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=None):
421
"""See VersionedFile.add_lines."""
422
idx = self._add(version_id, lines, map(self._lookup, parents))
423
return sha_strings(lines), sum(map(len, lines)), idx
425
def _add(self, version_id, lines, parents, sha1=None):
426
"""Add a single text on top of the weave.
428
Returns the index number of the newly added version.
431
Symbolic name for this version.
432
(Typically the revision-id of the revision that added it.)
435
List or set of direct parent version numbers.
438
Sequence of lines to be added in the new version.
441
assert isinstance(version_id, basestring)
442
self._check_lines_not_unicode(lines)
443
self._check_lines_are_lines(lines)
445
sha1 = sha_strings(lines)
446
if version_id in self._name_map:
447
return self._check_repeated_add(version_id, parents, lines, sha1)
449
self._check_versions(parents)
450
## self._check_lines(lines)
451
new_version = len(self._parents)
453
# if we abort after here the (in-memory) weave will be corrupt because only
454
# some fields are updated
455
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
456
# - Robert Collins 20060226
457
self._parents.append(parents[:])
458
self._sha1s.append(sha1)
459
self._names.append(version_id)
460
self._name_map[version_id] = new_version
464
# special case; adding with no parents revision; can do
465
# this more quickly by just appending unconditionally.
466
# even more specially, if we're adding an empty text we
467
# need do nothing at all.
469
self._weave.append(('{', new_version))
470
self._weave.extend(lines)
471
self._weave.append(('}', None))
474
if len(parents) == 1:
475
pv = list(parents)[0]
476
if sha1 == self._sha1s[pv]:
477
# special case: same as the single parent
481
ancestors = self._inclusions(parents)
485
# basis a list of (origin, lineno, line)
488
for origin, lineno, line in self._extract(ancestors):
489
basis_lineno.append(lineno)
490
basis_lines.append(line)
492
# another small special case: a merge, producing the same text
494
if lines == basis_lines:
497
# add a sentinel, because we can also match against the final line
498
basis_lineno.append(len(self._weave))
500
# XXX: which line of the weave should we really consider
501
# matches the end of the file? the current code says it's the
502
# last line of the weave?
504
#print 'basis_lines:', basis_lines
505
#print 'new_lines: ', lines
507
s = self._matcher(None, basis_lines, lines)
509
# offset gives the number of lines that have been inserted
510
# into the weave up to the current point; if the original edit instruction
511
# says to change line A then we actually change (A+offset)
514
for tag, i1, i2, j1, j2 in s.get_opcodes():
515
# i1,i2 are given in offsets within basis_lines; we need to map them
516
# back to offsets within the entire weave
517
#print 'raw match', tag, i1, i2, j1, j2
521
i1 = basis_lineno[i1]
522
i2 = basis_lineno[i2]
524
assert 0 <= j1 <= j2 <= len(lines)
526
#print tag, i1, i2, j1, j2
528
# the deletion and insertion are handled separately.
529
# first delete the region.
531
self._weave.insert(i1+offset, ('[', new_version))
532
self._weave.insert(i2+offset+1, (']', new_version))
536
# there may have been a deletion spanning up to
537
# i2; we want to insert after this region to make sure
538
# we don't destroy ourselves
540
self._weave[i:i] = ([('{', new_version)]
543
offset += 2 + (j2 - j1)
546
def _clone_text(self, new_version_id, old_version_id, parents):
547
"""See VersionedFile.clone_text."""
548
old_lines = self.get_text(old_version_id)
549
self.add_lines(new_version_id, parents, old_lines)
551
def _inclusions(self, versions):
552
"""Return set of all ancestors of given version(s)."""
553
if not len(versions):
556
for v in xrange(max(versions), 0, -1):
558
# include all its parents
559
i.update(self._parents[v])
561
## except IndexError:
562
## raise ValueError("version %d not present in weave" % v)
564
def get_ancestry(self, version_ids, topo_sorted=True):
565
"""See VersionedFile.get_ancestry."""
