50
50
# have slight specializations for different ways its used: annotate,
51
51
# basis for add, get, etc.
53
# TODO: Perhaps the API should work only in names to hide the integer
54
# indexes from the user?
53
# TODO: Probably the API should work only in names to hide the integer
54
# indexes from the user.
56
56
# TODO: Is there any potential performance win by having an add()
57
57
# variant that is passed a pre-cooked version of the single basis
60
# TODO: Have a way to go back and insert a revision V1 that is a parent
61
# of an already-stored revision V2. This means some lines previously
62
# counted as new in V2 will be discovered to have actually come from V1.
63
# It is probably necessary to insert V1, then compute a whole new diff
64
# from the mashed ancestors to V2. This must be repeated for every
65
# direct child of V1. The deltas from V2 to its descendents won't change,
66
# although their location within the weave may change. It may be possible
67
# to just adjust the location of those instructions rather than
68
# re-weaving the whole thing. This is expected to be a fairly rare
69
# operation, only used when inserting data that was previously a ghost.
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
75
from difflib import SequenceMatcher
80
class WeaveError(Exception):
81
"""Exception in processing weave"""
84
class WeaveFormatError(WeaveError):
85
"""Weave invariant violated"""
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):
89
101
"""weave - versioned text file storage.
91
103
A Weave manages versions of line-based text files, keeping track
205
224
return self._parents == other._parents \
206
225
and self._weave == other._weave \
207
226
and self._sha1s == other._sha1s
210
228
def __ne__(self, other):
211
229
return not self.__eq__(other)
214
def maybe_lookup(self, name_or_index):
215
"""Convert possible symbolic name to index, or pass through indexes."""
216
if isinstance(name_or_index, (int, long)):
219
return self.lookup(name_or_index)
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)
222
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):
223
249
"""Convert symbolic version name to index."""
250
self.check_not_reserved_id(name)
225
252
return self._name_map[name]
227
raise WeaveError("name %r not present in weave %r" %
228
(name, self._weave_name))
254
raise RevisionNotPresent(name, self._weave_name)
256
@deprecated_method(zero_eight)
231
257
def iter_names(self):
232
"""Yield a list of all names in this weave."""
233
return iter(self._names)
235
def idx_to_name(self, version):
236
return self._names[version]
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)])
239
436
def _check_repeated_add(self, name, parents, text, sha1):
240
437
"""Check that a duplicated add is OK.
242
439
If it is, return the (old) index; otherwise raise an exception.
244
idx = self.lookup(name)
245
if sorted(self._parents[idx]) != sorted(parents):
246
raise WeaveError("name \"%s\" already present in weave "
247
"with different parents" % name)
248
if sha1 != self._sha1s[idx]:
249
raise WeaveError("name \"%s\" already present in weave "
250
"with different text" % name)
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
"""See VersionedFile.add_lines."""
454
return self._add(version_id, lines, map(self._lookup, parents))
456
@deprecated_method(zero_eight)
255
457
def add(self, name, parents, text, sha1=None):
458
"""See VersionedFile.add_lines for the non deprecated api."""
459
return self._add(name, text, map(self._maybe_lookup, parents), sha1)
461
def _add(self, version_id, lines, parents, sha1=None):
256
462
"""Add a single text on top of the weave.
258
464
Returns the index number of the newly added version.
261
467
Symbolic name for this version.
262
468
(Typically the revision-id of the revision that added it.)
265
471
List or set of direct parent version numbers.
268
474
Sequence of lines to be added in the new version.
270
sha -- SHA-1 of the file, if known. This is trusted to be
273
from bzrlib.osutils import sha_strings
275
assert isinstance(name, basestring)
277
sha1 = sha_strings(text)
278
if name in self._name_map:
279
return self._check_repeated_add(name, parents, text, sha1)
281
parents = map(self.maybe_lookup, parents)
477
assert isinstance(version_id, basestring)
478
self._check_lines_not_unicode(lines)
479
self._check_lines_are_lines(lines)
481
sha1 = sha_strings(lines)
482
if version_id in self._name_map:
483
return self._check_repeated_add(version_id, parents, lines, sha1)
282
485
self._check_versions(parents)
283
## self._check_lines(text)
486
## self._check_lines(lines)
284
487
new_version = len(self._parents)
287
489
# if we abort after here the (in-memory) weave will be corrupt because only
288
490
# some fields are updated
491
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
492
# - Robert Collins 20060226
289
493
self._parents.append(parents[:])
290
494
self._sha1s.append(sha1)
291
self._names.append(name)
292
self._name_map[name] = new_version
495
self._names.append(version_id)
496
self._name_map[version_id] = new_version
440
633
except IndexError:
441
634
raise IndexError("invalid version number %r" % i)
636
def _compatible_parents(self, my_parents, other_parents):
637
"""During join check that other_parents are joinable with my_parents.
