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
def get_parents(self, version_id):
262
"""See VersionedFile.get_parent."""
263
return map(self._idx_to_name, self._parents[self._lookup(version_id)])
380
265
def _check_repeated_add(self, name, parents, text, sha1):
381
266
"""Check that a duplicated add is OK.
388
273
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):
276
@deprecated_method(zero_eight)
277
def add_identical(self, old_rev_id, new_rev_id, parents):
278
"""Please use Weave.clone_text now."""
279
return self.clone_text(new_rev_id, old_rev_id, parents)
281
def add_lines(self, version_id, parents, lines):
393
282
"""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):
283
return self._add(version_id, lines, map(self._lookup, parents))
285
@deprecated_method(zero_eight)
286
def add(self, name, parents, text, sha1=None):
287
"""See VersionedFile.add_lines for the non deprecated api."""
288
return self._add(name, text, map(self._maybe_lookup, parents), sha1)
290
def _add(self, version_id, lines, parents, sha1=None):
399
291
"""Add a single text on top of the weave.
401
293
Returns the index number of the newly added version.
404
296
Symbolic name for this version.
405
297
(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
300
List or set of direct parent version numbers.
412
303
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)
306
assert isinstance(version_id, basestring)
419
308
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
309
if version_id in self._name_map:
425
310
return self._check_repeated_add(version_id, parents, lines, sha1)
554
456
except IndexError:
555
457
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
459
def annotate(self, version_id):
566
"""Return a list of (version-id, line) tuples for version_id.
460
if isinstance(version_id, int):
461
warn('Weave.annotate(int) is deprecated. Please use version names'
462
' in all circumstances as of 0.8',
467
for origin, lineno, text in self._extract([version_id]):
468
result.append((origin, text))
471
return super(Weave, self).annotate(version_id)
473
def annotate_iter(self, version_id):
474
"""Yield list of (version-id, line) pairs for the specified version.
568
476
The index indicates when the line originated in the weave."""
569
477
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."""
478
for origin, lineno, text in self._extract(incls):
479
yield self._idx_to_name(origin), text
481
@deprecated_method(zero_eight)
483
"""_walk has become walk, a supported api."""
486
def walk(self, version_ids=None):
487
"""See VersionedFile.walk."""
592
492
lineno = 0 # line of weave, 0-based
594
494
for l in self._weave:
595
if l.__class__ == tuple:
495
if isinstance(l, tuple):
599
istack.append(self._names[v])
499
istack.append(self._idx_to_name(v))
603
dset.add(self._names[v])
503
assert self._idx_to_name(v) not in dset
504
dset.add(self._idx_to_name(v))
605
dset.remove(self._names[v])
506
dset.remove(self._idx_to_name(v))
607
508
raise WeaveFormatError('unexpected instruction %r' % v)
609
yield lineno, istack[-1], frozenset(dset), l
510
assert isinstance(l, basestring)
512
yield lineno, istack[-1], dset.copy(), l
616
519
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
522
def _extract(self, versions):
664
523
"""Yield annotation of lines in included set.
688
546
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
548
for l in self._weave:
718
if l.__class__ == tuple:
549
if isinstance(l, tuple):
553
assert v not in istack
725
iset.remove(istack.pop())
727
558
if v in included:
730
563
if v in included:
733
raise AssertionError()
567
assert isinstance(l, basestring)
735
568
if isactive is None:
736
569
isactive = (not dset) and istack and (istack[-1] in included)
756
599
return self._lookup(name_or_index)
601
def _get_iter(self, version_id):
602
"""Yield lines for the specified version."""
603
incls = [self._maybe_lookup(version_id)]
608
# We don't have sha1 sums for multiple entries
610
for origin, lineno, line in self._extract(incls):
615
expected_sha1 = self._sha1s[index]
616
measured_sha1 = cur_sha.hexdigest()
617
if measured_sha1 != expected_sha1:
618
raise errors.WeaveInvalidChecksum(
619
'file %s, revision %s, expected: %s, measured %s'
620
% (self._weave_name, self._names[index],
621
expected_sha1, measured_sha1))
623
@deprecated_method(zero_eight)
624
def get(self, version_id):
625
"""Please use either Weave.get_text or Weave.get_lines as desired."""
626
return self.get_lines(version_id)
758
628
def get_lines(self, version_id):
759
629
"""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?"""
630
return list(self._get_iter(version_id))
632
def get_sha1(self, name):
633
"""Get the stored sha1 sum for the given revision.
635
:param name: The name of the version to lookup
637
return self._sha1s[self._lookup(name)]
639
def numversions(self):
780
640
l = len(self._parents)
641
assert l == len(self._sha1s)
783
__len__ = num_versions
644
__len__ = numversions
785
646
def check(self, progress_bar=None):
786
647
# 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()):
648
# check no circular inclusions
649
for version in range(self.numversions()):
790
650
inclusions = list(self._parents[version])
792
652
inclusions.sort()
848
706
# no lines outside of insertion blocks, that deletions are
849
707
# properly paired, etc.
710
def join(self, other, pb=None, msg=None, version_ids=None):
712
"""Integrate versions from other into this weave.
