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 dirname, sha_strings, split_lines
90
import bzrlib.patiencediff
91
from bzrlib.revision import NULL_REVISION
92
from bzrlib.symbol_versioning import *
93
from bzrlib.trace import mutter
94
from bzrlib.tsort import topo_sort
95
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[-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__(access_mode)
233
self._weave_name = weave_name
235
self._matcher = bzrlib.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 = topo_sort(parents)
317
new_versions.extend(set(versions).difference(set(parents)))
318
versions = new_versions
319
for version in versions:
321
yield WeaveContentFactory(version, self)
323
yield AbsentContentFactory((version,))
325
def get_parent_map(self, version_ids):
326
"""See VersionedFile.get_parent_map."""
328
for version_id in version_ids:
329
if version_id == NULL_REVISION:
334
map(self._idx_to_name,
335
self._parents[self._lookup(version_id)]))
336
except RevisionNotPresent:
338
result[version_id] = parents
341
def get_parents_with_ghosts(self, version_id):
342
raise NotImplementedError(self.get_parents_with_ghosts)
344
def insert_record_stream(self, stream):
345
"""Insert a record stream into this versioned file.
347
:param stream: A stream of records to insert.
349
:seealso VersionedFile.get_record_stream:
352
for record in stream:
353
# Raise an error when a record is missing.
354
if record.storage_kind == 'absent':
355
raise RevisionNotPresent([record.key[0]], self)
356
# adapt to non-tuple interface
357
parents = [parent[0] for parent in record.parents]
358
if record.storage_kind == 'fulltext':
359
self.add_lines(record.key[0], parents,
360
split_lines(record.get_bytes_as('fulltext')))
362
adapter_key = record.storage_kind, 'fulltext'
364
adapter = adapters[adapter_key]
366
adapter_factory = adapter_registry.get(adapter_key)
367
adapter = adapter_factory(self)
368
adapters[adapter_key] = adapter
369
lines = split_lines(adapter.get_bytes(
370
record, record.get_bytes_as(record.storage_kind)))
372
self.add_lines(record.key[0], parents, lines)
373
except RevisionAlreadyPresent:
376
def _check_repeated_add(self, name, parents, text, sha1):
377
"""Check that a duplicated add is OK.
379
If it is, return the (old) index; otherwise raise an exception.
381
idx = self._lookup(name)
382
if sorted(self._parents[idx]) != sorted(parents) \
383
or sha1 != self._sha1s[idx]:
384
raise RevisionAlreadyPresent(name, self._weave_name)
387
def _add_lines(self, version_id, parents, lines, parent_texts,
388
left_matching_blocks, nostore_sha, random_id, check_content):
389
"""See VersionedFile.add_lines."""
390
idx = self._add(version_id, lines, map(self._lookup, parents),
391
nostore_sha=nostore_sha)
392
return sha_strings(lines), sum(map(len, lines)), idx
394
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
395
"""Add a single text on top of the weave.
397
Returns the index number of the newly added version.
400
Symbolic name for this version.
401
(Typically the revision-id of the revision that added it.)
404
List or set of direct parent version numbers.
407
Sequence of lines to be added in the new version.
409
:param nostore_sha: See VersionedFile.add_lines.
411
self._check_lines_not_unicode(lines)
412
self._check_lines_are_lines(lines)
414
sha1 = sha_strings(lines)
415
if sha1 == nostore_sha:
416
raise errors.ExistingContent
417
if version_id in self._name_map:
418
return self._check_repeated_add(version_id, parents, lines, sha1)
420
self._check_versions(parents)
421
## self._check_lines(lines)
422
new_version = len(self._parents)
424
# if we abort after here the (in-memory) weave will be corrupt because only
425
# some fields are updated
426
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
427
# - Robert Collins 20060226
428
self._parents.append(parents[:])
429
self._sha1s.append(sha1)
430
self._names.append(version_id)
