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
"""See VersionedFile.add_lines."""
421
return self._add(version_id, lines, map(self._lookup, parents))
423
def _add(self, version_id, lines, parents, sha1=None):
424
"""Add a single text on top of the weave.
426
Returns the index number of the newly added version.
429
Symbolic name for this version.
430
(Typically the revision-id of the revision that added it.)
433
List or set of direct parent version numbers.
436
Sequence of lines to be added in the new version.
439
assert isinstance(version_id, basestring)
440
self._check_lines_not_unicode(lines)
441
self._check_lines_are_lines(lines)
443
sha1 = sha_strings(lines)
444
if version_id in self._name_map:
445
return self._check_repeated_add(version_id, parents, lines, sha1)
447
self._check_versions(parents)
448
## self._check_lines(lines)
449
new_version = len(self._parents)
451
# if we abort after here the (in-memory) weave will be corrupt because only
452
# some fields are updated
453
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
454
# - Robert Collins 20060226
455
self._parents.append(parents[:])
456
self._sha1s.append(sha1)
457
self._names.append(version_id)
458
self._name_map[version_id] = new_version
462
# special case; adding with no parents revision; can do
463
# this more quickly by just appending unconditionally.
464
# even more specially, if we're adding an empty text we
465
# need do nothing at all.
467
self._weave.append(('{', new_version))
468
self._weave.extend(lines)
469
self._weave.append(('}', None))
472
if len(parents) == 1:
473
pv = list(parents)[0]
474
if sha1 == self._sha1s[pv]:
475
# special case: same as the single parent
479
ancestors = self._inclusions(parents)
483
# basis a list of (origin, lineno, line)
486
for origin, lineno, line in self._extract(ancestors):
487
basis_lineno.append(lineno)
488
basis_lines.append(line)
490
# another small special case: a merge, producing the same text
492
if lines == basis_lines:
495
# add a sentinel, because we can also match against the final line
496
basis_lineno.append(len(self._weave))
498
# XXX: which line of the weave should we really consider
499
# matches the end of the file? the current code says it's the
500
# last line of the weave?
502
#print 'basis_lines:', basis_lines
503
#print 'new_lines: ', lines
505
s = self._matcher(None, basis_lines, lines)
507
# offset gives the number of lines that have been inserted
508
# into the weave up to the current point; if the original edit instruction
509
# says to change line A then we actually change (A+offset)
512
for tag, i1, i2, j1, j2 in s.get_opcodes():
513
# i1,i2 are given in offsets within basis_lines; we need to map them
514
# back to offsets within the entire weave
515
#print 'raw match', tag, i1, i2, j1, j2
519
i1 = basis_lineno[i1]
520
i2 = basis_lineno[i2]
522
assert 0 <= j1 <= j2 <= len(lines)
524
#print tag, i1, i2, j1, j2
526
# the deletion and insertion are handled separately.
527
# first delete the region.
529
self._weave.insert(i1+offset, ('[', new_version))
530
self._weave.insert(i2+offset+1, (']', new_version))
534
# there may have been a deletion spanning up to
535
# i2; we want to insert after this region to make sure
536
# we don't destroy ourselves
538
self._weave[i:i] = ([('{', new_version)]
541
offset += 2 + (j2 - j1)
544
def _clone_text(self, new_version_id, old_version_id, parents):
545
"""See VersionedFile.clone_text."""
546
old_lines = self.get_text(old_version_id)
547
self.add_lines(new_version_id, parents, old_lines)
549
def _inclusions(self, versions):
550
"""Return set of all ancestors of given version(s)."""
551
if not len(versions):
554
for v in xrange(max(versions), 0, -1):
556
# include all its parents
557
i.update(self._parents[v])
559
## except IndexError:
560
## raise ValueError("version %d not present in weave" % v)
562
def get_ancestry(self, version_ids, topo_sorted=True):
563
"""See VersionedFile.get_ancestry."""