714
The resulting weave contains all the history of both weaves;
715
any version you could retrieve from either self or other can be
716
retrieved from self after this call.
718
It is illegal for the two weaves to contain different values
719
or different parents for any version. See also reweave().
721
:param other: The other weave to pull into this one
722
:param pb: An optional progress bar
723
:param msg: An optional message to display for progress
726
self._join(other, pb, msg, version_ids)
727
except errors.WeaveParentMismatch:
728
self._reweave(other, pb, msg)
730
def _join(self, other, pb, msg, version_ids):
731
"""Worker routine for join()."""
732
if not other.versions():
733
return # nothing to update, easy
736
for version_id in version_ids:
737
if not self.has_version(version_id):
738
raise RevisionNotPresent(version_id, self._weave_name)
739
assert version_ids == None
741
# two loops so that we do not change ourselves before verifying it
743
# work through in index order to make sure we get all dependencies
746
for other_idx, name in enumerate(other._names):
747
self._check_version_consistent(other, other_idx, name)
748
sha1 = other._sha1s[other_idx]
752
if name in self._name_map:
753
idx = self._lookup(name)
754
n1 = set(map(other._idx_to_name, other._parents[other_idx]))
755
n2 = set(map(self._idx_to_name, self._parents[idx]))
756
if sha1 == self._sha1s[idx] and n1 == n2:
759
names_to_join.append((other_idx, name))
766
for other_idx, name in names_to_join:
767
# TODO: If all the parents of the other version are already
768
# present then we can avoid some work by just taking the delta
769
# and adjusting the offsets.
770
new_parents = self._imported_parents(other, other_idx)
771
sha1 = other._sha1s[other_idx]
776
pb.update(msg, merged, len(names_to_join))
778
lines = other.get_lines(other_idx)
779
self._add(name, lines, new_parents, sha1)
781
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
782
merged, processed, self._weave_name, time.time()-time0))
851
784
def _imported_parents(self, other, other_idx):
852
785
"""Return list of parents in self corresponding to indexes in other."""
854
787
for parent_idx in other._parents[other_idx]:
855
788
parent_name = other._names[parent_idx]
856
if parent_name not in self._name_map:
789
if parent_name not in self._names:
857
790
# should not be possible
858
raise WeaveError("missing parent {%s} of {%s} in %r"
791
raise WeaveError("missing parent {%s} of {%s} in %r"
859
792
% (parent_name, other._name_map[other_idx], self))
860
793
new_parents.append(self._name_map[parent_name])
861
794
return new_parents
896
834
:param msg: An optional message for the progress
898
836
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."""
903
837
for attr in self.__slots__:
904
838
if attr != '_weave_name':
905
setattr(self, attr, copy(getattr(otherweave, attr)))
839
setattr(self, attr, getattr(new_weave, attr))
908
842
class WeaveFile(Weave):
909
843
"""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)
845
def __init__(self, name, transport, mode=None):
846
super(WeaveFile, self).__init__(name)
920
847
self._transport = transport
921
self._filemode = filemode
923
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
850
_read_weave_v5(self._transport.get(name), self)
924
851
except errors.NoSuchFile:
930
def _add_lines(self, version_id, parents, lines, parent_texts,
931
left_matching_blocks, nostore_sha, random_id, check_content):
855
def add_lines(self, version_id, parents, lines):
932
856
"""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,
857
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
861
"""Save the weave."""
950
self._check_write_ok()
952
863
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)
865
self._transport.put(self._weave_name, sio, self._mode)
867
def join(self, other, pb=None, msg=None, version_ids=None):
868
"""Join other into self and save."""
869
super(WeaveFile, self).join(other, pb, msg, version_ids)
873
@deprecated_function(zero_eight)
874
def reweave(wa, wb, pb=None, msg=None):
875
"""reweaving is deprecation, please just use weave.join()."""
876
_reweave(wa, wb, pb, msg)
972
878
def _reweave(wa, wb, pb=None, msg=None):
973
879
"""Combine two weaves and return the result.
