3
# Copyright (C) 2005 Canonical Ltd
5
# This program is free software; you can redistribute it and/or modify
6
# it under the terms of the GNU General Public License as published by
7
# the Free Software Foundation; either version 2 of the License, or
8
# (at your option) any later version.
10
# This program is distributed in the hope that it will be useful,
11
# but WITHOUT ANY WARRANTY; without even the implied warranty of
12
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13
# GNU General Public License for more details.
15
# You should have received a copy of the GNU General Public License
16
# along with this program; if not, write to the Free Software
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
19
# Author: Martin Pool <mbp@canonical.com>
22
"""Weave - storage of related text file versions"""
25
# XXX: If we do weaves this way, will a merge still behave the same
26
# way if it's done in a different order? That's a pretty desirable
29
# TODO: Nothing here so far assumes the lines are really \n newlines,
30
# rather than being split up in some other way. We could accommodate
31
# binaries, perhaps by naively splitting on \n or perhaps using
32
# something like a rolling checksum.
34
# TODO: End marker for each version so we can stop reading?
36
# TODO: Check that no insertion occurs inside a deletion that was
37
# active in the version of the insertion.
39
# TODO: In addition to the SHA-1 check, perhaps have some code that
40
# checks structural constraints of the weave: ie that insertions are
41
# properly nested, that there is no text outside of an insertion, that
42
# insertions or deletions are not repeated, etc.
44
# TODO: Parallel-extract that passes back each line along with a
45
# description of which revisions include it. Nice for checking all
46
# shas or calculating stats in parallel.
48
# TODO: Using a single _extract routine and then processing the output
49
# is probably inefficient. It's simple enough that we can afford to
50
# have slight specializations for different ways its used: annotate,
51
# basis for add, get, etc.
53
# TODO: Probably the API should work only in names to hide the integer
54
# indexes from the user.
56
# TODO: Is there any potential performance win by having an add()
57
# variant that is passed a pre-cooked version of the single basis
60
# TODO: Reweave can possibly be made faster by remembering diffs
61
# where the basis and destination are unchanged.
63
# FIXME: Sometimes we will be given a parents list for a revision
64
# that includes some redundant parents (i.e. already a parent of
65
# something in the list.) We should eliminate them. This can
66
# be done fairly efficiently because the sequence numbers constrain
67
# the possible relationships.
69
# FIXME: the conflict markers should be *7* characters
72
from cStringIO import StringIO
78
from bzrlib.trace import mutter
79
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
80
RevisionAlreadyPresent,
82
WeaveRevisionAlreadyPresent,
83
WeaveRevisionNotPresent,
85
import bzrlib.errors as errors
86
from bzrlib.osutils import sha_strings
87
import bzrlib.patiencediff
88
from bzrlib.symbol_versioning import (deprecated_method,
92
from bzrlib.tsort import topo_sort
93
from bzrlib.versionedfile import VersionedFile, InterVersionedFile
94
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
97
class Weave(VersionedFile):
98
"""weave - versioned text file storage.
100
A Weave manages versions of line-based text files, keeping track
101
of the originating version for each line.
103
To clients the "lines" of the file are represented as a list of strings.
104
These strings will typically have terminal newline characters, but
105
this is not required. In particular files commonly do not have a newline
106
at the end of the file.
108
Texts can be identified in either of two ways:
110
* a nonnegative index number.
112
* a version-id string.
114
Typically the index number will be valid only inside this weave and
115
the version-id is used to reference it in the larger world.
117
The weave is represented as a list mixing edit instructions and
118
literal text. Each entry in _weave can be either a string (or
119
unicode), or a tuple. If a string, it means that the given line
120
should be output in the currently active revisions.
122
If a tuple, it gives a processing instruction saying in which
123
revisions the enclosed lines are active. The tuple has the form
124
(instruction, version).
126
The instruction can be '{' or '}' for an insertion block, and '['
127
and ']' for a deletion block respectively. The version is the
128
integer version index. There is no replace operator, only deletes
129
and inserts. For '}', the end of an insertion, there is no
130
version parameter because it always closes the most recently
135
* A later version can delete lines that were introduced by any
136
number of ancestor versions; this implies that deletion
137
instructions can span insertion blocks without regard to the
138
insertion block's nesting.
140
* Similarly, deletions need not be properly nested with regard to
141
each other, because they might have been generated by
142
independent revisions.
144
* Insertions are always made by inserting a new bracketed block
145
into a single point in the previous weave. This implies they
146
can nest but not overlap, and the nesting must always have later
147
insertions on the inside.
149
* It doesn't seem very useful to have an active insertion
150
inside an inactive insertion, but it might happen.
152
* Therefore, all instructions are always"considered"; that
153
is passed onto and off the stack. An outer inactive block
154
doesn't disable an inner block.
156
* Lines are enabled if the most recent enclosing insertion is
157
active and none of the enclosing deletions are active.