431
self._name_map[version_id] = new_version
435
# special case; adding with no parents revision; can do
436
# this more quickly by just appending unconditionally.
437
# even more specially, if we're adding an empty text we
438
# need do nothing at all.
440
self._weave.append(('{', new_version))
441
self._weave.extend(lines)
442
self._weave.append(('}', None))
445
if len(parents) == 1:
446
pv = list(parents)[0]
447
if sha1 == self._sha1s[pv]:
448
# special case: same as the single parent
452
ancestors = self._inclusions(parents)
456
# basis a list of (origin, lineno, line)
459
for origin, lineno, line in self._extract(ancestors):
460
basis_lineno.append(lineno)
461
basis_lines.append(line)
463
# another small special case: a merge, producing the same text
465
if lines == basis_lines:
468
# add a sentinel, because we can also match against the final line
469
basis_lineno.append(len(self._weave))
471
# XXX: which line of the weave should we really consider
472
# matches the end of the file? the current code says it's the
473
# last line of the weave?
475
#print 'basis_lines:', basis_lines
476
#print 'new_lines: ', lines
478
s = self._matcher(None, basis_lines, lines)
480
# offset gives the number of lines that have been inserted
481
# into the weave up to the current point; if the original edit instruction
482
# says to change line A then we actually change (A+offset)
485
for tag, i1, i2, j1, j2 in s.get_opcodes():
486
# i1,i2 are given in offsets within basis_lines; we need to map them
487
# back to offsets within the entire weave
488
#print 'raw match', tag, i1, i2, j1, j2
491
i1 = basis_lineno[i1]
492
i2 = basis_lineno[i2]
493
# the deletion and insertion are handled separately.
494
# first delete the region.
496
self._weave.insert(i1+offset, ('[', new_version))
497
self._weave.insert(i2+offset+1, (']', new_version))
501
# there may have been a deletion spanning up to
502
# i2; we want to insert after this region to make sure
503
# we don't destroy ourselves
505
self._weave[i:i] = ([('{', new_version)]
508
offset += 2 + (j2 - j1)
511
def _inclusions(self, versions):
512
"""Return set of all ancestors of given version(s)."""
513
if not len(versions):
516
for v in xrange(max(versions), 0, -1):
518
# include all its parents
519
i.update(self._parents[v])
521
## except IndexError:
522
## raise ValueError("version %d not present in weave" % v)
524
def get_ancestry(self, version_ids, topo_sorted=True):
525
"""See VersionedFile.get_ancestry."""
526
if isinstance(version_ids, basestring):
527
version_ids = [version_ids]
528
i = self._inclusions([self._lookup(v) for v in version_ids])
529
return [self._idx_to_name(v) for v in i]
531
def _check_lines(self, text):
532
if not isinstance(text, list):
533
raise ValueError("text should be a list, not %s" % type(text))
536
if not isinstance(l, basestring):
537
raise ValueError("text line should be a string or unicode, not %s"
542
def _check_versions(self, indexes):
543
"""Check everything in the sequence of indexes is valid"""
548
raise IndexError("invalid version number %r" % i)
550
def _compatible_parents(self, my_parents, other_parents):
551
"""During join check that other_parents are joinable with my_parents.
553
Joinable is defined as 'is a subset of' - supersets may require
554
regeneration of diffs, but subsets do not.
556
return len(other_parents.difference(my_parents)) == 0
558
def annotate(self, version_id):
559
"""Return a list of (version-id, line) tuples for version_id.
561
The index indicates when the line originated in the weave."""
562
incls = [self._lookup(version_id)]
563
return [(self._idx_to_name(origin), text) for origin, lineno, text in
564
self._extract(incls)]
566
def iter_lines_added_or_present_in_versions(self, version_ids=None,
568
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
569
if version_ids is None:
570
version_ids = self.versions()
571
version_ids = set(version_ids)
572
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
573
# if inserted not in version_ids then it was inserted before the
574
# versions we care about, but because weaves cannot represent ghosts
575
# properly, we do not filter down to that
576
# if inserted not in version_ids: continue
578
yield line + '\n', inserted
582
def _walk_internal(self, version_ids=None):
583
"""Helper method for weave actions."""
