71
69
from copy import copy
72
70
from cStringIO import StringIO
75
from bzrlib.lazy_import import lazy_import
76
lazy_import(globals(), """
77
from bzrlib import tsort
78
79
from bzrlib import (
81
from bzrlib.trace import mutter
82
84
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
83
85
RevisionAlreadyPresent,
84
86
RevisionNotPresent,
87
UnavailableRepresentation,
85
88
WeaveRevisionAlreadyPresent,
86
89
WeaveRevisionNotPresent,
88
import bzrlib.errors as errors
89
from bzrlib.osutils import sha_strings
91
from bzrlib.osutils import dirname, sha, sha_strings, split_lines
90
92
import bzrlib.patiencediff
91
from bzrlib.tsort import topo_sort
92
from bzrlib.versionedfile import VersionedFile, InterVersionedFile
93
from bzrlib.revision import NULL_REVISION
94
from bzrlib.symbol_versioning import *
95
from bzrlib.trace import mutter
96
from bzrlib.versionedfile import (
93
103
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
106
class WeaveContentFactory(ContentFactory):
107
"""Content factory for streaming from weaves.
109
:seealso ContentFactory:
112
def __init__(self, version, weave):
113
"""Create a WeaveContentFactory for version from weave."""
114
ContentFactory.__init__(self)
115
self.sha1 = weave.get_sha1s([version])[version]
116
self.key = (version,)
117
parents = weave.get_parent_map([version])[version]
118
self.parents = tuple((parent,) for parent in parents)
119
self.storage_kind = 'fulltext'
122
def get_bytes_as(self, storage_kind):
123
if storage_kind == 'fulltext':
124
return self._weave.get_text(self.key[-1])
125
elif storage_kind == 'chunked':
126
return self._weave.get_lines(self.key[-1])
128
raise UnavailableRepresentation(self.key, storage_kind, 'fulltext')
96
131
class Weave(VersionedFile):
97
132
"""weave - versioned text file storage.
99
134
A Weave manages versions of line-based text files, keeping track
100
135
of the originating version for each line.
246
302
__contains__ = has_version
304
def get_record_stream(self, versions, ordering, include_delta_closure):
305
"""Get a stream of records for versions.
307
:param versions: The versions to include. Each version is a tuple
309
:param ordering: Either 'unordered' or 'topological'. A topologically
310
sorted stream has compression parents strictly before their
312
:param include_delta_closure: If True then the closure across any
313
compression parents will be included (in the opaque data).
314
:return: An iterator of ContentFactory objects, each of which is only
315
valid until the iterator is advanced.
317
versions = [version[-1] for version in versions]
318
if ordering == 'topological':
319
parents = self.get_parent_map(versions)
320
new_versions = tsort.topo_sort(parents)
321
new_versions.extend(set(versions).difference(set(parents)))
322
versions = new_versions
323
elif ordering == 'groupcompress':
324
parents = self.get_parent_map(versions)
325
new_versions = sort_groupcompress(parents)
326
new_versions.extend(set(versions).difference(set(parents)))
327
versions = new_versions
328
for version in versions:
330
yield WeaveContentFactory(version, self)
332
yield AbsentContentFactory((version,))
248
334
def get_parent_map(self, version_ids):
249
335
"""See VersionedFile.get_parent_map."""
251
337
for version_id in version_ids:
253
result[version_id] = tuple(
254
map(self._idx_to_name, self._parents[self._lookup(version_id)]))
255
except RevisionNotPresent:
338
if version_id == NULL_REVISION:
343
map(self._idx_to_name,
344
self._parents[self._lookup(version_id)]))
345
except RevisionNotPresent:
347
result[version_id] = parents
259
350
def get_parents_with_ghosts(self, version_id):
260
351
raise NotImplementedError(self.get_parents_with_ghosts)
353
def insert_record_stream(self, stream):
354
"""Insert a record stream into this versioned file.
356
:param stream: A stream of records to insert.