159
* There is no point having a deletion directly inside its own
160
insertion; you might as well just not write it. And there
161
should be no way to get an earlier version deleting a later
165
Text of the weave; list of control instruction tuples and strings.
168
List of parents, indexed by version number.
169
It is only necessary to store the minimal set of parents for
170
each version; the parent's parents are implied.
173
List of hex SHA-1 of each version.
176
List of symbolic names for each version. Each should be unique.
179
For each name, the version number.
182
Descriptive name of this weave; typically the filename if known.
186
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
187
'_weave_name', '_matcher']
189
def __init__(self, weave_name=None, access_mode='w', matcher=None):
190
super(Weave, self).__init__(access_mode)
196
self._weave_name = weave_name
198
self._matcher = bzrlib.patiencediff.PatienceSequenceMatcher
200
self._matcher = matcher
203
return "Weave(%r)" % self._weave_name
206
"""Return a deep copy of self.
208
The copy can be modified without affecting the original weave."""
210
other._weave = self._weave[:]
211
other._parents = self._parents[:]
212
other._sha1s = self._sha1s[:]
213
other._names = self._names[:]
214
other._name_map = self._name_map.copy()
215
other._weave_name = self._weave_name
218
def __eq__(self, other):
219
if not isinstance(other, Weave):
221
return self._parents == other._parents \
222
and self._weave == other._weave \
223
and self._sha1s == other._sha1s
225
def __ne__(self, other):
226
return not self.__eq__(other)
228
@deprecated_method(zero_eight)
229
def idx_to_name(self, index):
230
"""Old public interface, the public interface is all names now."""
233
def _idx_to_name(self, version):
234
return self._names[version]
236
@deprecated_method(zero_eight)
237
def lookup(self, name):
238
"""Backwards compatibility thunk:
240
Return name, as name is valid in the api now, and spew deprecation
245
def _lookup(self, name):
246
"""Convert symbolic version name to index."""
248
return self._name_map[name]
250
raise RevisionNotPresent(name, self._weave_name)
252
@deprecated_method(zero_eight)
253
def iter_names(self):
254
"""Deprecated convenience function, please see VersionedFile.names()."""
255
return iter(self.names())
257
@deprecated_method(zero_eight)
259
"""See Weave.versions for the current api."""
260
return self.versions()
263
"""See VersionedFile.versions."""
264
return self._names[:]
266
def has_version(self, version_id):
267
"""See VersionedFile.has_version."""
268
return self._name_map.has_key(version_id)
270
__contains__ = has_version
272
def get_delta(self, version_id):
273
"""See VersionedFile.get_delta."""
274
return self.get_deltas([version_id])[version_id]
276
def get_deltas(self, version_ids):
277
"""See VersionedFile.get_deltas."""
278
version_ids = self.get_ancestry(version_ids)
279
for version_id in version_ids:
280
if not self.has_version(version_id):
281
raise RevisionNotPresent(version_id, self)
282
# try extracting all versions; parallel extraction is used
283
nv = self.num_versions()
289
last_parent_lines = {}
291
parent_inclusions = {}
296
# its simplest to generate a full set of prepared variables.
298
name = self._names[i]
299
sha1s[name] = self.get_sha1(name)
300
parents_list = self.get_parents(name)
302
parent = parents_list[0]
303
parents[name] = parent
304
parent_inclusions[name] = inclusions[parent]
307
parent_inclusions[name] = set()
308
# we want to emit start, finish, replacement_length, replacement_lines tuples.
309
diff_hunks[name] = []
310
current_hunks[name] = [0, 0, 0, []] # #start, finish, repl_length, repl_tuples
311
parent_linenums[name] = 0
313
parent_noeols[name] = False
314
last_parent_lines[name] = None
315
new_inc = set([name])
316
for p in self._parents[i]:
317
new_inc.update(inclusions[self._idx_to_name(p)])
318
# debug only, known good so far.
319
#assert set(new_inc) == set(self.get_ancestry(name)), \
320
# 'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
321
inclusions[name] = new_inc
323
nlines = len(self._weave)
325
for lineno, inserted, deletes, line in self._walk_internal():
326
# a line is active in a version if:
327
# insert is in the versions inclusions
329
# deleteset & the versions inclusions is an empty set.
330
# so - if we have a included by mapping - version is included by
331
# children, we get a list of children to examine for deletes affect
332
# ing them, which is less than the entire set of children.
333
for version_id in version_ids:
