1
# Copyright (C) 2005, 2009 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# 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
# Author: Martin Pool <mbp@canonical.com>
19
"""Weave - storage of related text file versions"""
21
from __future__ import absolute_import
23
# XXX: If we do weaves this way, will a merge still behave the same
24
# way if it's done in a different order? That's a pretty desirable
27
# 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
29
# binaries, perhaps by naively splitting on \n or perhaps using
30
# something like a rolling checksum.
32
# TODO: End marker for each version so we can stop reading?
34
# TODO: Check that no insertion occurs inside a deletion that was
35
# 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
72
from .lazy_import import lazy_import
73
lazy_import(globals(), """
74
from breezy import tsort
80
from .errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
81
RevisionAlreadyPresent,
83
UnavailableRepresentation,
85
from .osutils import dirname, sha, sha_strings, split_lines
86
from . import patiencediff
87
from .revision import NULL_REVISION
91
from .trace import mutter
92
from .versionedfile import (
99
from .weavefile import _read_weave_v5, write_weave_v5
102
class WeaveContentFactory(ContentFactory):
103
"""Content factory for streaming from weaves.
105
:seealso ContentFactory:
108
def __init__(self, version, weave):
109
"""Create a WeaveContentFactory for version from weave."""
110
ContentFactory.__init__(self)
111
self.sha1 = weave.get_sha1s([version])[version]
112
self.key = (version,)
113
parents = weave.get_parent_map([version])[version]
114
self.parents = tuple((parent,) for parent in parents)
115
self.storage_kind = 'fulltext'
118
def get_bytes_as(self, storage_kind):
119
if storage_kind == 'fulltext':
120
return self._weave.get_text(self.key[-1])
121
elif storage_kind == 'chunked':
122
return self._weave.get_lines(self.key[-1])
124
raise UnavailableRepresentation(self.key, storage_kind, 'fulltext')
127
class Weave(VersionedFile):
128
"""weave - versioned text file storage.
130
A Weave manages versions of line-based text files, keeping track
131
of the originating version for each line.
133
To clients the "lines" of the file are represented as a list of strings.
134
These strings will typically have terminal newline characters, but
135
this is not required. In particular files commonly do not have a newline
136
at the end of the file.
138
Texts can be identified in either of two ways:
140
* a nonnegative index number.
142
* a version-id string.
144
Typically the index number will be valid only inside this weave and
145
the version-id is used to reference it in the larger world.
147
The weave is represented as a list mixing edit instructions and
148
literal text. Each entry in _weave can be either a string (or
149
unicode), or a tuple. If a string, it means that the given line
150
should be output in the currently active revisions.
152
If a tuple, it gives a processing instruction saying in which
153
revisions the enclosed lines are active. The tuple has the form
154
(instruction, version).
156
The instruction can be '{' or '}' for an insertion block, and '['
157
and ']' for a deletion block respectively. The version is the
158
integer version index. There is no replace operator, only deletes
159
and inserts. For '}', the end of an insertion, there is no
160
version parameter because it always closes the most recently
165
* A later version can delete lines that were introduced by any
166
number of ancestor versions; this implies that deletion
167
instructions can span insertion blocks without regard to the
168
insertion block's nesting.
170
* Similarly, deletions need not be properly nested with regard to
171
each other, because they might have been generated by
172
independent revisions.
174
* Insertions are always made by inserting a new bracketed block
175
into a single point in the previous weave. This implies they
176
can nest but not overlap, and the nesting must always have later
177
insertions on the inside.
179
* It doesn't seem very useful to have an active insertion
180
inside an inactive insertion, but it might happen.
182
* Therefore, all instructions are always"considered"; that
183
is passed onto and off the stack. An outer inactive block
184
doesn't disable an inner block.
186
* Lines are enabled if the most recent enclosing insertion is
187
active and none of the enclosing deletions are active.
189
* There is no point having a deletion directly inside its own
190
insertion; you might as well just not write it. And there
191
should be no way to get an earlier version deleting a later
195
Text of the weave; list of control instruction tuples and strings.
198
List of parents, indexed by version number.
199
It is only necessary to store the minimal set of parents for
200
each version; the parent's parents are implied.
203
List of hex SHA-1 of each version.
206
List of symbolic names for each version. Each should be unique.
209
For each name, the version number.