358
:seealso VersionedFile.get_record_stream:
361
for record in stream:
362
# Raise an error when a record is missing.
363
if record.storage_kind == 'absent':
364
raise RevisionNotPresent([record.key[0]], self)
365
# adapt to non-tuple interface
366
parents = [parent[0] for parent in record.parents]
367
if (record.storage_kind == 'fulltext'
368
or record.storage_kind == 'chunked'):
369
self.add_lines(record.key[0], parents,
370
osutils.chunks_to_lines(record.get_bytes_as('chunked')))
372
adapter_key = record.storage_kind, 'fulltext'
374
adapter = adapters[adapter_key]
376
adapter_factory = adapter_registry.get(adapter_key)
377
adapter = adapter_factory(self)
378
adapters[adapter_key] = adapter
379
lines = split_lines(adapter.get_bytes(record))
381
self.add_lines(record.key[0], parents, lines)
382
except RevisionAlreadyPresent:
262
385
def _check_repeated_add(self, name, parents, text, sha1):
263
386
"""Check that a duplicated add is OK.
280
403
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
281
404
"""Add a single text on top of the weave.
283
406
Returns the index number of the newly added version.
286
409
Symbolic name for this version.
287
410
(Typically the revision-id of the revision that added it.)
411
If None, a name will be allocated based on the hash. (sha1:SHAHASH)
290
414
List or set of direct parent version numbers.
293
417
Sequence of lines to be added in the new version.
295
419
:param nostore_sha: See VersionedFile.add_lines.
297
assert isinstance(version_id, basestring)
298
421
self._check_lines_not_unicode(lines)
299
422
self._check_lines_are_lines(lines)
301
424
sha1 = sha_strings(lines)
302
425
if sha1 == nostore_sha:
303
426
raise errors.ExistingContent
427
if version_id is None:
428
version_id = "sha1:" + sha1
304
429
if version_id in self._name_map:
305
430
return self._check_repeated_add(version_id, parents, lines, sha1)
448
562
def _compatible_parents(self, my_parents, other_parents):
449
563
"""During join check that other_parents are joinable with my_parents.
451
Joinable is defined as 'is a subset of' - supersets may require
565
Joinable is defined as 'is a subset of' - supersets may require
452
566
regeneration of diffs, but subsets do not.
454
568
return len(other_parents.difference(my_parents)) == 0
456
def annotate_iter(self, version_id):
457
"""Yield list of (version-id, line) pairs for the specified version.
570
def annotate(self, version_id):
571
"""Return a list of (version-id, line) tuples for version_id.
459
573
The index indicates when the line originated in the weave."""
460
574
incls = [self._lookup(version_id)]
461
for origin, lineno, text in self._extract(incls):
462
yield self._idx_to_name(origin), text
575
return [(self._idx_to_name(origin), text) for origin, lineno, text in
576
self._extract(incls)]
464
578
def iter_lines_added_or_present_in_versions(self, version_ids=None,
598
704
# we're still spending ~1/4 of the method in isinstance though.
599
705
# so lets hard code the acceptable string classes we expect:
600
706
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
601
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
707
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
603
709
# yay, down to ~1/4 the initial extract time, and our inline time
604
710
# has shrunk again, with isinstance no longer dominating.
605
711
# tweaking the stack inclusion test to use a set gives:
606
712
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
607
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
713
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
609
715
# - a 5% win, or possibly just noise. However with large istacks that
610
716
# 'in' test could dominate, so I'm leaving this change in place -
611
717
# when its fast enough to consider profiling big datasets we can review.
616
722
for l in self._weave:
617
723
if l.__class__ == tuple:
625
730
iset.remove(istack.pop())
627
732
if v in included:
632
735
if v in included:
738
raise AssertionError()
636
assert l.__class__ in (str, unicode)
637
740
if isactive is None:
638
741
isactive = (not dset) and istack and (istack[-1] in included)
750
853
# no lines outside of insertion blocks, that deletions are
751
854
# properly paired, etc.
753
def _join(self, other, pb, msg, version_ids, ignore_missing):
754
"""Worker routine for join()."""
