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.symbol_versioning import (deprecated_method,
95
from bzrlib.tsort import topo_sort
96
from bzrlib.versionedfile import VersionedFile, InterVersionedFile
97
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
100
class Weave(VersionedFile):
101
"""weave - versioned text file storage.
103
A Weave manages versions of line-based text files, keeping track
104
of the originating version for each line.
106
To clients the "lines" of the file are represented as a list of strings.
107
These strings will typically have terminal newline characters, but
108
this is not required. In particular files commonly do not have a newline
109
at the end of the file.
111
Texts can be identified in either of two ways:
113
* a nonnegative index number.
115
* a version-id string.
117
Typically the index number will be valid only inside this weave and
118
the version-id is used to reference it in the larger world.
120
The weave is represented as a list mixing edit instructions and
121
literal text. Each entry in _weave can be either a string (or
122
unicode), or a tuple. If a string, it means that the given line
123
should be output in the currently active revisions.
125
If a tuple, it gives a processing instruction saying in which
126
revisions the enclosed lines are active. The tuple has the form
127
(instruction, version).
129
The instruction can be '{' or '}' for an insertion block, and '['
130
and ']' for a deletion block respectively. The version is the
131
integer version index. There is no replace operator, only deletes
132
and inserts. For '}', the end of an insertion, there is no
133
version parameter because it always closes the most recently
138
* A later version can delete lines that were introduced by any
139
number of ancestor versions; this implies that deletion
140
instructions can span insertion blocks without regard to the
141
insertion block's nesting.
143
* Similarly, deletions need not be properly nested with regard to
144
each other, because they might have been generated by
145
independent revisions.
147
* Insertions are always made by inserting a new bracketed block
148
into a single point in the previous weave. This implies they
149
can nest but not overlap, and the nesting must always have later
150
insertions on the inside.
152
* It doesn't seem very useful to have an active insertion
153
inside an inactive insertion, but it might happen.
155
* Therefore, all instructions are always"considered"; that
156
is passed onto and off the stack. An outer inactive block
157
doesn't disable an inner block.
159
* Lines are enabled if the most recent enclosing insertion is
160
active and none of the enclosing deletions are active.
162
* There is no point having a deletion directly inside its own
163
insertion; you might as well just not write it. And there
164
should be no way to get an earlier version deleting a later
168
Text of the weave; list of control instruction tuples and strings.
171
List of parents, indexed by version number.
172
It is only necessary to store the minimal set of parents for
173
each version; the parent's parents are implied.
176
List of hex SHA-1 of each version.
179
List of symbolic names for each version. Each should be unique.
182
For each name, the version number.
185
Descriptive name of this weave; typically the filename if known.
189
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
190
'_weave_name', '_matcher']
192
def __init__(self, weave_name=None, access_mode='w', matcher=None):
193
super(Weave, self).__init__(access_mode)
199
self._weave_name = weave_name
201
self._matcher = bzrlib.patiencediff.PatienceSequenceMatcher
203
self._matcher = matcher
206
return "Weave(%r)" % self._weave_name
209
"""Return a deep copy of self.
211
The copy can be modified without affecting the original weave."""
213
other._weave = self._weave[:]
214
other._parents = self._parents[:]
215
other._sha1s = self._sha1s[:]
216
other._names = self._names[:]
217
other._name_map = self._name_map.copy()
218
other._weave_name = self._weave_name
221
def __eq__(self, other):
222
if not isinstance(other, Weave):
224
return self._parents == other._parents \
225
and self._weave == other._weave \
226
and self._sha1s == other._sha1s
228
def __ne__(self, other):
229
return not self.__eq__(other)
231
@deprecated_method(zero_eight)
232
def idx_to_name(self, index):
233
"""Old public interface, the public interface is all names now."""
236
def _idx_to_name(self, version):
237
return self._names[version]
239
@deprecated_method(zero_eight)
240
def lookup(self, name):
241
"""Backwards compatibility thunk:
243
Return name, as name is valid in the api now, and spew deprecation
248
def _lookup(self, name):
249
"""Convert symbolic version name to index."""
250
self.check_not_reserved_id(name)
252
return self._name_map[name]
254
raise RevisionNotPresent(name, self._weave_name)
256
@deprecated_method(zero_eight)
257
def iter_names(self):
258
"""Deprecated convenience function, please see VersionedFile.names()."""
259
return iter(self.names())
261
@deprecated_method(zero_eight)
263
"""See Weave.versions for the current api."""
264
return self.versions()
267
"""See VersionedFile.versions."""