212
Descriptive name of this weave; typically the filename if known.
216
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
217
'_weave_name', '_matcher', '_allow_reserved']
219
def __init__(self, weave_name=None, access_mode='w', matcher=None,
220
get_scope=None, allow_reserved=False):
223
:param get_scope: A callable that returns an opaque object to be used
224
for detecting when this weave goes out of scope (should stop
225
answering requests or allowing mutation).
227
super(Weave, self).__init__()
233
self._weave_name = weave_name
235
self._matcher = patiencediff.PatienceSequenceMatcher
237
self._matcher = matcher
238
if get_scope is None:
239
get_scope = lambda:None
240
self._get_scope = get_scope
241
self._scope = get_scope()
242
self._access_mode = access_mode
243
self._allow_reserved = allow_reserved
246
return "Weave(%r)" % self._weave_name
248
def _check_write_ok(self):
249
"""Is the versioned file marked as 'finished' ? Raise if it is."""
250
if self._get_scope() != self._scope:
251
raise errors.OutSideTransaction()
252
if self._access_mode != 'w':
253
raise errors.ReadOnlyObjectDirtiedError(self)
256
"""Return a deep copy of self.
258
The copy can be modified without affecting the original weave."""
260
other._weave = self._weave[:]
261
other._parents = self._parents[:]
262
other._sha1s = self._sha1s[:]
263
other._names = self._names[:]
264
other._name_map = self._name_map.copy()
265
other._weave_name = self._weave_name
268
def __eq__(self, other):
269
if not isinstance(other, Weave):
271
return self._parents == other._parents \
272
and self._weave == other._weave \
273
and self._sha1s == other._sha1s
275
def __ne__(self, other):
276
return not self.__eq__(other)
278
def _idx_to_name(self, version):
279
return self._names[version]
281
def _lookup(self, name):
282
"""Convert symbolic version name to index."""
283
if not self._allow_reserved:
284
self.check_not_reserved_id(name)
286
return self._name_map[name]
288
raise RevisionNotPresent(name, self._weave_name)
291
"""See VersionedFile.versions."""
292
return self._names[:]
294
def has_version(self, version_id):
295
"""See VersionedFile.has_version."""
296
return (version_id in self._name_map)
298
__contains__ = has_version
300
def get_record_stream(self, versions, ordering, include_delta_closure):
301
"""Get a stream of records for versions.
303
:param versions: The versions to include. Each version is a tuple
305
:param ordering: Either 'unordered' or 'topological'. A topologically
306
sorted stream has compression parents strictly before their
308
:param include_delta_closure: If True then the closure across any
309
compression parents will be included (in the opaque data).
310
:return: An iterator of ContentFactory objects, each of which is only
311
valid until the iterator is advanced.
313
versions = [version[-1] for version in versions]
314
if ordering == 'topological':
315
parents = self.get_parent_map(versions)
316
new_versions = tsort.topo_sort(parents)
317
new_versions.extend(set(versions).difference(set(parents)))
318
versions = new_versions
319
elif ordering == 'groupcompress':
320
parents = self.get_parent_map(versions)
321
new_versions = sort_groupcompress(parents)
322
new_versions.extend(set(versions).difference(set(parents)))
323
versions = new_versions
324
for version in versions:
326
yield WeaveContentFactory(version, self)
328
yield AbsentContentFactory((version,))
330
def get_parent_map(self, version_ids):
331
"""See VersionedFile.get_parent_map."""
333
for version_id in version_ids:
334
if version_id == NULL_REVISION:
339
map(self._idx_to_name,
340
self._parents[self._lookup(version_id)]))
341
except RevisionNotPresent:
343
result[version_id] = parents
346
def get_parents_with_ghosts(self, version_id):
347
raise NotImplementedError(self.get_parents_with_ghosts)
349
def insert_record_stream(self, stream):
350
"""Insert a record stream into this versioned file.
352
:param stream: A stream of records to insert.
