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."""
251
return self._name_map[name]
253
raise RevisionNotPresent(name, self._weave_name)
255
@deprecated_method(zero_eight)
256
def iter_names(self):
257
"""Deprecated convenience function, please see VersionedFile.names()."""
258
return iter(self.names())
260
@deprecated_method(zero_eight)
262
"""See Weave.versions for the current api."""
263
return self.versions()
266
"""See VersionedFile.versions."""
267
return self._names[:]
269
def has_version(self, version_id):
270
"""See VersionedFile.has_version."""
271
return (version_id in self._name_map)
273
__contains__ = has_version
275
def get_delta(self, version_id):
276
"""See VersionedFile.get_delta."""
277
return self.get_deltas([version_id])[version_id]
279
def get_deltas(self, version_ids):
280
"""See VersionedFile.get_deltas."""
281
version_ids = self.get_ancestry(version_ids)
282
for version_id in version_ids:
283
if not self.has_version(version_id):
284
raise RevisionNotPresent(version_id, self)
285
# try extracting all versions; parallel extraction is used
286
nv = self.num_versions()
292
last_parent_lines = {}
294
parent_inclusions = {}
299
# its simplest to generate a full set of prepared variables.
301
name = self._names[i]
302
sha1s[name] = self.get_sha1(name)
303
parents_list = self.get_parents(name)
305
parent = parents_list[0]
306
parents[name] = parent
307
parent_inclusions[name] = inclusions[parent]
310
parent_inclusions[name] = set()
311
# we want to emit start, finish, replacement_length, replacement_lines tuples.
312
diff_hunks[name] = []
313
current_hunks[name] = [0, 0, 0, []] # #start, finish, repl_length, repl_tuples
314
parent_linenums[name] = 0
316
parent_noeols[name] = False
317
last_parent_lines[name] = None
318
new_inc = set([name])
319
for p in self._parents[i]:
320
new_inc.update(inclusions[self._idx_to_name(p)])
321
# debug only, known good so far.
322
#assert set(new_inc) == set(self.get_ancestry(name)), \
323
# 'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
324
inclusions[name] = new_inc
326
nlines = len(self._weave)
328
for lineno, inserted, deletes, line in self._walk_internal():
329
# a line is active in a version if:
330
# insert is in the versions inclusions
332
# deleteset & the versions inclusions is an empty set.
333
# so - if we have a included by mapping - version is included by
334
# children, we get a list of children to examine for deletes affect
335
# ing them, which is less than the entire set of children.
336
for version_id in version_ids:
337
# The active inclusion must be an ancestor,
338
# and no ancestors must have deleted this line,
339
# because we don't support resurrection.
340
parent_inclusion = parent_inclusions[version_id]
341
inclusion = inclusions[version_id]
342
parent_active = inserted in parent_inclusion and not (deletes & parent_inclusion)
343
version_active = inserted in inclusion and not (deletes & inclusion)
344
if not parent_active and not version_active:
345
# unrelated line of ancestry
347
elif parent_active and version_active:
349
parent_linenum = parent_linenums[version_id]
350
if current_hunks[version_id] != [parent_linenum, parent_linenum, 0, []]:
351
diff_hunks[version_id].append(tuple(current_hunks[version_id]))
353
current_hunks[version_id] = [parent_linenum, parent_linenum, 0, []]
354
parent_linenums[version_id] = parent_linenum
357
noeols[version_id] = True
360
elif parent_active and not version_active:
362
current_hunks[version_id][1] += 1
363
parent_linenums[version_id] += 1
364
last_parent_lines[version_id] = line
365
elif not parent_active and version_active:
367
# noeol only occurs at the end of a file because we
368
# diff linewise. We want to show noeol changes as a
369
# empty diff unless the actual eol-less content changed.
372
if last_parent_lines[version_id][-1] != '\n':
373
parent_noeols[version_id] = True
374
except (TypeError, IndexError):
377
if theline[-1] != '\n':
378
noeols[version_id] = True
382
parent_should_go = False
384
if parent_noeols[version_id] == noeols[version_id]:
