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
78
from bzrlib.lazy_import import lazy_import
79
lazy_import(globals(), """
80
from bzrlib import tsort
85
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
86
RevisionAlreadyPresent,
88
UnavailableRepresentation,
89
WeaveRevisionAlreadyPresent,
90
WeaveRevisionNotPresent,
92
import bzrlib.errors as errors
93
from bzrlib.osutils import dirname, sha_strings, split_lines
94
import bzrlib.patiencediff
95
from bzrlib.revision import NULL_REVISION
96
from bzrlib.symbol_versioning import *
97
from bzrlib.trace import mutter
98
from bzrlib.tsort import topo_sort
99
from bzrlib.versionedfile import (
100
AbsentContentFactory,
105
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
108
class WeaveContentFactory(ContentFactory):
109
"""Content factory for streaming from weaves.
111
:seealso ContentFactory:
114
def __init__(self, version, weave):
115
"""Create a WeaveContentFactory for version from weave."""
116
ContentFactory.__init__(self)
117
self.sha1 = weave.get_sha1s([version])[version]
118
self.key = (version,)
119
parents = weave.get_parent_map([version])[version]
120
self.parents = tuple((parent,) for parent in parents)
121
self.storage_kind = 'fulltext'
124
def get_bytes_as(self, storage_kind):
125
if storage_kind == 'fulltext':
126
return self._weave.get_text(self.key[-1])
128
raise UnavailableRepresentation(self.key, storage_kind, 'fulltext')
131
class Weave(VersionedFile):
132
"""weave - versioned text file storage.
134
A Weave manages versions of line-based text files, keeping track
135
of the originating version for each line.
137
To clients the "lines" of the file are represented as a list of strings.
138
These strings will typically have terminal newline characters, but
139
this is not required. In particular files commonly do not have a newline
140
at the end of the file.
142
Texts can be identified in either of two ways:
144
* a nonnegative index number.
146
* a version-id string.
148
Typically the index number will be valid only inside this weave and
149
the version-id is used to reference it in the larger world.
151
The weave is represented as a list mixing edit instructions and
152
literal text. Each entry in _weave can be either a string (or
153
unicode), or a tuple. If a string, it means that the given line
154
should be output in the currently active revisions.
156
If a tuple, it gives a processing instruction saying in which
157
revisions the enclosed lines are active. The tuple has the form
158
(instruction, version).
160
The instruction can be '{' or '}' for an insertion block, and '['
161
and ']' for a deletion block respectively. The version is the
162
integer version index. There is no replace operator, only deletes
163
and inserts. For '}', the end of an insertion, there is no
164
version parameter because it always closes the most recently
169
* A later version can delete lines that were introduced by any
170
number of ancestor versions; this implies that deletion
171
instructions can span insertion blocks without regard to the
172
insertion block's nesting.
174
* Similarly, deletions need not be properly nested with regard to
175
each other, because they might have been generated by
176
independent revisions.
178
* Insertions are always made by inserting a new bracketed block
179
into a single point in the previous weave. This implies they
180
can nest but not overlap, and the nesting must always have later
181
insertions on the inside.
183
* It doesn't seem very useful to have an active insertion
184
inside an inactive insertion, but it might happen.
186
* Therefore, all instructions are always"considered"; that
187
is passed onto and off the stack. An outer inactive block
188
doesn't disable an inner block.
190
* Lines are enabled if the most recent enclosing insertion is
191
active and none of the enclosing deletions are active.
193
* There is no point having a deletion directly inside its own
194
insertion; you might as well just not write it. And there
195
should be no way to get an earlier version deleting a later
199
Text of the weave; list of control instruction tuples and strings.
202
List of parents, indexed by version number.
203
It is only necessary to store the minimal set of parents for
204
each version; the parent's parents are implied.
207
List of hex SHA-1 of each version.
210
List of symbolic names for each version. Each should be unique.
