1
# Copyright (C) 2005, 2009 Canonical Ltd
3
# Copyright (C) 2005 Canonical Ltd
3
5
# This program is free software; you can redistribute it and/or modify
4
6
# it under the terms of the GNU General Public License as published by
5
7
# the Free Software Foundation; either version 2 of the License, or
6
8
# (at your option) any later version.
8
10
# This program is distributed in the hope that it will be useful,
9
11
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
12
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
13
# GNU General Public License for more details.
13
15
# You should have received a copy of the GNU General Public License
14
16
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
19
# Author: Martin Pool <mbp@canonical.com>
20
22
"""Weave - storage of related text file versions"""
24
# TODO: Perhaps have copy method for Weave instances?
23
26
# XXX: If we do weaves this way, will a merge still behave the same
24
27
# way if it's done in a different order? That's a pretty desirable
30
# TODO: How to write these to disk? One option is cPickle, which
31
# would be fast but less friendly to C, and perhaps not portable. Another is
27
33
# TODO: Nothing here so far assumes the lines are really \n newlines,
28
# rather than being split up in some other way. We could accommodate
34
# rather than being split up in some other way. We could accomodate
29
35
# binaries, perhaps by naively splitting on \n or perhaps using
30
36
# something like a rolling checksum.
38
# TODO: Perhaps track SHA-1 in the header for protection? This would
39
# be redundant with it being stored in the inventory, but perhaps
42
# TODO: Track version names as well as indexes.
44
# TODO: Probably do transitive expansion when specifying parents?
46
# TODO: Separate out some code to read and write weaves.
32
48
# TODO: End marker for each version so we can stop reading?
34
50
# TODO: Check that no insertion occurs inside a deletion that was
35
51
# active in the version of the insertion.
37
# TODO: In addition to the SHA-1 check, perhaps have some code that
38
# checks structural constraints of the weave: ie that insertions are
39
# properly nested, that there is no text outside of an insertion, that
40
# insertions or deletions are not repeated, etc.
42
# TODO: Parallel-extract that passes back each line along with a
43
# description of which revisions include it. Nice for checking all
44
# shas or calculating stats in parallel.
46
# TODO: Using a single _extract routine and then processing the output
47
# is probably inefficient. It's simple enough that we can afford to
48
# have slight specializations for different ways its used: annotate,
49
# basis for add, get, etc.
51
# TODO: Probably the API should work only in names to hide the integer
52
# indexes from the user.
54
# TODO: Is there any potential performance win by having an add()
55
# 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
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 *
95
from bzrlib.trace import mutter
96
from bzrlib.versionedfile import (
103
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
106
class WeaveContentFactory(ContentFactory):
107
"""Content factory for streaming from weaves.
109
:seealso ContentFactory:
112
def __init__(self, version, weave):
113
"""Create a WeaveContentFactory for version from weave."""
114
ContentFactory.__init__(self)
115
self.sha1 = weave.get_sha1s([version])[version]
116
self.key = (version,)
117
parents = weave.get_parent_map([version])[version]
118
self.parents = tuple((parent,) for parent in parents)
119
self.storage_kind = 'fulltext'
122
def get_bytes_as(self, storage_kind):
123
if storage_kind == 'fulltext':
124
return self._weave.get_text(self.key[-1])
125
elif storage_kind == 'chunked':
126
return self._weave.get_lines(self.key[-1])
128
raise UnavailableRepresentation(self.key, storage_kind, 'fulltext')
131
class Weave(VersionedFile):
53
# TODO: Perhaps a special slower check() method that verifies more
54
# nesting constraints and the MD5 of each version?
62
from sets import Set, ImmutableSet
64
frozenset = ImmutableSet
68
class WeaveError(Exception):
69
"""Exception in processing weave"""
72
class WeaveFormatError(WeaveError):
73
"""Weave invariant violated"""
132
77
"""weave - versioned text file storage.
134
79
A Weave manages versions of line-based text files, keeping track
135
80
of the originating version for each line.
195
138
should be no way to get an earlier version deleting a later
199
Text of the weave; list of control instruction tuples and strings.
202
145
List of parents, indexed by version number.
203
146
It is only necessary to store the minimal set of parents for
204
147
each version; the parent's parents are implied.