268
return self._names[:]
270
def has_version(self, version_id):
271
"""See VersionedFile.has_version."""
272
return (version_id in self._name_map)
274
__contains__ = has_version
276
def get_delta(self, version_id):
277
"""See VersionedFile.get_delta."""
278
return self.get_deltas([version_id])[version_id]
280
def get_deltas(self, version_ids):
281
"""See VersionedFile.get_deltas."""
282
version_ids = self.get_ancestry(version_ids)
283
for version_id in version_ids:
284
if not self.has_version(version_id):
285
raise RevisionNotPresent(version_id, self)
286
# try extracting all versions; parallel extraction is used
287
nv = self.num_versions()
293
last_parent_lines = {}
295
parent_inclusions = {}
300
# its simplest to generate a full set of prepared variables.
302
name = self._names[i]
303
sha1s[name] = self.get_sha1(name)
304
parents_list = self.get_parents(name)
306
parent = parents_list[0]
307
parents[name] = parent
308
parent_inclusions[name] = inclusions[parent]
311
parent_inclusions[name] = set()
312
# we want to emit start, finish, replacement_length, replacement_lines tuples.
313
diff_hunks[name] = []
314
current_hunks[name] = [0, 0, 0, []] # #start, finish, repl_length, repl_tuples
315
parent_linenums[name] = 0
317
parent_noeols[name] = False
318
last_parent_lines[name] = None
319
new_inc = set([name])
320
for p in self._parents[i]:
321
new_inc.update(inclusions[self._idx_to_name(p)])
322
# debug only, known good so far.
323
#assert set(new_inc) == set(self.get_ancestry(name)), \
324
# 'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
325
inclusions[name] = new_inc
327
nlines = len(self._weave)
329
for lineno, inserted, deletes, line in self._walk_internal():
330
# a line is active in a version if:
331
# insert is in the versions inclusions
333
# deleteset & the versions inclusions is an empty set.
334
# so - if we have a included by mapping - version is included by
335
# children, we get a list of children to examine for deletes affect
336
# ing them, which is less than the entire set of children.
337
for version_id in version_ids:
338
# The active inclusion must be an ancestor,
339
# and no ancestors must have deleted this line,
340
# because we don't support resurrection.
341
parent_inclusion = parent_inclusions[version_id]
342
inclusion = inclusions[version_id]
343
parent_active = inserted in parent_inclusion and not (deletes & parent_inclusion)
344
version_active = inserted in inclusion and not (deletes & inclusion)
345
if not parent_active and not version_active:
346
# unrelated line of ancestry
348
elif parent_active and version_active:
350
parent_linenum = parent_linenums[version_id]
351
if current_hunks[version_id] != [parent_linenum, parent_linenum, 0, []]:
352
diff_hunks[version_id].append(tuple(current_hunks[version_id]))
354
current_hunks[version_id] = [parent_linenum, parent_linenum, 0, []]
355
parent_linenums[version_id] = parent_linenum
358
noeols[version_id] = True
361
elif parent_active and not version_active:
363
current_hunks[version_id][1] += 1
364
parent_linenums[version_id] += 1
365
last_parent_lines[version_id] = line
366
elif not parent_active and version_active:
368
# noeol only occurs at the end of a file because we
369
# diff linewise. We want to show noeol changes as a
370
# empty diff unless the actual eol-less content changed.
373
if last_parent_lines[version_id][-1] != '\n':
374
parent_noeols[version_id] = True
375
except (TypeError, IndexError):
378
if theline[-1] != '\n':
379
noeols[version_id] = True
383
parent_should_go = False
385
if parent_noeols[version_id] == noeols[version_id]:
386
# no noeol toggle, so trust the weaves statement
387
# that this line is changed.
389
if parent_noeols[version_id]:
390
theline = theline + '\n'
391
elif parent_noeols[version_id]:
392
# parent has no eol, we do:
393
# our line is new, report as such..
395
elif noeols[version_id]:
396
# append a eol so that it looks like
398
theline = theline + '\n'
399
if parents[version_id] is not None:
400
#if last_parent_lines[version_id] is not None:
401
parent_should_go = True
402
if last_parent_lines[version_id] != theline:
405
#parent_should_go = False
407
current_hunks[version_id][2] += 1
408
current_hunks[version_id][3].append((inserted, theline))
410
# last hunk last parent line is not eaten
411
current_hunks[version_id][1] -= 1
412
if current_hunks[version_id][1] < 0:
413
current_hunks[version_id][1] = 0
414
# import pdb;pdb.set_trace()
415
# assert current_hunks[version_id][1] >= 0
419
version = self._idx_to_name(i)
420
if current_hunks[version] != [0, 0, 0, []]:
421
diff_hunks[version].append(tuple(current_hunks[version]))
423
for version_id in version_ids:
424
result[version_id] = (
428
diff_hunks[version_id],
432
def get_parents(self, version_id):
433
"""See VersionedFile.get_parent."""
