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.trace import mutter
82
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
83
RevisionAlreadyPresent,
85
WeaveRevisionAlreadyPresent,
86
WeaveRevisionNotPresent,
88
import bzrlib.errors as errors
89
from bzrlib.osutils import sha_strings
90
import bzrlib.patiencediff
91
from bzrlib.tsort import topo_sort
92
from bzrlib.versionedfile import VersionedFile, InterVersionedFile
93
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
96
class Weave(VersionedFile):
97
"""weave - versioned text file storage.
99
A Weave manages versions of line-based text files, keeping track
100
of the originating version for each line.
102
To clients the "lines" of the file are represented as a list of strings.
103
These strings will typically have terminal newline characters, but
104
this is not required. In particular files commonly do not have a newline
105
at the end of the file.
107
Texts can be identified in either of two ways:
109
* a nonnegative index number.
111
* a version-id string.
113
Typically the index number will be valid only inside this weave and
114
the version-id is used to reference it in the larger world.
116
The weave is represented as a list mixing edit instructions and
117
literal text. Each entry in _weave can be either a string (or
118
unicode), or a tuple. If a string, it means that the given line
119
should be output in the currently active revisions.
121
If a tuple, it gives a processing instruction saying in which
122
revisions the enclosed lines are active. The tuple has the form
123
(instruction, version).
125
The instruction can be '{' or '}' for an insertion block, and '['
126
and ']' for a deletion block respectively. The version is the
127
integer version index. There is no replace operator, only deletes
128
and inserts. For '}', the end of an insertion, there is no
129
version parameter because it always closes the most recently
134
* A later version can delete lines that were introduced by any
135
number of ancestor versions; this implies that deletion
136
instructions can span insertion blocks without regard to the
137
insertion block's nesting.
139
* Similarly, deletions need not be properly nested with regard to
140
each other, because they might have been generated by
141
independent revisions.
143
* Insertions are always made by inserting a new bracketed block
144
into a single point in the previous weave. This implies they
145
can nest but not overlap, and the nesting must always have later
146
insertions on the inside.
148
* It doesn't seem very useful to have an active insertion
149
inside an inactive insertion, but it might happen.
151
* Therefore, all instructions are always"considered"; that
152
is passed onto and off the stack. An outer inactive block
153
doesn't disable an inner block.
155
* Lines are enabled if the most recent enclosing insertion is
156
active and none of the enclosing deletions are active.
158
* There is no point having a deletion directly inside its own
159
insertion; you might as well just not write it. And there
160
should be no way to get an earlier version deleting a later
164
Text of the weave; list of control instruction tuples and strings.
167
List of parents, indexed by version number.
168
It is only necessary to store the minimal set of parents for
169
each version; the parent's parents are implied.
172
List of hex SHA-1 of each version.
175
List of symbolic names for each version. Each should be unique.
178
For each name, the version number.
181
Descriptive name of this weave; typically the filename if known.
185
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
186
'_weave_name', '_matcher']
188
def __init__(self, weave_name=None, access_mode='w', matcher=None):
189
super(Weave, self).__init__(access_mode)
195
self._weave_name = weave_name
197
self._matcher = bzrlib.patiencediff.PatienceSequenceMatcher
199
self._matcher = matcher
202
return "Weave(%r)" % self._weave_name
205
"""Return a deep copy of self.
207
The copy can be modified without affecting the original weave."""
209
other._weave = self._weave[:]
210
other._parents = self._parents[:]
211
other._sha1s = self._sha1s[:]
212
other._names = self._names[:]
213
other._name_map = self._name_map.copy()
214
other._weave_name = self._weave_name
217
def __eq__(self, other):
218
if not isinstance(other, Weave):
220
return self._parents == other._parents \
221
and self._weave == other._weave \
222
and self._sha1s == other._sha1s
224
def __ne__(self, other):
225
return not self.__eq__(other)
227
def _idx_to_name(self, version):
228
return self._names[version]
230
def _lookup(self, name):
231
"""Convert symbolic version name to index."""
232
self.check_not_reserved_id(name)
234
return self._name_map[name]
236
raise RevisionNotPresent(name, self._weave_name)
239
"""See VersionedFile.versions."""
240
return self._names[:]
242
def has_version(self, version_id):
243
"""See VersionedFile.has_version."""
