3
# Copyright (C) 2005 Canonical Ltd
1
# Copyright (C) 2005, 2009 Canonical Ltd
5
3
# This program is free software; you can redistribute it and/or modify
6
4
# it under the terms of the GNU General Public License as published by
7
5
# the Free Software Foundation; either version 2 of the License, or
8
6
# (at your option) any later version.
10
8
# This program is distributed in the hope that it will be useful,
11
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
12
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13
11
# GNU General Public License for more details.
15
13
# You should have received a copy of the GNU General Public License
16
14
# along with this program; if not, write to the Free Software
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19
17
# Author: Martin Pool <mbp@canonical.com>
61
59
# where the basis and destination are unchanged.
63
61
# 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
62
# that includes some redundant parents (i.e. already a parent of
63
# something in the list.) We should eliminate them. This can
66
64
# be done fairly efficiently because the sequence numbers constrain
67
65
# the possible relationships.
67
# FIXME: the conflict markers should be *7* characters
70
from cStringIO import StringIO
73
from difflib import SequenceMatcher
75
from bzrlib.trace import mutter
73
from bzrlib.lazy_import import lazy_import
74
lazy_import(globals(), """
75
from bzrlib import tsort
76
81
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
77
82
RevisionAlreadyPresent,
78
83
RevisionNotPresent,
79
WeaveRevisionAlreadyPresent,
80
WeaveRevisionNotPresent,
84
UnavailableRepresentation,
82
import bzrlib.errors as errors
86
from bzrlib.osutils import dirname, sha, sha_strings, split_lines
87
import bzrlib.patiencediff
88
from bzrlib.revision import NULL_REVISION
83
89
from bzrlib.symbol_versioning import *
84
from bzrlib.tsort import topo_sort
85
from bzrlib.versionedfile import VersionedFile
90
from bzrlib.trace import mutter
91
from bzrlib.versionedfile import (
98
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')
88
126
class Weave(VersionedFile):
89
127
"""weave - versioned text file storage.
91
129
A Weave manages versions of line-based text files, keeping track
92
130
of the originating version for each line.
177
215
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
180
def __init__(self, weave_name=None):
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__()
182
228
self._parents = []
185
231
self._name_map = {}
186
232
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
188
244
def __repr__(self):
189
245
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)
193
255
"""Return a deep copy of self.
195
257
The copy can be modified without affecting the original weave."""
197
259
other._weave = self._weave[:]
208
270
return self._parents == other._parents \
209
271
and self._weave == other._weave \
210
and self._sha1s == other._sha1s
272
and self._sha1s == other._sha1s
213
274
def __ne__(self, other):
214
275
return not self.__eq__(other)
216
@deprecated_method(zero_eight)
217
def idx_to_name(self, index):
218
"""Old public interface, the public interface is all names now."""
221
277
def _idx_to_name(self, version):
222
278
return self._names[version]
224
@deprecated_method(zero_eight)
225
def lookup(self, name):
226
"""Backwards compatability thunk:
228
Return name, as name is valid in the api now, and spew deprecation
233
280
def _lookup(self, name):
234
281
"""Convert symbolic version name to index."""
282
if not self._allow_reserved:
283
self.check_not_reserved_id(name)
236
285
return self._name_map[name]
238
287
raise RevisionNotPresent(name, self._weave_name)
240
@deprecated_method(zero_eight)
241
def iter_names(self):
242
"""Deprecated convenience function, please see VersionedFile.names()."""
243
return iter(self.names())
245
@deprecated_method(zero_eight)
247
"""See Weave.versions for the current api."""
248
return self.versions()
250
289
def versions(self):
251
290
"""See VersionedFile.versions."""
252
291
return self._names[:]
254
293
def has_version(self, version_id):
255
294
"""See VersionedFile.has_version."""