975
This works even if a revision R has different parents in
881
This works even if a revision R has different parents in
976
882
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
884
This is done by just building up a new weave, maintaining ordering
979
885
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
886
might be possible but it should only be necessary to do
887
this operation rarely, when a new previously ghost version is
984
890
:param pb: An optional progress bar, indicating how far done we are
1029
934
p = combined.setdefault(name, set())
1030
935
p.update(map(weave._idx_to_name, weave._parents[idx]))
940
"""Show the weave's table-of-contents"""
941
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
942
for i in (6, 50, 10, 10):
945
for i in range(w.numversions()):
948
parent_str = ' '.join(map(str, w._parents[i]))
949
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
953
def weave_stats(weave_file, pb):
954
from bzrlib.weavefile import read_weave
956
wf = file(weave_file, 'rb')
957
w = read_weave(wf, WeaveVersionedFile)
958
# FIXME: doesn't work on pipes
959
weave_size = wf.tell()
963
for i in range(vers):
964
pb.update('checking sizes', i, vers)
965
for origin, lineno, line in w._extract([i]):
970
print 'versions %9d' % vers
971
print 'weave file %9d bytes' % weave_size
972
print 'total contents %9d bytes' % total
973
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
976
print 'average size %9d bytes' % avg
977
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
981
print """bzr weave tool
983
Experimental tool for weave algorithm.
987
Create an empty weave file
988
weave get WEAVEFILE VERSION
989
Write out specified version.
990
weave check WEAVEFILE
991
Check consistency of all versions.
993
Display table of contents.
994
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
995
Add NEWTEXT, with specified parent versions.
996
weave annotate WEAVEFILE VERSION
997
Display origin of each line.
998
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
999
Auto-merge two versions and display conflicts.
1000
weave diff WEAVEFILE VERSION1 VERSION2
1001
Show differences between two versions.
1005
% weave init foo.weave
1007
% weave add foo.weave ver0 < foo.txt
1010
(create updated version)
1012
% weave get foo.weave 0 | diff -u - foo.txt
1013
% weave add foo.weave ver1 0 < foo.txt
1016
% weave get foo.weave 0 > foo.txt (create forked version)
1018
% weave add foo.weave ver2 0 < foo.txt
1021
% weave merge foo.weave 1 2 > foo.txt (merge them)
1022
% vi foo.txt (resolve conflicts)
1023
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1035
# in case we're run directly from the subdirectory
1036
sys.path.append('..')
1038
from bzrlib.weavefile import write_weave, read_weave
1039
from bzrlib.progress import ProgressBar
1054
return read_weave(file(argv[2], 'rb'))
1060
# at the moment, based on everything in the file
1062
parents = map(int, argv[4:])
1063
lines = sys.stdin.readlines()
1064
ver = w.add(name, parents, lines)
1065
write_weave(w, file(argv[2], 'wb'))
1066
print 'added version %r %d' % (name, ver)
1069
if os.path.exists(fn):
1070
raise IOError("file exists")
1072
write_weave(w, file(fn, 'wb'))
1073
elif cmd == 'get': # get one version
1075
sys.stdout.writelines(w.get_iter(int(argv[3])))
1078
from difflib import unified_diff
1081
v1, v2 = map(int, argv[3:5])
1084
diff_gen = unified_diff(lines1, lines2,
1085
'%s version %d' % (fn, v1),
1086
'%s version %d' % (fn, v2))
1087
sys.stdout.writelines(diff_gen)
1089
elif cmd == 'annotate':
1091
# newline is added to all lines regardless; too hard to get
1092
# reasonable formatting otherwise
1094
for origin, text in w.annotate(int(argv[3])):
1095
text = text.rstrip('\r\n')
1097
print ' | %s' % (text)
1099
print '%5d | %s' % (origin, text)
1105
elif cmd == 'stats':
1106
weave_stats(argv[2], ProgressBar())
1108
elif cmd == 'check':
1113
print '%d versions ok' % w.numversions()
1115
elif cmd == 'inclusions':
1117
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1119
elif cmd == 'parents':
1121
print ' '.join(map(str, w._parents[int(argv[3])]))
1123
elif cmd == 'plan-merge':
1125
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1127
print '%14s | %s' % (state, line),
1129
elif cmd == 'merge':
1131
p = w.plan_merge(int(argv[3]), int(argv[4]))
1132
sys.stdout.writelines(w.weave_merge(p))
1135
raise ValueError('unknown command %r' % cmd)
1139
def profile_main(argv):
1140
import tempfile, hotshot, hotshot.stats
1142
prof_f = tempfile.NamedTemporaryFile()
1144
prof = hotshot.Profile(prof_f.name)
1146
ret = prof.runcall(main, argv)
1149
stats = hotshot.stats.load(prof_f.name)
1151
stats.sort_stats('cumulative')
1152
## XXX: Might like to write to stderr or the trace file instead but
1153
## print_stats seems hardcoded to stdout
1154
stats.print_stats(20)
1159
def lsprofile_main(argv):
1160
from bzrlib.lsprof import profile
1161
ret,stats = profile(main, argv)
1167
if __name__ == '__main__':
1169
if '--profile' in sys.argv:
1171
args.remove('--profile')
1172
sys.exit(profile_main(args))
1173
elif '--lsprof' in sys.argv:
1175
args.remove('--lsprof')
1176
sys.exit(lsprofile_main(args))
1178
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
