3
# Copyright (C) 2005 Canonical Ltd
5
# This program is free software; you can redistribute it and/or modify
6
# it under the terms of the GNU General Public License as published by
7
# the Free Software Foundation; either version 2 of the License, or
8
# (at your option) any later version.
10
# This program is distributed in the hope that it will be useful,
11
# but WITHOUT ANY WARRANTY; without even the implied warranty of
12
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13
# GNU General Public License for more details.
15
# You should have received a copy of the GNU General Public License
16
# along with this program; if not, write to the Free Software
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
19
# Author: Martin Pool <mbp@canonical.com>
22
"""Weave - storage of related text file versions"""
25
# XXX: If we do weaves this way, will a merge still behave the same
26
# way if it's done in a different order? That's a pretty desirable
29
# TODO: Nothing here so far assumes the lines are really \n newlines,
30
# rather than being split up in some other way. We could accommodate
31
# binaries, perhaps by naively splitting on \n or perhaps using
32
# something like a rolling checksum.
34
# TODO: End marker for each version so we can stop reading?
36
# TODO: Check that no insertion occurs inside a deletion that was
37
# active in the version of the insertion.
39
# TODO: In addition to the SHA-1 check, perhaps have some code that
40
# checks structural constraints of the weave: ie that insertions are
41
# properly nested, that there is no text outside of an insertion, that
42
# insertions or deletions are not repeated, etc.
44
# TODO: Parallel-extract that passes back each line along with a
45
# description of which revisions include it. Nice for checking all
46
# shas or calculating stats in parallel.
48
# TODO: Using a single _extract routine and then processing the output
49
# is probably inefficient. It's simple enough that we can afford to
50
# have slight specializations for different ways its used: annotate,
51
# basis for add, get, etc.
53
# TODO: Probably the API should work only in names to hide the integer
54
# indexes from the user.
56
# TODO: Is there any potential performance win by having an add()
57
# variant that is passed a pre-cooked version of the single basis
60
# TODO: Reweave can possibly be made faster by remembering diffs
61
# where the basis and destination are unchanged.
63
# FIXME: Sometimes we will be given a parents list for a revision
64
# that includes some redundant parents (i.e. already a parent of
65
# something in the list.) We should eliminate them. This can
66
# be done fairly efficiently because the sequence numbers constrain
67
# the possible relationships.
69
# FIXME: the conflict markers should be *7* characters
72
from cStringIO import StringIO
81
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
82
RevisionAlreadyPresent,
84
UnavailableRepresentation,
85
WeaveRevisionAlreadyPresent,
86
WeaveRevisionNotPresent,
88
import bzrlib.errors as errors
89
from bzrlib.osutils import sha_strings, split_lines
90
import bzrlib.patiencediff
91
from bzrlib.symbol_versioning import *
92
from bzrlib.trace import mutter
93
from bzrlib.tsort import topo_sort
94
from bzrlib.versionedfile import (
101
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
104
class WeaveContentFactory(ContentFactory):
105
"""Content factory for streaming from weaves.
107
:seealso ContentFactory:
110
def __init__(self, version, weave):
111
"""Create a WeaveContentFactory for version from weave."""
112
ContentFactory.__init__(self)
113
self.sha1 = weave.get_sha1s([version])[0]
114
self.key = (version,)
115
parents = weave.get_parent_map([version])[version]
116
self.parents = tuple((parent,) for parent in parents)
117
self.storage_kind = 'fulltext'
120
def get_bytes_as(self, storage_kind):
121
if storage_kind == 'fulltext':
122
return self._weave.get_text(self.key[0])
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']
219
def __init__(self, weave_name=None, access_mode='w', matcher=None, get_scope=None):
222
:param get_scope: A callable that returns an opaque object to be used
223
for detecting when this weave goes out of scope (should stop
224
answering requests or allowing mutation).