354
:seealso VersionedFile.get_record_stream:
357
for record in stream:
358
# Raise an error when a record is missing.
359
if record.storage_kind == 'absent':
360
raise RevisionNotPresent([record.key[0]], self)
361
# adapt to non-tuple interface
362
parents = [parent[0] for parent in record.parents]
363
if (record.storage_kind == 'fulltext'
364
or record.storage_kind == 'chunked'):
365
self.add_lines(record.key[0], parents,
366
osutils.chunks_to_lines(record.get_bytes_as('chunked')))
368
adapter_key = record.storage_kind, 'fulltext'
370
adapter = adapters[adapter_key]
372
adapter_factory = adapter_registry.get(adapter_key)
373
adapter = adapter_factory(self)
374
adapters[adapter_key] = adapter
375
lines = split_lines(adapter.get_bytes(record))
377
self.add_lines(record.key[0], parents, lines)
378
except RevisionAlreadyPresent:
381
def _check_repeated_add(self, name, parents, text, sha1):
382
"""Check that a duplicated add is OK.
384
If it is, return the (old) index; otherwise raise an exception.
386
idx = self._lookup(name)
387
if sorted(self._parents[idx]) != sorted(parents) \
388
or sha1 != self._sha1s[idx]:
389
raise RevisionAlreadyPresent(name, self._weave_name)
392
def _add_lines(self, version_id, parents, lines, parent_texts,
393
left_matching_blocks, nostore_sha, random_id, check_content):
394
"""See VersionedFile.add_lines."""
395
idx = self._add(version_id, lines, list(map(self._lookup, parents)),
396
nostore_sha=nostore_sha)
397
return sha_strings(lines), sum(map(len, lines)), idx
399
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
400
"""Add a single text on top of the weave.
402
Returns the index number of the newly added version.
405
Symbolic name for this version.
406
(Typically the revision-id of the revision that added it.)
407
If None, a name will be allocated based on the hash. (sha1:SHAHASH)
410
List or set of direct parent version numbers.
413
Sequence of lines to be added in the new version.
415
:param nostore_sha: See VersionedFile.add_lines.
417
self._check_lines_not_unicode(lines)
418
self._check_lines_are_lines(lines)
420
sha1 = sha_strings(lines)
421
if sha1 == nostore_sha:
422
raise errors.ExistingContent
423
if version_id is None:
424
version_id = "sha1:" + sha1
425
if version_id in self._name_map:
426
return self._check_repeated_add(version_id, parents, lines, sha1)
428
self._check_versions(parents)
429
new_version = len(self._parents)
431
# if we abort after here the (in-memory) weave will be corrupt because only
432
# some fields are updated
433
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
434
# - Robert Collins 20060226
435
self._parents.append(parents[:])
436
self._sha1s.append(sha1)
437
self._names.append(version_id)
438
self._name_map[version_id] = new_version
442
# special case; adding with no parents revision; can do
443
# this more quickly by just appending unconditionally.
444
# even more specially, if we're adding an empty text we
445
# need do nothing at all.
447
self._weave.append(('{', new_version))
448
self._weave.extend(lines)
449
self._weave.append(('}', None))
452
if len(parents) == 1:
453
pv = list(parents)[0]
454
if sha1 == self._sha1s[pv]:
455
# special case: same as the single parent
459
ancestors = self._inclusions(parents)
463
# basis a list of (origin, lineno, line)
466
for origin, lineno, line in self._extract(ancestors):
467
basis_lineno.append(lineno)
468
basis_lines.append(line)
470
# another small special case: a merge, producing the same text
472
if lines == basis_lines:
475
# add a sentinel, because we can also match against the final line
476
basis_lineno.append(len(self._weave))
478
# XXX: which line of the weave should we really consider
479
# matches the end of the file? the current code says it's the
480
# last line of the weave?
482
#print 'basis_lines:', basis_lines
483
#print 'new_lines: ', lines
485
s = self._matcher(None, basis_lines, lines)
487
# offset gives the number of lines that have been inserted
488
# into the weave up to the current point; if the original edit instruction
489
# says to change line A then we actually change (A+offset)
492
for tag, i1, i2, j1, j2 in s.get_opcodes():
493
# i1,i2 are given in offsets within basis_lines; we need to map them
494
# back to offsets within the entire weave
495
#print 'raw match', tag, i1, i2, j1, j2
498
i1 = basis_lineno[i1]
499
i2 = basis_lineno[i2]
500
# the deletion and insertion are handled separately.
501
# first delete the region.
503
self._weave.insert(i1+offset, ('[', new_version))
504
self._weave.insert(i2+offset+1, (']', new_version))
508
# there may have been a deletion spanning up to
509
# i2; we want to insert after this region to make sure
510
# we don't destroy ourselves
512
self._weave[i:i] = ([('{', new_version)]
515
offset += 2 + (j2 - j1)
518
def _inclusions(self, versions):
519
"""Return set of all ancestors of given version(s)."""
