1
# Copyright (C) 2005, 2009 Canonical Ltd
3
# Copyright (C) 2005 Canonical Ltd
3
5
# This program is free software; you can redistribute it and/or modify
4
6
# it under the terms of the GNU General Public License as published by
5
7
# the Free Software Foundation; either version 2 of the License, or
6
8
# (at your option) any later version.
8
10
# This program is distributed in the hope that it will be useful,
9
11
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
12
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
13
# GNU General Public License for more details.
13
15
# You should have received a copy of the GNU General Public License
14
16
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
19
# Author: Martin Pool <mbp@canonical.com>
59
61
# where the basis and destination are unchanged.
61
63
# FIXME: Sometimes we will be given a parents list for a revision
62
# that includes some redundant parents (i.e. already a parent of
63
# something in the list.) We should eliminate them. This can
64
# that includes some redundant parents (i.e. already a parent of
65
# something in the list.) We should eliminate them. This can
64
66
# be done fairly efficiently because the sequence numbers constrain
65
67
# the possible relationships.
67
# FIXME: the conflict markers should be *7* characters
70
70
from cStringIO import StringIO
71
from difflib import SequenceMatcher
73
from bzrlib.lazy_import import lazy_import
74
lazy_import(globals(), """
75
from bzrlib import tsort
76
from bzrlib.trace import mutter
81
77
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
82
78
RevisionAlreadyPresent,
83
79
RevisionNotPresent,
84
UnavailableRepresentation,
80
WeaveRevisionAlreadyPresent,
81
WeaveRevisionNotPresent,
86
from bzrlib.osutils import dirname, sha, sha_strings, split_lines
87
import bzrlib.patiencediff
88
from bzrlib.revision import NULL_REVISION
83
import bzrlib.errors as errors
84
from bzrlib.osutils import sha_strings
89
85
from bzrlib.symbol_versioning import *
90
from bzrlib.trace import mutter
91
from bzrlib.versionedfile import (
86
from bzrlib.tsort import topo_sort
87
from bzrlib.versionedfile import VersionedFile
98
88
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
101
class WeaveContentFactory(ContentFactory):
102
"""Content factory for streaming from weaves.
104
:seealso ContentFactory:
107
def __init__(self, version, weave):
108
"""Create a WeaveContentFactory for version from weave."""
109
ContentFactory.__init__(self)
110
self.sha1 = weave.get_sha1s([version])[version]
111
self.key = (version,)
112
parents = weave.get_parent_map([version])[version]
113
self.parents = tuple((parent,) for parent in parents)
114
self.storage_kind = 'fulltext'
117
def get_bytes_as(self, storage_kind):
118
if storage_kind == 'fulltext':
119
return self._weave.get_text(self.key[-1])
120
elif storage_kind == 'chunked':
121
return self._weave.get_lines(self.key[-1])
123
raise UnavailableRepresentation(self.key, storage_kind, 'fulltext')
126
91
class Weave(VersionedFile):
127
92
"""weave - versioned text file storage.
129
94
A Weave manages versions of line-based text files, keeping track
130
95
of the originating version for each line.
215
180
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
216
'_weave_name', '_matcher', '_allow_reserved']
218
def __init__(self, weave_name=None, access_mode='w', matcher=None,
219
get_scope=None, allow_reserved=False):
222
:param get_scope: A callable that returns an opaque object to be used
223
for detecting when this weave goes out of scope (should stop
224
answering requests or allowing mutation).
226
super(Weave, self).__init__()
183
def __init__(self, weave_name=None):
228
185
self._parents = []
231
188
self._name_map = {}
232
189
self._weave_name = weave_name
234
self._matcher = bzrlib.patiencediff.PatienceSequenceMatcher
236
self._matcher = matcher
237
if get_scope is None:
238
get_scope = lambda:None
239
self._get_scope = get_scope
240
self._scope = get_scope()
241
self._access_mode = access_mode
242
self._allow_reserved = allow_reserved
244
191
def __repr__(self):
245
192
return "Weave(%r)" % self._weave_name
247
def _check_write_ok(self):
248
"""Is the versioned file marked as 'finished' ? Raise if it is."""
249
if self._get_scope() != self._scope:
250
raise errors.OutSideTransaction()
251
if self._access_mode != 'w':
252
raise errors.ReadOnlyObjectDirtiedError(self)
255
195
"""Return a deep copy of self.
