398
384
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
399
385
"""Add a single text on top of the weave.
401
387
Returns the index number of the newly added version.
404
390
Symbolic name for this version.
405
391
(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
394
List or set of direct parent version numbers.
412
397
Sequence of lines to be added in the new version.
414
399
:param nostore_sha: See VersionedFile.add_lines.
401
assert isinstance(version_id, basestring)
416
402
self._check_lines_not_unicode(lines)
417
403
self._check_lines_are_lines(lines)
419
405
sha1 = sha_strings(lines)
420
406
if sha1 == nostore_sha:
421
407
raise errors.ExistingContent
422
if version_id is None:
423
version_id = "sha1:" + sha1
424
408
if version_id in self._name_map:
425
409
return self._check_repeated_add(version_id, parents, lines, sha1)
699
697
# we're still spending ~1/4 of the method in isinstance though.
700
698
# so lets hard code the acceptable string classes we expect:
701
699
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
702
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
700
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
704
702
# yay, down to ~1/4 the initial extract time, and our inline time
705
703
# has shrunk again, with isinstance no longer dominating.
706
704
# tweaking the stack inclusion test to use a set gives:
707
705
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
708
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
706
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
710
708
# - a 5% win, or possibly just noise. However with large istacks that
711
709
# 'in' test could dominate, so I'm leaving this change in place -
712
710
# when its fast enough to consider profiling big datasets we can review.
717
715
for l in self._weave:
718
716
if l.__class__ == tuple:
725
724
iset.remove(istack.pop())
727
726
if v in included:
730
731
if v in included:
733
raise AssertionError()
735
assert l.__class__ in (str, unicode)
735
736
if isactive is None:
736
737
isactive = (not dset) and istack and (istack[-1] in included)
848
845
# no lines outside of insertion blocks, that deletions are
849
846
# properly paired, etc.
848
def _join(self, other, pb, msg, version_ids, ignore_missing):
849
"""Worker routine for join()."""
850
if not other.versions():
851
return # nothing to update, easy
854
# versions is never none, InterWeave checks this.
857
# two loops so that we do not change ourselves before verifying it
859
# work through in index order to make sure we get all dependencies
862
# get the selected versions only that are in other.versions.
863
version_ids = set(other.versions()).intersection(set(version_ids))
864
# pull in the referenced graph.
865
version_ids = other.get_ancestry(version_ids)
866
pending_parents = other.get_parent_map(version_ids)
867
pending_graph = pending_parents.items()
868
if len(pending_graph) != len(version_ids):
869
raise RevisionNotPresent(
870
set(version_ids) - set(pending_parents.keys()), self)
871
for name in topo_sort(pending_graph):
872
other_idx = other._name_map[name]
873
# returns True if we have it, False if we need it.
874
if not self._check_version_consistent(other, other_idx, name):
875
names_to_join.append((other_idx, name))
883
for other_idx, name in names_to_join:
884
# TODO: If all the parents of the other version are already
885
# present then we can avoid some work by just taking the delta
886
# and adjusting the offsets.
887
new_parents = self._imported_parents(other, other_idx)
888
sha1 = other._sha1s[other_idx]
893
pb.update(msg, merged, len(names_to_join))
895
lines = other.get_lines(other_idx)
896
self._add(name, lines, new_parents, sha1)
898
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
899
merged, processed, self._weave_name, time.time()-time0))
851
901
def _imported_parents(self, other, other_idx):
852
902
"""Return list of parents in self corresponding to indexes in other."""
951
1000
sio = StringIO()
952
1001
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)
1003
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
963
1008
def get_suffixes():
964
1009
"""See VersionedFile.get_suffixes()."""
965
1010
return [WeaveFile.WEAVE_SUFFIX]
967
def insert_record_stream(self, stream):
968
super(WeaveFile, self).insert_record_stream(stream)
1012
def join(self, other, pb=None, msg=None, version_ids=None,
1013
ignore_missing=False):
1014
"""Join other into self and save."""
1015
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
972
1019
def _reweave(wa, wb, pb=None, msg=None):
973
1020
"""Combine two weaves and return the result.