588
lineno = 0 # line of weave, 0-based
590
for l in self._weave:
591
if l.__class__ == tuple:
595
istack.append(self._names[v])
599
dset.add(self._names[v])
601
dset.remove(self._names[v])
603
raise WeaveFormatError('unexpected instruction %r' % v)
605
yield lineno, istack[-1], frozenset(dset), l
609
raise WeaveFormatError("unclosed insertion blocks "
610
"at end of weave: %s" % istack)
612
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
615
def plan_merge(self, ver_a, ver_b):
616
"""Return pseudo-annotation indicating how the two versions merge.
618
This is computed between versions a and b and their common
621
Weave lines present in none of them are skipped entirely.
623
inc_a = set(self.get_ancestry([ver_a]))
624
inc_b = set(self.get_ancestry([ver_b]))
625
inc_c = inc_a & inc_b
627
for lineno, insert, deleteset, line in self._walk_internal([ver_a, ver_b]):
628
if deleteset & inc_c:
629
# killed in parent; can't be in either a or b
630
# not relevant to our work
631
yield 'killed-base', line
632
elif insert in inc_c:
633
# was inserted in base
634
killed_a = bool(deleteset & inc_a)
635
killed_b = bool(deleteset & inc_b)
636
if killed_a and killed_b:
637
yield 'killed-both', line
639
yield 'killed-a', line
641
yield 'killed-b', line
643
yield 'unchanged', line
644
elif insert in inc_a:
645
if deleteset & inc_a:
646
yield 'ghost-a', line
650
elif insert in inc_b:
651
if deleteset & inc_b:
652
yield 'ghost-b', line
656
# not in either revision
657
yield 'irrelevant', line
659
def _extract(self, versions):
660
"""Yield annotation of lines in included set.
662
Yields a sequence of tuples (origin, lineno, text), where
663
origin is the origin version, lineno the index in the weave,
664
and text the text of the line.
666
The set typically but not necessarily corresponds to a version.
669
if not isinstance(i, int):
672
included = self._inclusions(versions)
678
lineno = 0 # line of weave, 0-based
684
WFE = WeaveFormatError
687
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
688
# +285282 0 1676.8040 1676.8040 +<isinstance>
689
# 1.6 seconds in 'isinstance'.
690
# changing the first isinstance:
691
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
692
# +140414 0 762.8050 762.8050 +<isinstance>
693
# note that the inline time actually dropped (less function calls)
694
# and total processing time was halved.
695
# we're still spending ~1/4 of the method in isinstance though.
696
# so lets hard code the acceptable string classes we expect:
697
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
698
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
700
# yay, down to ~1/4 the initial extract time, and our inline time
701
# has shrunk again, with isinstance no longer dominating.
702
# tweaking the stack inclusion test to use a set gives:
703
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
704
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
706
# - a 5% win, or possibly just noise. However with large istacks that
707
# 'in' test could dominate, so I'm leaving this change in place -
708
# when its fast enough to consider profiling big datasets we can review.
713
for l in self._weave:
714
if l.__class__ == tuple:
721
iset.remove(istack.pop())
729
raise AssertionError()
732
isactive = (not dset) and istack and (istack[-1] in included)
734
result.append((istack[-1], lineno, l))
737
raise WeaveFormatError("unclosed insertion blocks "
738
"at end of weave: %s" % istack)
740
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
744
def _maybe_lookup(self, name_or_index):
745
"""Convert possible symbolic name to index, or pass through indexes.
749
if isinstance(name_or_index, (int, long)):
752
return self._lookup(name_or_index)
754
def get_lines(self, version_id):
755
"""See VersionedFile.get_lines()."""
756
int_index = self._maybe_lookup(version_id)
757
result = [line for (origin, lineno, line) in self._extract([int_index])]
758
expected_sha1 = self._sha1s[int_index]
759
measured_sha1 = sha_strings(result)
760
if measured_sha1 != expected_sha1:
761
raise errors.WeaveInvalidChecksum(
762
'file %s, revision %s, expected: %s, measured %s'
763
% (self._weave_name, version_id,
764
expected_sha1, measured_sha1))
767
def get_sha1s(self, version_ids):
768
"""See VersionedFile.get_sha1s()."""