226
super(Weave, self).__init__(access_mode)
232
self._weave_name = weave_name
234
self._matcher = bzrlib.patiencediff.PatienceSequenceMatcher
236
self._matcher = matcher
237
if get_scope is None:
238
get_scope = lambda:None
239
self._get_scope = get_scope
240
self._scope = get_scope()
241
self._access_mode = access_mode
244
return "Weave(%r)" % self._weave_name
246
def _check_write_ok(self):
247
"""Is the versioned file marked as 'finished' ? Raise if it is."""
248
if self._get_scope() != self._scope:
249
raise errors.OutSideTransaction()
250
if self._access_mode != 'w':
251
raise errors.ReadOnlyObjectDirtiedError(self)
254
"""Return a deep copy of self.
256
The copy can be modified without affecting the original weave."""
258
other._weave = self._weave[:]
259
other._parents = self._parents[:]
260
other._sha1s = self._sha1s[:]
261
other._names = self._names[:]
262
other._name_map = self._name_map.copy()
263
other._weave_name = self._weave_name
266
def __eq__(self, other):
267
if not isinstance(other, Weave):
269
return self._parents == other._parents \
270
and self._weave == other._weave \
271
and self._sha1s == other._sha1s
273
def __ne__(self, other):
274
return not self.__eq__(other)
276
def _idx_to_name(self, version):
277
return self._names[version]
279
def _lookup(self, name):
280
"""Convert symbolic version name to index."""
281
self.check_not_reserved_id(name)
283
return self._name_map[name]
285
raise RevisionNotPresent(name, self._weave_name)
288
"""See VersionedFile.versions."""
289
return self._names[:]
291
def has_version(self, version_id):
292
"""See VersionedFile.has_version."""
293
return (version_id in self._name_map)
295
__contains__ = has_version
297
def get_record_stream(self, versions, ordering, include_delta_closure):
298
"""Get a stream of records for versions.
300
:param versions: The versions to include. Each version is a tuple
302
:param ordering: Either 'unordered' or 'topological'. A topologically
303
sorted stream has compression parents strictly before their
305
:param include_delta_closure: If True then the closure across any
306
compression parents will be included (in the opaque data).
307
:return: An iterator of ContentFactory objects, each of which is only
308
valid until the iterator is advanced.
310
if ordering == 'topological':
311
parents = self.get_parent_map(versions)
312
new_versions = topo_sort(parents)
313
new_versions.extend(set(versions).difference(set(parents)))
314
versions = new_versions
315
for version in versions:
317
yield WeaveContentFactory(version, self)
319
yield AbsentContentFactory((version,))
321
def get_parent_map(self, version_ids):
322
"""See VersionedFile.get_parent_map."""
324
for version_id in version_ids:
326
result[version_id] = tuple(
327
map(self._idx_to_name, self._parents[self._lookup(version_id)]))
328
except RevisionNotPresent:
332
def get_parents_with_ghosts(self, version_id):
333
raise NotImplementedError(self.get_parents_with_ghosts)
335
def insert_record_stream(self, stream):
336
"""Insert a record stream into this versioned file.
338
:param stream: A stream of records to insert.
340
:seealso VersionedFile.get_record_stream:
343
for record in stream:
344
# Raise an error when a record is missing.
345
if record.storage_kind == 'absent':
346
raise RevisionNotPresent([record.key[0]], self)
347
# adapt to non-tuple interface
348
parents = [parent[0] for parent in record.parents]
349
if record.storage_kind == 'fulltext':
350
self.add_lines(record.key[0], parents,
351
split_lines(record.get_bytes_as('fulltext')))
353
adapter_key = record.storage_kind, 'fulltext'
355
adapter = adapters[adapter_key]
357
adapter_factory = adapter_registry.get(adapter_key)
358
adapter = adapter_factory(self)
359
adapters[adapter_key] = adapter
360
lines = split_lines(adapter.get_bytes(
361
record, record.get_bytes_as(record.storage_kind)))
363
self.add_lines(record.key[0], parents, lines)
364
except RevisionAlreadyPresent:
367
def _check_repeated_add(self, name, parents, text, sha1):
368
"""Check that a duplicated add is OK.
