3
# Copyright (C) 2005 Canonical Ltd
5
# This program is free software; you can redistribute it and/or modify
6
# it under the terms of the GNU General Public License as published by
7
# the Free Software Foundation; either version 2 of the License, or
8
# (at your option) any later version.
10
# This program is distributed in the hope that it will be useful,
11
# but WITHOUT ANY WARRANTY; without even the implied warranty of
12
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13
# GNU General Public License for more details.
15
# You should have received a copy of the GNU General Public License
16
# along with this program; if not, write to the Free Software
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
19
# Author: Martin Pool <mbp@canonical.com>
22
"""Weave - storage of related text file versions"""
25
# XXX: If we do weaves this way, will a merge still behave the same
26
# way if it's done in a different order? That's a pretty desirable
29
# TODO: Nothing here so far assumes the lines are really \n newlines,
30
# rather than being split up in some other way. We could accomodate
31
# binaries, perhaps by naively splitting on \n or perhaps using
32
# something like a rolling checksum.
34
# TODO: End marker for each version so we can stop reading?
36
# TODO: Check that no insertion occurs inside a deletion that was
37
# active in the version of the insertion.
39
# TODO: In addition to the SHA-1 check, perhaps have some code that
40
# checks structural constraints of the weave: ie that insertions are
41
# properly nested, that there is no text outside of an insertion, that
42
# insertions or deletions are not repeated, etc.
44
# TODO: Parallel-extract that passes back each line along with a
45
# description of which revisions include it. Nice for checking all
46
# shas or calculating stats in parallel.
48
# TODO: Using a single _extract routine and then processing the output
49
# is probably inefficient. It's simple enough that we can afford to
50
# have slight specializations for different ways its used: annotate,
51
# basis for add, get, etc.
53
# TODO: Probably the API should work only in names to hide the integer
54
# indexes from the user.
56
# TODO: Is there any potential performance win by having an add()
57
# variant that is passed a pre-cooked version of the single basis
60
# TODO: Reweave can possibly be made faster by remembering diffs
61
# where the basis and destination are unchanged.
63
# 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
66
# be done fairly efficiently because the sequence numbers constrain
67
# the possible relationships.
71
from cStringIO import StringIO
72
from difflib import SequenceMatcher
77
from bzrlib.trace import mutter
78
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
79
RevisionAlreadyPresent,
81
WeaveRevisionAlreadyPresent,
82
WeaveRevisionNotPresent,
84
import bzrlib.errors as errors
85
from bzrlib.osutils import sha_strings
86
from bzrlib.symbol_versioning import *
87
from bzrlib.tsort import topo_sort
88
from bzrlib.versionedfile import VersionedFile, InterVersionedFile
89
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
92
class Weave(VersionedFile):
93
"""weave - versioned text file storage.
95
A Weave manages versions of line-based text files, keeping track
96
of the originating version for each line.
98
To clients the "lines" of the file are represented as a list of strings.
99
These strings will typically have terminal newline characters, but
100
this is not required. In particular files commonly do not have a newline
101
at the end of the file.
103
Texts can be identified in either of two ways:
105
* a nonnegative index number.
107
* a version-id string.
109
Typically the index number will be valid only inside this weave and
110
the version-id is used to reference it in the larger world.
112
The weave is represented as a list mixing edit instructions and
113
literal text. Each entry in _weave can be either a string (or
114
unicode), or a tuple. If a string, it means that the given line
115
should be output in the currently active revisions.
117
If a tuple, it gives a processing instruction saying in which
118
revisions the enclosed lines are active. The tuple has the form
119
(instruction, version).
121
The instruction can be '{' or '}' for an insertion block, and '['
122
and ']' for a deletion block respectively. The version is the
123
integer version index. There is no replace operator, only deletes
124
and inserts. For '}', the end of an insertion, there is no
125
version parameter because it always closes the most recently
130
* A later version can delete lines that were introduced by any
131
number of ancestor versions; this implies that deletion
132
instructions can span insertion blocks without regard to the
133
insertion block's nesting.
135
* Similarly, deletions need not be properly nested with regard to
136
each other, because they might have been generated by
137
independent revisions.
139
* Insertions are always made by inserting a new bracketed block
140
into a single point in the previous weave. This implies they
141
can nest but not overlap, and the nesting must always have later
142
insertions on the inside.