334
# The active inclusion must be an ancestor,
335
# and no ancestors must have deleted this line,
336
# because we don't support resurrection.
337
parent_inclusion = parent_inclusions[version_id]
338
inclusion = inclusions[version_id]
339
parent_active = inserted in parent_inclusion and not (deletes & parent_inclusion)
340
version_active = inserted in inclusion and not (deletes & inclusion)
341
if not parent_active and not version_active:
342
# unrelated line of ancestry
344
elif parent_active and version_active:
346
parent_linenum = parent_linenums[version_id]
347
if current_hunks[version_id] != [parent_linenum, parent_linenum, 0, []]:
348
diff_hunks[version_id].append(tuple(current_hunks[version_id]))
350
current_hunks[version_id] = [parent_linenum, parent_linenum, 0, []]
351
parent_linenums[version_id] = parent_linenum
354
noeols[version_id] = True
357
elif parent_active and not version_active:
359
current_hunks[version_id][1] += 1
360
parent_linenums[version_id] += 1
361
last_parent_lines[version_id] = line
362
elif not parent_active and version_active:
364
# noeol only occurs at the end of a file because we
365
# diff linewise. We want to show noeol changes as a
366
# empty diff unless the actual eol-less content changed.
369
if last_parent_lines[version_id][-1] != '\n':
370
parent_noeols[version_id] = True
371
except (TypeError, IndexError):
374
if theline[-1] != '\n':
375
noeols[version_id] = True
379
parent_should_go = False
381
if parent_noeols[version_id] == noeols[version_id]:
382
# no noeol toggle, so trust the weaves statement
383
# that this line is changed.
385
if parent_noeols[version_id]:
386
theline = theline + '\n'
387
elif parent_noeols[version_id]:
388
# parent has no eol, we do:
389
# our line is new, report as such..
391
elif noeols[version_id]:
392
# append a eol so that it looks like
394
theline = theline + '\n'
395
if parents[version_id] is not None:
396
#if last_parent_lines[version_id] is not None:
397
parent_should_go = True
398
if last_parent_lines[version_id] != theline:
401
#parent_should_go = False
403
current_hunks[version_id][2] += 1
404
current_hunks[version_id][3].append((inserted, theline))
406
# last hunk last parent line is not eaten
407
current_hunks[version_id][1] -= 1
408
if current_hunks[version_id][1] < 0:
409
current_hunks[version_id][1] = 0
410
# import pdb;pdb.set_trace()
411
# assert current_hunks[version_id][1] >= 0
415
version = self._idx_to_name(i)
416
if current_hunks[version] != [0, 0, 0, []]:
417
diff_hunks[version].append(tuple(current_hunks[version]))
419
for version_id in version_ids:
420
result[version_id] = (
424
diff_hunks[version_id],
428
def get_parents(self, version_id):
429
"""See VersionedFile.get_parent."""
430
return map(self._idx_to_name, self._parents[self._lookup(version_id)])
432
def _check_repeated_add(self, name, parents, text, sha1):
433
"""Check that a duplicated add is OK.
435
If it is, return the (old) index; otherwise raise an exception.
437
idx = self._lookup(name)
438
if sorted(self._parents[idx]) != sorted(parents) \
439
or sha1 != self._sha1s[idx]:
440
raise RevisionAlreadyPresent(name, self._weave_name)
443
@deprecated_method(zero_eight)
444
def add_identical(self, old_rev_id, new_rev_id, parents):
445
"""Please use Weave.clone_text now."""
446
return self.clone_text(new_rev_id, old_rev_id, parents)
448
def _add_lines(self, version_id, parents, lines, parent_texts):
449
"""See VersionedFile.add_lines."""
450
return self._add(version_id, lines, map(self._lookup, parents))
452
@deprecated_method(zero_eight)
453
def add(self, name, parents, text, sha1=None):
454
"""See VersionedFile.add_lines for the non deprecated api."""
455
return self._add(name, text, map(self._maybe_lookup, parents), sha1)
457
def _add(self, version_id, lines, parents, sha1=None):
458
"""Add a single text on top of the weave.
460
Returns the index number of the newly added version.
463
Symbolic name for this version.
464
(Typically the revision-id of the revision that added it.)
467
List or set of direct parent version numbers.
470
Sequence of lines to be added in the new version.
473
assert isinstance(version_id, basestring)
474
self._check_lines_not_unicode(lines)
475
self._check_lines_are_lines(lines)
477
sha1 = sha_strings(lines)
478
if version_id in self._name_map:
479
return self._check_repeated_add(version_id, parents, lines, sha1)
481
self._check_versions(parents)
482
## self._check_lines(lines)
483
new_version = len(self._parents)
485
# if we abort after here the (in-memory) weave will be corrupt because only
486
# some fields are updated
487
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
488
# - Robert Collins 20060226
489
self._parents.append(parents[:])
490
self._sha1s.append(sha1)
491
self._names.append(version_id)