755
if not other.versions():
756
return # nothing to update, easy
759
# versions is never none, InterWeave checks this.
762
# two loops so that we do not change ourselves before verifying it
764
# work through in index order to make sure we get all dependencies
767
# get the selected versions only that are in other.versions.
768
version_ids = set(other.versions()).intersection(set(version_ids))
769
# pull in the referenced graph.
770
version_ids = other.get_ancestry(version_ids)
771
pending_parents = other.get_parent_map(version_ids)
772
pending_graph = pending_parents.items()
773
if len(pending_graph) != len(version_ids):
774
raise RevisionNotPresent(
775
set(version_ids) - set(pending_parents.keys()), self)
776
for name in topo_sort(pending_graph):
777
other_idx = other._name_map[name]
778
# returns True if we have it, False if we need it.
779
if not self._check_version_consistent(other, other_idx, name):
780
names_to_join.append((other_idx, name))
788
for other_idx, name in names_to_join:
789
# TODO: If all the parents of the other version are already
790
# present then we can avoid some work by just taking the delta
791
# and adjusting the offsets.
792
new_parents = self._imported_parents(other, other_idx)
793
sha1 = other._sha1s[other_idx]
798
pb.update(msg, merged, len(names_to_join))
800
lines = other.get_lines(other_idx)
801
self._add(name, lines, new_parents, sha1)
803
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
804
merged, processed, self._weave_name, time.time()-time0))
806
856
def _imported_parents(self, other, other_idx):
807
857
"""Return list of parents in self corresponding to indexes in other."""
911
957
write_weave_v5(self, sio)
913
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
959
bytes = sio.getvalue()
960
path = self._weave_name + WeaveFile.WEAVE_SUFFIX
962
self._transport.put_bytes(path, bytes, self._filemode)
963
except errors.NoSuchFile:
964
self._transport.mkdir(dirname(path))
965
self._transport.put_bytes(path, bytes, self._filemode)
918
968
def get_suffixes():
919
969
"""See VersionedFile.get_suffixes()."""
920
970
return [WeaveFile.WEAVE_SUFFIX]
922
def join(self, other, pb=None, msg=None, version_ids=None,
923
ignore_missing=False):
924
"""Join other into self and save."""
925
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
972
def insert_record_stream(self, stream):
973
super(WeaveFile, self).insert_record_stream(stream)
929
977
def _reweave(wa, wb, pb=None, msg=None):
930
978
"""Combine two weaves and return the result.
932
This works even if a revision R has different parents in
980
This works even if a revision R has different parents in
933
981
wa and wb. In the resulting weave all the parents are given.
935
This is done by just building up a new weave, maintaining ordering
983
This is done by just building up a new weave, maintaining ordering
936
984
of the versions in the two inputs. More efficient approaches
937
might be possible but it should only be necessary to do
938
this operation rarely, when a new previously ghost version is
985
might be possible but it should only be necessary to do
986
this operation rarely, when a new previously ghost version is
941
989
:param pb: An optional progress bar, indicating how far done we are
985
1034
p = combined.setdefault(name, set())
986
1035
p.update(map(weave._idx_to_name, weave._parents[idx]))
991
"""Show the weave's table-of-contents"""
992
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
993
for i in (6, 50, 10, 10):
996
for i in range(w.num_versions()):
999
parent_str = ' '.join(map(str, w._parents[i]))
1000
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1004
def weave_stats(weave_file, pb):
1005
from bzrlib.weavefile import read_weave
1007
wf = file(weave_file, 'rb')
1009
# FIXME: doesn't work on pipes
1010
weave_size = wf.tell()
1014
for i in range(vers):
1015
pb.update('checking sizes', i, vers)
1016
for origin, lineno, line in w._extract([i]):
1021
print 'versions %9d' % vers
1022
print 'weave file %9d bytes' % weave_size
1023
print 'total contents %9d bytes' % total
1024
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1027
print 'average size %9d bytes' % avg
1028
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1032
print """bzr weave tool
1034
Experimental tool for weave algorithm.
