50
50
# have slight specializations for different ways its used: annotate,
51
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.
53
# TODO: Perhaps 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: 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
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.
78
from bzrlib.trace import mutter
79
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
80
RevisionAlreadyPresent,
82
WeaveRevisionAlreadyPresent,
83
WeaveRevisionNotPresent,
85
import bzrlib.errors as errors
86
from bzrlib.osutils import sha_strings
87
import bzrlib.patiencediff
88
from bzrlib.symbol_versioning import (deprecated_method,
92
from bzrlib.tsort import topo_sort
93
from bzrlib.versionedfile import VersionedFile, InterVersionedFile
94
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
97
class Weave(VersionedFile):
75
from difflib import SequenceMatcher
80
class WeaveError(Exception):
81
"""Exception in processing weave"""
84
class WeaveFormatError(WeaveError):
85
"""Weave invariant violated"""
98
89
"""weave - versioned text file storage.
100
91
A Weave manages versions of line-based text files, keeping track
221
205
return self._parents == other._parents \
222
206
and self._weave == other._weave \
223
207
and self._sha1s == other._sha1s
225
210
def __ne__(self, other):
226
211
return not self.__eq__(other)
228
@deprecated_method(zero_eight)
229
def idx_to_name(self, index):
230
"""Old public interface, the public interface is all names now."""
233
def _idx_to_name(self, version):
234
return self._names[version]
236
@deprecated_method(zero_eight)
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)
237
222
def lookup(self, name):
238
"""Backwards compatibility thunk:
240
Return name, as name is valid in the api now, and spew deprecation
245
def _lookup(self, name):
246
223
"""Convert symbolic version name to index."""
248
225
return self._name_map[name]
250
raise RevisionNotPresent(name, self._weave_name)
252
@deprecated_method(zero_eight)
227
raise WeaveError("name %r not present in weave %r" %
228
(name, self._weave_name))
253
231
def iter_names(self):
254
"""Deprecated convenience function, please see VersionedFile.names()."""
255
return iter(self.names())
257
@deprecated_method(zero_eight)
259
"""See Weave.versions for the current api."""
260
return self.versions()
263
"""See VersionedFile.versions."""
264
return self._names[:]
266
def has_version(self, version_id):
267
"""See VersionedFile.has_version."""
268
return (version_id in self._name_map)
270
__contains__ = has_version
272
def get_delta(self, version_id):
273
"""See VersionedFile.get_delta."""
274
return self.get_deltas([version_id])[version_id]
276
def get_deltas(self, version_ids):
277
"""See VersionedFile.get_deltas."""
278
version_ids = self.get_ancestry(version_ids)
279
for version_id in version_ids:
280
if not self.has_version(version_id):
281
raise RevisionNotPresent(version_id, self)
282
# try extracting all versions; parallel extraction is used
283
nv = self.num_versions()
289
last_parent_lines = {}
291
parent_inclusions = {}
296
# its simplest to generate a full set of prepared variables.
298
name = self._names[i]
299
sha1s[name] = self.get_sha1(name)
300
parents_list = self.get_parents(name)
302
parent = parents_list[0]
303
parents[name] = parent
304
parent_inclusions[name] = inclusions[parent]
307
parent_inclusions[name] = set()
308
# we want to emit start, finish, replacement_length, replacement_lines tuples.
309
diff_hunks[name] = []
310
current_hunks[name] = [0, 0, 0, []] # #start, finish, repl_length, repl_tuples
311
parent_linenums[name] = 0
313
parent_noeols[name] = False
314
last_parent_lines[name] = None
315
new_inc = set([name])
316
for p in self._parents[i]:
317
new_inc.update(inclusions[self._idx_to_name(p)])
318
# debug only, known good so far.
319
#assert set(new_inc) == set(self.get_ancestry(name)), \
320
# 'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
321
inclusions[name] = new_inc
323
nlines = len(self._weave)
325
for lineno, inserted, deletes, line in self._walk_internal():