492
self._name_map[version_id] = new_version
496
# special case; adding with no parents revision; can do
497
# this more quickly by just appending unconditionally.
498
# even more specially, if we're adding an empty text we
499
# need do nothing at all.
501
self._weave.append(('{', new_version))
502
self._weave.extend(lines)
503
self._weave.append(('}', None))
506
if len(parents) == 1:
507
pv = list(parents)[0]
508
if sha1 == self._sha1s[pv]:
509
# special case: same as the single parent
513
ancestors = self._inclusions(parents)
517
# basis a list of (origin, lineno, line)
520
for origin, lineno, line in self._extract(ancestors):
521
basis_lineno.append(lineno)
522
basis_lines.append(line)
524
# another small special case: a merge, producing the same text
526
if lines == basis_lines:
529
# add a sentinel, because we can also match against the final line
530
basis_lineno.append(len(self._weave))
532
# XXX: which line of the weave should we really consider
533
# matches the end of the file? the current code says it's the
534
# last line of the weave?
536
#print 'basis_lines:', basis_lines
537
#print 'new_lines: ', lines
539
s = self._matcher(None, basis_lines, lines)
541
# offset gives the number of lines that have been inserted
542
# into the weave up to the current point; if the original edit instruction
543
# says to change line A then we actually change (A+offset)
546
for tag, i1, i2, j1, j2 in s.get_opcodes():
547
# i1,i2 are given in offsets within basis_lines; we need to map them
548
# back to offsets within the entire weave
549
#print 'raw match', tag, i1, i2, j1, j2
553
i1 = basis_lineno[i1]
554
i2 = basis_lineno[i2]
556
assert 0 <= j1 <= j2 <= len(lines)
558
#print tag, i1, i2, j1, j2
560
# the deletion and insertion are handled separately.
561
# first delete the region.
563
self._weave.insert(i1+offset, ('[', new_version))
564
self._weave.insert(i2+offset+1, (']', new_version))
568
# there may have been a deletion spanning up to
569
# i2; we want to insert after this region to make sure
570
# we don't destroy ourselves
572
self._weave[i:i] = ([('{', new_version)]
575
offset += 2 + (j2 - j1)
578
def _clone_text(self, new_version_id, old_version_id, parents):
579
"""See VersionedFile.clone_text."""
580
old_lines = self.get_text(old_version_id)
581
self.add_lines(new_version_id, parents, old_lines)
583
def _inclusions(self, versions):
584
"""Return set of all ancestors of given version(s)."""
585
if not len(versions):
588
for v in xrange(max(versions), 0, -1):
590
# include all its parents
591
i.update(self._parents[v])
593
## except IndexError:
594
## raise ValueError("version %d not present in weave" % v)
596
@deprecated_method(zero_eight)
597
def inclusions(self, version_ids):
598
"""Deprecated - see VersionedFile.get_ancestry for the replacement."""
601
if isinstance(version_ids[0], int):
602
return [self._idx_to_name(v) for v in self._inclusions(version_ids)]
604
return self.get_ancestry(version_ids)
606
def get_ancestry(self, version_ids):
607
"""See VersionedFile.get_ancestry."""
608
if isinstance(version_ids, basestring):
609
version_ids = [version_ids]
610
i = self._inclusions([self._lookup(v) for v in version_ids])
611
return [self._idx_to_name(v) for v in i]
613
def _check_lines(self, text):
614
if not isinstance(text, list):
615
raise ValueError("text should be a list, not %s" % type(text))
618
if not isinstance(l, basestring):
619
raise ValueError("text line should be a string or unicode, not %s"
624
def _check_versions(self, indexes):
625
"""Check everything in the sequence of indexes is valid"""
630
raise IndexError("invalid version number %r" % i)
632
def _compatible_parents(self, my_parents, other_parents):
633
"""During join check that other_parents are joinable with my_parents.
635
Joinable is defined as 'is a subset of' - supersets may require
636
regeneration of diffs, but subsets do not.
638
return len(other_parents.difference(my_parents)) == 0
640
def annotate(self, version_id):
641
if isinstance(version_id, int):
642
warnings.warn('Weave.annotate(int) is deprecated. Please use version names'
643
' in all circumstances as of 0.8',
648
for origin, lineno, text in self._extract([version_id]):
649
result.append((origin, text))
652
return super(Weave, self).annotate(version_id)
654
def annotate_iter(self, version_id):
655
"""Yield list of (version-id, line) pairs for the specified version.
657
The index indicates when the line originated in the weave."""
658
incls = [self._lookup(version_id)]
659
for origin, lineno, text in self._extract(incls):
660
yield self._idx_to_name(origin), text
662
@deprecated_method(zero_eight)
664
"""_walk has become visit, a supported api."""
665
return self._walk_internal()
667
def iter_lines_added_or_present_in_versions(self, version_ids=None):
668
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
669
if version_ids is None:
670
version_ids = self.versions()
671
version_ids = set(version_ids)
672
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
673
# if inserted not in version_ids then it was inserted before the
674
# versions we care about, but because weaves cannot represent ghosts
675
# properly, we do not filter down to that
676
# if inserted not in version_ids: continue
682
#@deprecated_method(zero_eight)
683
def walk(self, version_ids=None):
684
"""See VersionedFile.walk."""
685
return self._walk_internal(version_ids)
687
def _walk_internal(self, version_ids=None):
688
"""Helper method for weave actions."""