566
if isinstance(version_ids, basestring):
567
version_ids = [version_ids]
568
i = self._inclusions([self._lookup(v) for v in version_ids])
569
return [self._idx_to_name(v) for v in i]
571
def _check_lines(self, text):
572
if not isinstance(text, list):
573
raise ValueError("text should be a list, not %s" % type(text))
576
if not isinstance(l, basestring):
577
raise ValueError("text line should be a string or unicode, not %s"
582
def _check_versions(self, indexes):
583
"""Check everything in the sequence of indexes is valid"""
588
raise IndexError("invalid version number %r" % i)
590
def _compatible_parents(self, my_parents, other_parents):
591
"""During join check that other_parents are joinable with my_parents.
593
Joinable is defined as 'is a subset of' - supersets may require
594
regeneration of diffs, but subsets do not.
596
return len(other_parents.difference(my_parents)) == 0
598
def annotate_iter(self, version_id):
599
"""Yield list of (version-id, line) pairs for the specified version.
601
The index indicates when the line originated in the weave."""
602
incls = [self._lookup(version_id)]
603
for origin, lineno, text in self._extract(incls):
604
yield self._idx_to_name(origin), text
606
def iter_lines_added_or_present_in_versions(self, version_ids=None,
608
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
609
if version_ids is None:
610
version_ids = self.versions()
611
version_ids = set(version_ids)
612
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
613
# if inserted not in version_ids then it was inserted before the
614
# versions we care about, but because weaves cannot represent ghosts
615
# properly, we do not filter down to that
616
# if inserted not in version_ids: continue
622
def _walk_internal(self, version_ids=None):
623
"""Helper method for weave actions."""
628
lineno = 0 # line of weave, 0-based
630
for l in self._weave:
631
if l.__class__ == tuple:
635
istack.append(self._names[v])
639
assert self._names[v] not in dset
640
dset.add(self._names[v])
642
dset.remove(self._names[v])
644
raise WeaveFormatError('unexpected instruction %r' % v)
646
assert l.__class__ in (str, unicode)
648
yield lineno, istack[-1], frozenset(dset), l
652
raise WeaveFormatError("unclosed insertion blocks "
653
"at end of weave: %s" % istack)
655
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
658
def plan_merge(self, ver_a, ver_b):
659
"""Return pseudo-annotation indicating how the two versions merge.
661
This is computed between versions a and b and their common
664
Weave lines present in none of them are skipped entirely.
666
inc_a = set(self.get_ancestry([ver_a]))
667
inc_b = set(self.get_ancestry([ver_b]))
668
inc_c = inc_a & inc_b
670
for lineno, insert, deleteset, line in self._walk_internal([ver_a, ver_b]):
671
if deleteset & inc_c:
672
# killed in parent; can't be in either a or b
673
# not relevant to our work
674
yield 'killed-base', line
675
elif insert in inc_c:
676
# was inserted in base
677
killed_a = bool(deleteset & inc_a)
678
killed_b = bool(deleteset & inc_b)
679
if killed_a and killed_b:
680
yield 'killed-both', line
682
yield 'killed-a', line
684
yield 'killed-b', line
686
yield 'unchanged', line
687
elif insert in inc_a:
688
if deleteset & inc_a:
689
yield 'ghost-a', line
693
elif insert in inc_b:
694
if deleteset & inc_b:
695
yield 'ghost-b', line
699
# not in either revision
700
yield 'irrelevant', line
702
yield 'unchanged', '' # terminator
704
def _extract(self, versions):
705
"""Yield annotation of lines in included set.
707
Yields a sequence of tuples (origin, lineno, text), where
708
origin is the origin version, lineno the index in the weave,
709
and text the text of the line.
711
The set typically but not necessarily corresponds to a version.