257
197
The copy can be modified without affecting the original weave."""
259
199
other._weave = self._weave[:]
270
210
return self._parents == other._parents \
271
211
and self._weave == other._weave \
272
and self._sha1s == other._sha1s
212
and self._sha1s == other._sha1s
274
214
def __ne__(self, other):
275
215
return not self.__eq__(other)
217
@deprecated_method(zero_eight)
218
def idx_to_name(self, index):
219
"""Old public interface, the public interface is all names now."""
277
222
def _idx_to_name(self, version):
278
223
return self._names[version]
225
@deprecated_method(zero_eight)
226
def lookup(self, name):
227
"""Backwards compatability thunk:
229
Return name, as name is valid in the api now, and spew deprecation
280
234
def _lookup(self, name):
281
235
"""Convert symbolic version name to index."""
282
if not self._allow_reserved:
283
self.check_not_reserved_id(name)
285
237
return self._name_map[name]
287
239
raise RevisionNotPresent(name, self._weave_name)
241
@deprecated_method(zero_eight)
242
def iter_names(self):
243
"""Deprecated convenience function, please see VersionedFile.names()."""
244
return iter(self.names())
246
@deprecated_method(zero_eight)
248
"""See Weave.versions for the current api."""
249
return self.versions()
289
251
def versions(self):
290
252
"""See VersionedFile.versions."""
291
253
return self._names[:]
293
255
def has_version(self, version_id):
294
256
"""See VersionedFile.has_version."""
295
return (version_id in self._name_map)
257
return self._name_map.has_key(version_id)
297
259
__contains__ = has_version
299
def get_record_stream(self, versions, ordering, include_delta_closure):
300
"""Get a stream of records for versions.
302
:param versions: The versions to include. Each version is a tuple
304
:param ordering: Either 'unordered' or 'topological'. A topologically
305
sorted stream has compression parents strictly before their
307
:param include_delta_closure: If True then the closure across any
308
compression parents will be included (in the opaque data).
309
:return: An iterator of ContentFactory objects, each of which is only
310
valid until the iterator is advanced.
312
versions = [version[-1] for version in versions]
313
if ordering == 'topological':
314
parents = self.get_parent_map(versions)
315
new_versions = tsort.topo_sort(parents)
316
new_versions.extend(set(versions).difference(set(parents)))
317
versions = new_versions
318
elif ordering == 'groupcompress':
319
parents = self.get_parent_map(versions)
320
new_versions = sort_groupcompress(parents)
321
new_versions.extend(set(versions).difference(set(parents)))
322
versions = new_versions
323
for version in versions:
325
yield WeaveContentFactory(version, self)
327
yield AbsentContentFactory((version,))
329
def get_parent_map(self, version_ids):
330
"""See VersionedFile.get_parent_map."""
332
for version_id in version_ids:
333
if version_id == NULL_REVISION:
338
map(self._idx_to_name,
339
self._parents[self._lookup(version_id)]))
340
except RevisionNotPresent:
342
result[version_id] = parents
345
def get_parents_with_ghosts(self, version_id):
346
raise NotImplementedError(self.get_parents_with_ghosts)
348
def insert_record_stream(self, stream):
349
"""Insert a record stream into this versioned file.
351
:param stream: A stream of records to insert.
353
:seealso VersionedFile.get_record_stream:
356
for record in stream:
357
# Raise an error when a record is missing.
358
if record.storage_kind == 'absent':
359
raise RevisionNotPresent([record.key[0]], self)
360
# adapt to non-tuple interface
361
parents = [parent[0] for parent in record.parents]
362
if (record.storage_kind == 'fulltext'
363
or record.storage_kind == 'chunked'):
364
self.add_lines(record.key[0], parents,
365
osutils.chunks_to_lines(record.get_bytes_as('chunked')))
367
adapter_key = record.storage_kind, 'fulltext'
369
adapter = adapters[adapter_key]
371
adapter_factory = adapter_registry.get(adapter_key)
372
adapter = adapter_factory(self)
373
adapters[adapter_key] = adapter
374
lines = split_lines(adapter.get_bytes(record))
376
self.add_lines(record.key[0], parents, lines)
377
except RevisionAlreadyPresent:
261
@deprecated_method(zero_eight)
262
def parent_names(self, version):
263
"""Return version names for parents of a version.