693
lineno = 0 # line of weave, 0-based
695
for l in self._weave:
696
if l.__class__ == tuple:
700
istack.append(self._names[v])
704
assert self._names[v] not in dset
705
dset.add(self._names[v])
707
dset.remove(self._names[v])
709
raise WeaveFormatError('unexpected instruction %r' % v)
711
assert l.__class__ in (str, unicode)
713
yield lineno, istack[-1], frozenset(dset), l
717
raise WeaveFormatError("unclosed insertion blocks "
718
"at end of weave: %s" % istack)
720
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
723
def plan_merge(self, ver_a, ver_b):
724
"""Return pseudo-annotation indicating how the two versions merge.
726
This is computed between versions a and b and their common
729
Weave lines present in none of them are skipped entirely.
731
inc_a = set(self.get_ancestry([ver_a]))
732
inc_b = set(self.get_ancestry([ver_b]))
733
inc_c = inc_a & inc_b
735
for lineno, insert, deleteset, line in\
736
self.walk([ver_a, ver_b]):
737
if deleteset & inc_c:
738
# killed in parent; can't be in either a or b
739
# not relevant to our work
740
yield 'killed-base', line
741
elif insert in inc_c:
742
# was inserted in base
743
killed_a = bool(deleteset & inc_a)
744
killed_b = bool(deleteset & inc_b)
745
if killed_a and killed_b:
746
yield 'killed-both', line
748
yield 'killed-a', line
750
yield 'killed-b', line
752
yield 'unchanged', line
753
elif insert in inc_a:
754
if deleteset & inc_a:
755
yield 'ghost-a', line
759
elif insert in inc_b:
760
if deleteset & inc_b:
761
yield 'ghost-b', line
765
# not in either revision
766
yield 'irrelevant', line
768
yield 'unchanged', '' # terminator
770
def _extract(self, versions):
771
"""Yield annotation of lines in included set.
773
Yields a sequence of tuples (origin, lineno, text), where
774
origin is the origin version, lineno the index in the weave,
775
and text the text of the line.
777
The set typically but not necessarily corresponds to a version.
780
if not isinstance(i, int):
783
included = self._inclusions(versions)
789
lineno = 0 # line of weave, 0-based
795
WFE = WeaveFormatError
798
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
799
# +285282 0 1676.8040 1676.8040 +<isinstance>
800
# 1.6 seconds in 'isinstance'.
801
# changing the first isinstance:
802
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
803
# +140414 0 762.8050 762.8050 +<isinstance>
804
# note that the inline time actually dropped (less function calls)
805
# and total processing time was halved.
806
# we're still spending ~1/4 of the method in isinstance though.
807
# so lets hard code the acceptable string classes we expect:
808
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
809
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
811
# yay, down to ~1/4 the initial extract time, and our inline time
812
# has shrunk again, with isinstance no longer dominating.
813
# tweaking the stack inclusion test to use a set gives:
814
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
815
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
817
# - a 5% win, or possibly just noise. However with large istacks that
818
# 'in' test could dominate, so I'm leaving this change in place -
819
# when its fast enough to consider profiling big datasets we can review.
824
for l in self._weave:
825
if l.__class__ == tuple:
833
iset.remove(istack.pop())
844
assert l.__class__ in (str, unicode)
846
isactive = (not dset) and istack and (istack[-1] in included)
848
result.append((istack[-1], lineno, l))
851
raise WeaveFormatError("unclosed insertion blocks "
852
"at end of weave: %s" % istack)
854
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
858
@deprecated_method(zero_eight)
859
def get_iter(self, name_or_index):
860
"""Deprecated, please do not use. Lookups are not not needed.
862
Please use get_lines now.
864
return iter(self.get_lines(self._maybe_lookup(name_or_index)))
866
@deprecated_method(zero_eight)
867
def maybe_lookup(self, name_or_index):
868
"""Deprecated, please do not use. Lookups are not not needed."""
869
return self._maybe_lookup(name_or_index)
871
def _maybe_lookup(self, name_or_index):
872
"""Convert possible symbolic name to index, or pass through indexes.
876
if isinstance(name_or_index, (int, long)):
879
return self._lookup(name_or_index)
881
@deprecated_method(zero_eight)
882
def get(self, version_id):
883
"""Please use either Weave.get_text or Weave.get_lines as desired."""
884
return self.get_lines(version_id)
886
def get_lines(self, version_id):
887
"""See VersionedFile.get_lines()."""
888
int_index = self._maybe_lookup(version_id)
889
result = [line for (origin, lineno, line) in self._extract([int_index])]
890
expected_sha1 = self._sha1s[int_index]
891
measured_sha1 = sha_strings(result)
892
if measured_sha1 != expected_sha1:
893
raise errors.WeaveInvalidChecksum(
894
'file %s, revision %s, expected: %s, measured %s'
895
% (self._weave_name, version_id,
896
expected_sha1, measured_sha1))
899
def get_sha1(self, version_id):
900
"""See VersionedFile.get_sha1()."""