975
This works even if a revision R has different parents in
1022
This works even if a revision R has different parents in
976
1023
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
1025
This is done by just building up a new weave, maintaining ordering
979
1026
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
1027
might be possible but it should only be necessary to do
1028
this operation rarely, when a new previously ghost version is
984
1031
:param pb: An optional progress bar, indicating how far done we are
1029
1075
p = combined.setdefault(name, set())
1030
1076
p.update(map(weave._idx_to_name, weave._parents[idx]))
1031
1077
return combined
1081
"""Show the weave's table-of-contents"""
1082
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1083
for i in (6, 50, 10, 10):
1086
for i in range(w.num_versions()):
1089
parent_str = ' '.join(map(str, w._parents[i]))
1090
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1094
def weave_stats(weave_file, pb):
1095
from bzrlib.weavefile import read_weave
1097
wf = file(weave_file, 'rb')
1099
# FIXME: doesn't work on pipes
1100
weave_size = wf.tell()
1104
for i in range(vers):
1105
pb.update('checking sizes', i, vers)
1106
for origin, lineno, line in w._extract([i]):
1111
print 'versions %9d' % vers
1112
print 'weave file %9d bytes' % weave_size
1113
print 'total contents %9d bytes' % total
1114
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1117
print 'average size %9d bytes' % avg
1118
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1122
print """bzr weave tool
1124
Experimental tool for weave algorithm.
1127
weave init WEAVEFILE
1128
Create an empty weave file
1129
weave get WEAVEFILE VERSION
1130
Write out specified version.
1131
weave check WEAVEFILE
1132
Check consistency of all versions.
1134
Display table of contents.
1135
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1136
Add NEWTEXT, with specified parent versions.
1137
weave annotate WEAVEFILE VERSION
1138
Display origin of each line.
1139
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1140
Auto-merge two versions and display conflicts.
1141
weave diff WEAVEFILE VERSION1 VERSION2
1142
Show differences between two versions.
1146
% weave init foo.weave
1148
% weave add foo.weave ver0 < foo.txt
1151
(create updated version)
1153
% weave get foo.weave 0 | diff -u - foo.txt
1154
% weave add foo.weave ver1 0 < foo.txt
1157
% weave get foo.weave 0 > foo.txt (create forked version)
1159
% weave add foo.weave ver2 0 < foo.txt
1162
% weave merge foo.weave 1 2 > foo.txt (merge them)
1163
% vi foo.txt (resolve conflicts)
1164
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1176
# in case we're run directly from the subdirectory
1177
sys.path.append('..')
1179
from bzrlib.weavefile import write_weave, read_weave
1180
from bzrlib.progress import ProgressBar
1195
return read_weave(file(argv[2], 'rb'))
1201
# at the moment, based on everything in the file
1203
parents = map(int, argv[4:])
1204
lines = sys.stdin.readlines()
1205
ver = w.add(name, parents, lines)
1206
write_weave(w, file(argv[2], 'wb'))
1207
print 'added version %r %d' % (name, ver)
1210
if os.path.exists(fn):
1211
raise IOError("file exists")
1213
write_weave(w, file(fn, 'wb'))
1214
elif cmd == 'get': # get one version
1216
sys.stdout.writelines(w.get_iter(int(argv[3])))
1221
v1, v2 = map(int, argv[3:5])
1224
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1225
'%s version %d' % (fn, v1),
1226
'%s version %d' % (fn, v2))
1227
sys.stdout.writelines(diff_gen)
1229
elif cmd == 'annotate':
1231
# newline is added to all lines regardless; too hard to get
1232
# reasonable formatting otherwise
1234
for origin, text in w.annotate(int(argv[3])):
1235
text = text.rstrip('\r\n')
1237
print ' | %s' % (text)
1239
print '%5d | %s' % (origin, text)
1245
elif cmd == 'stats':
1246
weave_stats(argv[2], ProgressBar())
1248
elif cmd == 'check':
1253
print '%d versions ok' % w.num_versions()
1255
elif cmd == 'inclusions':
1257
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1259
elif cmd == 'parents':
1261
print ' '.join(map(str, w._parents[int(argv[3])]))
1263
elif cmd == 'plan-merge':
1264
# replaced by 'bzr weave-plan-merge'
1266
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1268
print '%14s | %s' % (state, line),
1269
elif cmd == 'merge':
1270
# replaced by 'bzr weave-merge-text'
1272
p = w.plan_merge(int(argv[3]), int(argv[4]))
1273
sys.stdout.writelines(w.weave_merge(p))
1275
raise ValueError('unknown command %r' % cmd)
1278
if __name__ == '__main__':
1280
sys.exit(main(sys.argv))
1283
class InterWeave(InterVersionedFile):
1284
"""Optimised code paths for weave to weave operations."""
1286
_matching_file_from_factory = staticmethod(WeaveFile)
1287
_matching_file_to_factory = staticmethod(WeaveFile)
1290
def is_compatible(source, target):
1291
"""Be compatible with weaves."""
1293
return (isinstance(source, Weave) and
1294
isinstance(target, Weave))
1295
except AttributeError:
1298
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1299
"""See InterVersionedFile.join."""
1300
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1301
if self.target.versions() == [] and version_ids is None:
1302
self.target._copy_weave_content(self.source)
1304
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1307
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