370
If it is, return the (old) index; otherwise raise an exception.
372
idx = self._lookup(name)
373
if sorted(self._parents[idx]) != sorted(parents) \
374
or sha1 != self._sha1s[idx]:
375
raise RevisionAlreadyPresent(name, self._weave_name)
378
def _add_lines(self, version_id, parents, lines, parent_texts,
379
left_matching_blocks, nostore_sha, random_id, check_content):
380
"""See VersionedFile.add_lines."""
381
idx = self._add(version_id, lines, map(self._lookup, parents),
382
nostore_sha=nostore_sha)
383
return sha_strings(lines), sum(map(len, lines)), idx
385
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
386
"""Add a single text on top of the weave.
388
Returns the index number of the newly added version.
391
Symbolic name for this version.
392
(Typically the revision-id of the revision that added it.)
395
List or set of direct parent version numbers.
398
Sequence of lines to be added in the new version.
400
:param nostore_sha: See VersionedFile.add_lines.
402
self._check_lines_not_unicode(lines)
403
self._check_lines_are_lines(lines)
405
sha1 = sha_strings(lines)
406
if sha1 == nostore_sha:
407
raise errors.ExistingContent
408
if version_id in self._name_map:
409
return self._check_repeated_add(version_id, parents, lines, sha1)
411
self._check_versions(parents)
412
## self._check_lines(lines)
413
new_version = len(self._parents)
415
# if we abort after here the (in-memory) weave will be corrupt because only
416
# some fields are updated
417
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
418
# - Robert Collins 20060226
419
self._parents.append(parents[:])
420
self._sha1s.append(sha1)
421
self._names.append(version_id)
422
self._name_map[version_id] = new_version
426
# special case; adding with no parents revision; can do
427
# this more quickly by just appending unconditionally.
428
# even more specially, if we're adding an empty text we
429
# need do nothing at all.
431
self._weave.append(('{', new_version))
432
self._weave.extend(lines)
433
self._weave.append(('}', None))
436
if len(parents) == 1:
437
pv = list(parents)[0]
438
if sha1 == self._sha1s[pv]:
439
# special case: same as the single parent
443
ancestors = self._inclusions(parents)
447
# basis a list of (origin, lineno, line)
450
for origin, lineno, line in self._extract(ancestors):
451
basis_lineno.append(lineno)
452
basis_lines.append(line)
454
# another small special case: a merge, producing the same text
456
if lines == basis_lines:
459
# add a sentinel, because we can also match against the final line
460
basis_lineno.append(len(self._weave))
462
# XXX: which line of the weave should we really consider
463
# matches the end of the file? the current code says it's the
464
# last line of the weave?
466
#print 'basis_lines:', basis_lines
467
#print 'new_lines: ', lines
469
s = self._matcher(None, basis_lines, lines)
471
# offset gives the number of lines that have been inserted
472
# into the weave up to the current point; if the original edit instruction
473
# says to change line A then we actually change (A+offset)
476
for tag, i1, i2, j1, j2 in s.get_opcodes():
477
# i1,i2 are given in offsets within basis_lines; we need to map them
478
# back to offsets within the entire weave
479
#print 'raw match', tag, i1, i2, j1, j2
482
i1 = basis_lineno[i1]
483
i2 = basis_lineno[i2]
484
# the deletion and insertion are handled separately.
485
# first delete the region.
487
self._weave.insert(i1+offset, ('[', new_version))
488
self._weave.insert(i2+offset+1, (']', new_version))
492
# there may have been a deletion spanning up to
493
# i2; we want to insert after this region to make sure
494
# we don't destroy ourselves
496
self._weave[i:i] = ([('{', new_version)]
499
offset += 2 + (j2 - j1)
502
def _inclusions(self, versions):
503
"""Return set of all ancestors of given version(s)."""
504
if not len(versions):
507
for v in xrange(max(versions), 0, -1):
509
# include all its parents
510
i.update(self._parents[v])
512
## except IndexError:
513
## raise ValueError("version %d not present in weave" % v)
515
def get_ancestry(self, version_ids, topo_sorted=True):
516
"""See VersionedFile.get_ancestry."""