901
return self._sha1s[self._lookup(version_id)]
903
@deprecated_method(zero_eight)
904
def numversions(self):
905
"""How many versions are in this weave?
907
Deprecated in favour of num_versions.
909
return self.num_versions()
911
def num_versions(self):
912
"""How many versions are in this weave?"""
913
l = len(self._parents)
914
assert l == len(self._sha1s)
917
__len__ = num_versions
919
def check(self, progress_bar=None):
920
# TODO evaluate performance hit of using string sets in this routine.
921
# TODO: check no circular inclusions
922
# TODO: create a nested progress bar
923
for version in range(self.num_versions()):
924
inclusions = list(self._parents[version])
927
if inclusions[-1] >= version:
928
raise WeaveFormatError("invalid included version %d for index %d"
929
% (inclusions[-1], version))
931
# try extracting all versions; parallel extraction is used
932
nv = self.num_versions()
937
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
938
# The problem is that set membership is much more expensive
939
name = self._idx_to_name(i)
940
sha1s[name] = sha.new()
942
new_inc = set([name])
943
for p in self._parents[i]:
944
new_inc.update(inclusions[self._idx_to_name(p)])
946
assert set(new_inc) == set(self.get_ancestry(name)), \
947
'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
948
inclusions[name] = new_inc
950
nlines = len(self._weave)
952
update_text = 'checking weave'
954
short_name = os.path.basename(self._weave_name)
955
update_text = 'checking %s' % (short_name,)
956
update_text = update_text[:25]
958
for lineno, insert, deleteset, line in self._walk_internal():
960
progress_bar.update(update_text, lineno, nlines)
962
for name, name_inclusions in inclusions.items():
963
# The active inclusion must be an ancestor,
964
# and no ancestors must have deleted this line,
965
# because we don't support resurrection.
966
if (insert in name_inclusions) and not (deleteset & name_inclusions):
967
sha1s[name].update(line)
970
version = self._idx_to_name(i)
971
hd = sha1s[version].hexdigest()
972
expected = self._sha1s[i]
974
raise errors.WeaveInvalidChecksum(
975
"mismatched sha1 for version %s: "
976
"got %s, expected %s"
977
% (version, hd, expected))
979
# TODO: check insertions are properly nested, that there are
980
# no lines outside of insertion blocks, that deletions are
981
# properly paired, etc.
983
def _join(self, other, pb, msg, version_ids, ignore_missing):
984
"""Worker routine for join()."""
985
if not other.versions():
986
return # nothing to update, easy
989
# versions is never none, InterWeave checks this.
992
# two loops so that we do not change ourselves before verifying it
994
# work through in index order to make sure we get all dependencies
997
# get the selected versions only that are in other.versions.
998
version_ids = set(other.versions()).intersection(set(version_ids))
999
# pull in the referenced graph.
1000
version_ids = other.get_ancestry(version_ids)
1001
pending_graph = [(version, other.get_parents(version)) for
1002
version in version_ids]
1003
for name in topo_sort(pending_graph):
1004
other_idx = other._name_map[name]
1005
# returns True if we have it, False if we need it.
1006
if not self._check_version_consistent(other, other_idx, name):
1007
names_to_join.append((other_idx, name))
1016
for other_idx, name in names_to_join:
1017
# TODO: If all the parents of the other version are already
1018
# present then we can avoid some work by just taking the delta
1019
# and adjusting the offsets.
1020
new_parents = self._imported_parents(other, other_idx)
1021
sha1 = other._sha1s[other_idx]
1026
pb.update(msg, merged, len(names_to_join))
1028
lines = other.get_lines(other_idx)
1029
self._add(name, lines, new_parents, sha1)
1031
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
1032
merged, processed, self._weave_name, time.time()-time0))
1034
def _imported_parents(self, other, other_idx):
1035
"""Return list of parents in self corresponding to indexes in other."""
1037
for parent_idx in other._parents[other_idx]:
1038
parent_name = other._names[parent_idx]
1039
if parent_name not in self._name_map:
1040
# should not be possible
1041
raise WeaveError("missing parent {%s} of {%s} in %r"
1042
% (parent_name, other._name_map[other_idx], self))
1043
new_parents.append(self._name_map[parent_name])
1046
def _check_version_consistent(self, other, other_idx, name):
1047
"""Check if a version in consistent in this and other.
1049
To be consistent it must have:
1052
* the same direct parents (by name, not index, and disregarding
1055
If present & correct return True;
1056
if not present in self return False;
1057
if inconsistent raise error."""