265
See get_parents for the current api.
267
return self.get_parents(version)
269
def get_parents(self, version_id):
270
"""See VersionedFile.get_parent."""
271
return map(self._idx_to_name, self._parents[self._lookup(version_id)])
380
273
def _check_repeated_add(self, name, parents, text, sha1):
381
274
"""Check that a duplicated add is OK.
388
281
raise RevisionAlreadyPresent(name, self._weave_name)
391
def _add_lines(self, version_id, parents, lines, parent_texts,
392
left_matching_blocks, nostore_sha, random_id, check_content):
284
@deprecated_method(zero_eight)
285
def add_identical(self, old_rev_id, new_rev_id, parents):
286
"""Please use Weave.clone_text now."""
287
return self.clone_text(new_rev_id, old_rev_id, parents)
289
def add_lines(self, version_id, parents, lines):
393
290
"""See VersionedFile.add_lines."""
394
idx = self._add(version_id, lines, map(self._lookup, parents),
395
nostore_sha=nostore_sha)
396
return sha_strings(lines), sum(map(len, lines)), idx
398
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
291
return self._add(version_id, lines, map(self._lookup, parents))
293
@deprecated_method(zero_eight)
294
def add(self, name, parents, text, sha1=None):
295
"""See VersionedFile.add_lines for the non deprecated api."""
296
return self._add(name, text, map(self._maybe_lookup, parents), sha1)
298
def _add(self, version_id, lines, parents, sha1=None):
399
299
"""Add a single text on top of the weave.
401
301
Returns the index number of the newly added version.
404
304
Symbolic name for this version.
405
305
(Typically the revision-id of the revision that added it.)
406
If None, a name will be allocated based on the hash. (sha1:SHAHASH)
409
308
List or set of direct parent version numbers.
412
311
Sequence of lines to be added in the new version.
414
:param nostore_sha: See VersionedFile.add_lines.
416
self._check_lines_not_unicode(lines)
417
self._check_lines_are_lines(lines)
314
assert isinstance(version_id, basestring)
419
316
sha1 = sha_strings(lines)
420
if sha1 == nostore_sha:
421
raise errors.ExistingContent
422
if version_id is None:
423
version_id = "sha1:" + sha1
424
317
if version_id in self._name_map:
425
318
return self._check_repeated_add(version_id, parents, lines, sha1)
554
464
except IndexError:
555
465
raise IndexError("invalid version number %r" % i)
557
def _compatible_parents(self, my_parents, other_parents):
558
"""During join check that other_parents are joinable with my_parents.
560
Joinable is defined as 'is a subset of' - supersets may require
561
regeneration of diffs, but subsets do not.
563
return len(other_parents.difference(my_parents)) == 0
565
467
def annotate(self, version_id):
566
"""Return a list of (version-id, line) tuples for version_id.
468
if isinstance(version_id, int):
469
warn('Weave.annotate(int) is deprecated. Please use version names'
470
' in all circumstances as of 0.8',
475
for origin, lineno, text in self._extract([version_id]):
476
result.append((origin, text))
479
return super(Weave, self).annotate(version_id)
481
def annotate_iter(self, version_id):
482
"""Yield list of (version-id, line) pairs for the specified version.
568
484
The index indicates when the line originated in the weave."""
569
485
incls = [self._lookup(version_id)]
570
return [(self._idx_to_name(origin), text) for origin, lineno, text in
571
self._extract(incls)]
573
def iter_lines_added_or_present_in_versions(self, version_ids=None,
575
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
576
if version_ids is None:
577
version_ids = self.versions()
578
version_ids = set(version_ids)
579
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
580
if inserted not in version_ids: continue
582
yield line + '\n', inserted
586
def _walk_internal(self, version_ids=None):
587
"""Helper method for weave actions."""
486
for origin, lineno, text in self._extract(incls):
487
yield self._idx_to_name(origin), text
489
@deprecated_method(zero_eight)
491
"""_walk has become walk, a supported api."""
494
def walk(self, version_ids=None):
495
"""See VersionedFile.walk."""