244
return (version_id in self._name_map)
246
__contains__ = has_version
248
def get_delta(self, version_id):
249
"""See VersionedFile.get_delta."""
250
return self.get_deltas([version_id])[version_id]
252
def get_deltas(self, version_ids):
253
"""See VersionedFile.get_deltas."""
254
version_ids = self.get_ancestry(version_ids)
255
for version_id in version_ids:
256
if not self.has_version(version_id):
257
raise RevisionNotPresent(version_id, self)
258
# try extracting all versions; parallel extraction is used
259
nv = self.num_versions()
265
last_parent_lines = {}
267
parent_inclusions = {}
272
# its simplest to generate a full set of prepared variables.
274
name = self._names[i]
275
sha1s[name] = self.get_sha1(name)
276
parents_list = self.get_parents(name)
278
parent = parents_list[0]
279
parents[name] = parent
280
parent_inclusions[name] = inclusions[parent]
283
parent_inclusions[name] = set()
284
# we want to emit start, finish, replacement_length, replacement_lines tuples.
285
diff_hunks[name] = []
286
current_hunks[name] = [0, 0, 0, []] # #start, finish, repl_length, repl_tuples
287
parent_linenums[name] = 0
289
parent_noeols[name] = False
290
last_parent_lines[name] = None
291
new_inc = set([name])
292
for p in self._parents[i]:
293
new_inc.update(inclusions[self._idx_to_name(p)])
294
# debug only, known good so far.
295
#assert set(new_inc) == set(self.get_ancestry(name)), \
296
# 'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
297
inclusions[name] = new_inc
299
nlines = len(self._weave)
301
for lineno, inserted, deletes, line in self._walk_internal():
302
# a line is active in a version if:
303
# insert is in the versions inclusions
305
# deleteset & the versions inclusions is an empty set.
306
# so - if we have a included by mapping - version is included by
307
# children, we get a list of children to examine for deletes affect
308
# ing them, which is less than the entire set of children.
309
for version_id in version_ids:
310
# The active inclusion must be an ancestor,
311
# and no ancestors must have deleted this line,
312
# because we don't support resurrection.
313
parent_inclusion = parent_inclusions[version_id]
314
inclusion = inclusions[version_id]
315
parent_active = inserted in parent_inclusion and not (deletes & parent_inclusion)
316
version_active = inserted in inclusion and not (deletes & inclusion)
317
if not parent_active and not version_active:
318
# unrelated line of ancestry
320
elif parent_active and version_active:
322
parent_linenum = parent_linenums[version_id]
323
if current_hunks[version_id] != [parent_linenum, parent_linenum, 0, []]:
324
diff_hunks[version_id].append(tuple(current_hunks[version_id]))
326
current_hunks[version_id] = [parent_linenum, parent_linenum, 0, []]
327
parent_linenums[version_id] = parent_linenum
330
noeols[version_id] = True
333
elif parent_active and not version_active:
335
current_hunks[version_id][1] += 1
336
parent_linenums[version_id] += 1
337
last_parent_lines[version_id] = line
338
elif not parent_active and version_active:
340
# noeol only occurs at the end of a file because we
341
# diff linewise. We want to show noeol changes as a
342
# empty diff unless the actual eol-less content changed.
345
if last_parent_lines[version_id][-1] != '\n':
346
parent_noeols[version_id] = True
347
except (TypeError, IndexError):
350
if theline[-1] != '\n':
351
noeols[version_id] = True
355
parent_should_go = False
357
if parent_noeols[version_id] == noeols[version_id]:
358
# no noeol toggle, so trust the weaves statement
359
# that this line is changed.
361
if parent_noeols[version_id]:
362
theline = theline + '\n'
363
elif parent_noeols[version_id]:
364
# parent has no eol, we do:
365
# our line is new, report as such..
367
elif noeols[version_id]:
368
# append a eol so that it looks like
370
theline = theline + '\n'
371
if parents[version_id] is not None:
372
#if last_parent_lines[version_id] is not None:
373
parent_should_go = True
374
if last_parent_lines[version_id] != theline:
377
#parent_should_go = False
379
current_hunks[version_id][2] += 1
380
current_hunks[version_id][3].append((inserted, theline))
382
# last hunk last parent line is not eaten
383
current_hunks[version_id][1] -= 1
384
if current_hunks[version_id][1] < 0:
385
current_hunks[version_id][1] = 0
386
# import pdb;pdb.set_trace()
387
# assert current_hunks[version_id][1] >= 0
391
version = self._idx_to_name(i)
392
if current_hunks[version] != [0, 0, 0, []]:
393
diff_hunks[version].append(tuple(current_hunks[version]))
395
for version_id in version_ids:
396
result[version_id] = (
400
diff_hunks[version_id],
404
def get_parents(self, version_id):
405
"""See VersionedFile.get_parent."""