564
if isinstance(version_ids, basestring):
565
version_ids = [version_ids]
566
i = self._inclusions([self._lookup(v) for v in version_ids])
567
return [self._idx_to_name(v) for v in i]
569
def _check_lines(self, text):
570
if not isinstance(text, list):
571
raise ValueError("text should be a list, not %s" % type(text))
574
if not isinstance(l, basestring):
575
raise ValueError("text line should be a string or unicode, not %s"
580
def _check_versions(self, indexes):
581
"""Check everything in the sequence of indexes is valid"""
586
raise IndexError("invalid version number %r" % i)
588
def _compatible_parents(self, my_parents, other_parents):
589
"""During join check that other_parents are joinable with my_parents.
591
Joinable is defined as 'is a subset of' - supersets may require
592
regeneration of diffs, but subsets do not.
594
return len(other_parents.difference(my_parents)) == 0
596
def annotate_iter(self, version_id):
597
"""Yield list of (version-id, line) pairs for the specified version.
599
The index indicates when the line originated in the weave."""
600
incls = [self._lookup(version_id)]
601
for origin, lineno, text in self._extract(incls):
602
yield self._idx_to_name(origin), text
604
def iter_lines_added_or_present_in_versions(self, version_ids=None,
606
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
607
if version_ids is None:
608
version_ids = self.versions()
609
version_ids = set(version_ids)
610
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
611
# if inserted not in version_ids then it was inserted before the
612
# versions we care about, but because weaves cannot represent ghosts
613
# properly, we do not filter down to that
614
# if inserted not in version_ids: continue
620
def _walk_internal(self, version_ids=None):
621
"""Helper method for weave actions."""
626
lineno = 0 # line of weave, 0-based
628
for l in self._weave:
629
if l.__class__ == tuple:
633
istack.append(self._names[v])
637
assert self._names[v] not in dset
638
dset.add(self._names[v])
640
dset.remove(self._names[v])
642
raise WeaveFormatError('unexpected instruction %r' % v)
644
assert l.__class__ in (str, unicode)
646
yield lineno, istack[-1], frozenset(dset), l
650
raise WeaveFormatError("unclosed insertion blocks "
651
"at end of weave: %s" % istack)
653
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
656
def plan_merge(self, ver_a, ver_b):
657
"""Return pseudo-annotation indicating how the two versions merge.
659
This is computed between versions a and b and their common
662
Weave lines present in none of them are skipped entirely.
664
inc_a = set(self.get_ancestry([ver_a]))
665
inc_b = set(self.get_ancestry([ver_b]))
666
inc_c = inc_a & inc_b
668
for lineno, insert, deleteset, line in self._walk_internal([ver_a, ver_b]):
669
if deleteset & inc_c:
670
# killed in parent; can't be in either a or b
671
# not relevant to our work
672
yield 'killed-base', line
673
elif insert in inc_c:
674
# was inserted in base
675
killed_a = bool(deleteset & inc_a)
676
killed_b = bool(deleteset & inc_b)
677
if killed_a and killed_b:
678
yield 'killed-both', line
680
yield 'killed-a', line
682
yield 'killed-b', line
684
yield 'unchanged', line
685
elif insert in inc_a:
686
if deleteset & inc_a:
687
yield 'ghost-a', line
691
elif insert in inc_b:
692
if deleteset & inc_b:
693
yield 'ghost-b', line
697
# not in either revision
698
yield 'irrelevant', line
700
yield 'unchanged', '' # terminator
702
def _extract(self, versions):
703
"""Yield annotation of lines in included set.
705
Yields a sequence of tuples (origin, lineno, text), where
706
origin is the origin version, lineno the index in the weave,
707
and text the text of the line.
709
The set typically but not necessarily corresponds to a version.