769
return [self._sha1s[self._lookup(v)] for v in version_ids]
771
def num_versions(self):
772
"""How many versions are in this weave?"""
773
l = len(self._parents)
776
__len__ = num_versions
778
def check(self, progress_bar=None):
779
# TODO evaluate performance hit of using string sets in this routine.
780
# TODO: check no circular inclusions
781
# TODO: create a nested progress bar
782
for version in range(self.num_versions()):
783
inclusions = list(self._parents[version])
786
if inclusions[-1] >= version:
787
raise WeaveFormatError("invalid included version %d for index %d"
788
% (inclusions[-1], version))
790
# try extracting all versions; parallel extraction is used
791
nv = self.num_versions()
796
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
797
# The problem is that set membership is much more expensive
798
name = self._idx_to_name(i)
799
sha1s[name] = sha.new()
801
new_inc = set([name])
802
for p in self._parents[i]:
803
new_inc.update(inclusions[self._idx_to_name(p)])
805
if set(new_inc) != set(self.get_ancestry(name)):
806
raise AssertionError(
808
% (set(new_inc), set(self.get_ancestry(name))))
809
inclusions[name] = new_inc
811
nlines = len(self._weave)
813
update_text = 'checking weave'
815
short_name = os.path.basename(self._weave_name)
816
update_text = 'checking %s' % (short_name,)
817
update_text = update_text[:25]
819
for lineno, insert, deleteset, line in self._walk_internal():
821
progress_bar.update(update_text, lineno, nlines)
823
for name, name_inclusions in inclusions.items():
824
# The active inclusion must be an ancestor,
825
# and no ancestors must have deleted this line,
826
# because we don't support resurrection.
827
if (insert in name_inclusions) and not (deleteset & name_inclusions):
828
sha1s[name].update(line)
831
version = self._idx_to_name(i)
832
hd = sha1s[version].hexdigest()
833
expected = self._sha1s[i]
835
raise errors.WeaveInvalidChecksum(
836
"mismatched sha1 for version %s: "
837
"got %s, expected %s"
838
% (version, hd, expected))
840
# TODO: check insertions are properly nested, that there are
841
# no lines outside of insertion blocks, that deletions are
842
# properly paired, etc.
844
def _imported_parents(self, other, other_idx):
845
"""Return list of parents in self corresponding to indexes in other."""
847
for parent_idx in other._parents[other_idx]:
848
parent_name = other._names[parent_idx]
849
if parent_name not in self._name_map:
850
# should not be possible
851
raise WeaveError("missing parent {%s} of {%s} in %r"
852
% (parent_name, other._name_map[other_idx], self))
853
new_parents.append(self._name_map[parent_name])
856
def _check_version_consistent(self, other, other_idx, name):
857
"""Check if a version in consistent in this and other.
859
To be consistent it must have:
862
* the same direct parents (by name, not index, and disregarding
865
If present & correct return True;
866
if not present in self return False;
867
if inconsistent raise error."""
868
this_idx = self._name_map.get(name, -1)
870
if self._sha1s[this_idx] != other._sha1s[other_idx]:
871
raise errors.WeaveTextDiffers(name, self, other)
872
self_parents = self._parents[this_idx]
873
other_parents = other._parents[other_idx]
874
n1 = set([self._names[i] for i in self_parents])
875
n2 = set([other._names[i] for i in other_parents])
876
if not self._compatible_parents(n1, n2):
877
raise WeaveParentMismatch("inconsistent parents "
878
"for version {%s}: %s vs %s" % (name, n1, n2))
884
def _reweave(self, other, pb, msg):
885
"""Reweave self with other - internal helper for join().
887
:param other: The other weave to merge
888
:param pb: An optional progress bar, indicating how far done we are
889
:param msg: An optional message for the progress
891
new_weave = _reweave(self, other, pb=pb, msg=msg)
892
self._copy_weave_content(new_weave)
894
def _copy_weave_content(self, otherweave):
895
"""adsorb the content from otherweave."""