326
# a line is active in a version if:
327
# insert is in the versions inclusions
329
# deleteset & the versions inclusions is an empty set.
330
# so - if we have a included by mapping - version is included by
331
# children, we get a list of children to examine for deletes affect
332
# ing them, which is less than the entire set of children.
333
for version_id in version_ids:
334
# The active inclusion must be an ancestor,
335
# and no ancestors must have deleted this line,
336
# because we don't support resurrection.
337
parent_inclusion = parent_inclusions[version_id]
338
inclusion = inclusions[version_id]
339
parent_active = inserted in parent_inclusion and not (deletes & parent_inclusion)
340
version_active = inserted in inclusion and not (deletes & inclusion)
341
if not parent_active and not version_active:
342
# unrelated line of ancestry
344
elif parent_active and version_active:
346
parent_linenum = parent_linenums[version_id]
347
if current_hunks[version_id] != [parent_linenum, parent_linenum, 0, []]:
348
diff_hunks[version_id].append(tuple(current_hunks[version_id]))
350
current_hunks[version_id] = [parent_linenum, parent_linenum, 0, []]
351
parent_linenums[version_id] = parent_linenum
354
noeols[version_id] = True
357
elif parent_active and not version_active:
359
current_hunks[version_id][1] += 1
360
parent_linenums[version_id] += 1
361
last_parent_lines[version_id] = line
362
elif not parent_active and version_active:
364
# noeol only occurs at the end of a file because we
365
# diff linewise. We want to show noeol changes as a
366
# empty diff unless the actual eol-less content changed.
369
if last_parent_lines[version_id][-1] != '\n':
370
parent_noeols[version_id] = True
371
except (TypeError, IndexError):
374
if theline[-1] != '\n':
375
noeols[version_id] = True
379
parent_should_go = False
381
if parent_noeols[version_id] == noeols[version_id]:
382
# no noeol toggle, so trust the weaves statement
383
# that this line is changed.
385
if parent_noeols[version_id]:
386
theline = theline + '\n'
387
elif parent_noeols[version_id]:
388
# parent has no eol, we do:
389
# our line is new, report as such..
391
elif noeols[version_id]:
392
# append a eol so that it looks like
394
theline = theline + '\n'
395
if parents[version_id] is not None:
396
#if last_parent_lines[version_id] is not None:
397
parent_should_go = True
398
if last_parent_lines[version_id] != theline:
401
#parent_should_go = False
403
current_hunks[version_id][2] += 1
404
current_hunks[version_id][3].append((inserted, theline))
406
# last hunk last parent line is not eaten
407
current_hunks[version_id][1] -= 1
408
if current_hunks[version_id][1] < 0:
409
current_hunks[version_id][1] = 0
410
# import pdb;pdb.set_trace()
411
# assert current_hunks[version_id][1] >= 0
415
version = self._idx_to_name(i)
416
if current_hunks[version] != [0, 0, 0, []]:
417
diff_hunks[version].append(tuple(current_hunks[version]))
419
for version_id in version_ids:
420
result[version_id] = (
424
diff_hunks[version_id],
428
def get_parents(self, version_id):
429
"""See VersionedFile.get_parent."""
430
return map(self._idx_to_name, self._parents[self._lookup(version_id)])
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]
432
239
def _check_repeated_add(self, name, parents, text, sha1):
433
240
"""Check that a duplicated add is OK.
435
242
If it is, return the (old) index; otherwise raise an exception.
437
idx = self._lookup(name)
438
if sorted(self._parents[idx]) != sorted(parents) \
439
or sha1 != self._sha1s[idx]:
440
raise RevisionAlreadyPresent(name, self._weave_name)
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)
443
@deprecated_method(zero_eight)
444
def add_identical(self, old_rev_id, new_rev_id, parents):
445
"""Please use Weave.clone_text now."""
446
return self.clone_text(new_rev_id, old_rev_id, parents)
448
def _add_lines(self, version_id, parents, lines, parent_texts):
449
"""See VersionedFile.add_lines."""
450
return self._add(version_id, lines, map(self._lookup, parents))
452
@deprecated_method(zero_eight)
453
255
def add(self, name, parents, text, sha1=None):
454
"""See VersionedFile.add_lines for the non deprecated api."""
455
return self._add(name, text, map(self._maybe_lookup, parents), sha1)
457
def _add(self, version_id, lines, parents, sha1=None):
458
256
"""Add a single text on top of the weave.
460
258
Returns the index number of the newly added version.
463
261
Symbolic name for this version.