712
if not isinstance(i, int):
715
included = self._inclusions(versions)
721
lineno = 0 # line of weave, 0-based
727
WFE = WeaveFormatError
730
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
731
# +285282 0 1676.8040 1676.8040 +<isinstance>
732
# 1.6 seconds in 'isinstance'.
733
# changing the first isinstance:
734
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
735
# +140414 0 762.8050 762.8050 +<isinstance>
736
# note that the inline time actually dropped (less function calls)
737
# and total processing time was halved.
738
# we're still spending ~1/4 of the method in isinstance though.
739
# so lets hard code the acceptable string classes we expect:
740
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
741
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
743
# yay, down to ~1/4 the initial extract time, and our inline time
744
# has shrunk again, with isinstance no longer dominating.
745
# tweaking the stack inclusion test to use a set gives:
746
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
747
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
749
# - a 5% win, or possibly just noise. However with large istacks that
750
# 'in' test could dominate, so I'm leaving this change in place -
751
# when its fast enough to consider profiling big datasets we can review.
756
for l in self._weave:
757
if l.__class__ == tuple:
765
iset.remove(istack.pop())
776
assert l.__class__ in (str, unicode)
778
isactive = (not dset) and istack and (istack[-1] in included)
780
result.append((istack[-1], lineno, l))
783
raise WeaveFormatError("unclosed insertion blocks "
784
"at end of weave: %s" % istack)
786
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
790
def _maybe_lookup(self, name_or_index):
791
"""Convert possible symbolic name to index, or pass through indexes.
795
if isinstance(name_or_index, (int, long)):
798
return self._lookup(name_or_index)
800
def get_lines(self, version_id):
801
"""See VersionedFile.get_lines()."""
802
int_index = self._maybe_lookup(version_id)
803
result = [line for (origin, lineno, line) in self._extract([int_index])]
804
expected_sha1 = self._sha1s[int_index]
805
measured_sha1 = sha_strings(result)
806
if measured_sha1 != expected_sha1:
807
raise errors.WeaveInvalidChecksum(
808
'file %s, revision %s, expected: %s, measured %s'
809
% (self._weave_name, version_id,
810
expected_sha1, measured_sha1))
813
def get_sha1(self, version_id):
814
"""See VersionedFile.get_sha1()."""
815
return self._sha1s[self._lookup(version_id)]
817
def get_sha1s(self, version_ids):
818
"""See VersionedFile.get_sha1s()."""
819
return [self._sha1s[self._lookup(v)] for v in version_ids]
821
def num_versions(self):
822
"""How many versions are in this weave?"""
823
l = len(self._parents)
824
assert l == len(self._sha1s)
827
__len__ = num_versions
829
def check(self, progress_bar=None):
830
# TODO evaluate performance hit of using string sets in this routine.
831
# TODO: check no circular inclusions
832
# TODO: create a nested progress bar
833
for version in range(self.num_versions()):
834
inclusions = list(self._parents[version])
837
if inclusions[-1] >= version:
838
raise WeaveFormatError("invalid included version %d for index %d"
839
% (inclusions[-1], version))
841
# try extracting all versions; parallel extraction is used
842
nv = self.num_versions()
847
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
848
# The problem is that set membership is much more expensive
849
name = self._idx_to_name(i)
850
sha1s[name] = sha.new()
852
new_inc = set([name])
853
for p in self._parents[i]:
854
new_inc.update(inclusions[self._idx_to_name(p)])
856
assert set(new_inc) == set(self.get_ancestry(name)), \
857
'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
858
inclusions[name] = new_inc
860
nlines = len(self._weave)
862
update_text = 'checking weave'
864
short_name = os.path.basename(self._weave_name)
865
update_text = 'checking %s' % (short_name,)
866
update_text = update_text[:25]
868
for lineno, insert, deleteset, line in self._walk_internal():
870
progress_bar.update(update_text, lineno, nlines)
872
for name, name_inclusions in inclusions.items():
873
# The active inclusion must be an ancestor,
874
# and no ancestors must have deleted this line,
875
# because we don't support resurrection.
876
if (insert in name_inclusions) and not (deleteset & name_inclusions):
877
sha1s[name].update(line)
880
version = self._idx_to_name(i)
881
hd = sha1s[version].hexdigest()
882
expected = self._sha1s[i]
884
raise errors.WeaveInvalidChecksum(
885
"mismatched sha1 for version %s: "
886
"got %s, expected %s"
887
% (version, hd, expected))
889
# TODO: check insertions are properly nested, that there are
890
# no lines outside of insertion blocks, that deletions are
891
# properly paired, etc.
893
def _join(self, other, pb, msg, version_ids, ignore_missing):
894
"""Worker routine for join()."""