517
if isinstance(version_ids, basestring):
518
version_ids = [version_ids]
519
i = self._inclusions([self._lookup(v) for v in version_ids])
520
return [self._idx_to_name(v) for v in i]
522
def _check_lines(self, text):
523
if not isinstance(text, list):
524
raise ValueError("text should be a list, not %s" % type(text))
527
if not isinstance(l, basestring):
528
raise ValueError("text line should be a string or unicode, not %s"
533
def _check_versions(self, indexes):
534
"""Check everything in the sequence of indexes is valid"""
539
raise IndexError("invalid version number %r" % i)
541
def _compatible_parents(self, my_parents, other_parents):
542
"""During join check that other_parents are joinable with my_parents.
544
Joinable is defined as 'is a subset of' - supersets may require
545
regeneration of diffs, but subsets do not.
547
return len(other_parents.difference(my_parents)) == 0
549
def annotate(self, version_id):
550
"""Return a list of (version-id, line) tuples for version_id.
552
The index indicates when the line originated in the weave."""
553
incls = [self._lookup(version_id)]
554
return [(self._idx_to_name(origin), text) for origin, lineno, text in
555
self._extract(incls)]
557
def iter_lines_added_or_present_in_versions(self, version_ids=None,
559
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
560
if version_ids is None:
561
version_ids = self.versions()
562
version_ids = set(version_ids)
563
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
564
# if inserted not in version_ids then it was inserted before the
565
# versions we care about, but because weaves cannot represent ghosts
566
# properly, we do not filter down to that
567
# if inserted not in version_ids: continue
569
yield line + '\n', inserted
573
def _walk_internal(self, version_ids=None):
574
"""Helper method for weave actions."""
579
lineno = 0 # line of weave, 0-based
581
for l in self._weave:
582
if l.__class__ == tuple:
586
istack.append(self._names[v])
590
dset.add(self._names[v])
592
dset.remove(self._names[v])
594
raise WeaveFormatError('unexpected instruction %r' % v)
596
yield lineno, istack[-1], frozenset(dset), l
600
raise WeaveFormatError("unclosed insertion blocks "
601
"at end of weave: %s" % istack)
603
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
606
def plan_merge(self, ver_a, ver_b):
607
"""Return pseudo-annotation indicating how the two versions merge.
609
This is computed between versions a and b and their common
612
Weave lines present in none of them are skipped entirely.
614
inc_a = set(self.get_ancestry([ver_a]))
615
inc_b = set(self.get_ancestry([ver_b]))
616
inc_c = inc_a & inc_b
618
for lineno, insert, deleteset, line in self._walk_internal([ver_a, ver_b]):
619
if deleteset & inc_c:
620
# killed in parent; can't be in either a or b
621
# not relevant to our work
622
yield 'killed-base', line
623
elif insert in inc_c:
624
# was inserted in base
625
killed_a = bool(deleteset & inc_a)
626
killed_b = bool(deleteset & inc_b)
627
if killed_a and killed_b:
628
yield 'killed-both', line
630
yield 'killed-a', line
632
yield 'killed-b', line
634
yield 'unchanged', line
635
elif insert in inc_a:
636
if deleteset & inc_a:
637
yield 'ghost-a', line
641
elif insert in inc_b:
642
if deleteset & inc_b:
643
yield 'ghost-b', line
647
# not in either revision
648
yield 'irrelevant', line
650
yield 'unchanged', '' # terminator
652
def _extract(self, versions):
653
"""Yield annotation of lines in included set.
655
Yields a sequence of tuples (origin, lineno, text), where
656
origin is the origin version, lineno the index in the weave,
657
and text the text of the line.
659
The set typically but not necessarily corresponds to a version.