714
if not isinstance(i, int):
717
included = self._inclusions(versions)
723
lineno = 0 # line of weave, 0-based
729
WFE = WeaveFormatError
732
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
733
# +285282 0 1676.8040 1676.8040 +<isinstance>
734
# 1.6 seconds in 'isinstance'.
735
# changing the first isinstance:
736
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
737
# +140414 0 762.8050 762.8050 +<isinstance>
738
# note that the inline time actually dropped (less function calls)
739
# and total processing time was halved.
740
# we're still spending ~1/4 of the method in isinstance though.
741
# so lets hard code the acceptable string classes we expect:
742
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
743
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
745
# yay, down to ~1/4 the initial extract time, and our inline time
746
# has shrunk again, with isinstance no longer dominating.
747
# tweaking the stack inclusion test to use a set gives:
748
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
749
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
751
# - a 5% win, or possibly just noise. However with large istacks that
752
# 'in' test could dominate, so I'm leaving this change in place -
753
# when its fast enough to consider profiling big datasets we can review.
758
for l in self._weave:
759
if l.__class__ == tuple:
767
iset.remove(istack.pop())
778
assert l.__class__ in (str, unicode)
780
isactive = (not dset) and istack and (istack[-1] in included)
782
result.append((istack[-1], lineno, l))
785
raise WeaveFormatError("unclosed insertion blocks "
786
"at end of weave: %s" % istack)
788
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
792
def _maybe_lookup(self, name_or_index):
793
"""Convert possible symbolic name to index, or pass through indexes.
797
if isinstance(name_or_index, (int, long)):
800
return self._lookup(name_or_index)
802
def get_lines(self, version_id):
803
"""See VersionedFile.get_lines()."""
804
int_index = self._maybe_lookup(version_id)
805
result = [line for (origin, lineno, line) in self._extract([int_index])]
806
expected_sha1 = self._sha1s[int_index]
807
measured_sha1 = sha_strings(result)
808
if measured_sha1 != expected_sha1:
809
raise errors.WeaveInvalidChecksum(
810
'file %s, revision %s, expected: %s, measured %s'
811
% (self._weave_name, version_id,
812
expected_sha1, measured_sha1))
815
def get_sha1(self, version_id):
816
"""See VersionedFile.get_sha1()."""
817
return self._sha1s[self._lookup(version_id)]
819
def get_sha1s(self, version_ids):
820
"""See VersionedFile.get_sha1s()."""
821
return [self._sha1s[self._lookup(v)] for v in version_ids]
823
def num_versions(self):
824
"""How many versions are in this weave?"""
825
l = len(self._parents)
826
assert l == len(self._sha1s)
829
__len__ = num_versions
831
def check(self, progress_bar=None):
832
# TODO evaluate performance hit of using string sets in this routine.
833
# TODO: check no circular inclusions
834
# TODO: create a nested progress bar
835
for version in range(self.num_versions()):
836
inclusions = list(self._parents[version])
839
if inclusions[-1] >= version:
840
raise WeaveFormatError("invalid included version %d for index %d"
841
% (inclusions[-1], version))
843
# try extracting all versions; parallel extraction is used
844
nv = self.num_versions()
849
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
850
# The problem is that set membership is much more expensive
851
name = self._idx_to_name(i)
852
sha1s[name] = sha.new()
854
new_inc = set([name])
855
for p in self._parents[i]:
856
new_inc.update(inclusions[self._idx_to_name(p)])
858
assert set(new_inc) == set(self.get_ancestry(name)), \
859
'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
860
inclusions[name] = new_inc
862
nlines = len(self._weave)
864
update_text = 'checking weave'
866
short_name = os.path.basename(self._weave_name)
867
update_text = 'checking %s' % (short_name,)
868
update_text = update_text[:25]
870
for lineno, insert, deleteset, line in self._walk_internal():
872
progress_bar.update(update_text, lineno, nlines)
874
for name, name_inclusions in inclusions.items():
875
# The active inclusion must be an ancestor,
876
# and no ancestors must have deleted this line,
877
# because we don't support resurrection.
878
if (insert in name_inclusions) and not (deleteset & name_inclusions):
879
sha1s[name].update(line)
882
version = self._idx_to_name(i)
883
hd = sha1s[version].hexdigest()
884
expected = self._sha1s[i]
886
raise errors.WeaveInvalidChecksum(
887
"mismatched sha1 for version %s: "
888
"got %s, expected %s"
889
% (version, hd, expected))
891
# TODO: check insertions are properly nested, that there are
892
# no lines outside of insertion blocks, that deletions are
893
# properly paired, etc.
895
def _join(self, other, pb, msg, version_ids, ignore_missing):
896
"""Worker routine for join()."""