895
if not other.versions():
896
return # nothing to update, easy
899
# versions is never none, InterWeave checks this.
902
# two loops so that we do not change ourselves before verifying it
904
# work through in index order to make sure we get all dependencies
907
# get the selected versions only that are in other.versions.
908
version_ids = set(other.versions()).intersection(set(version_ids))
909
# pull in the referenced graph.
910
version_ids = other.get_ancestry(version_ids)
911
pending_graph = [(version, other.get_parents(version)) for
912
version in version_ids]
913
for name in topo_sort(pending_graph):
914
other_idx = other._name_map[name]
915
# returns True if we have it, False if we need it.
916
if not self._check_version_consistent(other, other_idx, name):
917
names_to_join.append((other_idx, name))
926
for other_idx, name in names_to_join:
927
# TODO: If all the parents of the other version are already
928
# present then we can avoid some work by just taking the delta
929
# and adjusting the offsets.
930
new_parents = self._imported_parents(other, other_idx)
931
sha1 = other._sha1s[other_idx]
936
pb.update(msg, merged, len(names_to_join))
938
lines = other.get_lines(other_idx)
939
self._add(name, lines, new_parents, sha1)
941
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
942
merged, processed, self._weave_name, time.time()-time0))
944
def _imported_parents(self, other, other_idx):
945
"""Return list of parents in self corresponding to indexes in other."""
947
for parent_idx in other._parents[other_idx]:
948
parent_name = other._names[parent_idx]
949
if parent_name not in self._name_map:
950
# should not be possible
951
raise WeaveError("missing parent {%s} of {%s} in %r"
952
% (parent_name, other._name_map[other_idx], self))
953
new_parents.append(self._name_map[parent_name])
956
def _check_version_consistent(self, other, other_idx, name):
957
"""Check if a version in consistent in this and other.
959
To be consistent it must have:
962
* the same direct parents (by name, not index, and disregarding
965
If present & correct return True;
966
if not present in self return False;
967
if inconsistent raise error."""
968
this_idx = self._name_map.get(name, -1)
970
if self._sha1s[this_idx] != other._sha1s[other_idx]:
971
raise errors.WeaveTextDiffers(name, self, other)
972
self_parents = self._parents[this_idx]
973
other_parents = other._parents[other_idx]
974
n1 = set([self._names[i] for i in self_parents])
975
n2 = set([other._names[i] for i in other_parents])
976
if not self._compatible_parents(n1, n2):
977
raise WeaveParentMismatch("inconsistent parents "
978
"for version {%s}: %s vs %s" % (name, n1, n2))
984
def _reweave(self, other, pb, msg):
985
"""Reweave self with other - internal helper for join().
987
:param other: The other weave to merge
988
:param pb: An optional progress bar, indicating how far done we are
989
:param msg: An optional message for the progress
991
new_weave = _reweave(self, other, pb=pb, msg=msg)
992
self._copy_weave_content(new_weave)
994
def _copy_weave_content(self, otherweave):
995
"""adsorb the content from otherweave."""
996
for attr in self.__slots__:
997
if attr != '_weave_name':
998
setattr(self, attr, copy(getattr(otherweave, attr)))
1001
class WeaveFile(Weave):
1002
"""A WeaveFile represents a Weave on disk and writes on change."""
1004
WEAVE_SUFFIX = '.weave'
1006
def __init__(self, name, transport, filemode=None, create=False, access_mode='w'):
1007
"""Create a WeaveFile.
1009
:param create: If not True, only open an existing knit.
1011
super(WeaveFile, self).__init__(name, access_mode)
1012
self._transport = transport
1013
self._filemode = filemode
1015
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
1016
except errors.NoSuchFile:
1022
def _add_lines(self, version_id, parents, lines, parent_texts):
1023
"""Add a version and save the weave."""