592
500
lineno = 0 # line of weave, 0-based
594
502
for l in self._weave:
595
if l.__class__ == tuple:
503
if isinstance(l, tuple):
599
istack.append(self._names[v])
507
istack.append(self._idx_to_name(v))
603
dset.add(self._names[v])
511
assert self._idx_to_name(v) not in dset
512
dset.add(self._idx_to_name(v))
605
dset.remove(self._names[v])
514
dset.remove(self._idx_to_name(v))
607
516
raise WeaveFormatError('unexpected instruction %r' % v)
609
yield lineno, istack[-1], frozenset(dset), l
518
assert isinstance(l, basestring)
520
yield lineno, istack[-1], dset.copy(), l
616
527
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
619
def plan_merge(self, ver_a, ver_b):
620
"""Return pseudo-annotation indicating how the two versions merge.
622
This is computed between versions a and b and their common
625
Weave lines present in none of them are skipped entirely.
627
inc_a = set(self.get_ancestry([ver_a]))
628
inc_b = set(self.get_ancestry([ver_b]))
629
inc_c = inc_a & inc_b
631
for lineno, insert, deleteset, line in self._walk_internal([ver_a, ver_b]):
632
if deleteset & inc_c:
633
# killed in parent; can't be in either a or b
634
# not relevant to our work
635
yield 'killed-base', line
636
elif insert in inc_c:
637
# was inserted in base
638
killed_a = bool(deleteset & inc_a)
639
killed_b = bool(deleteset & inc_b)
640
if killed_a and killed_b:
641
yield 'killed-both', line
643
yield 'killed-a', line
645
yield 'killed-b', line
647
yield 'unchanged', line
648
elif insert in inc_a:
649
if deleteset & inc_a:
650
yield 'ghost-a', line
654
elif insert in inc_b:
655
if deleteset & inc_b:
656
yield 'ghost-b', line
660
# not in either revision
661
yield 'irrelevant', line
663
530
def _extract(self, versions):
664
531
"""Yield annotation of lines in included set.
688
554
WFE = WeaveFormatError
691
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
692
# +285282 0 1676.8040 1676.8040 +<isinstance>
693
# 1.6 seconds in 'isinstance'.
694
# changing the first isinstance:
695
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
696
# +140414 0 762.8050 762.8050 +<isinstance>
697
# note that the inline time actually dropped (less function calls)
698
# and total processing time was halved.
699
# we're still spending ~1/4 of the method in isinstance though.
700
# so lets hard code the acceptable string classes we expect:
701
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
702
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
704
# yay, down to ~1/4 the initial extract time, and our inline time
705
# has shrunk again, with isinstance no longer dominating.
706
# tweaking the stack inclusion test to use a set gives:
707
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
708
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
710
# - a 5% win, or possibly just noise. However with large istacks that
711
# 'in' test could dominate, so I'm leaving this change in place -
712
# when its fast enough to consider profiling big datasets we can review.
717
556
for l in self._weave:
718
if l.__class__ == tuple:
557
if isinstance(l, tuple):
561
assert v not in istack
725
iset.remove(istack.pop())
727
566
if v in included:
730
571
if v in included:
733
raise AssertionError()
575
assert isinstance(l, basestring)
735
576
if isactive is None:
736
577
isactive = (not dset) and istack and (istack[-1] in included)
756
607
return self._lookup(name_or_index)
609
def _get_iter(self, version_id):
610
"""Yield lines for the specified version."""
611
incls = [self._maybe_lookup(version_id)]
616
# We don't have sha1 sums for multiple entries
618
for origin, lineno, line in self._extract(incls):
623
expected_sha1 = self._sha1s[index]
624
measured_sha1 = cur_sha.hexdigest()
625
if measured_sha1 != expected_sha1:
626
raise errors.WeaveInvalidChecksum(
627
'file %s, revision %s, expected: %s, measured %s'
628
% (self._weave_name, self._names[index],
629
expected_sha1, measured_sha1))
631
@deprecated_method(zero_eight)
632
def get(self, version_id):
633
"""Please use either Weave.get_text or Weave.get_lines as desired."""
634
return self.get_lines(version_id)
758
636
def get_lines(self, version_id):
759
637
"""See VersionedFile.get_lines()."""