434
return map(self._idx_to_name, self._parents[self._lookup(version_id)])
436
def _check_repeated_add(self, name, parents, text, sha1):
437
"""Check that a duplicated add is OK.
439
If it is, return the (old) index; otherwise raise an exception.
441
idx = self._lookup(name)
442
if sorted(self._parents[idx]) != sorted(parents) \
443
or sha1 != self._sha1s[idx]:
444
raise RevisionAlreadyPresent(name, self._weave_name)
447
@deprecated_method(zero_eight)
448
def add_identical(self, old_rev_id, new_rev_id, parents):
449
"""Please use Weave.clone_text now."""
450
return self.clone_text(new_rev_id, old_rev_id, parents)
452
def _add_lines(self, version_id, parents, lines, parent_texts,
453
left_matching_blocks=None):
454
"""See VersionedFile.add_lines."""
455
return self._add(version_id, lines, map(self._lookup, parents))
457
@deprecated_method(zero_eight)
458
def add(self, name, parents, text, sha1=None):
459
"""See VersionedFile.add_lines for the non deprecated api."""
460
return self._add(name, text, map(self._maybe_lookup, parents), sha1)
462
def _add(self, version_id, lines, parents, sha1=None):
463
"""Add a single text on top of the weave.
465
Returns the index number of the newly added version.
468
Symbolic name for this version.
469
(Typically the revision-id of the revision that added it.)
472
List or set of direct parent version numbers.
475
Sequence of lines to be added in the new version.
478
assert isinstance(version_id, basestring)
479
self._check_lines_not_unicode(lines)
480
self._check_lines_are_lines(lines)
482
sha1 = sha_strings(lines)
483
if version_id in self._name_map:
484
return self._check_repeated_add(version_id, parents, lines, sha1)
486
self._check_versions(parents)
487
## self._check_lines(lines)
488
new_version = len(self._parents)
490
# if we abort after here the (in-memory) weave will be corrupt because only
491
# some fields are updated
492
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
493
# - Robert Collins 20060226
494
self._parents.append(parents[:])
495
self._sha1s.append(sha1)
496
self._names.append(version_id)
497
self._name_map[version_id] = new_version
501
# special case; adding with no parents revision; can do
502
# this more quickly by just appending unconditionally.
503
# even more specially, if we're adding an empty text we
504
# need do nothing at all.
506
self._weave.append(('{', new_version))
507
self._weave.extend(lines)
508
self._weave.append(('}', None))
511
if len(parents) == 1:
512
pv = list(parents)[0]
513
if sha1 == self._sha1s[pv]:
514
# special case: same as the single parent
518
ancestors = self._inclusions(parents)
522
# basis a list of (origin, lineno, line)
525
for origin, lineno, line in self._extract(ancestors):
526
basis_lineno.append(lineno)
527
basis_lines.append(line)
529
# another small special case: a merge, producing the same text
531
if lines == basis_lines:
534
# add a sentinel, because we can also match against the final line
535
basis_lineno.append(len(self._weave))
537
# XXX: which line of the weave should we really consider
538
# matches the end of the file? the current code says it's the
539
# last line of the weave?
541
#print 'basis_lines:', basis_lines
542
#print 'new_lines: ', lines
544
s = self._matcher(None, basis_lines, lines)
546
# offset gives the number of lines that have been inserted
547
# into the weave up to the current point; if the original edit instruction
548
# says to change line A then we actually change (A+offset)
551
for tag, i1, i2, j1, j2 in s.get_opcodes():
552
# i1,i2 are given in offsets within basis_lines; we need to map them
553
# back to offsets within the entire weave
554
#print 'raw match', tag, i1, i2, j1, j2
558
i1 = basis_lineno[i1]
559
i2 = basis_lineno[i2]
561
assert 0 <= j1 <= j2 <= len(lines)
563
#print tag, i1, i2, j1, j2
565
# the deletion and insertion are handled separately.
566
# first delete the region.
568
self._weave.insert(i1+offset, ('[', new_version))
569
self._weave.insert(i2+offset+1, (']', new_version))
573
# there may have been a deletion spanning up to
574
# i2; we want to insert after this region to make sure
575
# we don't destroy ourselves
577
self._weave[i:i] = ([('{', new_version)]
580
offset += 2 + (j2 - j1)
583
def _clone_text(self, new_version_id, old_version_id, parents):
584
"""See VersionedFile.clone_text."""