464
262
(Typically the revision-id of the revision that added it.)
467
265
List or set of direct parent version numbers.
470
268
Sequence of lines to be added in the new version.
270
sha -- SHA-1 of the file, if known. This is trusted to be
473
assert isinstance(version_id, basestring)
474
self._check_lines_not_unicode(lines)
475
self._check_lines_are_lines(lines)
477
sha1 = sha_strings(lines)
478
if version_id in self._name_map:
479
return self._check_repeated_add(version_id, parents, lines, sha1)
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)
481
282
self._check_versions(parents)
482
## self._check_lines(lines)
283
## self._check_lines(text)
483
284
new_version = len(self._parents)
485
287
# if we abort after here the (in-memory) weave will be corrupt because only
486
288
# some fields are updated
487
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
488
# - Robert Collins 20060226
489
289
self._parents.append(parents[:])
490
290
self._sha1s.append(sha1)
491
self._names.append(version_id)
492
self._name_map[version_id] = new_version
291
self._names.append(name)
292
self._name_map[name] = new_version
629
440
except IndexError:
630
441
raise IndexError("invalid version number %r" % i)
632
def _compatible_parents(self, my_parents, other_parents):
633
"""During join check that other_parents are joinable with my_parents.
635
Joinable is defined as 'is a subset of' - supersets may require
636
regeneration of diffs, but subsets do not.
638
return len(other_parents.difference(my_parents)) == 0
640
def annotate(self, version_id):
641
if isinstance(version_id, int):
642
warnings.warn('Weave.annotate(int) is deprecated. Please use version names'
643
' in all circumstances as of 0.8',
648
for origin, lineno, text in self._extract([version_id]):
649
result.append((origin, text))
652
return super(Weave, self).annotate(version_id)
654
def annotate_iter(self, version_id):
655
"""Yield list of (version-id, line) pairs for the specified version.
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.
657
451
The index indicates when the line originated in the weave."""
658
incls = [self._lookup(version_id)]
452
incls = [self.maybe_lookup(name_or_index)]
659
453
for origin, lineno, text in self._extract(incls):
660
yield self._idx_to_name(origin), text
662
@deprecated_method(zero_eight)
664
"""_walk has become visit, a supported api."""
665
return self._walk_internal()
667
def iter_lines_added_or_present_in_versions(self, version_ids=None):
668
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
669
if version_ids is None:
670
version_ids = self.versions()
671
version_ids = set(version_ids)
672
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
673
# if inserted not in version_ids then it was inserted before the
674
# versions we care about, but because weaves cannot represent ghosts
675
# properly, we do not filter down to that
676
# if inserted not in version_ids: continue
682
#@deprecated_method(zero_eight)
683
def walk(self, version_ids=None):
684
"""See VersionedFile.walk."""
685
return self._walk_internal(version_ids)
687
def _walk_internal(self, version_ids=None):
688
"""Helper method for weave actions."""
461
(lineno, insert, deletes, text)
462
for each literal line.
844
assert l.__class__ in (str, unicode)
539
assert isinstance(l, basestring)
845
540
if isactive is None:
846
541
isactive = (not dset) and istack and (istack[-1] in included)
848
543
result.append((istack[-1], lineno, l))
851
raise WeaveFormatError("unclosed insertion blocks "
852
"at end of weave: %s" % istack)
547
raise WFE("unclosed insertion blocks at end of weave",
854
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
550
raise WFE("unclosed deletion blocks at end of weave",
858
@deprecated_method(zero_eight)
859
557
def get_iter(self, name_or_index):
860
"""Deprecated, please do not use. Lookups are not not needed.
862
Please use get_lines now.
864
return iter(self.get_lines(self._maybe_lookup(name_or_index)))
866
@deprecated_method(zero_eight)
867
def maybe_lookup(self, name_or_index):
868
"""Deprecated, please do not use. Lookups are not not needed."""
869
return self._maybe_lookup(name_or_index)
871
def _maybe_lookup(self, name_or_index):
872
"""Convert possible symbolic name to index, or pass through indexes.
876
if isinstance(name_or_index, (int, long)):
879
return self._lookup(name_or_index)
881
@deprecated_method(zero_eight)
882
def get(self, version_id):
883
"""Please use either Weave.get_text or Weave.get_lines as desired."""
