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
258
__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
260
def get_parent_map(self, version_ids):
330
261
"""See VersionedFile.get_parent_map."""
332
263
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
265
result[version_id] = tuple(
266
map(self._idx_to_name, self._parents[self._lookup(version_id)]))
267
except RevisionNotPresent:
345
271
def get_parents_with_ghosts(self, version_id):
346
272
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
274
def _check_repeated_add(self, name, parents, text, sha1):
381
275
"""Check that a duplicated add is OK.
398
292
def _add(self, version_id, lines, parents, sha1=None, nostore_sha=None):
399
293
"""Add a single text on top of the weave.
401
295
Returns the index number of the newly added version.
404
298
Symbolic name for this version.
405
299
(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
302
List or set of direct parent version numbers.
412
305
Sequence of lines to be added in the new version.
414
307
:param nostore_sha: See VersionedFile.add_lines.
309
assert isinstance(version_id, basestring)
416
310
self._check_lines_not_unicode(lines)
417
311
self._check_lines_are_lines(lines)
419
313
sha1 = sha_strings(lines)
420
314
if sha1 == nostore_sha:
421
315
raise errors.ExistingContent
422
if version_id is None:
423
version_id = "sha1:" + sha1
424
316
if version_id in self._name_map:
425
317
return self._check_repeated_add(version_id, parents, lines, sha1)
557
460
def _compatible_parents(self, my_parents, other_parents):
558
461
"""During join check that other_parents are joinable with my_parents.
560
Joinable is defined as 'is a subset of' - supersets may require
463
Joinable is defined as 'is a subset of' - supersets may require
561
464
regeneration of diffs, but subsets do not.
563
466
return len(other_parents.difference(my_parents)) == 0
565
def annotate(self, version_id):
566
"""Return a list of (version-id, line) tuples for version_id.
468
def annotate_iter(self, version_id):
469
"""Yield list of (version-id, line) pairs for the specified version.
568
471
The index indicates when the line originated in the weave."""
569
472
incls = [self._lookup(version_id)]
570
return [(self._idx_to_name(origin), text) for origin, lineno, text in
571
self._extract(incls)]
473
for origin, lineno, text in self._extract(incls):
474
yield self._idx_to_name(origin), text
573
476
def iter_lines_added_or_present_in_versions(self, version_ids=None,
699
610
# we're still spending ~1/4 of the method in isinstance though.
700
611
# so lets hard code the acceptable string classes we expect:
701
612
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
702
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
613
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
704
615
# yay, down to ~1/4 the initial extract time, and our inline time
705
616
# has shrunk again, with isinstance no longer dominating.
706
617
# tweaking the stack inclusion test to use a set gives:
707
618
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
708
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
619
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
710
621
# - a 5% win, or possibly just noise. However with large istacks that
711
622
# 'in' test could dominate, so I'm leaving this change in place -
712
623
# when its fast enough to consider profiling big datasets we can review.
717
628
for l in self._weave:
718
629
if l.__class__ == tuple:
725
637
iset.remove(istack.pop())
727
639
if v in included:
730
644
if v in included:
733
raise AssertionError()
648
assert l.__class__ in (str, unicode)
735
649
if isactive is None:
736
650
isactive = (not dset) and istack and (istack[-1] in included)
848
762
# no lines outside of insertion blocks, that deletions are
849
763
# properly paired, etc.
765
def _join(self, other, pb, msg, version_ids, ignore_missing):
766
"""Worker routine for join()."""
767
if not other.versions():
768
return # nothing to update, easy
771
# versions is never none, InterWeave checks this.
774
# two loops so that we do not change ourselves before verifying it
776
# work through in index order to make sure we get all dependencies
779
# get the selected versions only that are in other.versions.
780
version_ids = set(other.versions()).intersection(set(version_ids))
781
# pull in the referenced graph.
782
version_ids = other.get_ancestry(version_ids)
783
pending_parents = other.get_parent_map(version_ids)
784
pending_graph = pending_parents.items()
785
if len(pending_graph) != len(version_ids):
786
raise RevisionNotPresent(
787
set(version_ids) - set(pending_parents.keys()), self)
788
for name in topo_sort(pending_graph):
789
other_idx = other._name_map[name]
790
# returns True if we have it, False if we need it.
791
if not self._check_version_consistent(other, other_idx, name):
792
names_to_join.append((other_idx, name))
800
for other_idx, name in names_to_join:
801
# TODO: If all the parents of the other version are already
802
# present then we can avoid some work by just taking the delta
803
# and adjusting the offsets.
804
new_parents = self._imported_parents(other, other_idx)
805
sha1 = other._sha1s[other_idx]
810
pb.update(msg, merged, len(names_to_join))
812
lines = other.get_lines(other_idx)
813
self._add(name, lines, new_parents, sha1)
815
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
816
merged, processed, self._weave_name, time.time()-time0))
851
818
def _imported_parents(self, other, other_idx):
852
819
"""Return list of parents in self corresponding to indexes in other."""
952
923
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)
925
self._transport.put_file(self._weave_name + WeaveFile.WEAVE_SUFFIX,
963
930
def get_suffixes():
964
931
"""See VersionedFile.get_suffixes()."""
965
932
return [WeaveFile.WEAVE_SUFFIX]
967
def insert_record_stream(self, stream):
968
super(WeaveFile, self).insert_record_stream(stream)
934
def join(self, other, pb=None, msg=None, version_ids=None,
935
ignore_missing=False):
936
"""Join other into self and save."""