207
List of hex SHA-1 of each version.
210
List of symbolic names for each version. Each should be unique.
213
For each name, the version number.
216
Descriptive name of this weave; typically the filename if known.
150
List of hex SHA-1 of each version, or None if not recorded.
220
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
221
'_weave_name', '_matcher', '_allow_reserved']
223
def __init__(self, weave_name=None, access_mode='w', matcher=None,
224
get_scope=None, allow_reserved=False):
227
:param get_scope: A callable that returns an opaque object to be used
228
for detecting when this weave goes out of scope (should stop
229
answering requests or allowing mutation).
231
super(Weave, self).__init__()
237
self._weave_name = weave_name
239
self._matcher = bzrlib.patiencediff.PatienceSequenceMatcher
241
self._matcher = matcher
242
if get_scope is None:
243
get_scope = lambda:None
244
self._get_scope = get_scope
245
self._scope = get_scope()
246
self._access_mode = access_mode
247
self._allow_reserved = allow_reserved
250
return "Weave(%r)" % self._weave_name
252
def _check_write_ok(self):
253
"""Is the versioned file marked as 'finished' ? Raise if it is."""
254
if self._get_scope() != self._scope:
255
raise errors.OutSideTransaction()
256
if self._access_mode != 'w':
257
raise errors.ReadOnlyObjectDirtiedError(self)
260
"""Return a deep copy of self.
262
The copy can be modified without affecting the original weave."""
264
other._weave = self._weave[:]
265
other._parents = self._parents[:]
266
other._sha1s = self._sha1s[:]
267
other._names = self._names[:]
268
other._name_map = self._name_map.copy()
269
other._weave_name = self._weave_name
272
158
def __eq__(self, other):
273
159
if not isinstance(other, Weave):
275
return self._parents == other._parents \
276
and self._weave == other._weave \
277
and self._sha1s == other._sha1s
161
return self._v == other._v \
162
and self._l == other._l
279
165
def __ne__(self, other):
280
166
return not self.__eq__(other)
282
def _idx_to_name(self, version):
283
return self._names[version]
285
def _lookup(self, name):
286
"""Convert symbolic version name to index."""
287
if not self._allow_reserved:
288
self.check_not_reserved_id(name)
290
return self._name_map[name]
292
raise RevisionNotPresent(name, self._weave_name)
295
"""See VersionedFile.versions."""
296
return self._names[:]
298
def has_version(self, version_id):
299
"""See VersionedFile.has_version."""
300
return (version_id in self._name_map)
302
__contains__ = has_version
304
def get_record_stream(self, versions, ordering, include_delta_closure):
305
"""Get a stream of records for versions.
307
:param versions: The versions to include. Each version is a tuple
309
:param ordering: Either 'unordered' or 'topological'. A topologically
310
sorted stream has compression parents strictly before their
312
:param include_delta_closure: If True then the closure across any
313
compression parents will be included (in the opaque data).
314
:return: An iterator of ContentFactory objects, each of which is only
315
valid until the iterator is advanced.
317
versions = [version[-1] for version in versions]
318
if ordering == 'topological':
319
parents = self.get_parent_map(versions)
320
new_versions = tsort.topo_sort(parents)
321
new_versions.extend(set(versions).difference(set(parents)))
322
versions = new_versions
323
elif ordering == 'groupcompress':
324
parents = self.get_parent_map(versions)
325
new_versions = sort_groupcompress(parents)
326
new_versions.extend(set(versions).difference(set(parents)))
327
versions = new_versions
328
for version in versions:
330
yield WeaveContentFactory(version, self)
332
yield AbsentContentFactory((version,))
334
def get_parent_map(self, version_ids):
335
"""See VersionedFile.get_parent_map."""
337
for version_id in version_ids:
338
if version_id == NULL_REVISION:
343
map(self._idx_to_name,
344
self._parents[self._lookup(version_id)]))
345
except RevisionNotPresent:
347
result[version_id] = parents
350
def get_parents_with_ghosts(self, version_id):
351
raise NotImplementedError(self.get_parents_with_ghosts)
353
def insert_record_stream(self, stream):
354
"""Insert a record stream into this versioned file.