884
return self.get_lines(version_id)
886
def get_lines(self, version_id):
887
"""See VersionedFile.get_lines()."""
888
int_index = self._maybe_lookup(version_id)
889
result = [line for (origin, lineno, line) in self._extract([int_index])]
890
expected_sha1 = self._sha1s[int_index]
891
measured_sha1 = sha_strings(result)
892
if measured_sha1 != expected_sha1:
893
raise errors.WeaveInvalidChecksum(
894
'file %s, revision %s, expected: %s, measured %s'
895
% (self._weave_name, version_id,
896
expected_sha1, measured_sha1))
899
def get_sha1(self, version_id):
900
"""See VersionedFile.get_sha1()."""
901
return self._sha1s[self._lookup(version_id)]
903
@deprecated_method(zero_eight)
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)
904
592
def numversions(self):
905
"""How many versions are in this weave?
907
Deprecated in favour of num_versions.
909
return self.num_versions()
911
def num_versions(self):
912
"""How many versions are in this weave?"""
913
593
l = len(self._parents)
914
594
assert l == len(self._sha1s)
917
__len__ = num_versions
599
return self.numversions()
919
602
def check(self, progress_bar=None):
920
# TODO evaluate performance hit of using string sets in this routine.
921
# TODO: check no circular inclusions
922
# TODO: create a nested progress bar
923
for version in range(self.num_versions()):
603
# check no circular inclusions
604
for version in range(self.numversions()):
924
605
inclusions = list(self._parents[version])
926
607
inclusions.sort()
928
609
raise WeaveFormatError("invalid included version %d for index %d"
929
610
% (inclusions[-1], version))
931
# try extracting all versions; parallel extraction is used
932
nv = self.num_versions()
937
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
938
# The problem is that set membership is much more expensive
939
name = self._idx_to_name(i)
940
sha1s[name] = sha.new()
942
new_inc = set([name])
943
for p in self._parents[i]:
944
new_inc.update(inclusions[self._idx_to_name(p)])
946
assert set(new_inc) == set(self.get_ancestry(name)), \
947
'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
948
inclusions[name] = new_inc
950
nlines = len(self._weave)
952
update_text = 'checking weave'
954
short_name = os.path.basename(self._weave_name)
955
update_text = 'checking %s' % (short_name,)
956
update_text = update_text[:25]
958
for lineno, insert, deleteset, line in self._walk_internal():
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):
960
progress_bar.update(update_text, lineno, nlines)
962
for name, name_inclusions in inclusions.items():
963
# The active inclusion must be an ancestor,
964
# and no ancestors must have deleted this line,
965
# because we don't support resurrection.
966
if (insert in name_inclusions) and not (deleteset & name_inclusions):
967
sha1s[name].update(line)
970
version = self._idx_to_name(i)
971
hd = sha1s[version].hexdigest()
972
expected = self._sha1s[i]
617
progress_bar.update('checking text', version, nv)
619
for l in self.get_iter(version):
622
expected = self._sha1s[version]
973
623
if hd != expected:
974
raise errors.WeaveInvalidChecksum(
975
"mismatched sha1 for version %s: "
976
"got %s, expected %s"
977
% (version, hd, expected))
624
raise WeaveError("mismatched sha1 for version %d; "
625
"got %s, expected %s"
626
% (version, hd, expected))
979
628
# TODO: check insertions are properly nested, that there are
980
629
# no lines outside of insertion blocks, that deletions are
981
630
# properly paired, etc.
983
def _join(self, other, pb, msg, version_ids, ignore_missing):
984
"""Worker routine for join()."""
985
if not other.versions():
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:
986
780
return # nothing to update, easy
989
# versions is never none, InterWeave checks this.
992
# two loops so that we do not change ourselves before verifying it
994
781
# work through in index order to make sure we get all dependencies
997
# get the selected versions only that are in other.versions.
998
version_ids = set(other.versions()).intersection(set(version_ids))
999
# pull in the referenced graph.
1000
version_ids = other.get_ancestry(version_ids)
1001
pending_graph = [(version, other.get_parents(version)) for
1002
version in version_ids]
1003
for name in topo_sort(pending_graph):
1004
other_idx = other._name_map[name]