385
# no noeol toggle, so trust the weaves statement
386
# that this line is changed.
388
if parent_noeols[version_id]:
389
theline = theline + '\n'
390
elif parent_noeols[version_id]:
391
# parent has no eol, we do:
392
# our line is new, report as such..
394
elif noeols[version_id]:
395
# append a eol so that it looks like
397
theline = theline + '\n'
398
if parents[version_id] is not None:
399
#if last_parent_lines[version_id] is not None:
400
parent_should_go = True
401
if last_parent_lines[version_id] != theline:
404
#parent_should_go = False
406
current_hunks[version_id][2] += 1
407
current_hunks[version_id][3].append((inserted, theline))
409
# last hunk last parent line is not eaten
410
current_hunks[version_id][1] -= 1
411
if current_hunks[version_id][1] < 0:
412
current_hunks[version_id][1] = 0
413
# import pdb;pdb.set_trace()
414
# assert current_hunks[version_id][1] >= 0
418
version = self._idx_to_name(i)
419
if current_hunks[version] != [0, 0, 0, []]:
420
diff_hunks[version].append(tuple(current_hunks[version]))
422
for version_id in version_ids:
423
result[version_id] = (
427
diff_hunks[version_id],
431
def get_parents(self, version_id):
432
"""See VersionedFile.get_parent."""
433
return map(self._idx_to_name, self._parents[self._lookup(version_id)])
435
def _check_repeated_add(self, name, parents, text, sha1):
436
"""Check that a duplicated add is OK.
438
If it is, return the (old) index; otherwise raise an exception.
440
idx = self._lookup(name)
441
if sorted(self._parents[idx]) != sorted(parents) \
442
or sha1 != self._sha1s[idx]:
443
raise RevisionAlreadyPresent(name, self._weave_name)
446
@deprecated_method(zero_eight)
447
def add_identical(self, old_rev_id, new_rev_id, parents):
448
"""Please use Weave.clone_text now."""
449
return self.clone_text(new_rev_id, old_rev_id, parents)
451
def _add_lines(self, version_id, parents, lines, parent_texts):
452
"""See VersionedFile.add_lines."""
453
return self._add(version_id, lines, map(self._lookup, parents))
455
@deprecated_method(zero_eight)
456
def add(self, name, parents, text, sha1=None):
457
"""See VersionedFile.add_lines for the non deprecated api."""
458
return self._add(name, text, map(self._maybe_lookup, parents), sha1)
460
def _add(self, version_id, lines, parents, sha1=None):
461
"""Add a single text on top of the weave.
463
Returns the index number of the newly added version.
466
Symbolic name for this version.
467
(Typically the revision-id of the revision that added it.)
470
List or set of direct parent version numbers.
473
Sequence of lines to be added in the new version.
476
assert isinstance(version_id, basestring)
477
self._check_lines_not_unicode(lines)
478
self._check_lines_are_lines(lines)
480
sha1 = sha_strings(lines)
481
if version_id in self._name_map:
482
return self._check_repeated_add(version_id, parents, lines, sha1)
484
self._check_versions(parents)
485
## self._check_lines(lines)
486
new_version = len(self._parents)
488
# if we abort after here the (in-memory) weave will be corrupt because only
489
# some fields are updated
490
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
491
# - Robert Collins 20060226
492
self._parents.append(parents[:])
493
self._sha1s.append(sha1)
494
self._names.append(version_id)
495
self._name_map[version_id] = new_version
499
# special case; adding with no parents revision; can do
500
# this more quickly by just appending unconditionally.
501
# even more specially, if we're adding an empty text we
502
# need do nothing at all.
504
self._weave.append(('{', new_version))
505
self._weave.extend(lines)
506
self._weave.append(('}', None))
509
if len(parents) == 1:
510
pv = list(parents)[0]
511
if sha1 == self._sha1s[pv]:
512
# special case: same as the single parent
516
ancestors = self._inclusions(parents)
520
# basis a list of (origin, lineno, line)
523
for origin, lineno, line in self._extract(ancestors):
524
basis_lineno.append(lineno)
525
basis_lines.append(line)
527
# another small special case: a merge, producing the same text
529
if lines == basis_lines:
532
# add a sentinel, because we can also match against the final line
533
basis_lineno.append(len(self._weave))