639
Joinable is defined as 'is a subset of' - supersets may require
640
regeneration of diffs, but subsets do not.
642
return len(other_parents.difference(my_parents)) == 0
644
def annotate(self, version_id):
645
if isinstance(version_id, int):
646
warnings.warn('Weave.annotate(int) is deprecated. Please use version names'
647
' in all circumstances as of 0.8',
652
for origin, lineno, text in self._extract([version_id]):
653
result.append((origin, text))
656
return super(Weave, self).annotate(version_id)
444
def annotate(self, name_or_index):
445
return list(self.annotate_iter(name_or_index))
448
def annotate_iter(self, name_or_index):
449
"""Yield list of (index-id, line) pairs for the specified version.
658
def annotate_iter(self, version_id):
659
"""Yield list of (version-id, line) pairs for the specified version.
451
661
The index indicates when the line originated in the weave."""
452
incls = [self.maybe_lookup(name_or_index)]
662
incls = [self._lookup(version_id)]
453
663
for origin, lineno, text in self._extract(incls):
664
yield self._idx_to_name(origin), text
666
@deprecated_method(zero_eight)
461
(lineno, insert, deletes, text)
462
for each literal line.
668
"""_walk has become visit, a supported api."""
669
return self._walk_internal()
671
def iter_lines_added_or_present_in_versions(self, version_ids=None,
673
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
674
if version_ids is None:
675
version_ids = self.versions()
676
version_ids = set(version_ids)
677
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
678
# if inserted not in version_ids then it was inserted before the
679
# versions we care about, but because weaves cannot represent ghosts
680
# properly, we do not filter down to that
681
# if inserted not in version_ids: continue
687
#@deprecated_method(zero_eight)
688
def walk(self, version_ids=None):
689
"""See VersionedFile.walk."""
690
return self._walk_internal(version_ids)
692
def _walk_internal(self, version_ids=None):
693
"""Helper method for weave actions."""
539
assert isinstance(l, basestring)
849
assert l.__class__ in (str, unicode)
540
850
if isactive is None:
541
851
isactive = (not dset) and istack and (istack[-1] in included)
543
853
result.append((istack[-1], lineno, l))
547
raise WFE("unclosed insertion blocks at end of weave",
856
raise WeaveFormatError("unclosed insertion blocks "
857
"at end of weave: %s" % istack)
550
raise WFE("unclosed deletion blocks at end of weave",
859
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
863
@deprecated_method(zero_eight)
557
864
def get_iter(self, name_or_index):
558
"""Yield lines for the specified version."""
559
incls = [self.maybe_lookup(name_or_index)]
560
for origin, lineno, line in self._extract(incls):
564
def get_text(self, name_or_index):
565
return ''.join(self.get_iter(name_or_index))
566
assert isinstance(version, int)
569
def get_lines(self, name_or_index):
570
return list(self.get_iter(name_or_index))
576
def mash_iter(self, included):
577
"""Return composed version of multiple included versions."""
578
included = map(self.maybe_lookup, included)
579
for origin, lineno, text in self._extract(included):
583
def dump(self, to_file):
584
from pprint import pprint
585
print >>to_file, "Weave._weave = ",
586
pprint(self._weave, to_file)
587
print >>to_file, "Weave._parents = ",
588
pprint(self._parents, to_file)
865
"""Deprecated, please do not use. Lookups are not not needed.
867
Please use get_lines now.
869
return iter(self.get_lines(self._maybe_lookup(name_or_index)))
871
@deprecated_method(zero_eight)
872
def maybe_lookup(self, name_or_index):
873
"""Deprecated, please do not use. Lookups are not not needed."""
874
return self._maybe_lookup(name_or_index)
876
def _maybe_lookup(self, name_or_index):
877
"""Convert possible symbolic name to index, or pass through indexes.