406
return map(self._idx_to_name, self._parents[self._lookup(version_id)])
408
def _check_repeated_add(self, name, parents, text, sha1):
409
"""Check that a duplicated add is OK.
411
If it is, return the (old) index; otherwise raise an exception.
413
idx = self._lookup(name)
414
if sorted(self._parents[idx]) != sorted(parents) \
415
or sha1 != self._sha1s[idx]:
416
raise RevisionAlreadyPresent(name, self._weave_name)
419
def _add_lines(self, version_id, parents, lines, parent_texts,
420
left_matching_blocks, nostore_sha):
421
"""See VersionedFile.add_lines."""
422
idx = self._add(version_id, lines, map(self._lookup, parents),
423
nostore_sha=nostore_sha)
424
return sha_strings(lines), sum(map(len, lines)), idx
426
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
427
"""Add a single text on top of the weave.
429
Returns the index number of the newly added version.
432
Symbolic name for this version.
433
(Typically the revision-id of the revision that added it.)
436
List or set of direct parent version numbers.
439
Sequence of lines to be added in the new version.
441
:param nostore_sha: See VersionedFile.add_lines.
443
assert isinstance(version_id, basestring)
444
self._check_lines_not_unicode(lines)
445
self._check_lines_are_lines(lines)
447
sha1 = sha_strings(lines)
448
if sha1 == nostore_sha:
449
raise errors.ExistingContent
450
if version_id in self._name_map:
451
return self._check_repeated_add(version_id, parents, lines, sha1)
453
self._check_versions(parents)
454
## self._check_lines(lines)
455
new_version = len(self._parents)
457
# if we abort after here the (in-memory) weave will be corrupt because only
458
# some fields are updated
459
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
460
# - Robert Collins 20060226
461
self._parents.append(parents[:])
462
self._sha1s.append(sha1)
463
self._names.append(version_id)
464
self._name_map[version_id] = new_version
468
# special case; adding with no parents revision; can do
469
# this more quickly by just appending unconditionally.
470
# even more specially, if we're adding an empty text we
471
# need do nothing at all.
473
self._weave.append(('{', new_version))
474
self._weave.extend(lines)
475
self._weave.append(('}', None))
478
if len(parents) == 1:
479
pv = list(parents)[0]
480
if sha1 == self._sha1s[pv]:
481
# special case: same as the single parent
485
ancestors = self._inclusions(parents)
489
# basis a list of (origin, lineno, line)
492
for origin, lineno, line in self._extract(ancestors):
493
basis_lineno.append(lineno)
494
basis_lines.append(line)
496
# another small special case: a merge, producing the same text
498
if lines == basis_lines:
501
# add a sentinel, because we can also match against the final line
502
basis_lineno.append(len(self._weave))
504
# XXX: which line of the weave should we really consider
505
# matches the end of the file? the current code says it's the
506
# last line of the weave?
508
#print 'basis_lines:', basis_lines
509
#print 'new_lines: ', lines
511
s = self._matcher(None, basis_lines, lines)
513
# offset gives the number of lines that have been inserted
514
# into the weave up to the current point; if the original edit instruction
515
# says to change line A then we actually change (A+offset)
518
for tag, i1, i2, j1, j2 in s.get_opcodes():
519
# i1,i2 are given in offsets within basis_lines; we need to map them
520
# back to offsets within the entire weave
521
#print 'raw match', tag, i1, i2, j1, j2
525
i1 = basis_lineno[i1]
526
i2 = basis_lineno[i2]
528
assert 0 <= j1 <= j2 <= len(lines)
530
#print tag, i1, i2, j1, j2
532
# the deletion and insertion are handled separately.
533
# first delete the region.
535
self._weave.insert(i1+offset, ('[', new_version))
536
self._weave.insert(i2+offset+1, (']', new_version))
540
# there may have been a deletion spanning up to
541
# i2; we want to insert after this region to make sure
542
# we don't destroy ourselves
544
self._weave[i:i] = ([('{', new_version)]
547
offset += 2 + (j2 - j1)
550
def _clone_text(self, new_version_id, old_version_id, parents):
551
"""See VersionedFile.clone_text."""