585
old_lines = self.get_text(old_version_id)
586
self.add_lines(new_version_id, parents, old_lines)
588
def _inclusions(self, versions):
589
"""Return set of all ancestors of given version(s)."""
590
if not len(versions):
593
for v in xrange(max(versions), 0, -1):
595
# include all its parents
596
i.update(self._parents[v])
598
## except IndexError:
599
## raise ValueError("version %d not present in weave" % v)
601
@deprecated_method(zero_eight)
602
def inclusions(self, version_ids):
603
"""Deprecated - see VersionedFile.get_ancestry for the replacement."""
606
if isinstance(version_ids[0], int):
607
return [self._idx_to_name(v) for v in self._inclusions(version_ids)]
609
return self.get_ancestry(version_ids)
611
def get_ancestry(self, version_ids, topo_sorted=True):
612
"""See VersionedFile.get_ancestry."""
613
if isinstance(version_ids, basestring):
614
version_ids = [version_ids]
615
i = self._inclusions([self._lookup(v) for v in version_ids])
616
return [self._idx_to_name(v) for v in i]
618
def _check_lines(self, text):
619
if not isinstance(text, list):
620
raise ValueError("text should be a list, not %s" % type(text))
623
if not isinstance(l, basestring):
624
raise ValueError("text line should be a string or unicode, not %s"
629
def _check_versions(self, indexes):
630
"""Check everything in the sequence of indexes is valid"""
635
raise IndexError("invalid version number %r" % i)
637
def _compatible_parents(self, my_parents, other_parents):
638
"""During join check that other_parents are joinable with my_parents.
640
Joinable is defined as 'is a subset of' - supersets may require
641
regeneration of diffs, but subsets do not.
643
return len(other_parents.difference(my_parents)) == 0
645
def annotate_iter(self, version_id):
646
"""Yield list of (version-id, line) pairs for the specified version.
648
The index indicates when the line originated in the weave."""
649
incls = [self._lookup(version_id)]
650
for origin, lineno, text in self._extract(incls):
651
yield self._idx_to_name(origin), text
653
@deprecated_method(zero_eight)
655
"""_walk has become visit, a supported api."""
656
return self._walk_internal()
658
def iter_lines_added_or_present_in_versions(self, version_ids=None,
660
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
661
if version_ids is None:
662
version_ids = self.versions()
663
version_ids = set(version_ids)
664
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
665
# if inserted not in version_ids then it was inserted before the
666
# versions we care about, but because weaves cannot represent ghosts
667
# properly, we do not filter down to that
668
# if inserted not in version_ids: continue
674
#@deprecated_method(zero_eight)
675
def walk(self, version_ids=None):
676
"""See VersionedFile.walk."""
677
return self._walk_internal(version_ids)
679
def _walk_internal(self, version_ids=None):
680
"""Helper method for weave actions."""
685
lineno = 0 # line of weave, 0-based
687
for l in self._weave:
688
if l.__class__ == tuple:
692
istack.append(self._names[v])
696
assert self._names[v] not in dset
697
dset.add(self._names[v])
699
dset.remove(self._names[v])
701
raise WeaveFormatError('unexpected instruction %r' % v)
703
assert l.__class__ in (str, unicode)
705
yield lineno, istack[-1], frozenset(dset), l
709
raise WeaveFormatError("unclosed insertion blocks "
710
"at end of weave: %s" % istack)
712
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
715
def plan_merge(self, ver_a, ver_b):
716
"""Return pseudo-annotation indicating how the two versions merge.
718
This is computed between versions a and b and their common
721
Weave lines present in none of them are skipped entirely.
723
inc_a = set(self.get_ancestry([ver_a]))
724
inc_b = set(self.get_ancestry([ver_b]))
725
inc_c = inc_a & inc_b
727
for lineno, insert, deleteset, line in\
728
self.walk([ver_a, ver_b]):
729
if deleteset & inc_c:
730
# killed in parent; can't be in either a or b
731
# not relevant to our work
732
yield 'killed-base', line
733
elif insert in inc_c:
734
# was inserted in base
735
killed_a = bool(deleteset & inc_a)
736
killed_b = bool(deleteset & inc_b)
737
if killed_a and killed_b:
738
yield 'killed-both', line
740
yield 'killed-a', line
742
yield 'killed-b', line
744
yield 'unchanged', line
745
elif insert in inc_a:
746
if deleteset & inc_a:
747
yield 'ghost-a', line
751
elif insert in inc_b:
752
if deleteset & inc_b:
753
yield 'ghost-b', line
757
# not in either revision
758
yield 'irrelevant', line
760
yield 'unchanged', '' # terminator
762
def _extract(self, versions):
763
"""Yield annotation of lines in included set.