1058
this_idx = self._name_map.get(name, -1)
1060
if self._sha1s[this_idx] != other._sha1s[other_idx]:
1061
raise errors.WeaveTextDiffers(name, self, other)
1062
self_parents = self._parents[this_idx]
1063
other_parents = other._parents[other_idx]
1064
n1 = set([self._names[i] for i in self_parents])
1065
n2 = set([other._names[i] for i in other_parents])
1066
if not self._compatible_parents(n1, n2):
1067
raise WeaveParentMismatch("inconsistent parents "
1068
"for version {%s}: %s vs %s" % (name, n1, n2))
1074
@deprecated_method(zero_eight)
1075
def reweave(self, other, pb=None, msg=None):
1076
"""reweave has been superseded by plain use of join."""
1077
return self.join(other, pb, msg)
1079
def _reweave(self, other, pb, msg):
1080
"""Reweave self with other - internal helper for join().
1082
:param other: The other weave to merge
1083
:param pb: An optional progress bar, indicating how far done we are
1084
:param msg: An optional message for the progress
1086
new_weave = _reweave(self, other, pb=pb, msg=msg)
1087
self._copy_weave_content(new_weave)
1089
def _copy_weave_content(self, otherweave):
1090
"""adsorb the content from otherweave."""
1091
for attr in self.__slots__:
1092
if attr != '_weave_name':
1093
setattr(self, attr, copy(getattr(otherweave, attr)))
1096
class WeaveFile(Weave):
1097
"""A WeaveFile represents a Weave on disk and writes on change."""
1099
WEAVE_SUFFIX = '.weave'
1101
def __init__(self, name, transport, filemode=None, create=False, access_mode='w'):
1102
"""Create a WeaveFile.
1104
:param create: If not True, only open an existing knit.
1106
super(WeaveFile, self).__init__(name, access_mode)
1107
self._transport = transport
1108
self._filemode = filemode
1110
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
1111
except errors.NoSuchFile:
1117
def _add_lines(self, version_id, parents, lines, parent_texts):
1118
"""Add a version and save the weave."""
1119
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
1124
def _clone_text(self, new_version_id, old_version_id, parents):
1125
"""See VersionedFile.clone_text."""
1126
super(WeaveFile, self)._clone_text(new_version_id, old_version_id, parents)
1129
def copy_to(self, name, transport):
1130
"""See VersionedFile.copy_to()."""
1131
# as we are all in memory always, just serialise to the new place.
1133
write_weave_v5(self, sio)
1135
transport.put(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
1137
def create_empty(self, name, transport, filemode=None):
1138
return WeaveFile(name, transport, filemode, create=True)
1141
"""Save the weave."""
1142
self._check_write_ok()
1144
write_weave_v5(self, sio)
1146
self._transport.put(self._weave_name + WeaveFile.WEAVE_SUFFIX,
1152
"""See VersionedFile.get_suffixes()."""
1153
return [WeaveFile.WEAVE_SUFFIX]
1155
def join(self, other, pb=None, msg=None, version_ids=None,
1156
ignore_missing=False):
1157
"""Join other into self and save."""
1158
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1162
@deprecated_function(zero_eight)
1163
def reweave(wa, wb, pb=None, msg=None):
1164
"""reweaving is deprecation, please just use weave.join()."""
1165
_reweave(wa, wb, pb, msg)
1167
def _reweave(wa, wb, pb=None, msg=None):
1168
"""Combine two weaves and return the result.
1170
This works even if a revision R has different parents in
1171
wa and wb. In the resulting weave all the parents are given.
1173
This is done by just building up a new weave, maintaining ordering
1174
of the versions in the two inputs. More efficient approaches
1175
might be possible but it should only be necessary to do
1176
this operation rarely, when a new previously ghost version is
1179
:param pb: An optional progress bar, indicating how far done we are
1180
:param msg: An optional message for the progress
1184
queue_a = range(wa.num_versions())
1185
queue_b = range(wb.num_versions())
1186
# first determine combined parents of all versions
1187
# map from version name -> all parent names
1188
combined_parents = _reweave_parent_graphs(wa, wb)
1189
mutter("combined parents: %r", combined_parents)
1190
order = topo_sort(combined_parents.iteritems())
1191
mutter("order to reweave: %r", order)
1196
for idx, name in enumerate(order):
1198
pb.update(msg, idx, len(order))
1199
if name in wa._name_map:
1200
lines = wa.get_lines(name)
1201
if name in wb._name_map:
1202
lines_b = wb.get_lines(name)
1203
if lines != lines_b:
1204
mutter('Weaves differ on content. rev_id {%s}', name)
1205
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1207
lines = list(difflib.unified_diff(lines, lines_b,
1208
wa._weave_name, wb._weave_name))
1209
mutter('lines:\n%s', ''.join(lines))
1210
raise errors.WeaveTextDiffers(name, wa, wb)
1212
lines = wb.get_lines(name)
1213
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1216
def _reweave_parent_graphs(wa, wb):
1217
"""Return combined parent ancestry for two weaves.