897
if not other.versions():
898
return # nothing to update, easy
901
# versions is never none, InterWeave checks this.
904
# two loops so that we do not change ourselves before verifying it
906
# work through in index order to make sure we get all dependencies
909
# get the selected versions only that are in other.versions.
910
version_ids = set(other.versions()).intersection(set(version_ids))
911
# pull in the referenced graph.
912
version_ids = other.get_ancestry(version_ids)
913
pending_graph = [(version, other.get_parents(version)) for
914
version in version_ids]
915
for name in topo_sort(pending_graph):
916
other_idx = other._name_map[name]
917
# returns True if we have it, False if we need it.
918
if not self._check_version_consistent(other, other_idx, name):
919
names_to_join.append((other_idx, name))
928
for other_idx, name in names_to_join:
929
# TODO: If all the parents of the other version are already
930
# present then we can avoid some work by just taking the delta
931
# and adjusting the offsets.
932
new_parents = self._imported_parents(other, other_idx)
933
sha1 = other._sha1s[other_idx]
938
pb.update(msg, merged, len(names_to_join))
940
lines = other.get_lines(other_idx)
941
self._add(name, lines, new_parents, sha1)
943
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
944
merged, processed, self._weave_name, time.time()-time0))
946
def _imported_parents(self, other, other_idx):
947
"""Return list of parents in self corresponding to indexes in other."""
949
for parent_idx in other._parents[other_idx]:
950
parent_name = other._names[parent_idx]
951
if parent_name not in self._name_map:
952
# should not be possible
953
raise WeaveError("missing parent {%s} of {%s} in %r"
954
% (parent_name, other._name_map[other_idx], self))
955
new_parents.append(self._name_map[parent_name])
958
def _check_version_consistent(self, other, other_idx, name):
959
"""Check if a version in consistent in this and other.
961
To be consistent it must have:
964
* the same direct parents (by name, not index, and disregarding
967
If present & correct return True;
968
if not present in self return False;
969
if inconsistent raise error."""
970
this_idx = self._name_map.get(name, -1)
972
if self._sha1s[this_idx] != other._sha1s[other_idx]:
973
raise errors.WeaveTextDiffers(name, self, other)
974
self_parents = self._parents[this_idx]
975
other_parents = other._parents[other_idx]
976
n1 = set([self._names[i] for i in self_parents])
977
n2 = set([other._names[i] for i in other_parents])
978
if not self._compatible_parents(n1, n2):
979
raise WeaveParentMismatch("inconsistent parents "
980
"for version {%s}: %s vs %s" % (name, n1, n2))
986
def _reweave(self, other, pb, msg):
987
"""Reweave self with other - internal helper for join().
989
:param other: The other weave to merge
990
:param pb: An optional progress bar, indicating how far done we are
991
:param msg: An optional message for the progress
993
new_weave = _reweave(self, other, pb=pb, msg=msg)
994
self._copy_weave_content(new_weave)
996
def _copy_weave_content(self, otherweave):
997
"""adsorb the content from otherweave."""
998
for attr in self.__slots__:
999
if attr != '_weave_name':
1000
setattr(self, attr, copy(getattr(otherweave, attr)))
1003
class WeaveFile(Weave):
1004
"""A WeaveFile represents a Weave on disk and writes on change."""
1006
WEAVE_SUFFIX = '.weave'
1008
def __init__(self, name, transport, filemode=None, create=False, access_mode='w'):
1009
"""Create a WeaveFile.
1011
:param create: If not True, only open an existing knit.
1013
super(WeaveFile, self).__init__(name, access_mode)
1014
self._transport = transport
1015
self._filemode = filemode
1017
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
1018
except errors.NoSuchFile:
1024
def _add_lines(self, version_id, parents, lines, parent_texts,
1025
left_matching_blocks=None):
1026
"""Add a version and save the weave."""