765
Yields a sequence of tuples (origin, lineno, text), where
766
origin is the origin version, lineno the index in the weave,
767
and text the text of the line.
769
The set typically but not necessarily corresponds to a version.
772
if not isinstance(i, int):
775
included = self._inclusions(versions)
781
lineno = 0 # line of weave, 0-based
787
WFE = WeaveFormatError
790
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
791
# +285282 0 1676.8040 1676.8040 +<isinstance>
792
# 1.6 seconds in 'isinstance'.
793
# changing the first isinstance:
794
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
795
# +140414 0 762.8050 762.8050 +<isinstance>
796
# note that the inline time actually dropped (less function calls)
797
# and total processing time was halved.
798
# we're still spending ~1/4 of the method in isinstance though.
799
# so lets hard code the acceptable string classes we expect:
800
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
801
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
803
# yay, down to ~1/4 the initial extract time, and our inline time
804
# has shrunk again, with isinstance no longer dominating.
805
# tweaking the stack inclusion test to use a set gives:
806
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
807
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
809
# - a 5% win, or possibly just noise. However with large istacks that
810
# 'in' test could dominate, so I'm leaving this change in place -
811
# when its fast enough to consider profiling big datasets we can review.
816
for l in self._weave:
817
if l.__class__ == tuple:
825
iset.remove(istack.pop())
836
assert l.__class__ in (str, unicode)
838
isactive = (not dset) and istack and (istack[-1] in included)
840
result.append((istack[-1], lineno, l))
843
raise WeaveFormatError("unclosed insertion blocks "
844
"at end of weave: %s" % istack)
846
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
850
@deprecated_method(zero_eight)
851
def get_iter(self, name_or_index):
852
"""Deprecated, please do not use. Lookups are not not needed.
854
Please use get_lines now.
856
return iter(self.get_lines(self._maybe_lookup(name_or_index)))
858
@deprecated_method(zero_eight)
859
def maybe_lookup(self, name_or_index):
860
"""Deprecated, please do not use. Lookups are not not needed."""
861
return self._maybe_lookup(name_or_index)
863
def _maybe_lookup(self, name_or_index):
864
"""Convert possible symbolic name to index, or pass through indexes.
868
if isinstance(name_or_index, (int, long)):
871
return self._lookup(name_or_index)
873
@deprecated_method(zero_eight)
874
def get(self, version_id):
875
"""Please use either Weave.get_text or Weave.get_lines as desired."""
876
return self.get_lines(version_id)
878
def get_lines(self, version_id):
879
"""See VersionedFile.get_lines()."""
880
int_index = self._maybe_lookup(version_id)
881
result = [line for (origin, lineno, line) in self._extract([int_index])]
882
expected_sha1 = self._sha1s[int_index]
883
measured_sha1 = sha_strings(result)
884
if measured_sha1 != expected_sha1:
885
raise errors.WeaveInvalidChecksum(
886
'file %s, revision %s, expected: %s, measured %s'
887
% (self._weave_name, version_id,
888
expected_sha1, measured_sha1))
891
def get_sha1(self, version_id):
892
"""See VersionedFile.get_sha1()."""
893
return self._sha1s[self._lookup(version_id)]
895
def get_sha1s(self, version_ids):
896
"""See VersionedFile.get_sha1s()."""
897
return [self._sha1s[self._lookup(v)] for v in version_ids]
899
@deprecated_method(zero_eight)
900
def numversions(self):
901
"""How many versions are in this weave?
903
Deprecated in favour of num_versions.
905
return self.num_versions()
907
def num_versions(self):
908
"""How many versions are in this weave?"""
909
l = len(self._parents)
910
assert l == len(self._sha1s)
913
__len__ = num_versions
915
def check(self, progress_bar=None):