1024
self.check_not_reserved_id(version_id)
1025
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
1030
def _clone_text(self, new_version_id, old_version_id, parents):
1031
"""See VersionedFile.clone_text."""
1032
super(WeaveFile, self)._clone_text(new_version_id, old_version_id, parents)
1035
def copy_to(self, name, transport):
1036
"""See VersionedFile.copy_to()."""
1037
# as we are all in memory always, just serialise to the new place.
1039
write_weave_v5(self, sio)
1041
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
1043
def create_empty(self, name, transport, filemode=None):
1044
return WeaveFile(name, transport, filemode, create=True)
1047
"""Save the weave."""
1048
self._check_write_ok()
1050
write_weave_v5(self, sio)
1052
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
1058
"""See VersionedFile.get_suffixes()."""
1059
return [WeaveFile.WEAVE_SUFFIX]
1061
def join(self, other, pb=None, msg=None, version_ids=None,
1062
ignore_missing=False):
1063
"""Join other into self and save."""
1064
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1068
def _reweave(wa, wb, pb=None, msg=None):
1069
"""Combine two weaves and return the result.
1071
This works even if a revision R has different parents in
1072
wa and wb. In the resulting weave all the parents are given.
1074
This is done by just building up a new weave, maintaining ordering
1075
of the versions in the two inputs. More efficient approaches
1076
might be possible but it should only be necessary to do
1077
this operation rarely, when a new previously ghost version is
1080
:param pb: An optional progress bar, indicating how far done we are
1081
:param msg: An optional message for the progress
1085
queue_a = range(wa.num_versions())
1086
queue_b = range(wb.num_versions())
1087
# first determine combined parents of all versions
1088
# map from version name -> all parent names
1089
combined_parents = _reweave_parent_graphs(wa, wb)
1090
mutter("combined parents: %r", combined_parents)
1091
order = topo_sort(combined_parents.iteritems())
1092
mutter("order to reweave: %r", order)
1097
for idx, name in enumerate(order):
1099
pb.update(msg, idx, len(order))
1100
if name in wa._name_map:
1101
lines = wa.get_lines(name)
1102
if name in wb._name_map:
1103
lines_b = wb.get_lines(name)
1104
if lines != lines_b:
1105
mutter('Weaves differ on content. rev_id {%s}', name)
1106
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1108
lines = list(difflib.unified_diff(lines, lines_b,
1109
wa._weave_name, wb._weave_name))
1110
mutter('lines:\n%s', ''.join(lines))
1111
raise errors.WeaveTextDiffers(name, wa, wb)
1113
lines = wb.get_lines(name)
1114
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1117
def _reweave_parent_graphs(wa, wb):
1118
"""Return combined parent ancestry for two weaves.
1120
Returned as a list of (version_name, set(parent_names))"""
1122
for weave in [wa, wb]:
1123
for idx, name in enumerate(weave._names):
1124
p = combined.setdefault(name, set())
1125
p.update(map(weave._idx_to_name, weave._parents[idx]))
1130
"""Show the weave's table-of-contents"""
1131
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1132
for i in (6, 50, 10, 10):
1135
for i in range(w.num_versions()):
1138
parent_str = ' '.join(map(str, w._parents[i]))
1139
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1143
def weave_stats(weave_file, pb):
1144
from bzrlib.weavefile import read_weave
1146
wf = file(weave_file, 'rb')
1148
# FIXME: doesn't work on pipes
1149
weave_size = wf.tell()
1153
for i in range(vers):
1154
pb.update('checking sizes', i, vers)
1155
for origin, lineno, line in w._extract([i]):
1160
print 'versions %9d' % vers
1161
print 'weave file %9d bytes' % weave_size
1162
print 'total contents %9d bytes' % total
1163
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1166
print 'average size %9d bytes' % avg
1167
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1171
print """bzr weave tool
1173
Experimental tool for weave algorithm.