1037
weave init WEAVEFILE
1038
Create an empty weave file
1039
weave get WEAVEFILE VERSION
1040
Write out specified version.
1041
weave check WEAVEFILE
1042
Check consistency of all versions.
1044
Display table of contents.
1045
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1046
Add NEWTEXT, with specified parent versions.
1047
weave annotate WEAVEFILE VERSION
1048
Display origin of each line.
1049
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1050
Auto-merge two versions and display conflicts.
1051
weave diff WEAVEFILE VERSION1 VERSION2
1052
Show differences between two versions.
1056
% weave init foo.weave
1058
% weave add foo.weave ver0 < foo.txt
1061
(create updated version)
1063
% weave get foo.weave 0 | diff -u - foo.txt
1064
% weave add foo.weave ver1 0 < foo.txt
1067
% weave get foo.weave 0 > foo.txt (create forked version)
1069
% weave add foo.weave ver2 0 < foo.txt
1072
% weave merge foo.weave 1 2 > foo.txt (merge them)
1073
% vi foo.txt (resolve conflicts)
1074
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1086
# in case we're run directly from the subdirectory
1087
sys.path.append('..')
1089
from bzrlib.weavefile import write_weave, read_weave
1090
from bzrlib.progress import ProgressBar
1105
return read_weave(file(argv[2], 'rb'))
1111
# at the moment, based on everything in the file
1113
parents = map(int, argv[4:])
1114
lines = sys.stdin.readlines()
1115
ver = w.add(name, parents, lines)
1116
write_weave(w, file(argv[2], 'wb'))
1117
print 'added version %r %d' % (name, ver)
1120
if os.path.exists(fn):
1121
raise IOError("file exists")
1123
write_weave(w, file(fn, 'wb'))
1124
elif cmd == 'get': # get one version
1126
sys.stdout.writelines(w.get_iter(int(argv[3])))
1131
v1, v2 = map(int, argv[3:5])
1134
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1135
'%s version %d' % (fn, v1),
1136
'%s version %d' % (fn, v2))
1137
sys.stdout.writelines(diff_gen)
1139
elif cmd == 'annotate':
1141
# newline is added to all lines regardless; too hard to get
1142
# reasonable formatting otherwise
1144
for origin, text in w.annotate(int(argv[3])):
1145
text = text.rstrip('\r\n')
1147
print ' | %s' % (text)
1149
print '%5d | %s' % (origin, text)
1155
elif cmd == 'stats':
1156
weave_stats(argv[2], ProgressBar())
1158
elif cmd == 'check':
1163
print '%d versions ok' % w.num_versions()
1165
elif cmd == 'inclusions':
1167
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1169
elif cmd == 'parents':
1171
print ' '.join(map(str, w._parents[int(argv[3])]))
1173
elif cmd == 'plan-merge':
1174
# replaced by 'bzr weave-plan-merge'
1176
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1178
print '%14s | %s' % (state, line),
1179
elif cmd == 'merge':
1180
# replaced by 'bzr weave-merge-text'
1182
p = w.plan_merge(int(argv[3]), int(argv[4]))
1183
sys.stdout.writelines(w.weave_merge(p))
1185
raise ValueError('unknown command %r' % cmd)
1188
if __name__ == '__main__':
1190
sys.exit(main(sys.argv))
1193
class InterWeave(InterVersionedFile):
1194
"""Optimised code paths for weave to weave operations."""
1196
_matching_file_from_factory = staticmethod(WeaveFile)
1197
_matching_file_to_factory = staticmethod(WeaveFile)
1200
def is_compatible(source, target):
1201
"""Be compatible with weaves."""
1203
return (isinstance(source, Weave) and
1204
isinstance(target, Weave))
1205
except AttributeError:
1208
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1209
"""See InterVersionedFile.join."""
1210
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1211
if self.target.versions() == [] and version_ids is None:
1212
self.target._copy_weave_content(self.source)
1214
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1217
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
