumulative')
1175
## XXX: Might like to write to stderr or the trace file instead but
1176
## print_stats seems hardcoded to stdout
1177
stats.print_stats(20)
1182
if __name__ == '__main__':
1184
if '--profile' in sys.argv:
1186
args.remove('--profile')
1187
sys.exit(profile_main(args))
1189
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