356
:param stream: A stream of records to insert.
358
:seealso VersionedFile.get_record_stream:
361
for record in stream:
362
# Raise an error when a record is missing.
363
if record.storage_kind == 'absent':
364
raise RevisionNotPresent([record.key[0]], self)
365
# adapt to non-tuple interface
366
parents = [parent[0] for parent in record.parents]
367
if (record.storage_kind == 'fulltext'
368
or record.storage_kind == 'chunked'):
369
self.add_lines(record.key[0], parents,
370
osutils.chunks_to_lines(record.get_bytes_as('chunked')))
372
adapter_key = record.storage_kind, 'fulltext'
374
adapter = adapters[adapter_key]
376
adapter_factory = adapter_registry.get(adapter_key)
377
adapter = adapter_factory(self)
378
adapters[adapter_key] = adapter
379
lines = split_lines(adapter.get_bytes(record))
381
self.add_lines(record.key[0], parents, lines)
382
except RevisionAlreadyPresent:
385
def _check_repeated_add(self, name, parents, text, sha1):
386
"""Check that a duplicated add is OK.
388
If it is, return the (old) index; otherwise raise an exception.
390
idx = self._lookup(name)
391
if sorted(self._parents[idx]) != sorted(parents) \
392
or sha1 != self._sha1s[idx]:
393
raise RevisionAlreadyPresent(name, self._weave_name)
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):
169
def add(self, parents, text):
404
170
"""Add a single text on top of the weave.
406
172
Returns the index number of the newly added version.
409
Symbolic name for this version.
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)
414
175
List or set of direct parent version numbers.
417
Sequence of lines to be added in the new version.
419
:param nostore_sha: See VersionedFile.add_lines.
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)
432
self._check_versions(parents)
433
## self._check_lines(lines)
434
new_version = len(self._parents)
436
# if we abort after here the (in-memory) weave will be corrupt because only
437
# some fields are updated
438
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
439
# - Robert Collins 20060226
440
self._parents.append(parents[:])
178
Sequence of lines to be added in the new version."""
179
## self._check_versions(parents)
180
## self._check_lines(text)
191
ancestors = self.inclusions(parents)
192
delta = self._delta(ancestors, text)
194
# offset gives the number of lines that have been inserted
195
# into the weave up to the current point; if the original edit instruction
196
# says to change line A then we actually change (A+offset)
199
for i1, i2, newlines in delta:
202
assert i2 <= len(self._l)
204
# the deletion and insertion are handled separately.
205
# first delete the region.
207
self._l.insert(i1+offset, ('[', idx))
208
self._l.insert(i2+offset+1, (']', idx))
213
# there may have been a deletion spanning up to
214
# i2; we want to insert after this region to make sure
215
# we don't destroy ourselves
217
self._l[i:i] = [('{', idx)] \
220
offset += 2 + len(newlines)
222
self._addversion(parents)
224
# special case; adding with no parents revision; can do this
225
# more quickly by just appending unconditionally
226
self._l.append(('{', idx))
228
self._l.append(('}', idx))
230
self._addversion(None)
441
232
self._sha1s.append(sha1)
442
self._names.append(version_id)
443
self._name_map[version_id] = new_version
447
# special case; adding with no parents revision; can do
448
# this more quickly by just appending unconditionally.
449
# even more specially, if we're adding an empty text we
450
# need do nothing at all.
452
self._weave.append(('{', new_version))
453
self._weave.extend(lines)
454
self._weave.append(('}', None))
457
if len(parents) == 1:
458
pv = list(parents)[0]
459
if sha1 == self._sha1s[pv]:
460
# special case: same as the single parent
464
ancestors = self._inclusions(parents)
468
# basis a list of (origin, lineno, line)
471
for origin, lineno, line in self._extract(ancestors):
472
basis_lineno.append(lineno)
473
basis_lines.append(line)
475
# another small special case: a merge, producing the same text
477
if lines == basis_lines:
480
# add a sentinel, because we can also match against the final line
481
basis_lineno.append(len(self._weave))