552
old_lines = self.get_text(old_version_id)
553
self.add_lines(new_version_id, parents, old_lines)
555
def _inclusions(self, versions):
556
"""Return set of all ancestors of given version(s)."""
557
if not len(versions):
560
for v in xrange(max(versions), 0, -1):
562
# include all its parents
563
i.update(self._parents[v])
565
## except IndexError:
566
## raise ValueError("version %d not present in weave" % v)
568
def get_ancestry(self, version_ids, topo_sorted=True):
569
"""See VersionedFile.get_ancestry."""
570
if isinstance(version_ids, basestring):
571
version_ids = [version_ids]
572
i = self._inclusions([self._lookup(v) for v in version_ids])
573
return [self._idx_to_name(v) for v in i]
575
def _check_lines(self, text):
576
if not isinstance(text, list):
577
raise ValueError("text should be a list, not %s" % type(text))
580
if not isinstance(l, basestring):
581
raise ValueError("text line should be a string or unicode, not %s"
586
def _check_versions(self, indexes):
587
"""Check everything in the sequence of indexes is valid"""
592
raise IndexError("invalid version number %r" % i)
594
def _compatible_parents(self, my_parents, other_parents):
595
"""During join check that other_parents are joinable with my_parents.
597
Joinable is defined as 'is a subset of' - supersets may require
598
regeneration of diffs, but subsets do not.
600
return len(other_parents.difference(my_parents)) == 0
602
def annotate_iter(self, version_id):
603
"""Yield list of (version-id, line) pairs for the specified version.
605
The index indicates when the line originated in the weave."""
606
incls = [self._lookup(version_id)]
607
for origin, lineno, text in self._extract(incls):
608
yield self._idx_to_name(origin), text
610
def iter_lines_added_or_present_in_versions(self, version_ids=None,
612
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
613
if version_ids is None:
614
version_ids = self.versions()
615
version_ids = set(version_ids)
616
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
617
# if inserted not in version_ids then it was inserted before the
618
# versions we care about, but because weaves cannot represent ghosts
619
# properly, we do not filter down to that
620
# if inserted not in version_ids: continue
626
def _walk_internal(self, version_ids=None):
627
"""Helper method for weave actions."""
632
lineno = 0 # line of weave, 0-based
634
for l in self._weave:
635
if l.__class__ == tuple:
639
istack.append(self._names[v])
643
assert self._names[v] not in dset
644
dset.add(self._names[v])
646
dset.remove(self._names[v])
648
raise WeaveFormatError('unexpected instruction %r' % v)
650
assert l.__class__ in (str, unicode)
652
yield lineno, istack[-1], frozenset(dset), l
656
raise WeaveFormatError("unclosed insertion blocks "
657
"at end of weave: %s" % istack)
659
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
662
def plan_merge(self, ver_a, ver_b):
663
"""Return pseudo-annotation indicating how the two versions merge.
665
This is computed between versions a and b and their common
668
Weave lines present in none of them are skipped entirely.
670
inc_a = set(self.get_ancestry([ver_a]))
671
inc_b = set(self.get_ancestry([ver_b]))
672
inc_c = inc_a & inc_b
674
for lineno, insert, deleteset, line in self._walk_internal([ver_a, ver_b]):
675
if deleteset & inc_c:
676
# killed in parent; can't be in either a or b
677
# not relevant to our work
678
yield 'killed-base', line
679
elif insert in inc_c:
680
# was inserted in base
681
killed_a = bool(deleteset & inc_a)
682
killed_b = bool(deleteset & inc_b)
683
if killed_a and killed_b:
684
yield 'killed-both', line
686
yield 'killed-a', line
688
yield 'killed-b', line
690
yield 'unchanged', line
691
elif insert in inc_a:
692
if deleteset & inc_a:
693
yield 'ghost-a', line
697
elif insert in inc_b:
698
if deleteset & inc_b:
699
yield 'ghost-b', line
703
# not in either revision
704
yield 'irrelevant', line
706
yield 'unchanged', '' # terminator
708
def _extract(self, versions):
709
"""Yield annotation of lines in included set.