213
For each name, the version number.
216
Descriptive name of this weave; typically the filename if known.
220
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
221
'_weave_name', '_matcher', '_allow_reserved']
223
def __init__(self, weave_name=None, access_mode='w', matcher=None,
224
get_scope=None, allow_reserved=False):
227
:param get_scope: A callable that returns an opaque object to be used
228
for detecting when this weave goes out of scope (should stop
229
answering requests or allowing mutation).
231
super(Weave, self).__init__(access_mode)
237
self._weave_name = weave_name
239
self._matcher = bzrlib.patiencediff.PatienceSequenceMatcher
241
self._matcher = matcher
242
if get_scope is None:
243
get_scope = lambda:None
244
self._get_scope = get_scope
245
self._scope = get_scope()
246
self._access_mode = access_mode
247
self._allow_reserved = allow_reserved
250
return "Weave(%r)" % self._weave_name
252
def _check_write_ok(self):
253
"""Is the versioned file marked as 'finished' ? Raise if it is."""
254
if self._get_scope() != self._scope:
255
raise errors.OutSideTransaction()
256
if self._access_mode != 'w':
257
raise errors.ReadOnlyObjectDirtiedError(self)
260
"""Return a deep copy of self.
262
The copy can be modified without affecting the original weave."""
264
other._weave = self._weave[:]
265
other._parents = self._parents[:]
266
other._sha1s = self._sha1s[:]
267
other._names = self._names[:]
268
other._name_map = self._name_map.copy()
269
other._weave_name = self._weave_name
272
def __eq__(self, other):
273
if not isinstance(other, Weave):
275
return self._parents == other._parents \
276
and self._weave == other._weave \
277
and self._sha1s == other._sha1s
279
def __ne__(self, other):
280
return not self.__eq__(other)
282
def _idx_to_name(self, version):
283
return self._names[version]
285
def _lookup(self, name):
286
"""Convert symbolic version name to index."""
287
if not self._allow_reserved:
288
self.check_not_reserved_id(name)
290
return self._name_map[name]
292
raise RevisionNotPresent(name, self._weave_name)
295
"""See VersionedFile.versions."""
296
return self._names[:]
298
def has_version(self, version_id):
299
"""See VersionedFile.has_version."""
300
return (version_id in self._name_map)
302
__contains__ = has_version
304
def get_record_stream(self, versions, ordering, include_delta_closure):
305
"""Get a stream of records for versions.
307
:param versions: The versions to include. Each version is a tuple
309
:param ordering: Either 'unordered' or 'topological'. A topologically
310
sorted stream has compression parents strictly before their
312
:param include_delta_closure: If True then the closure across any
313
compression parents will be included (in the opaque data).
314
:return: An iterator of ContentFactory objects, each of which is only
315
valid until the iterator is advanced.
317
versions = [version[-1] for version in versions]
318
if ordering == 'topological':
319
parents = self.get_parent_map(versions)
320
new_versions = topo_sort(parents)
321
new_versions.extend(set(versions).difference(set(parents)))
322
versions = new_versions
323
for version in versions:
325
yield WeaveContentFactory(version, self)
327
yield AbsentContentFactory((version,))
329
def get_parent_map(self, version_ids):
330
"""See VersionedFile.get_parent_map."""
332
for version_id in version_ids:
333
if version_id == NULL_REVISION:
338
map(self._idx_to_name,
339
self._parents[self._lookup(version_id)]))
340
except RevisionNotPresent:
342
result[version_id] = parents
345
def get_parents_with_ghosts(self, version_id):
346
raise NotImplementedError(self.get_parents_with_ghosts)
348
def insert_record_stream(self, stream):
349
"""Insert a record stream into this versioned file.
351
:param stream: A stream of records to insert.