760
int_index = self._maybe_lookup(version_id)
761
result = [line for (origin, lineno, line) in self._extract([int_index])]
762
expected_sha1 = self._sha1s[int_index]
763
measured_sha1 = sha_strings(result)
764
if measured_sha1 != expected_sha1:
765
raise errors.WeaveInvalidChecksum(
766
'file %s, revision %s, expected: %s, measured %s'
767
% (self._weave_name, version_id,
768
expected_sha1, measured_sha1))
771
def get_sha1s(self, version_ids):
772
"""See VersionedFile.get_sha1s()."""
774
for v in version_ids:
775
result[v] = self._sha1s[self._lookup(v)]
778
def num_versions(self):
779
"""How many versions are in this weave?"""
638
return list(self._get_iter(version_id))
640
def get_sha1(self, name):
641
"""Get the stored sha1 sum for the given revision.
643
:param name: The name of the version to lookup
645
return self._sha1s[self._lookup(name)]
647
def numversions(self):
780
648
l = len(self._parents)
649
assert l == len(self._sha1s)
783
__len__ = num_versions
652
__len__ = numversions
785
654
def check(self, progress_bar=None):
786
655
# TODO evaluate performance hit of using string sets in this routine.
787
# TODO: check no circular inclusions
788
# TODO: create a nested progress bar
789
for version in range(self.num_versions()):
656
# check no circular inclusions
657
for version in range(self.numversions()):
790
658
inclusions = list(self._parents[version])
792
660
inclusions.sort()
848
714
# no lines outside of insertion blocks, that deletions are
849
715
# properly paired, etc.
718
def join(self, other, pb=None, msg=None, version_ids=None):
720
"""Integrate versions from other into this weave.
722
The resulting weave contains all the history of both weaves;
723
any version you could retrieve from either self or other can be
724
retrieved from self after this call.
726
It is illegal for the two weaves to contain different values
727
or different parents for any version. See also reweave().
729
:param other: The other weave to pull into this one
730
:param pb: An optional progress bar
731
:param msg: An optional message to display for progress
733
if not other.versions():
734
return # nothing to update, easy
737
for version_id in version_ids:
738
if not self.has_version(version_id):
739
raise RevisionNotPresent(version_id, self._weave_name)
740
assert version_ids == None
742
# two loops so that we do not change ourselves before verifying it
744
# work through in index order to make sure we get all dependencies
747
for other_idx, name in enumerate(other._names):
748
self._check_version_consistent(other, other_idx, name)
749
sha1 = other._sha1s[other_idx]
753
if name in self._name_map:
754
idx = self._lookup(name)
755
n1 = set(map(other._idx_to_name, other._parents[other_idx]))
756
n2 = set(map(self._idx_to_name, self._parents[idx]))
757
if sha1 == self._sha1s[idx] and n1 == n2:
760
names_to_join.append((other_idx, name))
767
for other_idx, name in names_to_join:
768
# TODO: If all the parents of the other version are already
769
# present then we can avoid some work by just taking the delta
770
# and adjusting the offsets.
771
new_parents = self._imported_parents(other, other_idx)
772
sha1 = other._sha1s[other_idx]
777
pb.update(msg, merged, len(names_to_join))
779
lines = other.get_lines(other_idx)
780
self._add(name, lines, new_parents, sha1)
782
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
783
merged, processed, self._weave_name, time.time()-time0))
851
785
def _imported_parents(self, other, other_idx):
852
786
"""Return list of parents in self corresponding to indexes in other."""
854
788
for parent_idx in other._parents[other_idx]:
855
789
parent_name = other._names[parent_idx]
856
if parent_name not in self._name_map:
790
if parent_name not in self._names:
857
791
# should not be possible
858
raise WeaveError("missing parent {%s} of {%s} in %r"
792
raise WeaveError("missing parent {%s} of {%s} in %r"
859
793
% (parent_name, other._name_map[other_idx], self))
860
794
new_parents.append(self._name_map[parent_name])
861
795
return new_parents
891
def _reweave(self, other, pb, msg):
892
"""Reweave self with other - internal helper for join().
827
def reweave(self, other, pb=None, msg=None):
828
"""Reweave self with other.
894
830
:param other: The other weave to merge
895
831
:param pb: An optional progress bar, indicating how far done we are
896
832
:param msg: An optional message for the progress
898
new_weave = _reweave(self, other, pb=pb, msg=msg)
899
self._copy_weave_content(new_weave)
901
def _copy_weave_content(self, otherweave):
902
"""adsorb the content from otherweave."""