937
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
972
941
def _reweave(wa, wb, pb=None, msg=None):
973
942
"""Combine two weaves and return the result.
975
This works even if a revision R has different parents in
944
This works even if a revision R has different parents in
976
945
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
947
This is done by just building up a new weave, maintaining ordering
979
948
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
949
might be possible but it should only be necessary to do
950
this operation rarely, when a new previously ghost version is
984
953
:param pb: An optional progress bar, indicating how far done we are
1029
997
p = combined.setdefault(name, set())
1030
998
p.update(map(weave._idx_to_name, weave._parents[idx]))
1003
"""Show the weave's table-of-contents"""
1004
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1005
for i in (6, 50, 10, 10):
1008
for i in range(w.num_versions()):
1011
parent_str = ' '.join(map(str, w._parents[i]))
1012
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1016
def weave_stats(weave_file, pb):
1017
from bzrlib.weavefile import read_weave
1019
wf = file(weave_file, 'rb')
1021
# FIXME: doesn't work on pipes
1022
weave_size = wf.tell()
1026
for i in range(vers):
1027
pb.update('checking sizes', i, vers)
1028
for origin, lineno, line in w._extract([i]):
1033
print 'versions %9d' % vers
1034
print 'weave file %9d bytes' % weave_size
1035
print 'total contents %9d bytes' % total
1036
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1039
print 'average size %9d bytes' % avg
1040
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1044
print """bzr weave tool
1046
Experimental tool for weave algorithm.
1049
weave init WEAVEFILE
1050
Create an empty weave file
1051
weave get WEAVEFILE VERSION
1052
Write out specified version.
1053
weave check WEAVEFILE
1054
Check consistency of all versions.
1056
Display table of contents.
1057
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1058
Add NEWTEXT, with specified parent versions.
1059
weave annotate WEAVEFILE VERSION
1060
Display origin of each line.
1061
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1062
Auto-merge two versions and display conflicts.
1063
weave diff WEAVEFILE VERSION1 VERSION2
1064
Show differences between two versions.
1068
% weave init foo.weave
1070
% weave add foo.weave ver0 < foo.txt
1073
(create updated version)
1075
% weave get foo.weave 0 | diff -u - foo.txt
1076
% weave add foo.weave ver1 0 < foo.txt
1079
% weave get foo.weave 0 > foo.txt (create forked version)
1081
% weave add foo.weave ver2 0 < foo.txt
1084
% weave merge foo.weave 1 2 > foo.txt (merge them)
1085
% vi foo.txt (resolve conflicts)
1086
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1098
# in case we're run directly from the subdirectory
1099
sys.path.append('..')
1101
from bzrlib.weavefile import write_weave, read_weave
1102
from bzrlib.progress import ProgressBar
1117
return read_weave(file(argv[2], 'rb'))
1123
# at the moment, based on everything in the file
1125
parents = map(int, argv[4:])
1126
lines = sys.stdin.readlines()
1127
ver = w.add(name, parents, lines)
1128
write_weave(w, file(argv[2], 'wb'))
1129
print 'added version %r %d' % (name, ver)
1132
if os.path.exists(fn):
1133
raise IOError("file exists")
1135
write_weave(w, file(fn, 'wb'))
1136
elif cmd == 'get': # get one version
1138
sys.stdout.writelines(w.get_iter(int(argv[3])))
1143
v1, v2 = map(int, argv[3:5])
1146
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1147
'%s version %d' % (fn, v1),
1148
'%s version %d' % (fn, v2))
1149
sys.stdout.writelines(diff_gen)
1151
elif cmd == 'annotate':
1153
# newline is added to all lines regardless; too hard to get
1154
# reasonable formatting otherwise
1156
for origin, text in w.annotate(int(argv[3])):
1157
text = text.rstrip('\r\n')
1159
print ' | %s' % (text)
1161
print '%5d | %s' % (origin, text)
1167
elif cmd == 'stats':
1168
weave_stats(argv[2], ProgressBar())
1170
elif cmd == 'check':
1175
print '%d versions ok' % w.num_versions()
1177
elif cmd == 'inclusions':
1179
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1181
elif cmd == 'parents':
1183
print ' '.join(map(str, w._parents[int(argv[3])]))
1185
elif cmd == 'plan-merge':
1186
# replaced by 'bzr weave-plan-merge'
1188
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1190
print '%14s | %s' % (state, line),
1191
elif cmd == 'merge':
1192
# replaced by 'bzr weave-merge-text'
1194
p = w.plan_merge(int(argv[3]), int(argv[4]))
1195
sys.stdout.writelines(w.weave_merge(p))
1197
raise ValueError('unknown command %r' % cmd)
1200
if __name__ == '__main__':
1202
sys.exit(main(sys.argv))
1205
class InterWeave(InterVersionedFile):
1206
"""Optimised code paths for weave to weave operations."""
1208
_matching_file_from_factory = staticmethod(WeaveFile)
1209
_matching_file_to_factory = staticmethod(WeaveFile)
1212
def is_compatible(source, target):
1213
"""Be compatible with weaves."""
1215
return (isinstance(source, Weave) and
1216
isinstance(target, Weave))
1217
except AttributeError:
1220
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1221
"""See InterVersionedFile.join."""
1222
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1223
if self.target.versions() == [] and version_ids is None:
1224
self.target._copy_weave_content(self.source)
1226
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1229
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