662
if not isinstance(i, int):
665
included = self._inclusions(versions)
671
lineno = 0 # line of weave, 0-based
677
WFE = WeaveFormatError
680
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
681
# +285282 0 1676.8040 1676.8040 +<isinstance>
682
# 1.6 seconds in 'isinstance'.
683
# changing the first isinstance:
684
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
685
# +140414 0 762.8050 762.8050 +<isinstance>
686
# note that the inline time actually dropped (less function calls)
687
# and total processing time was halved.
688
# we're still spending ~1/4 of the method in isinstance though.
689
# so lets hard code the acceptable string classes we expect:
690
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
691
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
693
# yay, down to ~1/4 the initial extract time, and our inline time
694
# has shrunk again, with isinstance no longer dominating.
695
# tweaking the stack inclusion test to use a set gives:
696
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
697
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
699
# - a 5% win, or possibly just noise. However with large istacks that
700
# 'in' test could dominate, so I'm leaving this change in place -
701
# when its fast enough to consider profiling big datasets we can review.
706
for l in self._weave:
707
if l.__class__ == tuple:
714
iset.remove(istack.pop())
722
raise AssertionError()
725
isactive = (not dset) and istack and (istack[-1] in included)
727
result.append((istack[-1], lineno, l))
730
raise WeaveFormatError("unclosed insertion blocks "
731
"at end of weave: %s" % istack)
733
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
737
def _maybe_lookup(self, name_or_index):
738
"""Convert possible symbolic name to index, or pass through indexes.
742
if isinstance(name_or_index, (int, long)):
745
return self._lookup(name_or_index)
747
def get_lines(self, version_id):
748
"""See VersionedFile.get_lines()."""
749
int_index = self._maybe_lookup(version_id)
750
result = [line for (origin, lineno, line) in self._extract([int_index])]
751
expected_sha1 = self._sha1s[int_index]
752
measured_sha1 = sha_strings(result)
753
if measured_sha1 != expected_sha1:
754
raise errors.WeaveInvalidChecksum(
755
'file %s, revision %s, expected: %s, measured %s'
756
% (self._weave_name, version_id,
757
expected_sha1, measured_sha1))
760
def get_sha1s(self, version_ids):
761
"""See VersionedFile.get_sha1s()."""
762
return [self._sha1s[self._lookup(v)] for v in version_ids]
764
def num_versions(self):
765
"""How many versions are in this weave?"""
766
l = len(self._parents)
769
__len__ = num_versions
771
def check(self, progress_bar=None):
772
# TODO evaluate performance hit of using string sets in this routine.
773
# TODO: check no circular inclusions
774
# TODO: create a nested progress bar
775
for version in range(self.num_versions()):
776
inclusions = list(self._parents[version])
779
if inclusions[-1] >= version:
780
raise WeaveFormatError("invalid included version %d for index %d"
781
% (inclusions[-1], version))
783
# try extracting all versions; parallel extraction is used
784
nv = self.num_versions()
789
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
790
# The problem is that set membership is much more expensive
791
name = self._idx_to_name(i)
792
sha1s[name] = sha.new()
794
new_inc = set([name])
795
for p in self._parents[i]:
796
new_inc.update(inclusions[self._idx_to_name(p)])
798
if set(new_inc) != set(self.get_ancestry(name)):
799
raise AssertionError(
801
% (set(new_inc), set(self.get_ancestry(name))))
802
inclusions[name] = new_inc
804
nlines = len(self._weave)
806
update_text = 'checking weave'
808
short_name = os.path.basename(self._weave_name)
809
update_text = 'checking %s' % (short_name,)
810
update_text = update_text[:25]
812
for lineno, insert, deleteset, line in self._walk_internal():
814
progress_bar.update(update_text, lineno, nlines)
816
for name, name_inclusions in inclusions.items():
817
# The active inclusion must be an ancestor,
818
# and no ancestors must have deleted this line,
819
# because we don't support resurrection.
820
if (insert in name_inclusions) and not (deleteset & name_inclusions):
821
sha1s[name].update(line)
824
version = self._idx_to_name(i)
825
hd = sha1s[version].hexdigest()
826
expected = self._sha1s[i]
828
raise errors.WeaveInvalidChecksum(
829
"mismatched sha1 for version %s: "
830
"got %s, expected %s"
831
% (version, hd, expected))
833
# TODO: check insertions are properly nested, that there are
834
# no lines outside of insertion blocks, that deletions are
835
# properly paired, etc.
837
def _join(self, other, pb, msg, version_ids, ignore_missing):
838
"""Worker routine for join()."""