483
# XXX: which line of the weave should we really consider
484
# matches the end of the file? the current code says it's the
485
# last line of the weave?
487
#print 'basis_lines:', basis_lines
488
#print 'new_lines: ', lines
490
s = self._matcher(None, basis_lines, lines)
492
# offset gives the number of lines that have been inserted
493
# into the weave up to the current point; if the original edit instruction
494
# says to change line A then we actually change (A+offset)
497
for tag, i1, i2, j1, j2 in s.get_opcodes():
498
# i1,i2 are given in offsets within basis_lines; we need to map them
499
# back to offsets within the entire weave
500
#print 'raw match', tag, i1, i2, j1, j2
503
i1 = basis_lineno[i1]
504
i2 = basis_lineno[i2]
505
# the deletion and insertion are handled separately.
506
# first delete the region.
508
self._weave.insert(i1+offset, ('[', new_version))
509
self._weave.insert(i2+offset+1, (']', new_version))
513
# there may have been a deletion spanning up to
514
# i2; we want to insert after this region to make sure
515
# we don't destroy ourselves
517
self._weave[i:i] = ([('{', new_version)]
520
offset += 2 + (j2 - j1)
523
def _inclusions(self, versions):
237
def inclusions(self, versions):
524
238
"""Return set of all ancestors of given version(s)."""
525
if not len(versions):
244
# include all its parents
249
raise ValueError("version %d not present in weave" % v)
252
def minimal_parents(self, version):
253
"""Find the minimal set of parents for the version."""
254
included = self._v[version]
528
for v in xrange(max(versions), 0, -1):
530
# include all its parents
531
i.update(self._parents[v])
533
## except IndexError:
534
## raise ValueError("version %d not present in weave" % v)
536
def get_ancestry(self, version_ids, topo_sorted=True):
537
"""See VersionedFile.get_ancestry."""
538
if isinstance(version_ids, basestring):
539
version_ids = [version_ids]
540
i = self._inclusions([self._lookup(v) for v in version_ids])
541
return [self._idx_to_name(v) for v in i]
259
li.sort(reverse=True)
267
gotit.update(self.inclusions(pv))
269
assert mininc[0] >= 0
270
assert mininc[-1] < version
274
def _addversion(self, parents):
276
self._v.append(parents)
278
self._v.append(frozenset())
543
281
def _check_lines(self, text):
544
282
if not isinstance(text, list):
675
320
The set typically but not necessarily corresponds to a version.
678
if not isinstance(i, int):
681
included = self._inclusions(versions)
322
included = self.inclusions(versions)
687
327
lineno = 0 # line of weave, 0-based
693
331
WFE = WeaveFormatError
696
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
697
# +285282 0 1676.8040 1676.8040 +<isinstance>
698
# 1.6 seconds in 'isinstance'.
699
# changing the first isinstance:
700
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
701
# +140414 0 762.8050 762.8050 +<isinstance>
702
# note that the inline time actually dropped (less function calls)
703
# and total processing time was halved.
704
# we're still spending ~1/4 of the method in isinstance though.
705
# so lets hard code the acceptable string classes we expect:
706
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
707
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
709
# yay, down to ~1/4 the initial extract time, and our inline time
710
# has shrunk again, with isinstance no longer dominating.
711
# tweaking the stack inclusion test to use a set gives:
712
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
713
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
715
# - a 5% win, or possibly just noise. However with large istacks that
716
# 'in' test could dominate, so I'm leaving this change in place -
717
# when its fast enough to consider profiling big datasets we can review.
722
for l in self._weave:
723
if l.__class__ == tuple:
334
if isinstance(l, tuple):
338
assert v not in istack
730
iset.remove(istack.pop())
732
344
if v in included:
735
349
if v in included:
738
raise AssertionError()
353
assert isinstance(l, basestring)
740
354
if isactive is None:
741
355
isactive = (not dset) and istack and (istack[-1] in included)
743
result.append((istack[-1], lineno, l))
357
yield istack[-1], lineno, l
746
raise WeaveFormatError("unclosed insertion blocks "
747
"at end of weave: %s" % istack)
361
raise WFE("unclosed insertion blocks at end of weave",
749
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
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?"""
785
l = len(self._parents)
364
raise WFE("unclosed deletion blocks at end of weave",
368
def get_iter(self, version):
369
"""Yield lines for the specified version."""