881
if isinstance(name_or_index, (int, long)):
884
return self._lookup(name_or_index)
886
@deprecated_method(zero_eight)
887
def get(self, version_id):
888
"""Please use either Weave.get_text or Weave.get_lines as desired."""
889
return self.get_lines(version_id)
891
def get_lines(self, version_id):
892
"""See VersionedFile.get_lines()."""
893
int_index = self._maybe_lookup(version_id)
894
result = [line for (origin, lineno, line) in self._extract([int_index])]
895
expected_sha1 = self._sha1s[int_index]
896
measured_sha1 = sha_strings(result)
897
if measured_sha1 != expected_sha1:
898
raise errors.WeaveInvalidChecksum(
899
'file %s, revision %s, expected: %s, measured %s'
900
% (self._weave_name, version_id,
901
expected_sha1, measured_sha1))
904
def get_sha1(self, version_id):
905
"""See VersionedFile.get_sha1()."""
906
return self._sha1s[self._lookup(version_id)]
908
@deprecated_method(zero_eight)
592
909
def numversions(self):
910
"""How many versions are in this weave?
912
Deprecated in favour of num_versions.
914
return self.num_versions()
916
def num_versions(self):
917
"""How many versions are in this weave?"""
593
918
l = len(self._parents)
594
919
assert l == len(self._sha1s)
599
return self.numversions()
922
__len__ = num_versions
602
924
def check(self, progress_bar=None):
603
# check no circular inclusions
604
for version in range(self.numversions()):
925
# TODO evaluate performance hit of using string sets in this routine.
926
# TODO: check no circular inclusions
927
# TODO: create a nested progress bar
928
for version in range(self.num_versions()):
605
929
inclusions = list(self._parents[version])
607
931
inclusions.sort()
609
933
raise WeaveFormatError("invalid included version %d for index %d"
610
934
% (inclusions[-1], version))
612
# try extracting all versions; this is a bit slow and parallel
613
# extraction could be used
614
nv = self.numversions()
615
for version in range(nv):
936
# try extracting all versions; parallel extraction is used
937
nv = self.num_versions()
942
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
943
# The problem is that set membership is much more expensive
944
name = self._idx_to_name(i)
945
sha1s[name] = sha.new()
947
new_inc = set([name])
948
for p in self._parents[i]:
949
new_inc.update(inclusions[self._idx_to_name(p)])
951
assert set(new_inc) == set(self.get_ancestry(name)), \
952
'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
953
inclusions[name] = new_inc
955
nlines = len(self._weave)
957
update_text = 'checking weave'
959
short_name = os.path.basename(self._weave_name)
960
update_text = 'checking %s' % (short_name,)
961
update_text = update_text[:25]
963
for lineno, insert, deleteset, line in self._walk_internal():
617
progress_bar.update('checking text', version, nv)
619
for l in self.get_iter(version):
622
expected = self._sha1s[version]
965
progress_bar.update(update_text, lineno, nlines)
967
for name, name_inclusions in inclusions.items():
968
# The active inclusion must be an ancestor,
969
# and no ancestors must have deleted this line,
970
# because we don't support resurrection.
971
if (insert in name_inclusions) and not (deleteset & name_inclusions):
972
sha1s[name].update(line)
975
version = self._idx_to_name(i)
976
hd = sha1s[version].hexdigest()
977
expected = self._sha1s[i]
623
978
if hd != expected:
624
raise WeaveError("mismatched sha1 for version %d; "
625
"got %s, expected %s"
626
% (version, hd, expected))
979
raise errors.WeaveInvalidChecksum(
980
"mismatched sha1 for version %s: "
981
"got %s, expected %s"
982
% (version, hd, expected))
628
984
# TODO: check insertions are properly nested, that there are
629
985
# no lines outside of insertion blocks, that deletions are
630
986
# properly paired, etc.
634
def merge(self, merge_versions):
635
"""Automerge and mark conflicts between versions.
637
This returns a sequence, each entry describing alternatives
638
for a chunk of the file. Each of the alternatives is given as
641
If there is a chunk of the file where there's no diagreement,
642
only one alternative is given.
644
# approach: find the included versions common to all the
646
raise NotImplementedError()
650
def _delta(self, included, lines):
651
"""Return changes from basis to new revision.
653
The old text for comparison is the union of included revisions.
655
This is used in inserting a new text.
657
Delta is returned as a sequence of
658
(weave1, weave2, newlines).