711
Yields a sequence of tuples (origin, lineno, text), where
712
origin is the origin version, lineno the index in the weave,
713
and text the text of the line.
715
The set typically but not necessarily corresponds to a version.
718
if not isinstance(i, int):
721
included = self._inclusions(versions)
727
lineno = 0 # line of weave, 0-based
733
WFE = WeaveFormatError
736
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
737
# +285282 0 1676.8040 1676.8040 +<isinstance>
738
# 1.6 seconds in 'isinstance'.
739
# changing the first isinstance:
740
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
741
# +140414 0 762.8050 762.8050 +<isinstance>
742
# note that the inline time actually dropped (less function calls)
743
# and total processing time was halved.
744
# we're still spending ~1/4 of the method in isinstance though.
745
# so lets hard code the acceptable string classes we expect:
746
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
747
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
749
# yay, down to ~1/4 the initial extract time, and our inline time
750
# has shrunk again, with isinstance no longer dominating.
751
# tweaking the stack inclusion test to use a set gives:
752
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
753
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
755
# - a 5% win, or possibly just noise. However with large istacks that
756
# 'in' test could dominate, so I'm leaving this change in place -
757
# when its fast enough to consider profiling big datasets we can review.
762
for l in self._weave:
763
if l.__class__ == tuple:
771
iset.remove(istack.pop())
782
assert l.__class__ in (str, unicode)
784
isactive = (not dset) and istack and (istack[-1] in included)
786
result.append((istack[-1], lineno, l))
789
raise WeaveFormatError("unclosed insertion blocks "
790
"at end of weave: %s" % istack)
792
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
796
def _maybe_lookup(self, name_or_index):
797
"""Convert possible symbolic name to index, or pass through indexes.
801
if isinstance(name_or_index, (int, long)):
804
return self._lookup(name_or_index)
806
def get_lines(self, version_id):
807
"""See VersionedFile.get_lines()."""
808
int_index = self._maybe_lookup(version_id)
809
result = [line for (origin, lineno, line) in self._extract([int_index])]
810
expected_sha1 = self._sha1s[int_index]
811
measured_sha1 = sha_strings(result)
812
if measured_sha1 != expected_sha1:
813
raise errors.WeaveInvalidChecksum(
814
'file %s, revision %s, expected: %s, measured %s'
815
% (self._weave_name, version_id,
816
expected_sha1, measured_sha1))
819
def get_sha1(self, version_id):
820
"""See VersionedFile.get_sha1()."""
821
return self._sha1s[self._lookup(version_id)]
823
def get_sha1s(self, version_ids):
824
"""See VersionedFile.get_sha1s()."""
825
return [self._sha1s[self._lookup(v)] for v in version_ids]
827
def num_versions(self):
828
"""How many versions are in this weave?"""
829
l = len(self._parents)
830
assert l == len(self._sha1s)
833
__len__ = num_versions
835
def check(self, progress_bar=None):
836
# TODO evaluate performance hit of using string sets in this routine.
837
# TODO: check no circular inclusions
838
# TODO: create a nested progress bar
839
for version in range(self.num_versions()):
840
inclusions = list(self._parents[version])
843
if inclusions[-1] >= version:
844
raise WeaveFormatError("invalid included version %d for index %d"
845
% (inclusions[-1], version))
847
# try extracting all versions; parallel extraction is used
848
nv = self.num_versions()
853
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
854
# The problem is that set membership is much more expensive
855
name = self._idx_to_name(i)
856
sha1s[name] = sha.new()
858
new_inc = set([name])
859
for p in self._parents[i]:
860
new_inc.update(inclusions[self._idx_to_name(p)])
862
assert set(new_inc) == set(self.get_ancestry(name)), \
863
'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
864
inclusions[name] = new_inc
866
nlines = len(self._weave)
868
update_text = 'checking weave'
870
short_name = os.path.basename(self._weave_name)
871
update_text = 'checking %s' % (short_name,)
872
update_text = update_text[:25]
874
for lineno, insert, deleteset, line in self._walk_internal():
876
progress_bar.update(update_text, lineno, nlines)
878
for name, name_inclusions in inclusions.items():