834
new_weave = reweave(self, other, pb=pb, msg=msg)
903
835
for attr in self.__slots__:
904
if attr != '_weave_name':
905
setattr(self, attr, copy(getattr(otherweave, attr)))
836
setattr(self, attr, getattr(new_weave, attr))
908
839
class WeaveFile(Weave):
909
840
"""A WeaveFile represents a Weave on disk and writes on change."""
911
WEAVE_SUFFIX = '.weave'
913
def __init__(self, name, transport, filemode=None, create=False, access_mode='w', get_scope=None):
914
"""Create a WeaveFile.
916
:param create: If not True, only open an existing knit.
918
super(WeaveFile, self).__init__(name, access_mode, get_scope=get_scope,
919
allow_reserved=False)
842
def __init__(self, name, transport, mode=None):
843
super(WeaveFile, self).__init__(name)
920
844
self._transport = transport
921
self._filemode = filemode
923
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
847
_read_weave_v5(self._transport.get(name), self)
924
848
except errors.NoSuchFile:
930
def _add_lines(self, version_id, parents, lines, parent_texts,
931
left_matching_blocks, nostore_sha, random_id, check_content):
852
def add_lines(self, version_id, parents, lines):
932
853
"""Add a version and save the weave."""
933
self.check_not_reserved_id(version_id)
934
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
935
parent_texts, left_matching_blocks, nostore_sha, random_id,
854
super(WeaveFile, self).add_lines(version_id, parents, lines)
940
def copy_to(self, name, transport):
941
"""See VersionedFile.copy_to()."""
942
# as we are all in memory always, just serialise to the new place.
944
write_weave_v5(self, sio)
946
transport.put_file(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
949
858
"""Save the weave."""
950
self._check_write_ok()
952
860
write_weave_v5(self, sio)
954
bytes = sio.getvalue()
955
path = self._weave_name + WeaveFile.WEAVE_SUFFIX
957
self._transport.put_bytes(path, bytes, self._filemode)
958
except errors.NoSuchFile:
959
self._transport.mkdir(dirname(path))
960
self._transport.put_bytes(path, bytes, self._filemode)
964
"""See VersionedFile.get_suffixes()."""
965
return [WeaveFile.WEAVE_SUFFIX]
967
def insert_record_stream(self, stream):
968
super(WeaveFile, self).insert_record_stream(stream)
862
self._transport.put(self._weave_name, sio, self._mode)
864
def join(self, other, pb=None, msg=None, version_ids=None):
865
"""Join other into self and save."""
866
super(WeaveFile, self).join(other, pb, msg, version_ids)
972
def _reweave(wa, wb, pb=None, msg=None):
870
def reweave(wa, wb, pb=None, msg=None):
973
871
"""Combine two weaves and return the result.
975
This works even if a revision R has different parents in
873
This works even if a revision R has different parents in
976
874
wa and wb. In the resulting weave all the parents are given.
978
This is done by just building up a new weave, maintaining ordering
876
This is done by just building up a new weave, maintaining ordering
979
877
of the versions in the two inputs. More efficient approaches
980
might be possible but it should only be necessary to do
981
this operation rarely, when a new previously ghost version is
878
might be possible but it should only be necessary to do
879
this operation rarely, when a new previously ghost version is
984
882
:param pb: An optional progress bar, indicating how far done we are
1029
926
p = combined.setdefault(name, set())
1030
927
p.update(map(weave._idx_to_name, weave._parents[idx]))
932
"""Show the weave's table-of-contents"""
933
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
934
for i in (6, 50, 10, 10):
937
for i in range(w.numversions()):
940
parent_str = ' '.join(map(str, w._parents[i]))
941
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
945
def weave_stats(weave_file, pb):
946
from bzrlib.weavefile import read_weave
948
wf = file(weave_file, 'rb')
949
w = read_weave(wf, WeaveVersionedFile)
950
# FIXME: doesn't work on pipes
951
weave_size = wf.tell()
955
for i in range(vers):
956
pb.update('checking sizes', i, vers)
957
for origin, lineno, line in w._extract([i]):
962
print 'versions %9d' % vers
963
print 'weave file %9d bytes' % weave_size
964
print 'total contents %9d bytes' % total
965
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
968
print 'average size %9d bytes' % avg
969
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
973
print """bzr weave tool
975
Experimental tool for weave algorithm.
