69
71
from copy import copy
70
72
from cStringIO import StringIO
73
from bzrlib.lazy_import import lazy_import
74
lazy_import(globals(), """
75
from bzrlib import tsort
77
78
from bzrlib import (
81
from bzrlib.trace import mutter
81
82
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
82
83
RevisionAlreadyPresent,
83
84
RevisionNotPresent,
84
UnavailableRepresentation,
85
WeaveRevisionAlreadyPresent,
86
WeaveRevisionNotPresent,
86
from bzrlib.osutils import dirname, sha, sha_strings, split_lines
88
import bzrlib.errors as errors
89
from bzrlib.osutils import sha_strings
87
90
import bzrlib.patiencediff
88
from bzrlib.revision import NULL_REVISION
89
from bzrlib.symbol_versioning import *
90
from bzrlib.trace import mutter
91
from bzrlib.versionedfile import (
91
from bzrlib.tsort import topo_sort
92
from bzrlib.versionedfile import VersionedFile, InterVersionedFile
98
93
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
96
class Weave(VersionedFile):
127
97
"""weave - versioned text file storage.
129
99
A Weave manages versions of line-based text files, keeping track
130
100
of the originating version for each line.
297
264
__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
266
def get_parent_map(self, version_ids):
330
267
"""See VersionedFile.get_parent_map."""
332
269
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
271
result[version_id] = tuple(
272
map(self._idx_to_name, self._parents[self._lookup(version_id)]))
273
except RevisionNotPresent:
345
277
def get_parents_with_ghosts(self, version_id):
346
278
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:
380
280
def _check_repeated_add(self, name, parents, text, sha1):
381
281
"""Check that a duplicated add is OK.
848
747
# no lines outside of insertion blocks, that deletions are
849
748
# properly paired, etc.
750
def _join(self, other, pb, msg, version_ids, ignore_missing):
751
"""Worker routine for join()."""
752
if not other.versions():
753
return # nothing to update, easy
756
# versions is never none, InterWeave checks this.
759
# two loops so that we do not change ourselves before verifying it
761
# work through in index order to make sure we get all dependencies
764
# get the selected versions only that are in other.versions.
765
version_ids = set(other.versions()).intersection(set(version_ids))
766
# pull in the referenced graph.
767
version_ids = other.get_ancestry(version_ids)
768
pending_parents = other.get_parent_map(version_ids)
769
pending_graph = pending_parents.items()
770
if len(pending_graph) != len(version_ids):
771
raise RevisionNotPresent(
772
set(version_ids) - set(pending_parents.keys()), self)
773
for name in topo_sort(pending_graph):
774
other_idx = other._name_map[name]
775
# returns True if we have it, False if we need it.
776
if not self._check_version_consistent(other, other_idx, name):
777
names_to_join.append((other_idx, name))
785
for other_idx, name in names_to_join:
786
# TODO: If all the parents of the other version are already
787
# present then we can avoid some work by just taking the delta
788
# and adjusting the offsets.
789
new_parents = self._imported_parents(other, other_idx)
790
sha1 = other._sha1s[other_idx]
795
pb.update(msg, merged, len(names_to_join))
797
lines = other.get_lines(other_idx)
798
self._add(name, lines, new_parents, sha1)
800
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
801
merged, processed, self._weave_name, time.time()-time0))
851
803
def _imported_parents(self, other, other_idx):
852
804
"""Return list of parents in self corresponding to indexes in other."""
952
903
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)
905
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
963
910
def get_suffixes():
964
911
"""See VersionedFile.get_suffixes()."""
965
912
return [WeaveFile.WEAVE_SUFFIX]
967
def insert_record_stream(self, stream):
968
super(WeaveFile, self).insert_record_stream(stream)
914
def join(self, other, pb=None, msg=None, version_ids=None,
915
ignore_missing=False):
916
"""Join other into self and save."""
917
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
972
921
def _reweave(wa, wb, pb=None, msg=None):
973
922
"""Combine two weaves and return the result.