370
for origin, lineno, line in self._extract([version]):
374
def get(self, index):
375
return list(self.get_iter(index))
378
def mash_iter(self, included):
379
"""Return composed version of multiple included versions."""
380
included = frozenset(included)
381
for origin, lineno, text in self._extract(included):
385
def dump(self, to_file):
386
from pprint import pprint
387
print >>to_file, "Weave._l = ",
388
pprint(self._l, to_file)
389
print >>to_file, "Weave._v = ",
390
pprint(self._v, to_file)
394
def numversions(self):
396
assert l == len(self._sha1s)
788
__len__ = num_versions
790
400
def check(self, progress_bar=None):
791
# TODO evaluate performance hit of using string sets in this routine.
792
# TODO: check no circular inclusions
793
# TODO: create a nested progress bar
794
for version in range(self.num_versions()):
795
inclusions = list(self._parents[version])
401
# check no circular inclusions
402
for version in range(self.numversions()):
403
inclusions = list(self._v[version])
797
405
inclusions.sort()
798
406
if inclusions[-1] >= version:
799
407
raise WeaveFormatError("invalid included version %d for index %d"
800
408
% (inclusions[-1], version))
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():
410
# try extracting all versions; this is a bit slow and parallel
411
# extraction could be used
413
nv = self.numversions()
414
for version in range(nv):
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]
416
progress_bar.update('checking text', version, nv)
418
for l in self.get_iter(version):
421
expected = self._sha1s[version]
846
422
if hd != expected:
847
raise errors.WeaveInvalidChecksum(
848
"mismatched sha1 for version %s: "
849
"got %s, expected %s"
850
% (version, hd, expected))
423
raise WeaveError("mismatched sha1 for version %d; "
424
"got %s, expected %s"
425
% (version, hd, expected))
852
427
# TODO: check insertions are properly nested, that there are
853
428
# no lines outside of insertion blocks, that deletions are
854
429
# properly paired, etc.
856
def _imported_parents(self, other, other_idx):
857
"""Return list of parents in self corresponding to indexes in other."""
859
for parent_idx in other._parents[other_idx]:
860
parent_name = other._names[parent_idx]
861
if parent_name not in self._name_map:
862
# should not be possible
863
raise WeaveError("missing parent {%s} of {%s} in %r"
864
% (parent_name, other._name_map[other_idx], self))
865
new_parents.append(self._name_map[parent_name])
868
def _check_version_consistent(self, other, other_idx, name):
869
"""Check if a version in consistent in this and other.
871
To be consistent it must have:
874
* the same direct parents (by name, not index, and disregarding
877
If present & correct return True;
878
if not present in self return False;
879
if inconsistent raise error."""
880
this_idx = self._name_map.get(name, -1)
882
if self._sha1s[this_idx] != other._sha1s[other_idx]:
883
raise errors.WeaveTextDiffers(name, self, other)
884
self_parents = self._parents[this_idx]
885
other_parents = other._parents[other_idx]
886
n1 = set([self._names[i] for i in self_parents])
887
n2 = set([other._names[i] for i in other_parents])
888
if not self._compatible_parents(n1, n2):
889
raise WeaveParentMismatch("inconsistent parents "
890
"for version {%s}: %s vs %s" % (name, n1, n2))
433
def merge(self, merge_versions):
434
"""Automerge and mark conflicts between versions.
436
This returns a sequence, each entry describing alternatives
437
for a chunk of the file. Each of the alternatives is given as
440
If there is a chunk of the file where there's no diagreement,
441
only one alternative is given.
444
# approach: find the included versions common to all the
446
raise NotImplementedError()
450
def _delta(self, included, lines):
451
"""Return changes from basis to new revision.
453
The old text for comparison is the union of included revisions.
455
This is used in inserting a new text.
457
Delta is returned as a sequence of
458
(weave1, weave2, newlines).
460
This indicates that weave1:weave2 of the old weave should be
461
replaced by the sequence of lines in newlines. Note that
462
these line numbers are positions in the total weave and don't
463
correspond to the lines in any extracted version, or even the
464
extracted union of included versions.
466
If line1=line2, this is a pure insert; if newlines=[] this is a
467
pure delete. (Similar to difflib.)