879
# The active inclusion must be an ancestor,
880
# and no ancestors must have deleted this line,
881
# because we don't support resurrection.
882
if (insert in name_inclusions) and not (deleteset & name_inclusions):
883
sha1s[name].update(line)
886
version = self._idx_to_name(i)
887
hd = sha1s[version].hexdigest()
888
expected = self._sha1s[i]
890
raise errors.WeaveInvalidChecksum(
891
"mismatched sha1 for version %s: "
892
"got %s, expected %s"
893
% (version, hd, expected))
895
# TODO: check insertions are properly nested, that there are
896
# no lines outside of insertion blocks, that deletions are
897
# properly paired, etc.
899
def _join(self, other, pb, msg, version_ids, ignore_missing):
900
"""Worker routine for join()."""
901
if not other.versions():
902
return # nothing to update, easy
905
# versions is never none, InterWeave checks this.
908
# two loops so that we do not change ourselves before verifying it
910
# work through in index order to make sure we get all dependencies
913
# get the selected versions only that are in other.versions.
914
version_ids = set(other.versions()).intersection(set(version_ids))
915
# pull in the referenced graph.
916
version_ids = other.get_ancestry(version_ids)
917
pending_graph = [(version, other.get_parents(version)) for
918
version in version_ids]
919
for name in topo_sort(pending_graph):
920
other_idx = other._name_map[name]
921
# returns True if we have it, False if we need it.
922
if not self._check_version_consistent(other, other_idx, name):
923
names_to_join.append((other_idx, name))
932
for other_idx, name in names_to_join:
933
# TODO: If all the parents of the other version are already
934
# present then we can avoid some work by just taking the delta
935
# and adjusting the offsets.
936
new_parents = self._imported_parents(other, other_idx)
937
sha1 = other._sha1s[other_idx]
942
pb.update(msg, merged, len(names_to_join))
944
lines = other.get_lines(other_idx)
945
self._add(name, lines, new_parents, sha1)
947
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
948
merged, processed, self._weave_name, time.time()-time0))
950
def _imported_parents(self, other, other_idx):
951
"""Return list of parents in self corresponding to indexes in other."""
953
for parent_idx in other._parents[other_idx]:
954
parent_name = other._names[parent_idx]
955
if parent_name not in self._name_map:
956
# should not be possible
957
raise WeaveError("missing parent {%s} of {%s} in %r"
958
% (parent_name, other._name_map[other_idx], self))
959
new_parents.append(self._name_map[parent_name])
962
def _check_version_consistent(self, other, other_idx, name):
963
"""Check if a version in consistent in this and other.
965
To be consistent it must have:
968
* the same direct parents (by name, not index, and disregarding
971
If present & correct return True;
972
if not present in self return False;
973
if inconsistent raise error."""
974
this_idx = self._name_map.get(name, -1)
976
if self._sha1s[this_idx] != other._sha1s[other_idx]:
977
raise errors.WeaveTextDiffers(name, self, other)
978
self_parents = self._parents[this_idx]
979
other_parents = other._parents[other_idx]
980
n1 = set([self._names[i] for i in self_parents])
981
n2 = set([other._names[i] for i in other_parents])
982
if not self._compatible_parents(n1, n2):
983
raise WeaveParentMismatch("inconsistent parents "
984
"for version {%s}: %s vs %s" % (name, n1, n2))
990
def _reweave(self, other, pb, msg):
991
"""Reweave self with other - internal helper for join().
993
:param other: The other weave to merge
994
:param pb: An optional progress bar, indicating how far done we are
995
:param msg: An optional message for the progress
997
new_weave = _reweave(self, other, pb=pb, msg=msg)
998
self._copy_weave_content(new_weave)
1000
def _copy_weave_content(self, otherweave):
1001
"""adsorb the content from otherweave."""
1002
for attr in self.__slots__:
1003
if attr != '_weave_name':
1004
setattr(self, attr, copy(getattr(otherweave, attr)))
1007
class WeaveFile(Weave):
1008
"""A WeaveFile represents a Weave on disk and writes on change."""