256
return self._name_map.has_key(version_id)
295
return (version_id in self._name_map)
258
297
__contains__ = has_version
260
@deprecated_method(zero_eight)
261
def parent_names(self, version):
262
"""Return version names for parents of a version.
264
See get_parents for the current api.
266
return self.get_parents(version)
268
def get_parents(self, version_id):
269
"""See VersionedFile.get_parent."""
270
return map(self._idx_to_name, self._parents[self._lookup(version_id)])
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:
272
380
def _check_repeated_add(self, name, parents, text, sha1):
273
381
"""Check that a duplicated add is OK.
280
388
raise RevisionAlreadyPresent(name, self._weave_name)
283
@deprecated_method(zero_eight)
284
def add_identical(self, old_rev_id, new_rev_id, parents):
285
"""Please use Weave.clone_text now."""
286
return self.clone_text(new_rev_id, old_rev_id, parents, None)
288
def add_lines(self, version_id, parents, lines):
391
def _add_lines(self, version_id, parents, lines, parent_texts,
392
left_matching_blocks, nostore_sha, random_id, check_content):
289
393
"""See VersionedFile.add_lines."""
290
return self._add(version_id, lines, map(self._lookup, parents))
292
@deprecated_method(zero_eight)
293
def add(self, name, parents, text, sha1=None):
294
"""See VersionedFile.add_lines for the non deprecated api."""
295
return self._add(name, text, map(self._maybe_lookup, parents), sha1)
297
def _add(self, version_id, lines, parents, sha1=None):
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):
298
399
"""Add a single text on top of the weave.
300
401
Returns the index number of the newly added version.
303
404
Symbolic name for this version.
304
405
(Typically the revision-id of the revision that added it.)
406
If None, a name will be allocated based on the hash. (sha1:SHAHASH)
307
409
List or set of direct parent version numbers.
310
412
Sequence of lines to be added in the new version.
414
:param nostore_sha: See VersionedFile.add_lines.
312
from bzrlib.osutils import sha_strings
314
assert isinstance(version_id, basestring)
416
self._check_lines_not_unicode(lines)
417
self._check_lines_are_lines(lines)
316
419
sha1 = sha_strings(lines)
420
if sha1 == nostore_sha:
421
raise errors.ExistingContent
422
if version_id is None:
423
version_id = "sha1:" + sha1
317
424
if version_id in self._name_map:
318
425
return self._check_repeated_add(version_id, parents, lines, sha1)
465
554
except IndexError:
466
555
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
468
565
def annotate(self, version_id):
469
if isinstance(version_id, int):
470
warn('Weave.annotate(int) is deprecated. Please use version names'
471
' in all circumstances as of 0.8',
476
for origin, lineno, text in self._extract([version_id]):
477
result.append((origin, text))
480
return super(Weave, self).annotate(version_id)
482
def annotate_iter(self, version_id):
483
"""Yield list of (version-id, line) pairs for the specified version.
566
"""Return a list of (version-id, line) tuples for version_id.
485
568
The index indicates when the line originated in the weave."""
486
569
incls = [self._lookup(version_id)]
487
for origin, lineno, text in self._extract(incls):
488
yield self._idx_to_name(origin), text
490
@deprecated_method(zero_eight)
492
"""_walk has become walk, a supported api."""
495
def walk(self, version_ids=None):
496
"""See VersionedFile.walk."""
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."""
501
592
lineno = 0 # line of weave, 0-based
503
594
for l in self._weave:
504
if isinstance(l, tuple):
595
if l.__class__ == tuple:
508
istack.append(self._idx_to_name(v))
599
istack.append(self._names[v])
512
assert self._idx_to_name(v) not in dset
513
dset.add(self._idx_to_name(v))
603
dset.add(self._names[v])
515
dset.remove(self._idx_to_name(v))
605
dset.remove(self._names[v])
517
607
raise WeaveFormatError('unexpected instruction %r' % v)
519
assert isinstance(l, basestring)
521
yield lineno, istack[-1], dset.copy(), l
609
yield lineno, istack[-1], frozenset(dset), l
528
616
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
531
663
def _extract(self, versions):
532
664
"""Yield annotation of lines in included set.