839
if not other.versions():
840
return # nothing to update, easy
843
# versions is never none, InterWeave checks this.
846
# two loops so that we do not change ourselves before verifying it
848
# work through in index order to make sure we get all dependencies
851
# get the selected versions only that are in other.versions.
852
version_ids = set(other.versions()).intersection(set(version_ids))
853
# pull in the referenced graph.
854
version_ids = other.get_ancestry(version_ids)
855
pending_parents = other.get_parent_map(version_ids)
856
pending_graph = pending_parents.items()
857
if len(pending_graph) != len(version_ids):
858
raise RevisionNotPresent(
859
set(version_ids) - set(pending_parents.keys()), self)
860
for name in topo_sort(pending_graph):
861
other_idx = other._name_map[name]
862
# returns True if we have it, False if we need it.
863
if not self._check_version_consistent(other, other_idx, name):
864
names_to_join.append((other_idx, name))
872
for other_idx, name in names_to_join:
873
# TODO: If all the parents of the other version are already
874
# present then we can avoid some work by just taking the delta
875
# and adjusting the offsets.
876
new_parents = self._imported_parents(other, other_idx)
877
sha1 = other._sha1s[other_idx]
882
pb.update(msg, merged, len(names_to_join))
884
lines = other.get_lines(other_idx)
885
self._add(name, lines, new_parents, sha1)
887
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
888
merged, processed, self._weave_name, time.time()-time0))
890
def _imported_parents(self, other, other_idx):
891
"""Return list of parents in self corresponding to indexes in other."""
893
for parent_idx in other._parents[other_idx]:
894
parent_name = other._names[parent_idx]
895
if parent_name not in self._name_map:
896
# should not be possible
897
raise WeaveError("missing parent {%s} of {%s} in %r"
898
% (parent_name, other._name_map[other_idx], self))
899
new_parents.append(self._name_map[parent_name])
902
def _check_version_consistent(self, other, other_idx, name):
903
"""Check if a version in consistent in this and other.
905
To be consistent it must have:
908
* the same direct parents (by name, not index, and disregarding
911
If present & correct return True;
912
if not present in self return False;
913
if inconsistent raise error."""
914
this_idx = self._name_map.get(name, -1)
916
if self._sha1s[this_idx] != other._sha1s[other_idx]:
917
raise errors.WeaveTextDiffers(name, self, other)
918
self_parents = self._parents[this_idx]
919
other_parents = other._parents[other_idx]
920
n1 = set([self._names[i] for i in self_parents])
921
n2 = set([other._names[i] for i in other_parents])
922
if not self._compatible_parents(n1, n2):
923
raise WeaveParentMismatch("inconsistent parents "
924
"for version {%s}: %s vs %s" % (name, n1, n2))
930
def _reweave(self, other, pb, msg):
931
"""Reweave self with other - internal helper for join().
933
:param other: The other weave to merge
934
:param pb: An optional progress bar, indicating how far done we are
935
:param msg: An optional message for the progress
937
new_weave = _reweave(self, other, pb=pb, msg=msg)
938
self._copy_weave_content(new_weave)
940
def _copy_weave_content(self, otherweave):
941
"""adsorb the content from otherweave."""
942
for attr in self.__slots__:
943
if attr != '_weave_name':
944
setattr(self, attr, copy(getattr(otherweave, attr)))
947
class WeaveFile(Weave):
948
"""A WeaveFile represents a Weave on disk and writes on change."""
950
WEAVE_SUFFIX = '.weave'
952
def __init__(self, name, transport, filemode=None, create=False, access_mode='w', get_scope=None):
953
"""Create a WeaveFile.
955
:param create: If not True, only open an existing knit.