1010
WEAVE_SUFFIX = '.weave'
1012
def __init__(self, name, transport, filemode=None, create=False, access_mode='w'):
1013
"""Create a WeaveFile.
1015
:param create: If not True, only open an existing knit.
1017
super(WeaveFile, self).__init__(name, access_mode)
1018
self._transport = transport
1019
self._filemode = filemode
1021
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
1022
except errors.NoSuchFile:
1028
def _add_lines(self, version_id, parents, lines, parent_texts,
1029
left_matching_blocks, nostore_sha):
1030
"""Add a version and save the weave."""
1031
self.check_not_reserved_id(version_id)
1032
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
1033
parent_texts, left_matching_blocks, nostore_sha)
1037
def _clone_text(self, new_version_id, old_version_id, parents):
1038
"""See VersionedFile.clone_text."""
1039
super(WeaveFile, self)._clone_text(new_version_id, old_version_id, parents)
1042
def copy_to(self, name, transport):
1043
"""See VersionedFile.copy_to()."""
1044
# as we are all in memory always, just serialise to the new place.
1046
write_weave_v5(self, sio)
1048
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
1050
def create_empty(self, name, transport, filemode=None):
1051
return WeaveFile(name, transport, filemode, create=True)
1054
"""Save the weave."""
1055
self._check_write_ok()
1057
write_weave_v5(self, sio)
1059
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
1065
"""See VersionedFile.get_suffixes()."""
1066
return [WeaveFile.WEAVE_SUFFIX]
1068
def join(self, other, pb=None, msg=None, version_ids=None,
1069
ignore_missing=False):
1070
"""Join other into self and save."""
1071
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1075
def _reweave(wa, wb, pb=None, msg=None):
1076
"""Combine two weaves and return the result.
1078
This works even if a revision R has different parents in
1079
wa and wb. In the resulting weave all the parents are given.
1081
This is done by just building up a new weave, maintaining ordering
1082
of the versions in the two inputs. More efficient approaches
1083
might be possible but it should only be necessary to do
1084
this operation rarely, when a new previously ghost version is
1087
:param pb: An optional progress bar, indicating how far done we are
1088
:param msg: An optional message for the progress
1092
queue_a = range(wa.num_versions())
1093
queue_b = range(wb.num_versions())
1094
# first determine combined parents of all versions
1095
# map from version name -> all parent names
1096
combined_parents = _reweave_parent_graphs(wa, wb)
1097
mutter("combined parents: %r", combined_parents)
1098
order = topo_sort(combined_parents.iteritems())
1099
mutter("order to reweave: %r", order)
1104
for idx, name in enumerate(order):
1106
pb.update(msg, idx, len(order))
1107
if name in wa._name_map:
1108
lines = wa.get_lines(name)
1109
if name in wb._name_map:
1110
lines_b = wb.get_lines(name)
1111
if lines != lines_b:
1112
mutter('Weaves differ on content. rev_id {%s}', name)
1113
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1115
lines = list(difflib.unified_diff(lines, lines_b,
1116
wa._weave_name, wb._weave_name))
1117
mutter('lines:\n%s', ''.join(lines))
1118
raise errors.WeaveTextDiffers(name, wa, wb)
1120
lines = wb.get_lines(name)
1121
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1124
def _reweave_parent_graphs(wa, wb):
1125
"""Return combined parent ancestry for two weaves.
1127
Returned as a list of (version_name, set(parent_names))"""
1129
for weave in [wa, wb]:
1130
for idx, name in enumerate(weave._names):
1131
p = combined.setdefault(name, set())
1132
p.update(map(weave._idx_to_name, weave._parents[idx]))
1137
"""Show the weave's table-of-contents"""
1138
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1139
for i in (6, 50, 10, 10):
1142
for i in range(w.num_versions()):
1145
parent_str = ' '.join(map(str, w._parents[i]))
1146
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1150
def weave_stats(weave_file, pb):
1151
from bzrlib.weavefile import read_weave
1153
wf = file(weave_file, 'rb')
1155
# FIXME: doesn't work on pipes
1156
weave_size = wf.tell()
1160
for i in range(vers):
1161
pb.update('checking sizes', i, vers)
1162
for origin, lineno, line in w._extract([i]):
1167
print 'versions %9d' % vers
1168
print 'weave file %9d bytes' % weave_size
1169
print 'total contents %9d bytes' % total
1170
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1173
print 'average size %9d bytes' % avg
1174
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1178
print """bzr weave tool
1180
Experimental tool for weave algorithm.