896
for attr in self.__slots__:
897
if attr != '_weave_name':
898
setattr(self, attr, copy(getattr(otherweave, attr)))
901
class WeaveFile(Weave):
902
"""A WeaveFile represents a Weave on disk and writes on change."""
904
WEAVE_SUFFIX = '.weave'
906
def __init__(self, name, transport, filemode=None, create=False, access_mode='w', get_scope=None):
907
"""Create a WeaveFile.
909
:param create: If not True, only open an existing knit.
911
super(WeaveFile, self).__init__(name, access_mode, get_scope=get_scope,
912
allow_reserved=False)
913
self._transport = transport
914
self._filemode = filemode
916
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
917
except errors.NoSuchFile:
923
def _add_lines(self, version_id, parents, lines, parent_texts,
924
left_matching_blocks, nostore_sha, random_id, check_content):
925
"""Add a version and save the weave."""
926
self.check_not_reserved_id(version_id)
927
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
928
parent_texts, left_matching_blocks, nostore_sha, random_id,
933
def copy_to(self, name, transport):
934
"""See VersionedFile.copy_to()."""
935
# as we are all in memory always, just serialise to the new place.
937
write_weave_v5(self, sio)
939
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
942
"""Save the weave."""
943
self._check_write_ok()
945
write_weave_v5(self, sio)
947
bytes = sio.getvalue()
948
path = self._weave_name + WeaveFile.WEAVE_SUFFIX
950
self._transport.put_bytes(path, bytes, self._filemode)
951
except errors.NoSuchFile:
952
self._transport.mkdir(dirname(path))
953
self._transport.put_bytes(path, bytes, self._filemode)
957
"""See VersionedFile.get_suffixes()."""
958
return [WeaveFile.WEAVE_SUFFIX]
960
def insert_record_stream(self, stream):
961
super(WeaveFile, self).insert_record_stream(stream)
964
@deprecated_method(one_five)
965
def join(self, other, pb=None, msg=None, version_ids=None,
966
ignore_missing=False):
967
"""Join other into self and save."""
968
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
972
def _reweave(wa, wb, pb=None, msg=None):
973
"""Combine two weaves and return the result.
975
This works even if a revision R has different parents in
976
wa and wb. In the resulting weave all the parents are given.
978
This is done by just building up a new weave, maintaining ordering
979
of the versions in the two inputs. More efficient approaches
980
might be possible but it should only be necessary to do
981
this operation rarely, when a new previously ghost version is
984
:param pb: An optional progress bar, indicating how far done we are
985
:param msg: An optional message for the progress
989
queue_a = range(wa.num_versions())
990
queue_b = range(wb.num_versions())
991
# first determine combined parents of all versions
992
# map from version name -> all parent names
993
combined_parents = _reweave_parent_graphs(wa, wb)
994
mutter("combined parents: %r", combined_parents)
995
order = topo_sort(combined_parents.iteritems())
996
mutter("order to reweave: %r", order)
1001
for idx, name in enumerate(order):
1003
pb.update(msg, idx, len(order))
1004
if name in wa._name_map:
1005
lines = wa.get_lines(name)
1006
if name in wb._name_map:
1007
lines_b = wb.get_lines(name)
1008
if lines != lines_b:
1009
mutter('Weaves differ on content. rev_id {%s}', name)
1010
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1012
lines = list(difflib.unified_diff(lines, lines_b,
1013
wa._weave_name, wb._weave_name))
1014
mutter('lines:\n%s', ''.join(lines))
1015
raise errors.WeaveTextDiffers(name, wa, wb)
1017
lines = wb.get_lines(name)
1018
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1021
def _reweave_parent_graphs(wa, wb):
1022
"""Return combined parent ancestry for two weaves.