660
This indicates that weave1:weave2 of the old weave should be
661
replaced by the sequence of lines in newlines. Note that
662
these line numbers are positions in the total weave and don't
663
correspond to the lines in any extracted version, or even the
664
extracted union of included versions.
666
If line1=line2, this is a pure insert; if newlines=[] this is a
667
pure delete. (Similar to difflib.)
669
raise NotImplementedError()
672
def plan_merge(self, ver_a, ver_b):
673
"""Return pseudo-annotation indicating how the two versions merge.
675
This is computed between versions a and b and their common
678
Weave lines present in none of them are skipped entirely.
680
inc_a = self.inclusions([ver_a])
681
inc_b = self.inclusions([ver_b])
682
inc_c = inc_a & inc_b
684
for lineno, insert, deleteset, line in self._walk():
685
if deleteset & inc_c:
686
# killed in parent; can't be in either a or b
687
# not relevant to our work
688
yield 'killed-base', line
689
elif insert in inc_c:
690
# was inserted in base
691
killed_a = bool(deleteset & inc_a)
692
killed_b = bool(deleteset & inc_b)
693
if killed_a and killed_b:
694
yield 'killed-both', line
696
yield 'killed-a', line
698
yield 'killed-b', line
700
yield 'unchanged', line
701
elif insert in inc_a:
702
if deleteset & inc_a:
703
yield 'ghost-a', line
707
elif insert in inc_b:
708
if deleteset & inc_b:
709
yield 'ghost-b', line
713
# not in either revision
714
yield 'irrelevant', line
716
yield 'unchanged', '' # terminator
720
def weave_merge(self, plan):
725
for state, line in plan:
726
if state == 'unchanged' or state == 'killed-both':
727
# resync and flush queued conflicts changes if any
728
if not lines_a and not lines_b:
730
elif ch_a and not ch_b:
732
for l in lines_a: yield l
733
elif ch_b and not ch_a:
734
for l in lines_b: yield l
735
elif lines_a == lines_b:
736
for l in lines_a: yield l
739
for l in lines_a: yield l
741
for l in lines_b: yield l
748
if state == 'unchanged':
751
elif state == 'killed-a':
754
elif state == 'killed-b':
757
elif state == 'new-a':
760
elif state == 'new-b':
764
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
769
def join(self, other):
770
"""Integrate versions from other into this weave.
772
The resulting weave contains all the history of both weaves;
773
any version you could retrieve from either self or other can be
774
retrieved from self after this call.
776
It is illegal for the two weaves to contain different values
779
if other.numversions() == 0:
988
def _join(self, other, pb, msg, version_ids, ignore_missing):
989
"""Worker routine for join()."""
990
if not other.versions():
780
991
return # nothing to update, easy
994
# versions is never none, InterWeave checks this.
997
# two loops so that we do not change ourselves before verifying it
781
999
# work through in index order to make sure we get all dependencies
782
for other_idx, name in enumerate(other._names):
783
# TODO: If all the parents of the other version are already
1002
# get the selected versions only that are in other.versions.
1003
version_ids = set(other.versions()).intersection(set(version_ids))
1004
# pull in the referenced graph.
1005
version_ids = other.get_ancestry(version_ids)
1006
pending_graph = [(version, other.get_parents(version)) for
1007
version in version_ids]
1008
for name in topo_sort(pending_graph):
1009
other_idx = other._name_map[name]
1010
# returns True if we have it, False if we need it.
1011
if not self._check_version_consistent(other, other_idx, name):
1012
names_to_join.append((other_idx, name))
1021
for other_idx, name in names_to_join:
1022
# TODO: If all the parents of the other version are already
784
1023
# present then we can avoid some work by just taking the delta
785
1024
# and adjusting the offsets.
786
if self._check_version_consistent(other, other_idx, name):
788
1025
new_parents = self._imported_parents(other, other_idx)
1026
sha1 = other._sha1s[other_idx]
1031
pb.update(msg, merged, len(names_to_join))
789
1033
lines = other.get_lines(other_idx)
790
sha1 = other._sha1s[other_idx]
791
self.add(name, new_parents, lines, sha1)
1034
self._add(name, lines, new_parents, sha1)
1036
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
1037
merged, processed, self._weave_name, time.time()-time0))
794
1039
def _imported_parents(self, other, other_idx):
795
1040
"""Return list of parents in self corresponding to indexes in other."""