353
:seealso VersionedFile.get_record_stream:
356
for record in stream:
357
# Raise an error when a record is missing.
358
if record.storage_kind == 'absent':
359
raise RevisionNotPresent([record.key[0]], self)
360
# adapt to non-tuple interface
361
parents = [parent[0] for parent in record.parents]
362
if record.storage_kind == 'fulltext':
363
self.add_lines(record.key[0], parents,
364
split_lines(record.get_bytes_as('fulltext')))
366
adapter_key = record.storage_kind, 'fulltext'
368
adapter = adapters[adapter_key]
370
adapter_factory = adapter_registry.get(adapter_key)
371
adapter = adapter_factory(self)
372
adapters[adapter_key] = adapter
373
lines = split_lines(adapter.get_bytes(
374
record, record.get_bytes_as(record.storage_kind)))
376
self.add_lines(record.key[0], parents, lines)
377
except RevisionAlreadyPresent:
380
def _check_repeated_add(self, name, parents, text, sha1):
381
"""Check that a duplicated add is OK.
383
If it is, return the (old) index; otherwise raise an exception.
385
idx = self._lookup(name)
386
if sorted(self._parents[idx]) != sorted(parents) \
387
or sha1 != self._sha1s[idx]:
388
raise RevisionAlreadyPresent(name, self._weave_name)
391
def _add_lines(self, version_id, parents, lines, parent_texts,
392
left_matching_blocks, nostore_sha, random_id, check_content):
393
"""See VersionedFile.add_lines."""
394
idx = self._add(version_id, lines, map(self._lookup, parents),
395
nostore_sha=nostore_sha)
396
return sha_strings(lines), sum(map(len, lines)), idx
398
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
399
"""Add a single text on top of the weave.
401
Returns the index number of the newly added version.
404
Symbolic name for this version.
405
(Typically the revision-id of the revision that added it.)
408
List or set of direct parent version numbers.
411
Sequence of lines to be added in the new version.
413
:param nostore_sha: See VersionedFile.add_lines.
415
self._check_lines_not_unicode(lines)
416
self._check_lines_are_lines(lines)
418
sha1 = sha_strings(lines)
419
if sha1 == nostore_sha:
420
raise errors.ExistingContent
421
if version_id in self._name_map:
422
return self._check_repeated_add(version_id, parents, lines, sha1)
424
self._check_versions(parents)
425
## self._check_lines(lines)
426
new_version = len(self._parents)
428
# if we abort after here the (in-memory) weave will be corrupt because only
429
# some fields are updated
430
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
431
# - Robert Collins 20060226
432
self._parents.append(parents[:])
433
self._sha1s.append(sha1)
434
self._names.append(version_id)
435
self._name_map[version_id] = new_version
439
# special case; adding with no parents revision; can do
440
# this more quickly by just appending unconditionally.
441
# even more specially, if we're adding an empty text we
442
# need do nothing at all.
444
self._weave.append(('{', new_version))
445
self._weave.extend(lines)
446
self._weave.append(('}', None))
449
if len(parents) == 1:
450
pv = list(parents)[0]
451
if sha1 == self._sha1s[pv]:
452
# special case: same as the single parent
456
ancestors = self._inclusions(parents)
460
# basis a list of (origin, lineno, line)
463
for origin, lineno, line in self._extract(ancestors):
464
basis_lineno.append(lineno)
465
basis_lines.append(line)
467
# another small special case: a merge, producing the same text
469
if lines == basis_lines:
472
# add a sentinel, because we can also match against the final line
473
basis_lineno.append(len(self._weave))