1027
self.check_not_reserved_id(version_id)
1028
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
1033
def _clone_text(self, new_version_id, old_version_id, parents):
1034
"""See VersionedFile.clone_text."""
1035
super(WeaveFile, self)._clone_text(new_version_id, old_version_id, parents)
1038
def copy_to(self, name, transport):
1039
"""See VersionedFile.copy_to()."""
1040
# as we are all in memory always, just serialise to the new place.
1042
write_weave_v5(self, sio)
1044
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
1046
def create_empty(self, name, transport, filemode=None):
1047
return WeaveFile(name, transport, filemode, create=True)
1050
"""Save the weave."""
1051
self._check_write_ok()
1053
write_weave_v5(self, sio)
1055
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
1061
"""See VersionedFile.get_suffixes()."""
1062
return [WeaveFile.WEAVE_SUFFIX]
1064
def join(self, other, pb=None, msg=None, version_ids=None,
1065
ignore_missing=False):
1066
"""Join other into self and save."""
1067
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1071
def _reweave(wa, wb, pb=None, msg=None):
1072
"""Combine two weaves and return the result.
1074
This works even if a revision R has different parents in
1075
wa and wb. In the resulting weave all the parents are given.
1077
This is done by just building up a new weave, maintaining ordering
1078
of the versions in the two inputs. More efficient approaches
1079
might be possible but it should only be necessary to do
1080
this operation rarely, when a new previously ghost version is
1083
:param pb: An optional progress bar, indicating how far done we are
1084
:param msg: An optional message for the progress
1088
queue_a = range(wa.num_versions())
1089
queue_b = range(wb.num_versions())
1090
# first determine combined parents of all versions
1091
# map from version name -> all parent names
1092
combined_parents = _reweave_parent_graphs(wa, wb)
1093
mutter("combined parents: %r", combined_parents)
1094
order = topo_sort(combined_parents.iteritems())
1095
mutter("order to reweave: %r", order)
1100
for idx, name in enumerate(order):
1102
pb.update(msg, idx, len(order))
1103
if name in wa._name_map:
1104
lines = wa.get_lines(name)
1105
if name in wb._name_map:
1106
lines_b = wb.get_lines(name)
1107
if lines != lines_b:
1108
mutter('Weaves differ on content. rev_id {%s}', name)
1109
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1111
lines = list(difflib.unified_diff(lines, lines_b,
1112
wa._weave_name, wb._weave_name))
1113
mutter('lines:\n%s', ''.join(lines))
1114
raise errors.WeaveTextDiffers(name, wa, wb)
1116
lines = wb.get_lines(name)
1117
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1120
def _reweave_parent_graphs(wa, wb):
1121
"""Return combined parent ancestry for two weaves.
1123
Returned as a list of (version_name, set(parent_names))"""
1125
for weave in [wa, wb]:
1126
for idx, name in enumerate(weave._names):
1127
p = combined.setdefault(name, set())
1128
p.update(map(weave._idx_to_name, weave._parents[idx]))
1133
"""Show the weave's table-of-contents"""
1134
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1135
for i in (6, 50, 10, 10):
1138
for i in range(w.num_versions()):
1141
parent_str = ' '.join(map(str, w._parents[i]))
1142
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1146
def weave_stats(weave_file, pb):
1147
from bzrlib.weavefile import read_weave
1149
wf = file(weave_file, 'rb')
1151
# FIXME: doesn't work on pipes
1152
weave_size = wf.tell()
1156
for i in range(vers):
1157
pb.update('checking sizes', i, vers)
1158
for origin, lineno, line in w._extract([i]):
1163
print 'versions %9d' % vers
1164
print 'weave file %9d bytes' % weave_size
1165
print 'total contents %9d bytes' % total
1166
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1169
print 'average size %9d bytes' % avg
1170
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1174
print """bzr weave tool
1176
Experimental tool for weave algorithm.