144
* It doesn't seem very useful to have an active insertion
145
inside an inactive insertion, but it might happen.
147
* Therefore, all instructions are always"considered"; that
148
is passed onto and off the stack. An outer inactive block
149
doesn't disable an inner block.
151
* Lines are enabled if the most recent enclosing insertion is
152
active and none of the enclosing deletions are active.
154
* There is no point having a deletion directly inside its own
155
insertion; you might as well just not write it. And there
156
should be no way to get an earlier version deleting a later
160
Text of the weave; list of control instruction tuples and strings.
163
List of parents, indexed by version number.
164
It is only necessary to store the minimal set of parents for
165
each version; the parent's parents are implied.
168
List of hex SHA-1 of each version.
171
List of symbolic names for each version. Each should be unique.
174
For each name, the version number.
177
Descriptive name of this weave; typically the filename if known.
181
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
184
def __init__(self, weave_name=None, access_mode='w'):
185
super(Weave, self).__init__(access_mode)
191
self._weave_name = weave_name
194
return "Weave(%r)" % self._weave_name
197
"""Return a deep copy of self.
199
The copy can be modified without affecting the original weave."""
201
other._weave = self._weave[:]
202
other._parents = self._parents[:]
203
other._sha1s = self._sha1s[:]
204
other._names = self._names[:]
205
other._name_map = self._name_map.copy()
206
other._weave_name = self._weave_name
209
def __eq__(self, other):
210
if not isinstance(other, Weave):
212
return self._parents == other._parents \
213
and self._weave == other._weave \
214
and self._sha1s == other._sha1s
216
def __ne__(self, other):
217
return not self.__eq__(other)
219
@deprecated_method(zero_eight)
220
def idx_to_name(self, index):
221
"""Old public interface, the public interface is all names now."""
224
def _idx_to_name(self, version):
225
return self._names[version]
227
@deprecated_method(zero_eight)
228
def lookup(self, name):
229
"""Backwards compatability thunk:
231
Return name, as name is valid in the api now, and spew deprecation
236
def _lookup(self, name):
237
"""Convert symbolic version name to index."""
239
return self._name_map[name]
241
raise RevisionNotPresent(name, self._weave_name)
243
@deprecated_method(zero_eight)
244
def iter_names(self):
245
"""Deprecated convenience function, please see VersionedFile.names()."""
246
return iter(self.names())
248
@deprecated_method(zero_eight)
250
"""See Weave.versions for the current api."""
251
return self.versions()
254
"""See VersionedFile.versions."""
255
return self._names[:]
257
def has_version(self, version_id):
258
"""See VersionedFile.has_version."""
259
return self._name_map.has_key(version_id)
261
__contains__ = has_version
263
def get_delta(self, version_id):
264
"""See VersionedFile.get_delta."""
265
return self.get_deltas([version_id])[version_id]
267
def get_deltas(self, version_ids):
268
"""See VersionedFile.get_deltas."""
269
version_ids = self.get_ancestry(version_ids)
270
for version_id in version_ids:
271
if not self.has_version(version_id):
272
raise RevisionNotPresent(version_id, self)
273
# try extracting all versions; parallel extraction is used
274
nv = self.num_versions()
280
last_parent_lines = {}
282
parent_inclusions = {}
287
# its simplest to generate a full set of prepared variables.
289
name = self._names[i]
290
sha1s[name] = self.get_sha1(name)
291
parents_list = self.get_parents(name)
293
parent = parents_list[0]
294
parents[name] = parent
295
parent_inclusions[name] = inclusions[parent]
298
parent_inclusions[name] = set()
299
# we want to emit start, finish, replacement_length, replacement_lines tuples.
300
diff_hunks[name] = []
301
current_hunks[name] = [0, 0, 0, []] # #start, finish, repl_length, repl_tuples
302
parent_linenums[name] = 0
304
parent_noeols[name] = False
305
last_parent_lines[name] = None
306
new_inc = set([name])
307
for p in self._parents[i]:
308
new_inc.update(inclusions[self._idx_to_name(p)])