475
# XXX: which line of the weave should we really consider
476
# matches the end of the file? the current code says it's the
477
# last line of the weave?
479
#print 'basis_lines:', basis_lines
480
#print 'new_lines: ', lines
482
s = self._matcher(None, basis_lines, lines)
484
# offset gives the number of lines that have been inserted
485
# into the weave up to the current point; if the original edit instruction
486
# says to change line A then we actually change (A+offset)
489
for tag, i1, i2, j1, j2 in s.get_opcodes():
490
# i1,i2 are given in offsets within basis_lines; we need to map them
491
# back to offsets within the entire weave
492
#print 'raw match', tag, i1, i2, j1, j2
495
i1 = basis_lineno[i1]
496
i2 = basis_lineno[i2]
497
# the deletion and insertion are handled separately.
498
# first delete the region.
500
self._weave.insert(i1+offset, ('[', new_version))
501
self._weave.insert(i2+offset+1, (']', new_version))
505
# there may have been a deletion spanning up to
506
# i2; we want to insert after this region to make sure
507
# we don't destroy ourselves
509
self._weave[i:i] = ([('{', new_version)]
512
offset += 2 + (j2 - j1)
515
def _inclusions(self, versions):
516
"""Return set of all ancestors of given version(s)."""
517
if not len(versions):
520
for v in xrange(max(versions), 0, -1):
522
# include all its parents
523
i.update(self._parents[v])
525
## except IndexError:
526
## raise ValueError("version %d not present in weave" % v)
528
def get_ancestry(self, version_ids, topo_sorted=True):
529
"""See VersionedFile.get_ancestry."""
530
if isinstance(version_ids, basestring):
531
version_ids = [version_ids]
532
i = self._inclusions([self._lookup(v) for v in version_ids])
533
return [self._idx_to_name(v) for v in i]
535
def _check_lines(self, text):
536
if not isinstance(text, list):
537
raise ValueError("text should be a list, not %s" % type(text))
540
if not isinstance(l, basestring):
541
raise ValueError("text line should be a string or unicode, not %s"
546
def _check_versions(self, indexes):
547
"""Check everything in the sequence of indexes is valid"""
552
raise IndexError("invalid version number %r" % i)
554
def _compatible_parents(self, my_parents, other_parents):
555
"""During join check that other_parents are joinable with my_parents.
557
Joinable is defined as 'is a subset of' - supersets may require
558
regeneration of diffs, but subsets do not.
560
return len(other_parents.difference(my_parents)) == 0
562
def annotate(self, version_id):
563
"""Return a list of (version-id, line) tuples for version_id.
565
The index indicates when the line originated in the weave."""
566
incls = [self._lookup(version_id)]
567
return [(self._idx_to_name(origin), text) for origin, lineno, text in
568
self._extract(incls)]
570
def iter_lines_added_or_present_in_versions(self, version_ids=None,
572
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
573
if version_ids is None:
574
version_ids = self.versions()
575
version_ids = set(version_ids)
576
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
577
# if inserted not in version_ids then it was inserted before the
578
# versions we care about, but because weaves cannot represent ghosts
579
# properly, we do not filter down to that
580
# if inserted not in version_ids: continue
582
yield line + '\n', inserted
586
def _walk_internal(self, version_ids=None):
587
"""Helper method for weave actions."""
592
lineno = 0 # line of weave, 0-based
594
for l in self._weave:
595
if l.__class__ == tuple:
599
istack.append(self._names[v])
603
dset.add(self._names[v])
605
dset.remove(self._names[v])
607
raise WeaveFormatError('unexpected instruction %r' % v)
609
yield lineno, istack[-1], frozenset(dset), l
613
raise WeaveFormatError("unclosed insertion blocks "
614
"at end of weave: %s" % istack)
616
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
619
def plan_merge(self, ver_a, ver_b):
620
"""Return pseudo-annotation indicating how the two versions merge.
622
This is computed between versions a and b and their common
625
Weave lines present in none of them are skipped entirely.