796
1041
new_parents = []
797
1042
for parent_idx in other._parents[other_idx]:
798
1043
parent_name = other._names[parent_idx]
799
if parent_name not in self._names:
1044
if parent_name not in self._name_map:
800
1045
# should not be possible
801
1046
raise WeaveError("missing parent {%s} of {%s} in %r"
802
1047
% (parent_name, other._name_map[other_idx], self))
818
1063
this_idx = self._name_map.get(name, -1)
819
1064
if this_idx != -1:
820
1065
if self._sha1s[this_idx] != other._sha1s[other_idx]:
821
raise WeaveError("inconsistent texts for version {%s} "
822
"when joining weaves"
1066
raise errors.WeaveTextDiffers(name, self, other)
824
1067
self_parents = self._parents[this_idx]
825
1068
other_parents = other._parents[other_idx]
826
n1 = [self._names[i] for i in self_parents]
827
n2 = [other._names[i] for i in other_parents]
831
raise WeaveError("inconsistent parents for version {%s}: "
1069
n1 = set([self._names[i] for i in self_parents])
1070
n2 = set([other._names[i] for i in other_parents])
1071
if not self._compatible_parents(n1, n2):
1072
raise WeaveParentMismatch("inconsistent parents "
1073
"for version {%s}: %s vs %s" % (name, n1, n2))
835
1075
return True # ok!
1079
@deprecated_method(zero_eight)
1080
def reweave(self, other, pb=None, msg=None):
1081
"""reweave has been superseded by plain use of join."""
1082
return self.join(other, pb, msg)
1084
def _reweave(self, other, pb, msg):
1085
"""Reweave self with other - internal helper for join().
1087
:param other: The other weave to merge
1088
:param pb: An optional progress bar, indicating how far done we are
1089
:param msg: An optional message for the progress
1091
new_weave = _reweave(self, other, pb=pb, msg=msg)
1092
self._copy_weave_content(new_weave)
1094
def _copy_weave_content(self, otherweave):
1095
"""adsorb the content from otherweave."""
1096
for attr in self.__slots__:
1097
if attr != '_weave_name':
1098
setattr(self, attr, copy(getattr(otherweave, attr)))
1101
class WeaveFile(Weave):
1102
"""A WeaveFile represents a Weave on disk and writes on change."""
1104
WEAVE_SUFFIX = '.weave'
1106
def __init__(self, name, transport, filemode=None, create=False, access_mode='w'):
1107
"""Create a WeaveFile.
1109
:param create: If not True, only open an existing knit.
1111
super(WeaveFile, self).__init__(name, access_mode)
1112
self._transport = transport
1113
self._filemode = filemode
1115
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
1116
except errors.NoSuchFile:
1122
def _add_lines(self, version_id, parents, lines, parent_texts):
1123
"""Add a version and save the weave."""
1124
self.check_not_reserved_id(version_id)
1125
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
1130
def _clone_text(self, new_version_id, old_version_id, parents):
1131
"""See VersionedFile.clone_text."""
1132
super(WeaveFile, self)._clone_text(new_version_id, old_version_id, parents)
1135
def copy_to(self, name, transport):
1136
"""See VersionedFile.copy_to()."""
1137
# as we are all in memory always, just serialise to the new place.
1139
write_weave_v5(self, sio)
1141
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
1143
def create_empty(self, name, transport, filemode=None):
1144
return WeaveFile(name, transport, filemode, create=True)
1147
"""Save the weave."""
1148
self._check_write_ok()
1150
write_weave_v5(self, sio)
1152
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
1158
"""See VersionedFile.get_suffixes()."""
1159
return [WeaveFile.WEAVE_SUFFIX]
1161
def join(self, other, pb=None, msg=None, version_ids=None,
1162
ignore_missing=False):
1163
"""Join other into self and save."""
1164
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1168
@deprecated_function(zero_eight)
1169
def reweave(wa, wb, pb=None, msg=None):
1170
"""reweaving is deprecation, please just use weave.join()."""
1171
_reweave(wa, wb, pb, msg)
1173
def _reweave(wa, wb, pb=None, msg=None):
1174
"""Combine two weaves and return the result.
1176
This works even if a revision R has different parents in
1177
wa and wb. In the resulting weave all the parents are given.