1179
weave init WEAVEFILE
1180
Create an empty weave file
1181
weave get WEAVEFILE VERSION
1182
Write out specified version.
1183
weave check WEAVEFILE
1184
Check consistency of all versions.
1186
Display table of contents.
1187
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1188
Add NEWTEXT, with specified parent versions.
1189
weave annotate WEAVEFILE VERSION
1190
Display origin of each line.
1191
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1192
Auto-merge two versions and display conflicts.
1193
weave diff WEAVEFILE VERSION1 VERSION2
1194
Show differences between two versions.
1198
% weave init foo.weave
1200
% weave add foo.weave ver0 < foo.txt
1203
(create updated version)
1205
% weave get foo.weave 0 | diff -u - foo.txt
1206
% weave add foo.weave ver1 0 < foo.txt
1209
% weave get foo.weave 0 > foo.txt (create forked version)
1211
% weave add foo.weave ver2 0 < foo.txt
1214
% weave merge foo.weave 1 2 > foo.txt (merge them)
1215
% vi foo.txt (resolve conflicts)
1216
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1228
# in case we're run directly from the subdirectory
1229
sys.path.append('..')
1231
from bzrlib.weavefile import write_weave, read_weave
1232
from bzrlib.progress import ProgressBar
1247
return read_weave(file(argv[2], 'rb'))
1253
# at the moment, based on everything in the file
1255
parents = map(int, argv[4:])
1256
lines = sys.stdin.readlines()
1257
ver = w.add(name, parents, lines)
1258
write_weave(w, file(argv[2], 'wb'))
1259
print 'added version %r %d' % (name, ver)
1262
if os.path.exists(fn):
1263
raise IOError("file exists")
1265
write_weave(w, file(fn, 'wb'))
1266
elif cmd == 'get': # get one version
1268
sys.stdout.writelines(w.get_iter(int(argv[3])))
1273
v1, v2 = map(int, argv[3:5])
1276
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1277
'%s version %d' % (fn, v1),
1278
'%s version %d' % (fn, v2))
1279
sys.stdout.writelines(diff_gen)
1281
elif cmd == 'annotate':
1283
# newline is added to all lines regardless; too hard to get
1284
# reasonable formatting otherwise
1286
for origin, text in w.annotate(int(argv[3])):
1287
text = text.rstrip('\r\n')
1289
print ' | %s' % (text)
1291
print '%5d | %s' % (origin, text)
1297
elif cmd == 'stats':
1298
weave_stats(argv[2], ProgressBar())
1300
elif cmd == 'check':
1305
print '%d versions ok' % w.num_versions()
1307
elif cmd == 'inclusions':
1309
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1311
elif cmd == 'parents':
1313
print ' '.join(map(str, w._parents[int(argv[3])]))
1315
elif cmd == 'plan-merge':
1316
# replaced by 'bzr weave-plan-merge'
1318
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1320
print '%14s | %s' % (state, line),
1321
elif cmd == 'merge':
1322
# replaced by 'bzr weave-merge-text'
1324
p = w.plan_merge(int(argv[3]), int(argv[4]))
1325
sys.stdout.writelines(w.weave_merge(p))
1327
raise ValueError('unknown command %r' % cmd)
1330
if __name__ == '__main__':
1332
sys.exit(main(sys.argv))
1335
class InterWeave(InterVersionedFile):
1336
"""Optimised code paths for weave to weave operations."""
1338
_matching_file_from_factory = staticmethod(WeaveFile)
1339
_matching_file_to_factory = staticmethod(WeaveFile)
1342
def is_compatible(source, target):
1343
"""Be compatible with weaves."""
1345
return (isinstance(source, Weave) and
1346
isinstance(target, Weave))
1347
except AttributeError:
1350
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1351
"""See InterVersionedFile.join."""
1352
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1353
if self.target.versions() == [] and version_ids is None:
1354
self.target._copy_weave_content(self.source)
1357
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1358
except errors.WeaveParentMismatch:
1359
self.target._reweave(self.source, pb, msg)
1362
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