916
# TODO evaluate performance hit of using string sets in this routine.
917
# TODO: check no circular inclusions
918
# TODO: create a nested progress bar
919
for version in range(self.num_versions()):
920
inclusions = list(self._parents[version])
923
if inclusions[-1] >= version:
924
raise WeaveFormatError("invalid included version %d for index %d"
925
% (inclusions[-1], version))
927
# try extracting all versions; parallel extraction is used
928
nv = self.num_versions()
933
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
934
# The problem is that set membership is much more expensive
935
name = self._idx_to_name(i)
936
sha1s[name] = sha.new()
938
new_inc = set([name])
939
for p in self._parents[i]:
940
new_inc.update(inclusions[self._idx_to_name(p)])
942
assert set(new_inc) == set(self.get_ancestry(name)), \
943
'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
944
inclusions[name] = new_inc
946
nlines = len(self._weave)
948
update_text = 'checking weave'
950
short_name = os.path.basename(self._weave_name)
951
update_text = 'checking %s' % (short_name,)
952
update_text = update_text[:25]
954
for lineno, insert, deleteset, line in self._walk_internal():
956
progress_bar.update(update_text, lineno, nlines)
958
for name, name_inclusions in inclusions.items():
959
# The active inclusion must be an ancestor,
960
# and no ancestors must have deleted this line,
961
# because we don't support resurrection.
962
if (insert in name_inclusions) and not (deleteset & name_inclusions):
963
sha1s[name].update(line)
966
version = self._idx_to_name(i)
967
hd = sha1s[version].hexdigest()
968
expected = self._sha1s[i]
970
raise errors.WeaveInvalidChecksum(
971
"mismatched sha1 for version %s: "
972
"got %s, expected %s"
973
% (version, hd, expected))
975
# TODO: check insertions are properly nested, that there are
976
# no lines outside of insertion blocks, that deletions are
977
# properly paired, etc.
979
def _join(self, other, pb, msg, version_ids, ignore_missing):
980
"""Worker routine for join()."""
981
if not other.versions():
982
return # nothing to update, easy
985
# versions is never none, InterWeave checks this.
988
# two loops so that we do not change ourselves before verifying it
990
# work through in index order to make sure we get all dependencies
993
# get the selected versions only that are in other.versions.
994
version_ids = set(other.versions()).intersection(set(version_ids))
995
# pull in the referenced graph.
996
version_ids = other.get_ancestry(version_ids)
997
pending_graph = [(version, other.get_parents(version)) for
998
version in version_ids]
999
for name in topo_sort(pending_graph):
1000
other_idx = other._name_map[name]
1001
# returns True if we have it, False if we need it.
1002
if not self._check_version_consistent(other, other_idx, name):
1003
names_to_join.append((other_idx, name))
1012
for other_idx, name in names_to_join:
1013
# TODO: If all the parents of the other version are already
1014
# present then we can avoid some work by just taking the delta
1015
# and adjusting the offsets.
1016
new_parents = self._imported_parents(other, other_idx)
1017
sha1 = other._sha1s[other_idx]
1022
pb.update(msg, merged, len(names_to_join))
1024
lines = other.get_lines(other_idx)
1025
self._add(name, lines, new_parents, sha1)
1027
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
1028
merged, processed, self._weave_name, time.time()-time0))
1030
def _imported_parents(self, other, other_idx):
1031
"""Return list of parents in self corresponding to indexes in other."""
1033
for parent_idx in other._parents[other_idx]:
1034
parent_name = other._names[parent_idx]
1035
if parent_name not in self._name_map:
1036
# should not be possible
1037
raise WeaveError("missing parent {%s} of {%s} in %r"
1038
% (parent_name, other._name_map[other_idx], self))
1039
new_parents.append(self._name_map[parent_name])
1042
def _check_version_consistent(self, other, other_idx, name):
1043
"""Check if a version in consistent in this and other.
1045
To be consistent it must have:
1048
* the same direct parents (by name, not index, and disregarding
1051
If present & correct return True;
1052
if not present in self return False;
1053
if inconsistent raise error."""
1054
this_idx = self._name_map.get(name, -1)
1056
if self._sha1s[this_idx] != other._sha1s[other_idx]:
1057
raise errors.WeaveTextDiffers(name, self, other)
1058
self_parents = self._parents[this_idx]
1059
other_parents = other._parents[other_idx]
1060
n1 = set([self._names[i] for i in self_parents])
1061
n2 = set([other._names[i] for i in other_parents])
1062
if not self._compatible_parents(n1, n2):
1063
raise WeaveParentMismatch("inconsistent parents "
1064
"for version {%s}: %s vs %s" % (name, n1, n2))
1070
@deprecated_method(zero_eight)
1071
def reweave(self, other, pb=None, msg=None):
1072
"""reweave has been superseded by plain use of join."""
1073
return self.join(other, pb, msg)
1075
def _reweave(self, other, pb, msg):
1076
"""Reweave self with other - internal helper for join().