957
super(WeaveFile, self).__init__(name, access_mode, get_scope=get_scope)
958
self._transport = transport
959
self._filemode = filemode
961
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
962
except errors.NoSuchFile:
968
def _add_lines(self, version_id, parents, lines, parent_texts,
969
left_matching_blocks, nostore_sha, random_id, check_content):
970
"""Add a version and save the weave."""
971
self.check_not_reserved_id(version_id)
972
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
973
parent_texts, left_matching_blocks, nostore_sha, random_id,
978
def copy_to(self, name, transport):
979
"""See VersionedFile.copy_to()."""
980
# as we are all in memory always, just serialise to the new place.
982
write_weave_v5(self, sio)
984
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
987
"""Save the weave."""
988
self._check_write_ok()
990
write_weave_v5(self, sio)
992
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
998
"""See VersionedFile.get_suffixes()."""
999
return [WeaveFile.WEAVE_SUFFIX]
1001
def insert_record_stream(self, stream):
1002
super(WeaveFile, self).insert_record_stream(stream)
1005
@deprecated_method(one_five)
1006
def join(self, other, pb=None, msg=None, version_ids=None,
1007
ignore_missing=False):
1008
"""Join other into self and save."""
1009
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1013
def _reweave(wa, wb, pb=None, msg=None):
1014
"""Combine two weaves and return the result.
1016
This works even if a revision R has different parents in
1017
wa and wb. In the resulting weave all the parents are given.
1019
This is done by just building up a new weave, maintaining ordering
1020
of the versions in the two inputs. More efficient approaches
1021
might be possible but it should only be necessary to do
1022
this operation rarely, when a new previously ghost version is
1025
:param pb: An optional progress bar, indicating how far done we are
1026
:param msg: An optional message for the progress
1030
queue_a = range(wa.num_versions())
1031
queue_b = range(wb.num_versions())
1032
# first determine combined parents of all versions
1033
# map from version name -> all parent names
1034
combined_parents = _reweave_parent_graphs(wa, wb)
1035
mutter("combined parents: %r", combined_parents)
1036
order = topo_sort(combined_parents.iteritems())
1037
mutter("order to reweave: %r", order)
1042
for idx, name in enumerate(order):
1044
pb.update(msg, idx, len(order))
1045
if name in wa._name_map:
1046
lines = wa.get_lines(name)
1047
if name in wb._name_map:
1048
lines_b = wb.get_lines(name)
1049
if lines != lines_b:
1050
mutter('Weaves differ on content. rev_id {%s}', name)
1051
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1053
lines = list(difflib.unified_diff(lines, lines_b,
1054
wa._weave_name, wb._weave_name))
1055
mutter('lines:\n%s', ''.join(lines))
1056
raise errors.WeaveTextDiffers(name, wa, wb)
1058
lines = wb.get_lines(name)
1059
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1062
def _reweave_parent_graphs(wa, wb):
1063
"""Return combined parent ancestry for two weaves.
1065
Returned as a list of (version_name, set(parent_names))"""
1067
for weave in [wa, wb]:
1068
for idx, name in enumerate(weave._names):
1069
p = combined.setdefault(name, set())
1070
p.update(map(weave._idx_to_name, weave._parents[idx]))
1075
"""Show the weave's table-of-contents"""
1076
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1077
for i in (6, 50, 10, 10):
1080
for i in range(w.num_versions()):
1083
parent_str = ' '.join(map(str, w._parents[i]))
1084
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1088
def weave_stats(weave_file, pb):
1089
from bzrlib.weavefile import read_weave
1091
wf = file(weave_file, 'rb')
1093
# FIXME: doesn't work on pipes
1094
weave_size = wf.tell()
1098
for i in range(vers):
1099
pb.update('checking sizes', i, vers)
1100
for origin, lineno, line in w._extract([i]):
1105
print 'versions %9d' % vers
1106
print 'weave file %9d bytes' % weave_size
1107
print 'total contents %9d bytes' % total
1108
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1111
print 'average size %9d bytes' % avg
1112
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1116
print """bzr weave tool
1118
Experimental tool for weave algorithm.