627
inc_a = set(self.get_ancestry([ver_a]))
628
inc_b = set(self.get_ancestry([ver_b]))
629
inc_c = inc_a & inc_b
631
for lineno, insert, deleteset, line in self._walk_internal([ver_a, ver_b]):
632
if deleteset & inc_c:
633
# killed in parent; can't be in either a or b
634
# not relevant to our work
635
yield 'killed-base', line
636
elif insert in inc_c:
637
# was inserted in base
638
killed_a = bool(deleteset & inc_a)
639
killed_b = bool(deleteset & inc_b)
640
if killed_a and killed_b:
641
yield 'killed-both', line
643
yield 'killed-a', line
645
yield 'killed-b', line
647
yield 'unchanged', line
648
elif insert in inc_a:
649
if deleteset & inc_a:
650
yield 'ghost-a', line
654
elif insert in inc_b:
655
if deleteset & inc_b:
656
yield 'ghost-b', line
660
# not in either revision
661
yield 'irrelevant', line
663
def _extract(self, versions):
664
"""Yield annotation of lines in included set.
666
Yields a sequence of tuples (origin, lineno, text), where
667
origin is the origin version, lineno the index in the weave,
668
and text the text of the line.
670
The set typically but not necessarily corresponds to a version.
673
if not isinstance(i, int):
676
included = self._inclusions(versions)
682
lineno = 0 # line of weave, 0-based
688
WFE = WeaveFormatError
691
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
692
# +285282 0 1676.8040 1676.8040 +<isinstance>
693
# 1.6 seconds in 'isinstance'.
694
# changing the first isinstance:
695
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
696
# +140414 0 762.8050 762.8050 +<isinstance>
697
# note that the inline time actually dropped (less function calls)
698
# and total processing time was halved.
699
# we're still spending ~1/4 of the method in isinstance though.
700
# so lets hard code the acceptable string classes we expect:
701
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
702
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
704
# yay, down to ~1/4 the initial extract time, and our inline time
705
# has shrunk again, with isinstance no longer dominating.
706
# tweaking the stack inclusion test to use a set gives:
707
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
708
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
710
# - a 5% win, or possibly just noise. However with large istacks that
711
# 'in' test could dominate, so I'm leaving this change in place -
712
# when its fast enough to consider profiling big datasets we can review.
717
for l in self._weave:
718
if l.__class__ == tuple:
725
iset.remove(istack.pop())
733
raise AssertionError()
736
isactive = (not dset) and istack and (istack[-1] in included)
738
result.append((istack[-1], lineno, l))
741
raise WeaveFormatError("unclosed insertion blocks "
742
"at end of weave: %s" % istack)
744
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
748
def _maybe_lookup(self, name_or_index):
749
"""Convert possible symbolic name to index, or pass through indexes.
753
if isinstance(name_or_index, (int, long)):
756
return self._lookup(name_or_index)
758
def get_lines(self, version_id):
759
"""See VersionedFile.get_lines()."""
760
int_index = self._maybe_lookup(version_id)
761
result = [line for (origin, lineno, line) in self._extract([int_index])]
762
expected_sha1 = self._sha1s[int_index]
763
measured_sha1 = sha_strings(result)
764
if measured_sha1 != expected_sha1:
765
raise errors.WeaveInvalidChecksum(
766
'file %s, revision %s, expected: %s, measured %s'
767
% (self._weave_name, version_id,
768
expected_sha1, measured_sha1))
771
def get_sha1s(self, version_ids):
772
"""See VersionedFile.get_sha1s()."""
774
for v in version_ids:
775
result[v] = self._sha1s[self._lookup(v)]
778
def num_versions(self):
779
"""How many versions are in this weave?"""
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)
806
sha1s[name] = sha.new()
808
new_inc = set([name])
809
for p in self._parents[i]:
810
new_inc.update(inclusions[self._idx_to_name(p)])
812
if set(new_inc) != set(self.get_ancestry(name)):
813
raise AssertionError(
815
% (set(new_inc), set(self.get_ancestry(name))))
816
inclusions[name] = new_inc
818
nlines = len(self._weave)
820
update_text = 'checking weave'
822
short_name = os.path.basename(self._weave_name)
823
update_text = 'checking %s' % (short_name,)
824
update_text = update_text[:25]
826
for lineno, insert, deleteset, line in self._walk_internal():
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 = set([self._names[i] for i in self_parents])
882
n2 = set([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
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
924
except errors.NoSuchFile:
930
def _add_lines(self, version_id, parents, lines, parent_texts,
931
left_matching_blocks, nostore_sha, random_id, check_content):
932
"""Add a version and save the weave."""