535
# XXX: which line of the weave should we really consider
536
# matches the end of the file? the current code says it's the
537
# last line of the weave?
539
#print 'basis_lines:', basis_lines
540
#print 'new_lines: ', lines
542
s = self._matcher(None, basis_lines, lines)
544
# offset gives the number of lines that have been inserted
545
# into the weave up to the current point; if the original edit instruction
546
# says to change line A then we actually change (A+offset)
549
for tag, i1, i2, j1, j2 in s.get_opcodes():
550
# i1,i2 are given in offsets within basis_lines; we need to map them
551
# back to offsets within the entire weave
552
#print 'raw match', tag, i1, i2, j1, j2
556
i1 = basis_lineno[i1]
557
i2 = basis_lineno[i2]
559
assert 0 <= j1 <= j2 <= len(lines)
561
#print tag, i1, i2, j1, j2
563
# the deletion and insertion are handled separately.
564
# first delete the region.
566
self._weave.insert(i1+offset, ('[', new_version))
567
self._weave.insert(i2+offset+1, (']', new_version))
571
# there may have been a deletion spanning up to
572
# i2; we want to insert after this region to make sure
573
# we don't destroy ourselves
575
self._weave[i:i] = ([('{', new_version)]
578
offset += 2 + (j2 - j1)
581
def _clone_text(self, new_version_id, old_version_id, parents):
582
"""See VersionedFile.clone_text."""
583
old_lines = self.get_text(old_version_id)
584
self.add_lines(new_version_id, parents, old_lines)
586
def _inclusions(self, versions):
587
"""Return set of all ancestors of given version(s)."""
588
if not len(versions):
591
for v in xrange(max(versions), 0, -1):
593
# include all its parents
594
i.update(self._parents[v])
596
## except IndexError:
597
## raise ValueError("version %d not present in weave" % v)
599
@deprecated_method(zero_eight)
600
def inclusions(self, version_ids):
601
"""Deprecated - see VersionedFile.get_ancestry for the replacement."""
604
if isinstance(version_ids[0], int):
605
return [self._idx_to_name(v) for v in self._inclusions(version_ids)]
607
return self.get_ancestry(version_ids)
609
def get_ancestry(self, version_ids):
610
"""See VersionedFile.get_ancestry."""
611
if isinstance(version_ids, basestring):
612
version_ids = [version_ids]
613
i = self._inclusions([self._lookup(v) for v in version_ids])
614
return [self._idx_to_name(v) for v in i]
616
def _check_lines(self, text):
617
if not isinstance(text, list):
618
raise ValueError("text should be a list, not %s" % type(text))
621
if not isinstance(l, basestring):
622
raise ValueError("text line should be a string or unicode, not %s"
627
def _check_versions(self, indexes):
628
"""Check everything in the sequence of indexes is valid"""
633
raise IndexError("invalid version number %r" % i)
635
def _compatible_parents(self, my_parents, other_parents):
636
"""During join check that other_parents are joinable with my_parents.
638
Joinable is defined as 'is a subset of' - supersets may require
639
regeneration of diffs, but subsets do not.
641
return len(other_parents.difference(my_parents)) == 0
643
def annotate(self, version_id):
644
if isinstance(version_id, int):
645
warnings.warn('Weave.annotate(int) is deprecated. Please use version names'
646
' in all circumstances as of 0.8',
651
for origin, lineno, text in self._extract([version_id]):
652
result.append((origin, text))
655
return super(Weave, self).annotate(version_id)
657
def annotate_iter(self, version_id):
658
"""Yield list of (version-id, line) pairs for the specified version.
660
The index indicates when the line originated in the weave."""
661
incls = [self._lookup(version_id)]
662
for origin, lineno, text in self._extract(incls):
663
yield self._idx_to_name(origin), text
665
@deprecated_method(zero_eight)
667
"""_walk has become visit, a supported api."""