1121
weave init WEAVEFILE
1122
Create an empty weave file
1123
weave get WEAVEFILE VERSION
1124
Write out specified version.
1125
weave check WEAVEFILE
1126
Check consistency of all versions.
1128
Display table of contents.
1129
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1130
Add NEWTEXT, with specified parent versions.
1131
weave annotate WEAVEFILE VERSION
1132
Display origin of each line.
1133
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1134
Auto-merge two versions and display conflicts.
1135
weave diff WEAVEFILE VERSION1 VERSION2
1136
Show differences between two versions.
1140
% weave init foo.weave
1142
% weave add foo.weave ver0 < foo.txt
1145
(create updated version)
1147
% weave get foo.weave 0 | diff -u - foo.txt
1148
% weave add foo.weave ver1 0 < foo.txt
1151
% weave get foo.weave 0 > foo.txt (create forked version)
1153
% weave add foo.weave ver2 0 < foo.txt
1156
% weave merge foo.weave 1 2 > foo.txt (merge them)
1157
% vi foo.txt (resolve conflicts)
1158
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1170
# in case we're run directly from the subdirectory
1171
sys.path.append('..')
1173
from bzrlib.weavefile import write_weave, read_weave
1174
from bzrlib.progress import ProgressBar
1189
return read_weave(file(argv[2], 'rb'))
1195
# at the moment, based on everything in the file
1197
parents = map(int, argv[4:])
1198
lines = sys.stdin.readlines()
1199
ver = w.add(name, parents, lines)
1200
write_weave(w, file(argv[2], 'wb'))
1201
print 'added version %r %d' % (name, ver)
1204
if os.path.exists(fn):
1205
raise IOError("file exists")
1207
write_weave(w, file(fn, 'wb'))
1208
elif cmd == 'get': # get one version
1210
sys.stdout.writelines(w.get_iter(int(argv[3])))
1215
v1, v2 = map(int, argv[3:5])
1218
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1219
'%s version %d' % (fn, v1),
1220
'%s version %d' % (fn, v2))
1221
sys.stdout.writelines(diff_gen)
1223
elif cmd == 'annotate':
1225
# newline is added to all lines regardless; too hard to get
1226
# reasonable formatting otherwise
1228
for origin, text in w.annotate(int(argv[3])):
1229
text = text.rstrip('\r\n')
1231
print ' | %s' % (text)
1233
print '%5d | %s' % (origin, text)
1239
elif cmd == 'stats':
1240
weave_stats(argv[2], ProgressBar())
1242
elif cmd == 'check':
1247
print '%d versions ok' % w.num_versions()
1249
elif cmd == 'inclusions':
1251
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1253
elif cmd == 'parents':
1255
print ' '.join(map(str, w._parents[int(argv[3])]))
1257
elif cmd == 'plan-merge':
1258
# replaced by 'bzr weave-plan-merge'
1260
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1262
print '%14s | %s' % (state, line),
1263
elif cmd == 'merge':
1264
# replaced by 'bzr weave-merge-text'
1266
p = w.plan_merge(int(argv[3]), int(argv[4]))
1267
sys.stdout.writelines(w.weave_merge(p))
1269
raise ValueError('unknown command %r' % cmd)
1272
if __name__ == '__main__':
1274
sys.exit(main(sys.argv))
1277
class InterWeave(InterVersionedFile):
1278
"""Optimised code paths for weave to weave operations."""
1280
_matching_file_from_factory = staticmethod(WeaveFile)
1281
_matching_file_to_factory = staticmethod(WeaveFile)
1284
def is_compatible(source, target):
1285
"""Be compatible with weaves."""
1287
return (isinstance(source, Weave) and
1288
isinstance(target, Weave))
1289
except AttributeError:
1292
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1293
"""See InterVersionedFile.join."""
1294
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1295
if self.target.versions() == [] and version_ids is None:
1296
self.target._copy_weave_content(self.source)
1298
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1301
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