933
self.check_not_reserved_id(version_id)
934
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
935
parent_texts, left_matching_blocks, nostore_sha, random_id,
940
def copy_to(self, name, transport):
941
"""See VersionedFile.copy_to()."""
942
# as we are all in memory always, just serialise to the new place.
944
write_weave_v5(self, sio)
946
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
949
"""Save the weave."""
950
self._check_write_ok()
952
write_weave_v5(self, sio)
954
bytes = sio.getvalue()
955
path = self._weave_name + WeaveFile.WEAVE_SUFFIX
957
self._transport.put_bytes(path, bytes, self._filemode)
958
except errors.NoSuchFile:
959
self._transport.mkdir(dirname(path))
960
self._transport.put_bytes(path, bytes, self._filemode)
964
"""See VersionedFile.get_suffixes()."""
965
return [WeaveFile.WEAVE_SUFFIX]
967
def insert_record_stream(self, stream):
968
super(WeaveFile, self).insert_record_stream(stream)
971
@deprecated_method(one_five)
972
def join(self, other, pb=None, msg=None, version_ids=None,
973
ignore_missing=False):
974
"""Join other into self and save."""
975
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
979
def _reweave(wa, wb, pb=None, msg=None):
980
"""Combine two weaves and return the result.
982
This works even if a revision R has different parents in
983
wa and wb. In the resulting weave all the parents are given.
985
This is done by just building up a new weave, maintaining ordering
986
of the versions in the two inputs. More efficient approaches
987
might be possible but it should only be necessary to do
988
this operation rarely, when a new previously ghost version is
991
:param pb: An optional progress bar, indicating how far done we are
992
:param msg: An optional message for the progress
996
queue_a = range(wa.num_versions())
997
queue_b = range(wb.num_versions())
998
# first determine combined parents of all versions
999
# map from version name -> all parent names
1000
combined_parents = _reweave_parent_graphs(wa, wb)
1001
mutter("combined parents: %r", combined_parents)
1002
order = tsort.topo_sort(combined_parents.iteritems())
1003
mutter("order to reweave: %r", order)
1008
for idx, name in enumerate(order):
1010
pb.update(msg, idx, len(order))
1011
if name in wa._name_map:
1012
lines = wa.get_lines(name)
1013
if name in wb._name_map:
1014
lines_b = wb.get_lines(name)
1015
if lines != lines_b:
1016
mutter('Weaves differ on content. rev_id {%s}', name)
1017
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1019
lines = list(difflib.unified_diff(lines, lines_b,
1020
wa._weave_name, wb._weave_name))
1021
mutter('lines:\n%s', ''.join(lines))
1022
raise errors.WeaveTextDiffers(name, wa, wb)
1024
lines = wb.get_lines(name)
1025
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1028
def _reweave_parent_graphs(wa, wb):
1029
"""Return combined parent ancestry for two weaves.