1176
weave init WEAVEFILE
1177
Create an empty weave file
1178
weave get WEAVEFILE VERSION
1179
Write out specified version.
1180
weave check WEAVEFILE
1181
Check consistency of all versions.
1183
Display table of contents.
1184
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1185
Add NEWTEXT, with specified parent versions.
1186
weave annotate WEAVEFILE VERSION
1187
Display origin of each line.
1188
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1189
Auto-merge two versions and display conflicts.
1190
weave diff WEAVEFILE VERSION1 VERSION2
1191
Show differences between two versions.
1195
% weave init foo.weave
1197
% weave add foo.weave ver0 < foo.txt
1200
(create updated version)
1202
% weave get foo.weave 0 | diff -u - foo.txt
1203
% weave add foo.weave ver1 0 < foo.txt
1206
% weave get foo.weave 0 > foo.txt (create forked version)
1208
% weave add foo.weave ver2 0 < foo.txt
1211
% weave merge foo.weave 1 2 > foo.txt (merge them)
1212
% vi foo.txt (resolve conflicts)
1213
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1225
# in case we're run directly from the subdirectory
1226
sys.path.append('..')
1228
from bzrlib.weavefile import write_weave, read_weave
1229
from bzrlib.progress import ProgressBar
1244
return read_weave(file(argv[2], 'rb'))
1250
# at the moment, based on everything in the file
1252
parents = map(int, argv[4:])
1253
lines = sys.stdin.readlines()
1254
ver = w.add(name, parents, lines)
1255
write_weave(w, file(argv[2], 'wb'))
1256
print 'added version %r %d' % (name, ver)
1259
if os.path.exists(fn):
1260
raise IOError("file exists")
1262
write_weave(w, file(fn, 'wb'))
1263
elif cmd == 'get': # get one version
1265
sys.stdout.writelines(w.get_iter(int(argv[3])))
1270
v1, v2 = map(int, argv[3:5])
1273
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1274
'%s version %d' % (fn, v1),
1275
'%s version %d' % (fn, v2))
1276
sys.stdout.writelines(diff_gen)
1278
elif cmd == 'annotate':
1280
# newline is added to all lines regardless; too hard to get
1281
# reasonable formatting otherwise
1283
for origin, text in w.annotate(int(argv[3])):
1284
text = text.rstrip('\r\n')
1286
print ' | %s' % (text)
1288
print '%5d | %s' % (origin, text)
1294
elif cmd == 'stats':
1295
weave_stats(argv[2], ProgressBar())
1297
elif cmd == 'check':
1302
print '%d versions ok' % w.num_versions()
1304
elif cmd == 'inclusions':
1306
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1308
elif cmd == 'parents':
1310
print ' '.join(map(str, w._parents[int(argv[3])]))
1312
elif cmd == 'plan-merge':
1313
# replaced by 'bzr weave-plan-merge'
1315
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1317
print '%14s | %s' % (state, line),
1318
elif cmd == 'merge':
1319
# replaced by 'bzr weave-merge-text'
1321
p = w.plan_merge(int(argv[3]), int(argv[4]))
1322
sys.stdout.writelines(w.weave_merge(p))
1324
raise ValueError('unknown command %r' % cmd)
1327
if __name__ == '__main__':
1329
sys.exit(main(sys.argv))
1332
class InterWeave(InterVersionedFile):
1333
"""Optimised code paths for weave to weave operations."""
1335
_matching_file_from_factory = staticmethod(WeaveFile)
1336
_matching_file_to_factory = staticmethod(WeaveFile)
1339
def is_compatible(source, target):
1340
"""Be compatible with weaves."""
1342
return (isinstance(source, Weave) and
1343
isinstance(target, Weave))
1344
except AttributeError:
1347
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1348
"""See InterVersionedFile.join."""
1349
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1350
if self.target.versions() == [] and version_ids is None:
1351
self.target._copy_weave_content(self.source)
1354
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1355
except errors.WeaveParentMismatch:
1356
self.target._reweave(self.source, pb, msg)
1359
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