1183
weave init WEAVEFILE
1184
Create an empty weave file
1185
weave get WEAVEFILE VERSION
1186
Write out specified version.
1187
weave check WEAVEFILE
1188
Check consistency of all versions.
1190
Display table of contents.
1191
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1192
Add NEWTEXT, with specified parent versions.
1193
weave annotate WEAVEFILE VERSION
1194
Display origin of each line.
1195
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1196
Auto-merge two versions and display conflicts.
1197
weave diff WEAVEFILE VERSION1 VERSION2
1198
Show differences between two versions.
1202
% weave init foo.weave
1204
% weave add foo.weave ver0 < foo.txt
1207
(create updated version)
1209
% weave get foo.weave 0 | diff -u - foo.txt
1210
% weave add foo.weave ver1 0 < foo.txt
1213
% weave get foo.weave 0 > foo.txt (create forked version)
1215
% weave add foo.weave ver2 0 < foo.txt
1218
% weave merge foo.weave 1 2 > foo.txt (merge them)
1219
% vi foo.txt (resolve conflicts)
1220
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1232
# in case we're run directly from the subdirectory
1233
sys.path.append('..')
1235
from bzrlib.weavefile import write_weave, read_weave
1236
from bzrlib.progress import ProgressBar
1251
return read_weave(file(argv[2], 'rb'))
1257
# at the moment, based on everything in the file
1259
parents = map(int, argv[4:])
1260
lines = sys.stdin.readlines()
1261
ver = w.add(name, parents, lines)
1262
write_weave(w, file(argv[2], 'wb'))
1263
print 'added version %r %d' % (name, ver)
1266
if os.path.exists(fn):
1267
raise IOError("file exists")
1269
write_weave(w, file(fn, 'wb'))
1270
elif cmd == 'get': # get one version
1272
sys.stdout.writelines(w.get_iter(int(argv[3])))
1277
v1, v2 = map(int, argv[3:5])
1280
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1281
'%s version %d' % (fn, v1),
1282
'%s version %d' % (fn, v2))
1283
sys.stdout.writelines(diff_gen)
1285
elif cmd == 'annotate':
1287
# newline is added to all lines regardless; too hard to get
1288
# reasonable formatting otherwise
1290
for origin, text in w.annotate(int(argv[3])):
1291
text = text.rstrip('\r\n')
1293
print ' | %s' % (text)
1295
print '%5d | %s' % (origin, text)
1301
elif cmd == 'stats':
1302
weave_stats(argv[2], ProgressBar())
1304
elif cmd == 'check':
1309
print '%d versions ok' % w.num_versions()
1311
elif cmd == 'inclusions':
1313
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1315
elif cmd == 'parents':
1317
print ' '.join(map(str, w._parents[int(argv[3])]))
1319
elif cmd == 'plan-merge':
1320
# replaced by 'bzr weave-plan-merge'
1322
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1324
print '%14s | %s' % (state, line),
1325
elif cmd == 'merge':
1326
# replaced by 'bzr weave-merge-text'
1328
p = w.plan_merge(int(argv[3]), int(argv[4]))
1329
sys.stdout.writelines(w.weave_merge(p))
1331
raise ValueError('unknown command %r' % cmd)
1334
if __name__ == '__main__':
1336
sys.exit(main(sys.argv))
1339
class InterWeave(InterVersionedFile):
1340
"""Optimised code paths for weave to weave operations."""
1342
_matching_file_from_factory = staticmethod(WeaveFile)
1343
_matching_file_to_factory = staticmethod(WeaveFile)
1346
def is_compatible(source, target):
1347
"""Be compatible with weaves."""
1349
return (isinstance(source, Weave) and
1350
isinstance(target, Weave))
1351
except AttributeError:
1354
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1355
"""See InterVersionedFile.join."""
1356
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1357
if self.target.versions() == [] and version_ids is None:
1358
self.target._copy_weave_content(self.source)
1361
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1362
except errors.WeaveParentMismatch:
1363
self.target._reweave(self.source, pb, msg)
1366
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