1024
Returned as a list of (version_name, set(parent_names))"""
1026
for weave in [wa, wb]:
1027
for idx, name in enumerate(weave._names):
1028
p = combined.setdefault(name, set())
1029
p.update(map(weave._idx_to_name, weave._parents[idx]))
1034
"""Show the weave's table-of-contents"""
1035
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1036
for i in (6, 50, 10, 10):
1039
for i in range(w.num_versions()):
1042
parent_str = ' '.join(map(str, w._parents[i]))
1043
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1047
def weave_stats(weave_file, pb):
1048
from bzrlib.weavefile import read_weave
1050
wf = file(weave_file, 'rb')
1052
# FIXME: doesn't work on pipes
1053
weave_size = wf.tell()
1057
for i in range(vers):
1058
pb.update('checking sizes', i, vers)
1059
for origin, lineno, line in w._extract([i]):
1064
print 'versions %9d' % vers
1065
print 'weave file %9d bytes' % weave_size
1066
print 'total contents %9d bytes' % total
1067
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1070
print 'average size %9d bytes' % avg
1071
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1075
print """bzr weave tool
1077
Experimental tool for weave algorithm.
1080
weave init WEAVEFILE
1081
Create an empty weave file
1082
weave get WEAVEFILE VERSION
1083
Write out specified version.
1084
weave check WEAVEFILE
1085
Check consistency of all versions.
1087
Display table of contents.
1088
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1089
Add NEWTEXT, with specified parent versions.
1090
weave annotate WEAVEFILE VERSION
1091
Display origin of each line.
1092
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1093
Auto-merge two versions and display conflicts.
1094
weave diff WEAVEFILE VERSION1 VERSION2
1095
Show differences between two versions.
1099
% weave init foo.weave
1101
% weave add foo.weave ver0 < foo.txt
1104
(create updated version)
1106
% weave get foo.weave 0 | diff -u - foo.txt
1107
% weave add foo.weave ver1 0 < foo.txt
1110
% weave get foo.weave 0 > foo.txt (create forked version)
1112
% weave add foo.weave ver2 0 < foo.txt
1115
% weave merge foo.weave 1 2 > foo.txt (merge them)
1116
% vi foo.txt (resolve conflicts)
1117
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1129
# in case we're run directly from the subdirectory
1130
sys.path.append('..')
1132
from bzrlib.weavefile import write_weave, read_weave
1133
from bzrlib.progress import ProgressBar
1148
return read_weave(file(argv[2], 'rb'))
1154
# at the moment, based on everything in the file
1156
parents = map(int, argv[4:])
1157
lines = sys.stdin.readlines()
1158
ver = w.add(name, parents, lines)
1159
write_weave(w, file(argv[2], 'wb'))
1160
print 'added version %r %d' % (name, ver)
1163
if os.path.exists(fn):
1164
raise IOError("file exists")
1166
write_weave(w, file(fn, 'wb'))
1167
elif cmd == 'get': # get one version
1169
sys.stdout.writelines(w.get_iter(int(argv[3])))
1174
v1, v2 = map(int, argv[3:5])
1177
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1178
'%s version %d' % (fn, v1),
1179
'%s version %d' % (fn, v2))
1180
sys.stdout.writelines(diff_gen)
1182
elif cmd == 'annotate':
1184
# newline is added to all lines regardless; too hard to get
1185
# reasonable formatting otherwise
1187
for origin, text in w.annotate(int(argv[3])):
1188
text = text.rstrip('\r\n')
1190
print ' | %s' % (text)
1192
print '%5d | %s' % (origin, text)
1198
elif cmd == 'stats':
1199
weave_stats(argv[2], ProgressBar())
1201
elif cmd == 'check':
1206
print '%d versions ok' % w.num_versions()
1208
elif cmd == 'inclusions':
1210
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1212
elif cmd == 'parents':
1214
print ' '.join(map(str, w._parents[int(argv[3])]))
1216
elif cmd == 'plan-merge':
1217
# replaced by 'bzr weave-plan-merge'
1219
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1221
print '%14s | %s' % (state, line),
1222
elif cmd == 'merge':
1223
# replaced by 'bzr weave-merge-text'
1225
p = w.plan_merge(int(argv[3]), int(argv[4]))
1226
sys.stdout.writelines(w.weave_merge(p))
1228
raise ValueError('unknown command %r' % cmd)
1231
if __name__ == '__main__':
1233
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