668
return self._walk_internal()
670
def iter_lines_added_or_present_in_versions(self, version_ids=None,
672
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
673
if version_ids is None:
674
version_ids = self.versions()
675
version_ids = set(version_ids)
676
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
677
# if inserted not in version_ids then it was inserted before the
678
# versions we care about, but because weaves cannot represent ghosts
679
# properly, we do not filter down to that
680
# if inserted not in version_ids: continue
686
#@deprecated_method(zero_eight)
687
def walk(self, version_ids=None):
688
"""See VersionedFile.walk."""
689
return self._walk_internal(version_ids)
691
def _walk_internal(self, version_ids=None):
692
"""Helper method for weave actions."""
697
lineno = 0 # line of weave, 0-based
699
for l in self._weave:
700
if l.__class__ == tuple:
704
istack.append(self._names[v])
708
assert self._names[v] not in dset
709
dset.add(self._names[v])
711
dset.remove(self._names[v])
713
raise WeaveFormatError('unexpected instruction %r' % v)
715
assert l.__class__ in (str, unicode)
717
yield lineno, istack[-1], frozenset(dset), l
721
raise WeaveFormatError("unclosed insertion blocks "
722
"at end of weave: %s" % istack)
724
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
727
def plan_merge(self, ver_a, ver_b):
728
"""Return pseudo-annotation indicating how the two versions merge.
730
This is computed between versions a and b and their common
733
Weave lines present in none of them are skipped entirely.
735
inc_a = set(self.get_ancestry([ver_a]))
736
inc_b = set(self.get_ancestry([ver_b]))
737
inc_c = inc_a & inc_b
739
for lineno, insert, deleteset, line in\
740
self.walk([ver_a, ver_b]):
741
if deleteset & inc_c:
742
# killed in parent; can't be in either a or b
743
# not relevant to our work
744
yield 'killed-base', line
745
elif insert in inc_c:
746
# was inserted in base
747
killed_a = bool(deleteset & inc_a)
748
killed_b = bool(deleteset & inc_b)
749
if killed_a and killed_b:
750
yield 'killed-both', line
752
yield 'killed-a', line
754
yield 'killed-b', line
756
yield 'unchanged', line
757
elif insert in inc_a:
758
if deleteset & inc_a:
759
yield 'ghost-a', line
763
elif insert in inc_b:
764
if deleteset & inc_b:
765
yield 'ghost-b', line
769
# not in either revision
770
yield 'irrelevant', line
772
yield 'unchanged', '' # terminator
774
def _extract(self, versions):
775
"""Yield annotation of lines in included set.
777
Yields a sequence of tuples (origin, lineno, text), where
778
origin is the origin version, lineno the index in the weave,
779
and text the text of the line.
781
The set typically but not necessarily corresponds to a version.
784
if not isinstance(i, int):
787
included = self._inclusions(versions)
793
lineno = 0 # line of weave, 0-based
799
WFE = WeaveFormatError
802
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
803
# +285282 0 1676.8040 1676.8040 +<isinstance>
804
# 1.6 seconds in 'isinstance'.
805
# changing the first isinstance:
806
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
807
# +140414 0 762.8050 762.8050 +<isinstance>
808
# note that the inline time actually dropped (less function calls)
809
# and total processing time was halved.
810
# we're still spending ~1/4 of the method in isinstance though.
811
# so lets hard code the acceptable string classes we expect:
812
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
813
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
815
# yay, down to ~1/4 the initial extract time, and our inline time
816
# has shrunk again, with isinstance no longer dominating.
817
# tweaking the stack inclusion test to use a set gives:
818
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
819
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
821
# - a 5% win, or possibly just noise. However with large istacks that
822
# 'in' test could dominate, so I'm leaving this change in place -
823
# when its fast enough to consider profiling big datasets we can review.
828
for l in self._weave:
829
if l.__class__ == tuple:
837
iset.remove(istack.pop())
848
assert l.__class__ in (str, unicode)
850
isactive = (not dset) and istack and (istack[-1] in included)
852
result.append((istack[-1], lineno, l))
855
raise WeaveFormatError("unclosed insertion blocks "
856
"at end of weave: %s" % istack)
858
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
862
@deprecated_method(zero_eight)
863
def get_iter(self, name_or_index):
864
"""Deprecated, please do not use. Lookups are not not needed.