1078
:param other: The other weave to merge
1079
:param pb: An optional progress bar, indicating how far done we are
1080
:param msg: An optional message for the progress
1082
new_weave = _reweave(self, other, pb=pb, msg=msg)
1083
self._copy_weave_content(new_weave)
1085
def _copy_weave_content(self, otherweave):
1086
"""adsorb the content from otherweave."""
1087
for attr in self.__slots__:
1088
if attr != '_weave_name':
1089
setattr(self, attr, copy(getattr(otherweave, attr)))
1092
class WeaveFile(Weave):
1093
"""A WeaveFile represents a Weave on disk and writes on change."""
1095
WEAVE_SUFFIX = '.weave'
1097
def __init__(self, name, transport, filemode=None, create=False, access_mode='w'):
1098
"""Create a WeaveFile.
1100
:param create: If not True, only open an existing knit.
1102
super(WeaveFile, self).__init__(name, access_mode)
1103
self._transport = transport
1104
self._filemode = filemode
1106
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
1107
except errors.NoSuchFile:
1113
def _add_lines(self, version_id, parents, lines, parent_texts,
1114
left_matching_blocks=None):
1115
"""Add a version and save the weave."""
1116
self.check_not_reserved_id(version_id)
1117
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
1122
def _clone_text(self, new_version_id, old_version_id, parents):
1123
"""See VersionedFile.clone_text."""
1124
super(WeaveFile, self)._clone_text(new_version_id, old_version_id, parents)
1127
def copy_to(self, name, transport):
1128
"""See VersionedFile.copy_to()."""
1129
# as we are all in memory always, just serialise to the new place.
1131
write_weave_v5(self, sio)
1133
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
1135
def create_empty(self, name, transport, filemode=None):
1136
return WeaveFile(name, transport, filemode, create=True)
1139
"""Save the weave."""
1140
self._check_write_ok()
1142
write_weave_v5(self, sio)
1144
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
1150
"""See VersionedFile.get_suffixes()."""
1151
return [WeaveFile.WEAVE_SUFFIX]
1153
def join(self, other, pb=None, msg=None, version_ids=None,
1154
ignore_missing=False):
1155
"""Join other into self and save."""
1156
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1160
@deprecated_function(zero_eight)
1161
def reweave(wa, wb, pb=None, msg=None):
1162
"""reweaving is deprecation, please just use weave.join()."""
1163
_reweave(wa, wb, pb, msg)
1165
def _reweave(wa, wb, pb=None, msg=None):
1166
"""Combine two weaves and return the result.
1168
This works even if a revision R has different parents in
1169
wa and wb. In the resulting weave all the parents are given.
1171
This is done by just building up a new weave, maintaining ordering
1172
of the versions in the two inputs. More efficient approaches
1173
might be possible but it should only be necessary to do
1174
this operation rarely, when a new previously ghost version is
1177
:param pb: An optional progress bar, indicating how far done we are
1178
:param msg: An optional message for the progress
1182
queue_a = range(wa.num_versions())
1183
queue_b = range(wb.num_versions())
1184
# first determine combined parents of all versions
1185
# map from version name -> all parent names
1186
combined_parents = _reweave_parent_graphs(wa, wb)
1187
mutter("combined parents: %r", combined_parents)
1188
order = topo_sort(combined_parents.iteritems())
1189
mutter("order to reweave: %r", order)
1194
for idx, name in enumerate(order):
1196
pb.update(msg, idx, len(order))
1197
if name in wa._name_map:
1198
lines = wa.get_lines(name)
1199
if name in wb._name_map:
1200
lines_b = wb.get_lines(name)
1201
if lines != lines_b:
1202
mutter('Weaves differ on content. rev_id {%s}', name)
1203
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1205
lines = list(difflib.unified_diff(lines, lines_b,
1206
wa._weave_name, wb._weave_name))
1207
mutter('lines:\n%s', ''.join(lines))
1208
raise errors.WeaveTextDiffers(name, wa, wb)
1210
lines = wb.get_lines(name)
1211
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1214
def _reweave_parent_graphs(wa, wb):
1215
"""Return combined parent ancestry for two weaves.
1217
Returned as a list of (version_name, set(parent_names))"""
1219
for weave in [wa, wb]:
1220
for idx, name in enumerate(weave._names):
1221
p = combined.setdefault(name, set())
1222
p.update(map(weave._idx_to_name, weave._parents[idx]))
1227
"""Show the weave's table-of-contents"""
1228
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1229
for i in (6, 50, 10, 10):
1232
for i in range(w.num_versions()):
1235
parent_str = ' '.join(map(str, w._parents[i]))
1236
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1240
def weave_stats(weave_file, pb):
1241
from bzrlib.weavefile import read_weave
1243
wf = file(weave_file, 'rb')
1245
# FIXME: doesn't work on pipes
1246
weave_size = wf.tell()
1250
for i in range(vers):
1251
pb.update('checking sizes', i, vers)
1252
for origin, lineno, line in w._extract([i]):
1257
print 'versions %9d' % vers
1258
print 'weave file %9d bytes' % weave_size
1259
print 'total contents %9d bytes' % total
1260
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1263
print 'average size %9d bytes' % avg
1264
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1268
print """bzr weave tool
1270
Experimental tool for weave algorithm.