1031
Returned as a list of (version_name, set(parent_names))"""
1033
for weave in [wa, wb]:
1034
for idx, name in enumerate(weave._names):
1035
p = combined.setdefault(name, set())
1036
p.update(map(weave._idx_to_name, weave._parents[idx]))
1041
"""Show the weave's table-of-contents"""
1042
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1043
for i in (6, 50, 10, 10):
1046
for i in range(w.num_versions()):
1049
parent_str = ' '.join(map(str, w._parents[i]))
1050
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1054
def weave_stats(weave_file, pb):
1055
from bzrlib.weavefile import read_weave
1057
wf = file(weave_file, 'rb')
1059
# FIXME: doesn't work on pipes
1060
weave_size = wf.tell()
1064
for i in range(vers):
1065
pb.update('checking sizes', i, vers)
1066
for origin, lineno, line in w._extract([i]):
1071
print 'versions %9d' % vers
1072
print 'weave file %9d bytes' % weave_size
1073
print 'total contents %9d bytes' % total
1074
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1077
print 'average size %9d bytes' % avg
1078
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1082
print """bzr weave tool
1084
Experimental tool for weave algorithm.
1087
weave init WEAVEFILE
1088
Create an empty weave file
1089
weave get WEAVEFILE VERSION
1090
Write out specified version.
1091
weave check WEAVEFILE
1092
Check consistency of all versions.
1094
Display table of contents.
1095
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1096
Add NEWTEXT, with specified parent versions.
1097
weave annotate WEAVEFILE VERSION
1098
Display origin of each line.
1099
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1100
Auto-merge two versions and display conflicts.
1101
weave diff WEAVEFILE VERSION1 VERSION2
1102
Show differences between two versions.
1106
% weave init foo.weave
1108
% weave add foo.weave ver0 < foo.txt
1111
(create updated version)
1113
% weave get foo.weave 0 | diff -u - foo.txt
1114
% weave add foo.weave ver1 0 < foo.txt
1117
% weave get foo.weave 0 > foo.txt (create forked version)
1119
% weave add foo.weave ver2 0 < foo.txt
1122
% weave merge foo.weave 1 2 > foo.txt (merge them)
1123
% vi foo.txt (resolve conflicts)
1124
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1136
# in case we're run directly from the subdirectory
1137
sys.path.append('..')
1139
from bzrlib.weavefile import write_weave, read_weave
1140
from bzrlib.progress import ProgressBar
1155
return read_weave(file(argv[2], 'rb'))
1161
# at the moment, based on everything in the file
1163
parents = map(int, argv[4:])
1164
lines = sys.stdin.readlines()
1165
ver = w.add(name, parents, lines)
1166
write_weave(w, file(argv[2], 'wb'))
1167
print 'added version %r %d' % (name, ver)
1170
if os.path.exists(fn):
1171
raise IOError("file exists")
1173
write_weave(w, file(fn, 'wb'))
1174
elif cmd == 'get': # get one version
1176
sys.stdout.writelines(w.get_iter(int(argv[3])))
1181
v1, v2 = map(int, argv[3:5])
1184
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1185
'%s version %d' % (fn, v1),
1186
'%s version %d' % (fn, v2))
1187
sys.stdout.writelines(diff_gen)
1189
elif cmd == 'annotate':
1191
# newline is added to all lines regardless; too hard to get
1192
# reasonable formatting otherwise
1194
for origin, text in w.annotate(int(argv[3])):
1195
text = text.rstrip('\r\n')
1197
print ' | %s' % (text)
1199
print '%5d | %s' % (origin, text)
1205
elif cmd == 'stats':
1206
weave_stats(argv[2], ProgressBar())
1208
elif cmd == 'check':
1213
print '%d versions ok' % w.num_versions()
1215
elif cmd == 'inclusions':
1217
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1219
elif cmd == 'parents':
1221
print ' '.join(map(str, w._parents[int(argv[3])]))
1223
elif cmd == 'plan-merge':
1224
# replaced by 'bzr weave-plan-merge'
1226
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1228
print '%14s | %s' % (state, line),
1229
elif cmd == 'merge':
1230
# replaced by 'bzr weave-merge-text'
1232
p = w.plan_merge(int(argv[3]), int(argv[4]))
1233
sys.stdout.writelines(w.weave_merge(p))
1235
raise ValueError('unknown command %r' % cmd)
1238
if __name__ == '__main__':
1240
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