469
# basis a list of (origin, lineno, line)
472
for origin, lineno, line in self._extract(included):
473
basis_lineno.append(lineno)
474
basis_lines.append(line)
476
# add a sentinal, because we can also match against the final line
477
basis_lineno.append(len(self._l))
479
# XXX: which line of the weave should we really consider
480
# matches the end of the file? the current code says it's the
481
# last line of the weave?
483
from difflib import SequenceMatcher
484
s = SequenceMatcher(None, basis_lines, lines)
486
# TODO: Perhaps return line numbers from composed weave as well?
488
for tag, i1, i2, j1, j2 in s.get_opcodes():
489
##print tag, i1, i2, j1, j2
494
# i1,i2 are given in offsets within basis_lines; we need to map them
495
# back to offsets within the entire weave
496
real_i1 = basis_lineno[i1]
497
real_i2 = basis_lineno[i2]
501
assert j2 <= len(lines)
503
yield real_i1, real_i2, lines[j1:j2]
507
def weave_info(filename, out):
508
"""Show some text information about the weave."""
509
from weavefile import read_weave
510
wf = file(filename, 'rb')
512
# FIXME: doesn't work on pipes
513
weave_size = wf.tell()
514
print >>out, "weave file size %d bytes" % weave_size
515
print >>out, "weave contains %d versions" % len(w._v)
518
print '%6s %6s %8s %40s %20s' % ('ver', 'lines', 'bytes', 'sha1', 'parents')
519
for i in (6, 6, 8, 40, 20):
522
for i in range(len(w._v)):
525
bytes = sum((len(a) for a in text))
527
print '%6d %6d %8d %40s' % (i, lines, bytes, sha1),
533
print >>out, "versions total %d bytes" % total
534
print >>out, "compression ratio %.3f" % (float(total)/float(weave_size))
538
print """bzr weave tool
540
Experimental tool for weave algorithm.
544
Create an empty weave file
545
weave get WEAVEFILE VERSION
546
Write out specified version.
547
weave check WEAVEFILE
548
Check consistency of all versions.
550
Display table of contents.
551
weave add WEAVEFILE [BASE...] < NEWTEXT
552
Add NEWTEXT, with specified parent versions.
553
weave annotate WEAVEFILE VERSION
554
Display origin of each line.
555
weave mash WEAVEFILE VERSION...
556
Display composite of all selected versions.
557
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
558
Auto-merge two versions and display conflicts.
562
% weave init foo.weave
564
% weave add foo.weave < foo.txt
567
(create updated version)
569
% weave get foo.weave 0 | diff -u - foo.txt
570
% weave add foo.weave 0 < foo.txt
573
% weave get foo.weave 0 > foo.txt (create forked version)
575
% weave add foo.weave 0 < foo.txt
578
% weave merge foo.weave 1 2 > foo.txt (merge them)
579
% vi foo.txt (resolve conflicts)
580
% weave add foo.weave 1 2 < foo.txt (commit merged version)
589
from weavefile import write_weave, read_weave
590
from bzrlib.progress import ProgressBar
598
return read_weave(file(argv[2], 'rb'))
604
# at the moment, based on everything in the file
605
parents = map(int, argv[3:])
606
lines = sys.stdin.readlines()
607
ver = w.add(parents, lines)
608
write_weave(w, file(argv[2], 'wb'))
609
print 'added version %d' % ver
612
if os.path.exists(fn):
613
raise IOError("file exists")
615
write_weave(w, file(fn, 'wb'))
616
elif cmd == 'get': # get one version
618
sys.stdout.writelines(w.get_iter(int(argv[3])))
620
elif cmd == 'mash': # get composite
622
sys.stdout.writelines(w.mash_iter(map(int, argv[3:])))
624
elif cmd == 'annotate':
626
# newline is added to all lines regardless; too hard to get
627
# reasonable formatting otherwise
629
for origin, text in w.annotate(int(argv[3])):
630
text = text.rstrip('\r\n')
632
print ' | %s' % (text)
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."""
908
for attr in self.__slots__:
909
if attr != '_weave_name':
910
setattr(self, attr, copy(getattr(otherweave, attr)))
913
class WeaveFile(Weave):
914
"""A WeaveFile represents a Weave on disk and writes on change."""