866
Please use get_lines now.
868
return iter(self.get_lines(self._maybe_lookup(name_or_index)))
870
@deprecated_method(zero_eight)
871
def maybe_lookup(self, name_or_index):
872
"""Deprecated, please do not use. Lookups are not not needed."""
873
return self._maybe_lookup(name_or_index)
875
def _maybe_lookup(self, name_or_index):
876
"""Convert possible symbolic name to index, or pass through indexes.
880
if isinstance(name_or_index, (int, long)):
883
return self._lookup(name_or_index)
885
@deprecated_method(zero_eight)
886
def get(self, version_id):
887
"""Please use either Weave.get_text or Weave.get_lines as desired."""
888
return self.get_lines(version_id)
890
def get_lines(self, version_id):
891
"""See VersionedFile.get_lines()."""
892
int_index = self._maybe_lookup(version_id)
893
result = [line for (origin, lineno, line) in self._extract([int_index])]
894
expected_sha1 = self._sha1s[int_index]
895
measured_sha1 = sha_strings(result)
896
if measured_sha1 != expected_sha1:
897
raise errors.WeaveInvalidChecksum(
898
'file %s, revision %s, expected: %s, measured %s'
899
% (self._weave_name, version_id,
900
expected_sha1, measured_sha1))
903
def get_sha1(self, version_id):
904
"""See VersionedFile.get_sha1()."""
905
return self._sha1s[self._lookup(version_id)]
907
@deprecated_method(zero_eight)
908
def numversions(self):
909
"""How many versions are in this weave?
911
Deprecated in favour of num_versions.
913
return self.num_versions()
915
def num_versions(self):
916
"""How many versions are in this weave?"""
917
l = len(self._parents)
918
assert l == len(self._sha1s)
921
__len__ = num_versions
923
def check(self, progress_bar=None):
924
# TODO evaluate performance hit of using string sets in this routine.
925
# TODO: check no circular inclusions
926
# TODO: create a nested progress bar
927
for version in range(self.num_versions()):
928
inclusions = list(self._parents[version])
931
if inclusions[-1] >= version:
932
raise WeaveFormatError("invalid included version %d for index %d"
933
% (inclusions[-1], version))
935
# try extracting all versions; parallel extraction is used
936
nv = self.num_versions()
941
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
942
# The problem is that set membership is much more expensive
943
name = self._idx_to_name(i)
944
sha1s[name] = sha.new()
946
new_inc = set([name])
947
for p in self._parents[i]:
948
new_inc.update(inclusions[self._idx_to_name(p)])
950
assert set(new_inc) == set(self.get_ancestry(name)), \
951
'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
952
inclusions[name] = new_inc
954
nlines = len(self._weave)
956
update_text = 'checking weave'
958
short_name = os.path.basename(self._weave_name)
959
update_text = 'checking %s' % (short_name,)
960
update_text = update_text[:25]
962
for lineno, insert, deleteset, line in self._walk_internal():
964
progress_bar.update(update_text, lineno, nlines)
966
for name, name_inclusions in inclusions.items():
967
# The active inclusion must be an ancestor,
968
# and no ancestors must have deleted this line,
969
# because we don't support resurrection.
970
if (insert in name_inclusions) and not (deleteset & name_inclusions):
971
sha1s[name].update(line)
974
version = self._idx_to_name(i)
975
hd = sha1s[version].hexdigest()
976
expected = self._sha1s[i]
978
raise errors.WeaveInvalidChecksum(
979
"mismatched sha1 for version %s: "
980
"got %s, expected %s"
981
% (version, hd, expected))
983
# TODO: check insertions are properly nested, that there are
984
# no lines outside of insertion blocks, that deletions are
985
# properly paired, etc.
987
def _join(self, other, pb, msg, version_ids, ignore_missing):
988
"""Worker routine for join()."""