555
688
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.
557
717
for l in self._weave:
558
if isinstance(l, tuple):
718
if l.__class__ == tuple:
562
assert v not in istack
725
iset.remove(istack.pop())
567
727
if v in included:
572
730
if v in included:
733
raise AssertionError()
576
assert isinstance(l, basestring)
577
735
if isactive is None:
578
736
isactive = (not dset) and istack and (istack[-1] in included)
608
756
return self._lookup(name_or_index)
610
def _get_iter(self, version_id):
611
"""Yield lines for the specified version."""
612
incls = [self._maybe_lookup(version_id)]
617
# We don't have sha1 sums for multiple entries
619
for origin, lineno, line in self._extract(incls):
624
expected_sha1 = self._sha1s[index]
625
measured_sha1 = cur_sha.hexdigest()
626
if measured_sha1 != expected_sha1:
627
raise errors.WeaveInvalidChecksum(
628
'file %s, revision %s, expected: %s, measured %s'
629
% (self._weave_name, self._names[index],
630
expected_sha1, measured_sha1))
632
@deprecated_method(zero_eight)
633
def get(self, version_id):
634
"""Please use either Weave.get_text or Weave.get_lines as desired."""
635
return self.get_lines(version_id)
637
758
def get_lines(self, version_id):
638
759
"""See VersionedFile.get_lines()."""
639
return list(self._get_iter(version_id))
641
def get_sha1(self, name):
642
"""Get the stored sha1 sum for the given revision.
644
:param name: The name of the version to lookup
646
return self._sha1s[self._lookup(name)]
648
def numversions(self):
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?"""
649
780
l = len(self._parents)
650
assert l == len(self._sha1s)
653
__len__ = numversions
783
__len__ = num_versions
655
785
def check(self, progress_bar=None):
656
786
# TODO evaluate performance hit of using string sets in this routine.
657
# check no circular inclusions
658
for version in range(self.numversions()):
787
# TODO: check no circular inclusions
788
# TODO: create a nested progress bar
789
for version in range(self.num_versions()):
659
790
inclusions = list(self._parents[version])
661
792
inclusions.sort()
715
848
# no lines outside of insertion blocks, that deletions are
716
849
# properly paired, etc.
719
def join(self, other, pb=None, msg=None, version_ids=None):
721
"""Integrate versions from other into this weave.
723
The resulting weave contains all the history of both weaves;
724
any version you could retrieve from either self or other can be
725
retrieved from self after this call.
727
It is illegal for the two weaves to contain different values
728
or different parents for any version. See also reweave().
730
:param other: The other weave to pull into this one
731
:param pb: An optional progress bar
732
:param msg: An optional message to display for progress
734
if not other.versions():
735
return # nothing to update, easy
738
for version_id in version_ids:
739
if not self.has_version(version_id):
740
raise RevisionNotPresent(version_id, self._weave_name)
741
assert version_ids == None
743
# two loops so that we do not change ourselves before verifying it
745
# work through in index order to make sure we get all dependencies
748
for other_idx, name in enumerate(other._names):
749
self._check_version_consistent(other, other_idx, name)
750
sha1 = other._sha1s[other_idx]
754
if name in self._name_map:
755
idx = self._lookup(name)
756
n1 = set(map(other._idx_to_name, other._parents[other_idx]))
757
n2 = set(map(self._idx_to_name, self._parents[idx]))
758
if sha1 == self._sha1s[idx] and n1 == n2:
761
names_to_join.append((other_idx, name))
768
for other_idx, name in names_to_join:
769
# TODO: If all the parents of the other version are already
770
# present then we can avoid some work by just taking the delta
771
# and adjusting the offsets.
772
new_parents = self._imported_parents(other, other_idx)
773
sha1 = other._sha1s[other_idx]
778
pb.update(msg, merged, len(names_to_join))
780
lines = other.get_lines(other_idx)
781
self._add(name, lines, new_parents, sha1)
783
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
784
merged, processed, self._weave_name, time.time()-time0))
786
851
def _imported_parents(self, other, other_idx):
787
852
"""Return list of parents in self corresponding to indexes in other."""