1005
# returns True if we have it, False if we need it.
1006
if not self._check_version_consistent(other, other_idx, name):
1007
names_to_join.append((other_idx, name))
1016
for other_idx, name in names_to_join:
1017
# TODO: If all the parents of the other version are already
782
for other_idx, name in enumerate(other._names):
783
# TODO: If all the parents of the other version are already
1018
784
# present then we can avoid some work by just taking the delta
1019
785
# and adjusting the offsets.
786
if self._check_version_consistent(other, other_idx, name):
1020
788
new_parents = self._imported_parents(other, other_idx)
789
lines = other.get_lines(other_idx)
1021
790
sha1 = other._sha1s[other_idx]
1026
pb.update(msg, merged, len(names_to_join))
1028
lines = other.get_lines(other_idx)
1029
self._add(name, lines, new_parents, sha1)
1031
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
1032
merged, processed, self._weave_name, time.time()-time0))
791
self.add(name, new_parents, lines, sha1)
1034
794
def _imported_parents(self, other, other_idx):
1035
795
"""Return list of parents in self corresponding to indexes in other."""
1036
796
new_parents = []
1037
797
for parent_idx in other._parents[other_idx]:
1038
798
parent_name = other._names[parent_idx]
1039
if parent_name not in self._name_map:
799
if parent_name not in self._names:
1040
800
# should not be possible
1041
801
raise WeaveError("missing parent {%s} of {%s} in %r"
1042
802
% (parent_name, other._name_map[other_idx], self))
1058
818
this_idx = self._name_map.get(name, -1)
1059
819
if this_idx != -1:
1060
820
if self._sha1s[this_idx] != other._sha1s[other_idx]:
1061
raise errors.WeaveTextDiffers(name, self, other)
821
raise WeaveError("inconsistent texts for version {%s} "
822
"when joining weaves"
1062
824
self_parents = self._parents[this_idx]
1063
825
other_parents = other._parents[other_idx]
1064
n1 = set([self._names[i] for i in self_parents])
1065
n2 = set([other._names[i] for i in other_parents])
1066
if not self._compatible_parents(n1, n2):
1067
raise WeaveParentMismatch("inconsistent parents "
1068
"for version {%s}: %s vs %s" % (name, n1, n2))
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}: "
1070
835
return True # ok!
1074
@deprecated_method(zero_eight)
1075
def reweave(self, other, pb=None, msg=None):
1076
"""reweave has been superseded by plain use of join."""
1077
return self.join(other, pb, msg)
1079
def _reweave(self, other, pb, msg):
1080
"""Reweave self with other - internal helper for join().
1082
:param other: The other weave to merge
1083
:param pb: An optional progress bar, indicating how far done we are
1084
:param msg: An optional message for the progress
1086
new_weave = _reweave(self, other, pb=pb, msg=msg)
1087
self._copy_weave_content(new_weave)
1089
def _copy_weave_content(self, otherweave):
1090
"""adsorb the content from otherweave."""
1091
for attr in self.__slots__:
1092
if attr != '_weave_name':
1093
setattr(self, attr, copy(getattr(otherweave, attr)))
1096
class WeaveFile(Weave):
1097
"""A WeaveFile represents a Weave on disk and writes on change."""
1099
WEAVE_SUFFIX = '.weave'
1101
def __init__(self, name, transport, filemode=None, create=False, access_mode='w'):
1102
"""Create a WeaveFile.
1104
:param create: If not True, only open an existing knit.
1106
super(WeaveFile, self).__init__(name, access_mode)
1107
self._transport = transport
1108
self._filemode = filemode
1110
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
1111
except errors.NoSuchFile:
1117
def _add_lines(self, version_id, parents, lines, parent_texts):
1118
"""Add a version and save the weave."""
1119
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
1124
def _clone_text(self, new_version_id, old_version_id, parents):
1125
"""See VersionedFile.clone_text."""
1126
super(WeaveFile, self)._clone_text(new_version_id, old_version_id, parents)
1129
def copy_to(self, name, transport):
1130
"""See VersionedFile.copy_to()."""
1131
# as we are all in memory always, just serialise to the new place.
1133
write_weave_v5(self, sio)
1135
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
1137
def create_empty(self, name, transport, filemode=None):
1138
return WeaveFile(name, transport, filemode, create=True)
1141
"""Save the weave."""