989
if not other.versions():
990
return # nothing to update, easy
993
# versions is never none, InterWeave checks this.
996
# two loops so that we do not change ourselves before verifying it
998
# work through in index order to make sure we get all dependencies
1001
# get the selected versions only that are in other.versions.
1002
version_ids = set(other.versions()).intersection(set(version_ids))
1003
# pull in the referenced graph.
1004
version_ids = other.get_ancestry(version_ids)
1005
pending_graph = [(version, other.get_parents(version)) for
1006
version in version_ids]
1007
for name in topo_sort(pending_graph):
1008
other_idx = other._name_map[name]
1009
# returns True if we have it, False if we need it.
1010
if not self._check_version_consistent(other, other_idx, name):
1011
names_to_join.append((other_idx, name))
1020
for other_idx, name in names_to_join:
1021
# TODO: If all the parents of the other version are already
1022
# present then we can avoid some work by just taking the delta
1023
# and adjusting the offsets.
1024
new_parents = self._imported_parents(other, other_idx)
1025
sha1 = other._sha1s[other_idx]
1030
pb.update(msg, merged, len(names_to_join))
1032
lines = other.get_lines(other_idx)
1033
self._add(name, lines, new_parents, sha1)
1035
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
1036
merged, processed, self._weave_name, time.time()-time0))
1038
def _imported_parents(self, other, other_idx):
1039
"""Return list of parents in self corresponding to indexes in other."""
1041
for parent_idx in other._parents[other_idx]:
1042
parent_name = other._names[parent_idx]
1043
if parent_name not in self._name_map:
1044
# should not be possible
1045
raise WeaveError("missing parent {%s} of {%s} in %r"
1046
% (parent_name, other._name_map[other_idx], self))
1047
new_parents.append(self._name_map[parent_name])
1050
def _check_version_consistent(self, other, other_idx, name):
1051
"""Check if a version in consistent in this and other.
1053
To be consistent it must have:
1056
* the same direct parents (by name, not index, and disregarding
1059
If present & correct return True;
1060
if not present in self return False;
1061
if inconsistent raise error."""
1062
this_idx = self._name_map.get(name, -1)
1064
if self._sha1s[this_idx] != other._sha1s[other_idx]:
1065
raise errors.WeaveTextDiffers(name, self, other)
1066
self_parents = self._parents[this_idx]
1067
other_parents = other._parents[other_idx]
1068
n1 = set([self._names[i] for i in self_parents])
1069
n2 = set([other._names[i] for i in other_parents])
1070
if not self._compatible_parents(n1, n2):
1071
raise WeaveParentMismatch("inconsistent parents "
1072
"for version {%s}: %s vs %s" % (name, n1, n2))
1078
@deprecated_method(zero_eight)
1079
def reweave(self, other, pb=None, msg=None):
1080
"""reweave has been superseded by plain use of join."""
1081
return self.join(other, pb, msg)
1083
def _reweave(self, other, pb, msg):
1084
"""Reweave self with other - internal helper for join().
1086
:param other: The other weave to merge
1087
:param pb: An optional progress bar, indicating how far done we are
1088
:param msg: An optional message for the progress
1090
new_weave = _reweave(self, other, pb=pb, msg=msg)
1091
self._copy_weave_content(new_weave)
1093
def _copy_weave_content(self, otherweave):
1094
"""adsorb the content from otherweave."""
1095
for attr in self.__slots__:
1096
if attr != '_weave_name':
1097
setattr(self, attr, copy(getattr(otherweave, attr)))
1100
class WeaveFile(Weave):
1101
"""A WeaveFile represents a Weave on disk and writes on change."""
1103
WEAVE_SUFFIX = '.weave'
1105
def __init__(self, name, transport, filemode=None, create=False, access_mode='w'):
1106
"""Create a WeaveFile.
1108
:param create: If not True, only open an existing knit.
1110
super(WeaveFile, self).__init__(name, access_mode)
1111
self._transport = transport
1112
self._filemode = filemode
1114
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
1115
except errors.NoSuchFile:
1121
def _add_lines(self, version_id, parents, lines, parent_texts):
1122
"""Add a version and save the weave."""