1273
weave init WEAVEFILE
1274
Create an empty weave file
1275
weave get WEAVEFILE VERSION
1276
Write out specified version.
1277
weave check WEAVEFILE
1278
Check consistency of all versions.
1280
Display table of contents.
1281
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1282
Add NEWTEXT, with specified parent versions.
1283
weave annotate WEAVEFILE VERSION
1284
Display origin of each line.
1285
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1286
Auto-merge two versions and display conflicts.
1287
weave diff WEAVEFILE VERSION1 VERSION2
1288
Show differences between two versions.
1292
% weave init foo.weave
1294
% weave add foo.weave ver0 < foo.txt
1297
(create updated version)
1299
% weave get foo.weave 0 | diff -u - foo.txt
1300
% weave add foo.weave ver1 0 < foo.txt
1303
% weave get foo.weave 0 > foo.txt (create forked version)
1305
% weave add foo.weave ver2 0 < foo.txt
1308
% weave merge foo.weave 1 2 > foo.txt (merge them)
1309
% vi foo.txt (resolve conflicts)
1310
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1322
# in case we're run directly from the subdirectory
1323
sys.path.append('..')
1325
from bzrlib.weavefile import write_weave, read_weave
1326
from bzrlib.progress import ProgressBar
1341
return read_weave(file(argv[2], 'rb'))
1347
# at the moment, based on everything in the file
1349
parents = map(int, argv[4:])
1350
lines = sys.stdin.readlines()
1351
ver = w.add(name, parents, lines)
1352
write_weave(w, file(argv[2], 'wb'))
1353
print 'added version %r %d' % (name, ver)
1356
if os.path.exists(fn):
1357
raise IOError("file exists")
1359
write_weave(w, file(fn, 'wb'))
1360
elif cmd == 'get': # get one version
1362
sys.stdout.writelines(w.get_iter(int(argv[3])))
1367
v1, v2 = map(int, argv[3:5])
1370
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1371
'%s version %d' % (fn, v1),
1372
'%s version %d' % (fn, v2))
1373
sys.stdout.writelines(diff_gen)
1375
elif cmd == 'annotate':
1377
# newline is added to all lines regardless; too hard to get
1378
# reasonable formatting otherwise
1380
for origin, text in w.annotate(int(argv[3])):
1381
text = text.rstrip('\r\n')
1383
print ' | %s' % (text)
1385
print '%5d | %s' % (origin, text)
1391
elif cmd == 'stats':
1392
weave_stats(argv[2], ProgressBar())
1394
elif cmd == 'check':
1399
print '%d versions ok' % w.num_versions()
1401
elif cmd == 'inclusions':
1403
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1405
elif cmd == 'parents':
1407
print ' '.join(map(str, w._parents[int(argv[3])]))
1409
elif cmd == 'plan-merge':
1410
# replaced by 'bzr weave-plan-merge'
1412
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1414
print '%14s | %s' % (state, line),
1415
elif cmd == 'merge':
1416
# replaced by 'bzr weave-merge-text'
1418
p = w.plan_merge(int(argv[3]), int(argv[4]))
1419
sys.stdout.writelines(w.weave_merge(p))
1421
raise ValueError('unknown command %r' % cmd)
1424
if __name__ == '__main__':
1426
sys.exit(main(sys.argv))
1429
class InterWeave(InterVersionedFile):
1430
"""Optimised code paths for weave to weave operations."""
1432
_matching_file_from_factory = staticmethod(WeaveFile)
1433
_matching_file_to_factory = staticmethod(WeaveFile)
1436
def is_compatible(source, target):
1437
"""Be compatible with weaves."""
1439
return (isinstance(source, Weave) and
1440
isinstance(target, Weave))
1441
except AttributeError:
1444
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1445
"""See InterVersionedFile.join."""
1446
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1447
if self.target.versions() == [] and version_ids is None:
1448
self.target._copy_weave_content(self.source)
1451
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1452
except errors.WeaveParentMismatch:
1453
self.target._reweave(self.source, pb, msg)
1456
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