916
WEAVE_SUFFIX = '.weave'
918
def __init__(self, name, transport, filemode=None, create=False, access_mode='w', get_scope=None):
919
"""Create a WeaveFile.
921
:param create: If not True, only open an existing knit.
923
super(WeaveFile, self).__init__(name, access_mode, get_scope=get_scope,
924
allow_reserved=False)
925
self._transport = transport
926
self._filemode = filemode
928
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
929
except errors.NoSuchFile:
935
def _add_lines(self, version_id, parents, lines, parent_texts,
936
left_matching_blocks, nostore_sha, random_id, check_content):
937
"""Add a version and save the weave."""
938
self.check_not_reserved_id(version_id)
939
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
940
parent_texts, left_matching_blocks, nostore_sha, random_id,
945
def copy_to(self, name, transport):
946
"""See VersionedFile.copy_to()."""
947
# as we are all in memory always, just serialise to the new place.
949
write_weave_v5(self, sio)
951
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
954
"""Save the weave."""
955
self._check_write_ok()
957
write_weave_v5(self, sio)
959
bytes = sio.getvalue()
960
path = self._weave_name + WeaveFile.WEAVE_SUFFIX
962
self._transport.put_bytes(path, bytes, self._filemode)
963
except errors.NoSuchFile:
964
self._transport.mkdir(dirname(path))
965
self._transport.put_bytes(path, bytes, self._filemode)
969
"""See VersionedFile.get_suffixes()."""
970
return [WeaveFile.WEAVE_SUFFIX]
972
def insert_record_stream(self, stream):
973
super(WeaveFile, self).insert_record_stream(stream)
977
def _reweave(wa, wb, pb=None, msg=None):
978
"""Combine two weaves and return the result.
980
This works even if a revision R has different parents in
981
wa and wb. In the resulting weave all the parents are given.
983
This is done by just building up a new weave, maintaining ordering
984
of the versions in the two inputs. More efficient approaches
985
might be possible but it should only be necessary to do
986
this operation rarely, when a new previously ghost version is
989
:param pb: An optional progress bar, indicating how far done we are
990
:param msg: An optional message for the progress
994
queue_a = range(wa.num_versions())
995
queue_b = range(wb.num_versions())
996
# first determine combined parents of all versions
997
# map from version name -> all parent names
998
combined_parents = _reweave_parent_graphs(wa, wb)
999
mutter("combined parents: %r", combined_parents)
1000
order = tsort.topo_sort(combined_parents.iteritems())
1001
mutter("order to reweave: %r", order)
1006
for idx, name in enumerate(order):
1008
pb.update(msg, idx, len(order))
1009
if name in wa._name_map:
1010
lines = wa.get_lines(name)
1011
if name in wb._name_map:
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)
1022
lines = wb.get_lines(name)
1023
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1027
def _reweave_parent_graphs(wa, wb):
1028
"""Return combined parent ancestry for two weaves.
1030
Returned as a list of (version_name, set(parent_names))"""
1032
for weave in [wa, wb]:
1033
for idx, name in enumerate(weave._names):
1034
p = combined.setdefault(name, set())
1035
p.update(map(weave._idx_to_name, weave._parents[idx]))
634
print '%5d | %s' % (origin, text)
638
weave_info(argv[2], sys.stdout)
646
elif cmd == 'inclusions':
648
print ' '.join(map(str, w.inclusions([int(argv[3])])))
650
elif cmd == 'parents':
652
print ' '.join(map(str, w._v[int(argv[3])]))
660
v1, v2 = map(int, argv[3:5])
662
basis = w.inclusions([v1]).intersection(w.inclusions([v2]))
664
base_lines = list(w.mash_iter(basis))
665
a_lines = list(w.get(v1))
666
b_lines = list(w.get(v2))
668
from bzrlib.merge3 import Merge3
669
m3 = Merge3(base_lines, a_lines, b_lines)
671
name_a = 'version %d' % v1
672
name_b = 'version %d' % v2
673
sys.stdout.writelines(m3.merge_lines(name_a=name_a, name_b=name_b))
675
raise ValueError('unknown command %r' % cmd)
678
if __name__ == '__main__':
680
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