1123
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
1128
def _clone_text(self, new_version_id, old_version_id, parents):
1129
"""See VersionedFile.clone_text."""
1130
super(WeaveFile, self)._clone_text(new_version_id, old_version_id, parents)
1133
def copy_to(self, name, transport):
1134
"""See VersionedFile.copy_to()."""
1135
# as we are all in memory always, just serialise to the new place.
1137
write_weave_v5(self, sio)
1139
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
1141
def create_empty(self, name, transport, filemode=None):
1142
return WeaveFile(name, transport, filemode, create=True)
1145
"""Save the weave."""
1146
self._check_write_ok()
1148
write_weave_v5(self, sio)
1150
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
1156
"""See VersionedFile.get_suffixes()."""
1157
return [WeaveFile.WEAVE_SUFFIX]
1159
def join(self, other, pb=None, msg=None, version_ids=None,
1160
ignore_missing=False):
1161
"""Join other into self and save."""
1162
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1166
@deprecated_function(zero_eight)
1167
def reweave(wa, wb, pb=None, msg=None):
1168
"""reweaving is deprecation, please just use weave.join()."""
1169
_reweave(wa, wb, pb, msg)
1171
def _reweave(wa, wb, pb=None, msg=None):
1172
"""Combine two weaves and return the result.
1174
This works even if a revision R has different parents in
1175
wa and wb. In the resulting weave all the parents are given.
1177
This is done by just building up a new weave, maintaining ordering
1178
of the versions in the two inputs. More efficient approaches
1179
might be possible but it should only be necessary to do
1180
this operation rarely, when a new previously ghost version is
1183
:param pb: An optional progress bar, indicating how far done we are
1184
:param msg: An optional message for the progress
1188
queue_a = range(wa.num_versions())
1189
queue_b = range(wb.num_versions())
1190
# first determine combined parents of all versions
1191
# map from version name -> all parent names
1192
combined_parents = _reweave_parent_graphs(wa, wb)
1193
mutter("combined parents: %r", combined_parents)
1194
order = topo_sort(combined_parents.iteritems())
1195
mutter("order to reweave: %r", order)
1200
for idx, name in enumerate(order):
1202
pb.update(msg, idx, len(order))
1203
if name in wa._name_map:
1204
lines = wa.get_lines(name)
1205
if name in wb._name_map:
1206
lines_b = wb.get_lines(name)
1207
if lines != lines_b:
1208
mutter('Weaves differ on content. rev_id {%s}', name)
1209
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1211
lines = list(difflib.unified_diff(lines, lines_b,
1212
wa._weave_name, wb._weave_name))
1213
mutter('lines:\n%s', ''.join(lines))
1214
raise errors.WeaveTextDiffers(name, wa, wb)
1216
lines = wb.get_lines(name)
1217
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1220
def _reweave_parent_graphs(wa, wb):
1221
"""Return combined parent ancestry for two weaves.
1223
Returned as a list of (version_name, set(parent_names))"""
1225
for weave in [wa, wb]:
1226
for idx, name in enumerate(weave._names):
1227
p = combined.setdefault(name, set())
1228
p.update(map(weave._idx_to_name, weave._parents[idx]))
1233
"""Show the weave's table-of-contents"""
1234
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1235
for i in (6, 50, 10, 10):
1238
for i in range(w.num_versions()):
1241
parent_str = ' '.join(map(str, w._parents[i]))
1242
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1246
def weave_stats(weave_file, pb):
1247
from bzrlib.weavefile import read_weave
1249
wf = file(weave_file, 'rb')
1251
# FIXME: doesn't work on pipes
1252
weave_size = wf.tell()
1256
for i in range(vers):
1257
pb.update('checking sizes', i, vers)
1258
for origin, lineno, line in w._extract([i]):
1263
print 'versions %9d' % vers
1264
print 'weave file %9d bytes' % weave_size
1265
print 'total contents %9d bytes' % total
1266
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1269
print 'average size %9d bytes' % avg
1270
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1274
print """bzr weave tool
1276
Experimental tool for weave algorithm.