1179
This is done by just building up a new weave, maintaining ordering
1180
of the versions in the two inputs. More efficient approaches
1181
might be possible but it should only be necessary to do
1182
this operation rarely, when a new previously ghost version is
1185
:param pb: An optional progress bar, indicating how far done we are
1186
:param msg: An optional message for the progress
1190
queue_a = range(wa.num_versions())
1191
queue_b = range(wb.num_versions())
1192
# first determine combined parents of all versions
1193
# map from version name -> all parent names
1194
combined_parents = _reweave_parent_graphs(wa, wb)
1195
mutter("combined parents: %r", combined_parents)
1196
order = topo_sort(combined_parents.iteritems())
1197
mutter("order to reweave: %r", order)
1202
for idx, name in enumerate(order):
1204
pb.update(msg, idx, len(order))
1205
if name in wa._name_map:
1206
lines = wa.get_lines(name)
1207
if name in wb._name_map:
1208
lines_b = wb.get_lines(name)
1209
if lines != lines_b:
1210
mutter('Weaves differ on content. rev_id {%s}', name)
1211
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1213
lines = list(difflib.unified_diff(lines, lines_b,
1214
wa._weave_name, wb._weave_name))
1215
mutter('lines:\n%s', ''.join(lines))
1216
raise errors.WeaveTextDiffers(name, wa, wb)
1218
lines = wb.get_lines(name)
1219
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1222
def _reweave_parent_graphs(wa, wb):
1223
"""Return combined parent ancestry for two weaves.
1225
Returned as a list of (version_name, set(parent_names))"""
1227
for weave in [wa, wb]:
1228
for idx, name in enumerate(weave._names):
1229
p = combined.setdefault(name, set())
1230
p.update(map(weave._idx_to_name, weave._parents[idx]))
840
1234
def weave_toc(w):
841
1235
"""Show the weave's table-of-contents"""
1028
1415
print ' '.join(map(str, w._parents[int(argv[3])]))
1030
1417
elif cmd == 'plan-merge':
1418
# replaced by 'bzr weave-plan-merge'
1032
1420
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1034
1422
print '%14s | %s' % (state, line),
1036
1423
elif cmd == 'merge':
1424
# replaced by 'bzr weave-merge-text'
1038
1426
p = w.plan_merge(int(argv[3]), int(argv[4]))
1039
1427
sys.stdout.writelines(w.weave_merge(p))
1041
elif cmd == 'mash-merge':
1047
v1, v2 = map(int, argv[3:5])
1049
basis = w.inclusions([v1]).intersection(w.inclusions([v2]))
1051
base_lines = list(w.mash_iter(basis))
1052
a_lines = list(w.get(v1))
1053
b_lines = list(w.get(v2))
1055
from bzrlib.merge3 import Merge3
1056
m3 = Merge3(base_lines, a_lines, b_lines)
1058
name_a = 'version %d' % v1
1059
name_b = 'version %d' % v2
1060
sys.stdout.writelines(m3.merge_lines(name_a=name_a, name_b=name_b))
1062
1429
raise ValueError('unknown command %r' % cmd)
1066
def profile_main(argv):
1067
import tempfile, hotshot, hotshot.stats
1069
prof_f = tempfile.NamedTemporaryFile()
1071
prof = hotshot.Profile(prof_f.name)
1073
ret = prof.runcall(main, argv)
1076
stats = hotshot.stats.load(prof_f.name)
1078
stats.sort_stats('cumulative')
1079
## XXX: Might like to write to stderr or the trace file instead but
1080
## print_stats seems hardcoded to stdout
1081
stats.print_stats(20)
1086
1432
if __name__ == '__main__':
1088
if '--profile' in sys.argv:
1090
args.remove('--profile')
1091
sys.exit(profile_main(args))
1093
sys.exit(main(sys.argv))
1434
sys.exit(main(sys.argv))
1437
class InterWeave(InterVersionedFile):
1438
"""Optimised code paths for weave to weave operations."""
1440
_matching_file_from_factory = staticmethod(WeaveFile)
1441
_matching_file_to_factory = staticmethod(WeaveFile)
1444
def is_compatible(source, target):
1445
"""Be compatible with weaves."""
1447
return (isinstance(source, Weave) and
1448
isinstance(target, Weave))
1449
except AttributeError:
1452
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1453
"""See InterVersionedFile.join."""
1454
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1455
if self.target.versions() == [] and version_ids is None:
1456
self.target._copy_weave_content(self.source)
1459
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1460
except errors.WeaveParentMismatch:
1461
self.target._reweave(self.source, pb, msg)
1464
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