789
854
for parent_idx in other._parents[other_idx]:
790
855
parent_name = other._names[parent_idx]
791
if parent_name not in self._names:
856
if parent_name not in self._name_map:
792
857
# should not be possible
793
raise WeaveError("missing parent {%s} of {%s} in %r"
858
raise WeaveError("missing parent {%s} of {%s} in %r"
794
859
% (parent_name, other._name_map[other_idx], self))
795
860
new_parents.append(self._name_map[parent_name])
796
861
return new_parents
828
def reweave(self, other, pb=None, msg=None):
829
"""Reweave self with other.
891
def _reweave(self, other, pb, msg):
892
"""Reweave self with other - internal helper for join().
831
894
:param other: The other weave to merge
832
895
:param pb: An optional progress bar, indicating how far done we are
833
896
:param msg: An optional message for the progress
835
new_weave = reweave(self, other, pb=pb, msg=msg)
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."""
836
903
for attr in self.__slots__:
837
setattr(self, attr, getattr(new_weave, attr))
840
def reweave(wa, wb, pb=None, msg=None):
904
if attr != '_weave_name':
905
setattr(self, attr, copy(getattr(otherweave, attr)))
908
class WeaveFile(Weave):
909
"""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)
920
self._transport = transport
921
self._filemode = filemode
923
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
924
except errors.NoSuchFile:
930
def _add_lines(self, version_id, parents, lines, parent_texts,
931
left_matching_blocks, nostore_sha, random_id, check_content):
932
"""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,
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
"""Save the weave."""
950
self._check_write_ok()
952
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)
972
def _reweave(wa, wb, pb=None, msg=None):
841
973
"""Combine two weaves and return the result.
843
This works even if a revision R has different parents in
975
This works even if a revision R has different parents in
844
976
wa and wb. In the resulting weave all the parents are given.
846
This is done by just building up a new weave, maintaining ordering
978
This is done by just building up a new weave, maintaining ordering
847
979
of the versions in the two inputs. More efficient approaches
848
might be possible but it should only be necessary to do
849
this operation rarely, when a new previously ghost version is
980
might be possible but it should only be necessary to do
981
this operation rarely, when a new previously ghost version is
852
984
:param pb: An optional progress bar, indicating how far done we are
896
1029
p = combined.setdefault(name, set())
897
1030
p.update(map(weave._idx_to_name, weave._parents[idx]))
902
"""Show the weave's table-of-contents"""
903
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
904
for i in (6, 50, 10, 10):
907
for i in range(w.numversions()):
910
parent_str = ' '.join(map(str, w._parents[i]))
911
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
915
def weave_stats(weave_file, pb):
916
from bzrlib.weavefile import read_weave
918
wf = file(weave_file, 'rb')
919
w = read_weave(wf, WeaveVersionedFile)
920
# FIXME: doesn't work on pipes
921
weave_size = wf.tell()
925
for i in range(vers):
926
pb.update('checking sizes', i, vers)
927
for origin, lineno, line in w._extract([i]):
932
print 'versions %9d' % vers
933
print 'weave file %9d bytes' % weave_size
934
print 'total contents %9d bytes' % total
935
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
938
print 'average size %9d bytes' % avg
939
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
943
print """bzr weave tool
945
Experimental tool for weave algorithm.
949
Create an empty weave file
950
weave get WEAVEFILE VERSION
951
Write out specified version.
952
weave check WEAVEFILE
953
Check consistency of all versions.