1279
weave init WEAVEFILE
1280
Create an empty weave file
1281
weave get WEAVEFILE VERSION
1282
Write out specified version.
1283
weave check WEAVEFILE
1284
Check consistency of all versions.
1286
Display table of contents.
1287
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1288
Add NEWTEXT, with specified parent versions.
1289
weave annotate WEAVEFILE VERSION
1290
Display origin of each line.
1291
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1292
Auto-merge two versions and display conflicts.
1293
weave diff WEAVEFILE VERSION1 VERSION2
1294
Show differences between two versions.
1298
% weave init foo.weave
1300
% weave add foo.weave ver0 < foo.txt
1303
(create updated version)
1305
% weave get foo.weave 0 | diff -u - foo.txt
1306
% weave add foo.weave ver1 0 < foo.txt
1309
% weave get foo.weave 0 > foo.txt (create forked version)
1311
% weave add foo.weave ver2 0 < foo.txt
1314
% weave merge foo.weave 1 2 > foo.txt (merge them)
1315
% vi foo.txt (resolve conflicts)
1316
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1328
# in case we're run directly from the subdirectory
1329
sys.path.append('..')
1331
from bzrlib.weavefile import write_weave, read_weave
1332
from bzrlib.progress import ProgressBar
1347
return read_weave(file(argv[2], 'rb'))
1353
# at the moment, based on everything in the file
1355
parents = map(int, argv[4:])
1356
lines = sys.stdin.readlines()
1357
ver = w.add(name, parents, lines)
1358
write_weave(w, file(argv[2], 'wb'))
1359
print 'added version %r %d' % (name, ver)
1362
if os.path.exists(fn):
1363
raise IOError("file exists")
1365
write_weave(w, file(fn, 'wb'))
1366
elif cmd == 'get': # get one version
1368
sys.stdout.writelines(w.get_iter(int(argv[3])))
1373
v1, v2 = map(int, argv[3:5])
1376
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1377
'%s version %d' % (fn, v1),
1378
'%s version %d' % (fn, v2))
1379
sys.stdout.writelines(diff_gen)
1381
elif cmd == 'annotate':
1383
# newline is added to all lines regardless; too hard to get
1384
# reasonable formatting otherwise
1386
for origin, text in w.annotate(int(argv[3])):
1387
text = text.rstrip('\r\n')
1389
print ' | %s' % (text)
1391
print '%5d | %s' % (origin, text)
1397
elif cmd == 'stats':
1398
weave_stats(argv[2], ProgressBar())
1400
elif cmd == 'check':
1405
print '%d versions ok' % w.num_versions()
1407
elif cmd == 'inclusions':
1409
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1411
elif cmd == 'parents':
1413
print ' '.join(map(str, w._parents[int(argv[3])]))
1415
elif cmd == 'plan-merge':
1416
# replaced by 'bzr weave-plan-merge'
1418
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1420
print '%14s | %s' % (state, line),
1421
elif cmd == 'merge':
1422
# replaced by 'bzr weave-merge-text'
1424
p = w.plan_merge(int(argv[3]), int(argv[4]))
1425
sys.stdout.writelines(w.weave_merge(p))
1427
raise ValueError('unknown command %r' % cmd)
1430
if __name__ == '__main__':
1432
sys.exit(main(sys.argv))
1435
class InterWeave(InterVersionedFile):
1436
"""Optimised code paths for weave to weave operations."""
1438
_matching_file_from_factory = staticmethod(WeaveFile)
1439
_matching_file_to_factory = staticmethod(WeaveFile)
1442
def is_compatible(source, target):
1443
"""Be compatible with weaves."""
1445
return (isinstance(source, Weave) and
1446
isinstance(target, Weave))
1447
except AttributeError:
1450
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1451
"""See InterVersionedFile.join."""
1452
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1453
if self.target.versions() == [] and version_ids is None:
1454
self.target._copy_weave_content(self.source)
1457
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1458
except errors.WeaveParentMismatch:
1459
self.target._reweave(self.source, pb, msg)
1462
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