1219
Returned as a list of (version_name, set(parent_names))"""
1221
for weave in [wa, wb]:
1222
for idx, name in enumerate(weave._names):
1223
p = combined.setdefault(name, set())
1224
p.update(map(weave._idx_to_name, weave._parents[idx]))
1229
"""Show the weave's table-of-contents"""
1230
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1231
for i in (6, 50, 10, 10):
1234
for i in range(w.num_versions()):
1237
parent_str = ' '.join(map(str, w._parents[i]))
1238
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1242
def weave_stats(weave_file, pb):
1243
from bzrlib.weavefile import read_weave
1245
wf = file(weave_file, 'rb')
1247
# FIXME: doesn't work on pipes
1248
weave_size = wf.tell()
1252
for i in range(vers):
1253
pb.update('checking sizes', i, vers)
1254
for origin, lineno, line in w._extract([i]):
1259
print 'versions %9d' % vers
1260
print 'weave file %9d bytes' % weave_size
1261
print 'total contents %9d bytes' % total
1262
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1265
print 'average size %9d bytes' % avg
1266
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1270
print """bzr weave tool
1272
Experimental tool for weave algorithm.
1275
weave init WEAVEFILE
1276
Create an empty weave file
1277
weave get WEAVEFILE VERSION
1278
Write out specified version.
1279
weave check WEAVEFILE
1280
Check consistency of all versions.
1282
Display table of contents.
1283
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1284
Add NEWTEXT, with specified parent versions.
1285
weave annotate WEAVEFILE VERSION
1286
Display origin of each line.
1287
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1288
Auto-merge two versions and display conflicts.
1289
weave diff WEAVEFILE VERSION1 VERSION2
1290
Show differences between two versions.
1294
% weave init foo.weave
1296
% weave add foo.weave ver0 < foo.txt
1299
(create updated version)
1301
% weave get foo.weave 0 | diff -u - foo.txt
1302
% weave add foo.weave ver1 0 < foo.txt
1305
% weave get foo.weave 0 > foo.txt (create forked version)
1307
% weave add foo.weave ver2 0 < foo.txt
1310
% weave merge foo.weave 1 2 > foo.txt (merge them)
1311
% vi foo.txt (resolve conflicts)
1312
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1324
# in case we're run directly from the subdirectory
1325
sys.path.append('..')
1327
from bzrlib.weavefile import write_weave, read_weave
1328
from bzrlib.progress import ProgressBar
1343
return read_weave(file(argv[2], 'rb'))
1349
# at the moment, based on everything in the file
1351
parents = map(int, argv[4:])
1352
lines = sys.stdin.readlines()
1353
ver = w.add(name, parents, lines)
1354
write_weave(w, file(argv[2], 'wb'))
1355
print 'added version %r %d' % (name, ver)
1358
if os.path.exists(fn):
1359
raise IOError("file exists")
1361
write_weave(w, file(fn, 'wb'))
1362
elif cmd == 'get': # get one version
1364
sys.stdout.writelines(w.get_iter(int(argv[3])))
1369
v1, v2 = map(int, argv[3:5])
1372
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1373
'%s version %d' % (fn, v1),
1374
'%s version %d' % (fn, v2))
1375
sys.stdout.writelines(diff_gen)
1377
elif cmd == 'annotate':
1379
# newline is added to all lines regardless; too hard to get
1380
# reasonable formatting otherwise
1382
for origin, text in w.annotate(int(argv[3])):
1383
text = text.rstrip('\r\n')
1385
print ' | %s' % (text)
1387
print '%5d | %s' % (origin, text)
1393
elif cmd == 'stats':
1394
weave_stats(argv[2], ProgressBar())
1396
elif cmd == 'check':
1401
print '%d versions ok' % w.num_versions()
1403
elif cmd == 'inclusions':
1405
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1407
elif cmd == 'parents':
1409
print ' '.join(map(str, w._parents[int(argv[3])]))
1411
elif cmd == 'plan-merge':
1412
# replaced by 'bzr weave-plan-merge'
1414
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1416
print '%14s | %s' % (state, line),
1417
elif cmd == 'merge':
1418
# replaced by 'bzr weave-merge-text'
1420
p = w.plan_merge(int(argv[3]), int(argv[4]))
1421
sys.stdout.writelines(w.weave_merge(p))
1423
raise ValueError('unknown command %r' % cmd)
1426
if __name__ == '__main__':
1428
sys.exit(main(sys.argv))
1431
class InterWeave(InterVersionedFile):
1432
"""Optimised code paths for weave to weave operations."""
1434
_matching_file_from_factory = staticmethod(WeaveFile)
1435
_matching_file_to_factory = staticmethod(WeaveFile)
1438
def is_compatible(source, target):
1439
"""Be compatible with weaves."""
1441
return (isinstance(source, Weave) and
1442
isinstance(target, Weave))
1443
except AttributeError:
1446
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1447
"""See InterVersionedFile.join."""
1448
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1449
if self.target.versions() == [] and version_ids is None:
1450
self.target._copy_weave_content(self.source)
1453
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1454
except errors.WeaveParentMismatch:
1455
self.target._reweave(self.source, pb, msg)
1458
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