1142
self._check_write_ok()
1144
write_weave_v5(self, sio)
1146
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
1152
"""See VersionedFile.get_suffixes()."""
1153
return [WeaveFile.WEAVE_SUFFIX]
1155
def join(self, other, pb=None, msg=None, version_ids=None,
1156
ignore_missing=False):
1157
"""Join other into self and save."""
1158
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1162
@deprecated_function(zero_eight)
1163
def reweave(wa, wb, pb=None, msg=None):
1164
"""reweaving is deprecation, please just use weave.join()."""
1165
_reweave(wa, wb, pb, msg)
1167
def _reweave(wa, wb, pb=None, msg=None):
1168
"""Combine two weaves and return the result.
1170
This works even if a revision R has different parents in
1171
wa and wb. In the resulting weave all the parents are given.
1173
This is done by just building up a new weave, maintaining ordering
1174
of the versions in the two inputs. More efficient approaches
1175
might be possible but it should only be necessary to do
1176
this operation rarely, when a new previously ghost version is
1179
:param pb: An optional progress bar, indicating how far done we are
1180
:param msg: An optional message for the progress
1184
queue_a = range(wa.num_versions())
1185
queue_b = range(wb.num_versions())
1186
# first determine combined parents of all versions
1187
# map from version name -> all parent names
1188
combined_parents = _reweave_parent_graphs(wa, wb)
1189
mutter("combined parents: %r", combined_parents)
1190
order = topo_sort(combined_parents.iteritems())
1191
mutter("order to reweave: %r", order)
1196
for idx, name in enumerate(order):
1198
pb.update(msg, idx, len(order))
1199
if name in wa._name_map:
1200
lines = wa.get_lines(name)
1201
if name in wb._name_map:
1202
lines_b = wb.get_lines(name)
1203
if lines != lines_b:
1204
mutter('Weaves differ on content. rev_id {%s}', name)
1205
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1207
lines = list(difflib.unified_diff(lines, lines_b,
1208
wa._weave_name, wb._weave_name))
1209
mutter('lines:\n%s', ''.join(lines))
1210
raise errors.WeaveTextDiffers(name, wa, wb)
1212
lines = wb.get_lines(name)
1213
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1216
def _reweave_parent_graphs(wa, wb):
1217
"""Return combined parent ancestry for two weaves.
1219
Returned as a list of (version_name, set(parent_names))"""
1221
for weave in [wa, wb]:
1222
for idx, name in enumerate(weave._names):
1223
p = combined.setdefault(name, set())
1224
p.update(map(weave._idx_to_name, weave._parents[idx]))
1228
840
def weave_toc(w):
1229
841
"""Show the weave's table-of-contents"""
1409
1028
print ' '.join(map(str, w._parents[int(argv[3])]))
1411
1030
elif cmd == 'plan-merge':
1412
# replaced by 'bzr weave-plan-merge'
1414
1032
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1416
1034
print '%14s | %s' % (state, line),
1417
1036
elif cmd == 'merge':
1418
# replaced by 'bzr weave-merge-text'
1420
1038
p = w.plan_merge(int(argv[3]), int(argv[4]))
1421
1039
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))
1423
1062
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)
1426
1086
if __name__ == '__main__':
1428
sys.exit(main(sys.argv))
1431
class InterWeave(InterVersionedFile):
1432
"""Optimised code paths for weave to weave operations."""
1434
_matching_file_from_factory = staticmethod(WeaveFile)
1435
_matching_file_to_factory = staticmethod(WeaveFile)
1438
def is_compatible(source, target):
1439
"""Be compatible with weaves."""
1441
return (isinstance(source, Weave) and
1442
isinstance(target, Weave))
1443
except AttributeError:
1446
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1447
"""See InterVersionedFile.join."""
1448
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1449
if self.target.versions() == [] and version_ids is None:
1450
self.target._copy_weave_content(self.source)
1453
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1454
except errors.WeaveParentMismatch:
1455
self.target._reweave(self.source, pb, msg)
1458
InterVersionedFile.register_optimiser(InterWeave)
1088
if '--profile' in sys.argv:
1090
args.remove('--profile')
1091
sys.exit(profile_main(args))
1093
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