520
if not len(versions):
523
for v in range(max(versions), 0, -1):
525
# include all its parents
526
i.update(self._parents[v])
529
def get_ancestry(self, version_ids, topo_sorted=True):
530
"""See VersionedFile.get_ancestry."""
531
if isinstance(version_ids, basestring):
532
version_ids = [version_ids]
533
i = self._inclusions([self._lookup(v) for v in version_ids])
534
return [self._idx_to_name(v) for v in i]
536
def _check_lines(self, text):
537
if not isinstance(text, list):
538
raise ValueError("text should be a list, not %s" % type(text))
541
if not isinstance(l, basestring):
542
raise ValueError("text line should be a string or unicode, not %s"
547
def _check_versions(self, indexes):
548
"""Check everything in the sequence of indexes is valid"""
553
raise IndexError("invalid version number %r" % i)
555
def _compatible_parents(self, my_parents, other_parents):
556
"""During join check that other_parents are joinable with my_parents.
558
Joinable is defined as 'is a subset of' - supersets may require
559
regeneration of diffs, but subsets do not.
561
return len(other_parents.difference(my_parents)) == 0
563
def annotate(self, version_id):
564
"""Return a list of (version-id, line) tuples for version_id.
566
The index indicates when the line originated in the weave."""
567
incls = [self._lookup(version_id)]
568
return [(self._idx_to_name(origin), text) for origin, lineno, text in
569
self._extract(incls)]
571
def iter_lines_added_or_present_in_versions(self, version_ids=None,
573
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
574
if version_ids is None:
575
version_ids = self.versions()
576
version_ids = set(version_ids)
577
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
578
if inserted not in version_ids: continue
580
yield line + '\n', inserted
584
def _walk_internal(self, version_ids=None):
585
"""Helper method for weave actions."""
590
lineno = 0 # line of weave, 0-based
592
for l in self._weave:
593
if l.__class__ == tuple:
597
istack.append(self._names[v])
601
dset.add(self._names[v])
603
dset.remove(self._names[v])
605
raise WeaveFormatError('unexpected instruction %r' % v)
607
yield lineno, istack[-1], frozenset(dset), l
611
raise WeaveFormatError("unclosed insertion blocks "
612
"at end of weave: %s" % istack)
614
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
617
def plan_merge(self, ver_a, ver_b):
618
"""Return pseudo-annotation indicating how the two versions merge.
620
This is computed between versions a and b and their common
623
Weave lines present in none of them are skipped entirely.
625
inc_a = set(self.get_ancestry([ver_a]))
626
inc_b = set(self.get_ancestry([ver_b]))
627
inc_c = inc_a & inc_b
629
for lineno, insert, deleteset, line in self._walk_internal([ver_a, ver_b]):
630
if deleteset & inc_c:
631
# killed in parent; can't be in either a or b
632
# not relevant to our work
633
yield 'killed-base', line
634
elif insert in inc_c:
635
# was inserted in base
636
killed_a = bool(deleteset & inc_a)
637
killed_b = bool(deleteset & inc_b)
638
if killed_a and killed_b:
639
yield 'killed-both', line
641
yield 'killed-a', line
643
yield 'killed-b', line
645
yield 'unchanged', line
646
elif insert in inc_a:
647
if deleteset & inc_a:
648
yield 'ghost-a', line
652
elif insert in inc_b:
653
if deleteset & inc_b:
654
yield 'ghost-b', line
658
# not in either revision
659
yield 'irrelevant', line
661
def _extract(self, versions):
662
"""Yield annotation of lines in included set.
664
Yields a sequence of tuples (origin, lineno, text), where
665
origin is the origin version, lineno the index in the weave,
666
and text the text of the line.
668
The set typically but not necessarily corresponds to a version.