979
Create an empty weave file
980
weave get WEAVEFILE VERSION
981
Write out specified version.
982
weave check WEAVEFILE
983
Check consistency of all versions.
985
Display table of contents.
986
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
987
Add NEWTEXT, with specified parent versions.
988
weave annotate WEAVEFILE VERSION
989
Display origin of each line.
990
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
991
Auto-merge two versions and display conflicts.
992
weave diff WEAVEFILE VERSION1 VERSION2
993
Show differences between two versions.
997
% weave init foo.weave
999
% weave add foo.weave ver0 < foo.txt
1002
(create updated version)
1004
% weave get foo.weave 0 | diff -u - foo.txt
1005
% weave add foo.weave ver1 0 < foo.txt
1008
% weave get foo.weave 0 > foo.txt (create forked version)
1010
% weave add foo.weave ver2 0 < foo.txt
1013
% weave merge foo.weave 1 2 > foo.txt (merge them)
1014
% vi foo.txt (resolve conflicts)
1015
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1027
# in case we're run directly from the subdirectory
1028
sys.path.append('..')
1030
from bzrlib.weavefile import write_weave, read_weave
1031
from bzrlib.progress import ProgressBar
1046
return read_weave(file(argv[2], 'rb'))
1052
# at the moment, based on everything in the file
1054
parents = map(int, argv[4:])
1055
lines = sys.stdin.readlines()
1056
ver = w.add(name, parents, lines)
1057
write_weave(w, file(argv[2], 'wb'))
1058
print 'added version %r %d' % (name, ver)
1061
if os.path.exists(fn):
1062
raise IOError("file exists")
1064
write_weave(w, file(fn, 'wb'))
1065
elif cmd == 'get': # get one version
1067
sys.stdout.writelines(w.get_iter(int(argv[3])))
1070
from difflib import unified_diff
1073
v1, v2 = map(int, argv[3:5])
1076
diff_gen = unified_diff(lines1, lines2,
1077
'%s version %d' % (fn, v1),
1078
'%s version %d' % (fn, v2))
1079
sys.stdout.writelines(diff_gen)
1081
elif cmd == 'annotate':
1083
# newline is added to all lines regardless; too hard to get
1084
# reasonable formatting otherwise
1086
for origin, text in w.annotate(int(argv[3])):
1087
text = text.rstrip('\r\n')
1089
print ' | %s' % (text)
1091
print '%5d | %s' % (origin, text)
1097
elif cmd == 'stats':
1098
weave_stats(argv[2], ProgressBar())
1100
elif cmd == 'check':
1105
print '%d versions ok' % w.numversions()
1107
elif cmd == 'inclusions':
1109
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1111
elif cmd == 'parents':
1113
print ' '.join(map(str, w._parents[int(argv[3])]))
1115
elif cmd == 'plan-merge':
1117
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1119
print '%14s | %s' % (state, line),
1121
elif cmd == 'merge':
1123
p = w.plan_merge(int(argv[3]), int(argv[4]))
1124
sys.stdout.writelines(w.weave_merge(p))
1127
raise ValueError('unknown command %r' % cmd)
1131
def profile_main(argv):
1132
import tempfile, hotshot, hotshot.stats
1134
prof_f = tempfile.NamedTemporaryFile()
1136
prof = hotshot.Profile(prof_f.name)
1138
ret = prof.runcall(main, argv)
1141
stats = hotshot.stats.load(prof_f.name)
1143
stats.sort_stats('cumulative')
1144
## XXX: Might like to write to stderr or the trace file instead but
1145
## print_stats seems hardcoded to stdout
1146
stats.print_stats(20)
1151
def lsprofile_main(argv):
1152
from bzrlib.lsprof import profile
1153
ret,stats = profile(main, argv)
1159
if __name__ == '__main__':
1161
if '--profile' in sys.argv:
1163
args.remove('--profile')
1164
sys.exit(profile_main(args))
1165
elif '--lsprof' in sys.argv:
1167
args.remove('--lsprof')
1168
sys.exit(lsprofile_main(args))
1170
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