955
Display table of contents.
956
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
957
Add NEWTEXT, with specified parent versions.
958
weave annotate WEAVEFILE VERSION
959
Display origin of each line.
960
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
961
Auto-merge two versions and display conflicts.
962
weave diff WEAVEFILE VERSION1 VERSION2
963
Show differences between two versions.
967
% weave init foo.weave
969
% weave add foo.weave ver0 < foo.txt
972
(create updated version)
974
% weave get foo.weave 0 | diff -u - foo.txt
975
% weave add foo.weave ver1 0 < foo.txt
978
% weave get foo.weave 0 > foo.txt (create forked version)
980
% weave add foo.weave ver2 0 < foo.txt
983
% weave merge foo.weave 1 2 > foo.txt (merge them)
984
% vi foo.txt (resolve conflicts)
985
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
997
# in case we're run directly from the subdirectory
998
sys.path.append('..')
1000
from bzrlib.weavefile import write_weave, read_weave
1001
from bzrlib.progress import ProgressBar
1016
return read_weave(file(argv[2], 'rb'))
1022
# at the moment, based on everything in the file
1024
parents = map(int, argv[4:])
1025
lines = sys.stdin.readlines()
1026
ver = w.add(name, parents, lines)
1027
write_weave(w, file(argv[2], 'wb'))
1028
print 'added version %r %d' % (name, ver)
1031
if os.path.exists(fn):
1032
raise IOError("file exists")
1034
write_weave(w, file(fn, 'wb'))
1035
elif cmd == 'get': # get one version
1037
sys.stdout.writelines(w.get_iter(int(argv[3])))
1040
from difflib import unified_diff
1043
v1, v2 = map(int, argv[3:5])
1046
diff_gen = unified_diff(lines1, lines2,
1047
'%s version %d' % (fn, v1),
1048
'%s version %d' % (fn, v2))
1049
sys.stdout.writelines(diff_gen)
1051
elif cmd == 'annotate':
1053
# newline is added to all lines regardless; too hard to get
1054
# reasonable formatting otherwise
1056
for origin, text in w.annotate(int(argv[3])):
1057
text = text.rstrip('\r\n')
1059
print ' | %s' % (text)
1061
print '%5d | %s' % (origin, text)
1067
elif cmd == 'stats':
1068
weave_stats(argv[2], ProgressBar())
1070
elif cmd == 'check':
1075
print '%d versions ok' % w.numversions()
1077
elif cmd == 'inclusions':
1079
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1081
elif cmd == 'parents':
1083
print ' '.join(map(str, w._parents[int(argv[3])]))
1085
elif cmd == 'plan-merge':
1087
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1089
print '%14s | %s' % (state, line),
1091
elif cmd == 'merge':
1093
p = w.plan_merge(int(argv[3]), int(argv[4]))
1094
sys.stdout.writelines(w.weave_merge(p))
1097
raise ValueError('unknown command %r' % cmd)
1101
def profile_main(argv):
1102
import tempfile, hotshot, hotshot.stats
1104
prof_f = tempfile.NamedTemporaryFile()
1106
prof = hotshot.Profile(prof_f.name)
1108
ret = prof.runcall(main, argv)
1111
stats = hotshot.stats.load(prof_f.name)
1113
stats.sort_stats('cumulative')
1114
## XXX: Might like to write to stderr or the trace file instead but
1115
## print_stats seems hardcoded to stdout
1116
stats.print_stats(20)
1121
def lsprofile_main(argv):
1122
from bzrlib.lsprof import profile
1123
ret,stats = profile(main, argv)
1129
if __name__ == '__main__':
1131
if '--profile' in sys.argv:
1133
args.remove('--profile')
1134
sys.exit(profile_main(args))
1135
elif '--lsprof' in sys.argv:
1137
args.remove('--lsprof')
1138
sys.exit(lsprofile_main(args))
1140
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