975
This works even if a revision R has different parents in
924
This works even if a revision R has different parents in
976
925
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
927
This is done by just building up a new weave, maintaining ordering
979
928
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
929
might be possible but it should only be necessary to do
930
this operation rarely, when a new previously ghost version is
984
933
:param pb: An optional progress bar, indicating how far done we are
1029
977
p = combined.setdefault(name, set())
1030
978
p.update(map(weave._idx_to_name, weave._parents[idx]))
983
"""Show the weave's table-of-contents"""
984
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
985
for i in (6, 50, 10, 10):
988
for i in range(w.num_versions()):
991
parent_str = ' '.join(map(str, w._parents[i]))
992
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
996
def weave_stats(weave_file, pb):
997
from bzrlib.weavefile import read_weave
999
wf = file(weave_file, 'rb')
1001
# FIXME: doesn't work on pipes
1002
weave_size = wf.tell()
1006
for i in range(vers):
1007
pb.update('checking sizes', i, vers)
1008
for origin, lineno, line in w._extract([i]):
1013
print 'versions %9d' % vers
1014
print 'weave file %9d bytes' % weave_size
1015
print 'total contents %9d bytes' % total
1016
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1019
print 'average size %9d bytes' % avg
1020
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1024
print """bzr weave tool
1026
Experimental tool for weave algorithm.
1029
weave init WEAVEFILE
1030
Create an empty weave file
1031
weave get WEAVEFILE VERSION
1032
Write out specified version.
1033
weave check WEAVEFILE
1034
Check consistency of all versions.
1036
Display table of contents.
1037
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1038
Add NEWTEXT, with specified parent versions.
1039
weave annotate WEAVEFILE VERSION
1040
Display origin of each line.
1041
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1042
Auto-merge two versions and display conflicts.
1043
weave diff WEAVEFILE VERSION1 VERSION2
1044
Show differences between two versions.
1048
% weave init foo.weave
1050
% weave add foo.weave ver0 < foo.txt
1053
(create updated version)
1055
% weave get foo.weave 0 | diff -u - foo.txt
1056
% weave add foo.weave ver1 0 < foo.txt
1059
% weave get foo.weave 0 > foo.txt (create forked version)
1061
% weave add foo.weave ver2 0 < foo.txt
1064
% weave merge foo.weave 1 2 > foo.txt (merge them)
1065
% vi foo.txt (resolve conflicts)
1066
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1078
# in case we're run directly from the subdirectory
1079
sys.path.append('..')
1081
from bzrlib.weavefile import write_weave, read_weave
1082
from bzrlib.progress import ProgressBar
1097
return read_weave(file(argv[2], 'rb'))
1103
# at the moment, based on everything in the file
1105
parents = map(int, argv[4:])
1106
lines = sys.stdin.readlines()
1107
ver = w.add(name, parents, lines)
1108
write_weave(w, file(argv[2], 'wb'))
1109
print 'added version %r %d' % (name, ver)
1112
if os.path.exists(fn):
1113
raise IOError("file exists")
1115
write_weave(w, file(fn, 'wb'))
1116
elif cmd == 'get': # get one version
1118
sys.stdout.writelines(w.get_iter(int(argv[3])))
1123
v1, v2 = map(int, argv[3:5])
1126
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1127
'%s version %d' % (fn, v1),
1128
'%s version %d' % (fn, v2))
1129
sys.stdout.writelines(diff_gen)
1131
elif cmd == 'annotate':
1133
# newline is added to all lines regardless; too hard to get
1134
# reasonable formatting otherwise
1136
for origin, text in w.annotate(int(argv[3])):
1137
text = text.rstrip('\r\n')
1139
print ' | %s' % (text)
1141
print '%5d | %s' % (origin, text)
1147
elif cmd == 'stats':
1148
weave_stats(argv[2], ProgressBar())
1150
elif cmd == 'check':
1155
print '%d versions ok' % w.num_versions()
1157
elif cmd == 'inclusions':
1159
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1161
elif cmd == 'parents':
1163
print ' '.join(map(str, w._parents[int(argv[3])]))
1165
elif cmd == 'plan-merge':
1166
# replaced by 'bzr weave-plan-merge'
1168
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1170
print '%14s | %s' % (state, line),
1171
elif cmd == 'merge':
1172
# replaced by 'bzr weave-merge-text'
1174
p = w.plan_merge(int(argv[3]), int(argv[4]))
1175
sys.stdout.writelines(w.weave_merge(p))
1177
raise ValueError('unknown command %r' % cmd)
1180
if __name__ == '__main__':
1182
sys.exit(main(sys.argv))
1185
class InterWeave(InterVersionedFile):
1186
"""Optimised code paths for weave to weave operations."""
1188
_matching_file_from_factory = staticmethod(WeaveFile)
1189
_matching_file_to_factory = staticmethod(WeaveFile)
1192
def is_compatible(source, target):
1193
"""Be compatible with weaves."""
1195
return (isinstance(source, Weave) and
1196
isinstance(target, Weave))
1197
except AttributeError:
1200
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1201
"""See InterVersionedFile.join."""
1202
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1203
if self.target.versions() == [] and version_ids is None:
1204
self.target._copy_weave_content(self.source)
1206
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1209
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