671
if not isinstance(i, int):
674
included = self._inclusions(versions)
680
lineno = 0 # line of weave, 0-based
686
WFE = WeaveFormatError
689
# 449 0 4474.6820 2356.5590 breezy.weave:556(_extract)
690
# +285282 0 1676.8040 1676.8040 +<isinstance>
691
# 1.6 seconds in 'isinstance'.
692
# changing the first isinstance:
693
# 449 0 2814.2660 1577.1760 breezy.weave:556(_extract)
694
# +140414 0 762.8050 762.8050 +<isinstance>
695
# note that the inline time actually dropped (less function calls)
696
# and total processing time was halved.
697
# we're still spending ~1/4 of the method in isinstance though.
698
# so lets hard code the acceptable string classes we expect:
699
# 449 0 1202.9420 786.2930 breezy.weave:556(_extract)
700
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
702
# yay, down to ~1/4 the initial extract time, and our inline time
703
# has shrunk again, with isinstance no longer dominating.
704
# tweaking the stack inclusion test to use a set gives:
705
# 449 0 1122.8030 713.0080 breezy.weave:556(_extract)
706
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
708
# - a 5% win, or possibly just noise. However with large istacks that
709
# 'in' test could dominate, so I'm leaving this change in place -
710
# when its fast enough to consider profiling big datasets we can review.
715
for l in self._weave:
716
if l.__class__ == tuple:
723
iset.remove(istack.pop())
731
raise AssertionError()
734
isactive = (not dset) and istack and (istack[-1] in included)
736
result.append((istack[-1], lineno, l))
739
raise WeaveFormatError("unclosed insertion blocks "
740
"at end of weave: %s" % istack)
742
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
746
def _maybe_lookup(self, name_or_index):
747
"""Convert possible symbolic name to index, or pass through indexes.
751
# GZ 2017-04-01: This used to check for long as well, but I don't think
752
# there are python implementations with sys.maxsize > sys.maxint
753
if isinstance(name_or_index, int):
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)
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.
851
def _imported_parents(self, other, other_idx):
852
"""Return list of parents in self corresponding to indexes in other."""
854
for parent_idx in other._parents[other_idx]:
855
parent_name = other._names[parent_idx]
856
if parent_name not in self._name_map:
857
# should not be possible
858
raise WeaveError("missing parent {%s} of {%s} in %r"
859
% (parent_name, other._name_map[other_idx], self))
860
new_parents.append(self._name_map[parent_name])
863
def _check_version_consistent(self, other, other_idx, name):
864
"""Check if a version in consistent in this and other.
866
To be consistent it must have:
869
* the same direct parents (by name, not index, and disregarding
872
If present & correct return True;
873
if not present in self return False;
874
if inconsistent raise error."""
875
this_idx = self._name_map.get(name, -1)
877
if self._sha1s[this_idx] != other._sha1s[other_idx]:
878
raise errors.WeaveTextDiffers(name, self, other)
879
self_parents = self._parents[this_idx]
880
other_parents = other._parents[other_idx]
881
n1 = {self._names[i] for i in self_parents}
882
n2 = {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))
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(BytesIO(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
989
# first determine combined parents of all versions
990
# map from version name -> all parent names
991
combined_parents = _reweave_parent_graphs(wa, wb)
992
mutter("combined parents: %r", combined_parents)
993
order = tsort.topo_sort(combined_parents.items())
994
mutter("order to reweave: %r", order)
999
for idx, name in enumerate(order):
1001
pb.update(msg, idx, len(order))
1002
if name in wa._name_map:
1003
lines = wa.get_lines(name)
1004
if name in wb._name_map:
1005
lines_b = wb.get_lines(name)
1006
if lines != lines_b:
1007
mutter('Weaves differ on content. rev_id {%s}', name)
1008
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1010
lines = list(difflib.unified_diff(lines, lines_b,
1011
wa._weave_name, wb._weave_name))
1012
mutter('lines:\n%s', ''.join(lines))
1013
raise errors.WeaveTextDiffers(name, wa, wb)
1015
lines = wb.get_lines(name)
1016
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1020
def _reweave_parent_graphs(wa, wb):
1021
"""Return combined parent ancestry for two weaves.
1023
Returned as a list of (version_name, set(parent_names))"""
1025
for weave in [wa, wb]:
1026
for idx, name in enumerate(weave._names):
1027
p = combined.setdefault(name, set())
1028
p.update(map(weave._idx_to_name, weave._parents[idx]))
added
removed
breezy/weave.py