309
# debug only, known good so far.
310
#assert set(new_inc) == set(self.get_ancestry(name)), \
311
# 'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
312
inclusions[name] = new_inc
314
nlines = len(self._weave)
316
for lineno, inserted, deletes, line in self._walk_internal():
317
# a line is active in a version if:
318
# insert is in the versions inclusions
320
# deleteset & the versions inclusions is an empty set.
321
# so - if we have a included by mapping - version is included by
322
# children, we get a list of children to examine for deletes affect
323
# ing them, which is less than the entire set of children.
324
for version_id in version_ids:
325
# The active inclusion must be an ancestor,
326
# and no ancestors must have deleted this line,
327
# because we don't support resurrection.
328
parent_inclusion = parent_inclusions[version_id]
329
inclusion = inclusions[version_id]
330
parent_active = inserted in parent_inclusion and not (deletes & parent_inclusion)
331
version_active = inserted in inclusion and not (deletes & inclusion)
332
if not parent_active and not version_active:
333
# unrelated line of ancestry
335
elif parent_active and version_active:
337
parent_linenum = parent_linenums[version_id]
338
if current_hunks[version_id] != [parent_linenum, parent_linenum, 0, []]:
339
diff_hunks[version_id].append(tuple(current_hunks[version_id]))
341
current_hunks[version_id] = [parent_linenum, parent_linenum, 0, []]
342
parent_linenums[version_id] = parent_linenum
345
noeols[version_id] = True
348
elif parent_active and not version_active:
350
current_hunks[version_id][1] += 1
351
parent_linenums[version_id] += 1
352
last_parent_lines[version_id] = line
353
elif not parent_active and version_active:
355
# noeol only occurs at the end of a file because we
356
# diff linewise. We want to show noeol changes as a
357
# empty diff unless the actual eol-less content changed.
360
if last_parent_lines[version_id][-1] != '\n':
361
parent_noeols[version_id] = True
362
except (TypeError, IndexError):
365
if theline[-1] != '\n':
366
noeols[version_id] = True
370
parent_should_go = False
372
if parent_noeols[version_id] == noeols[version_id]:
373
# no noeol toggle, so trust the weaves statement
374
# that this line is changed.
376
if parent_noeols[version_id]:
377
theline = theline + '\n'
378
elif parent_noeols[version_id]:
379
# parent has no eol, we do:
380
# our line is new, report as such..
382
elif noeols[version_id]:
383
# append a eol so that it looks like
385
theline = theline + '\n'
386
if parents[version_id] is not None:
387
#if last_parent_lines[version_id] is not None:
388
parent_should_go = True
389
if last_parent_lines[version_id] != theline:
392
#parent_should_go = False
394
current_hunks[version_id][2] += 1
395
current_hunks[version_id][3].append((inserted, theline))
397
# last hunk last parent line is not eaten
398
current_hunks[version_id][1] -= 1
399
if current_hunks[version_id][1] < 0:
400
current_hunks[version_id][1] = 0
401
# import pdb;pdb.set_trace()
402
# assert current_hunks[version_id][1] >= 0
406
version = self._idx_to_name(i)
407
if current_hunks[version] != [0, 0, 0, []]:
408
diff_hunks[version].append(tuple(current_hunks[version]))
410
for version_id in version_ids:
411
result[version_id] = (
415
diff_hunks[version_id],
419
def get_parents(self, version_id):
420
"""See VersionedFile.get_parent."""
421
return map(self._idx_to_name, self._parents[self._lookup(version_id)])
423
def _check_repeated_add(self, name, parents, text, sha1):
424
"""Check that a duplicated add is OK.
426
If it is, return the (old) index; otherwise raise an exception.
428
idx = self._lookup(name)
429
if sorted(self._parents[idx]) != sorted(parents) \
430
or sha1 != self._sha1s[idx]:
431
raise RevisionAlreadyPresent(name, self._weave_name)
434
@deprecated_method(zero_eight)
435
def add_identical(self, old_rev_id, new_rev_id, parents):
436
"""Please use Weave.clone_text now."""
437
return self.clone_text(new_rev_id, old_rev_id, parents)
439
def _add_lines(self, version_id, parents, lines, parent_texts):
440
"""See VersionedFile.add_lines."""
441
return self._add(version_id, lines, map(self._lookup, parents))
443
@deprecated_method(zero_eight)
444
def add(self, name, parents, text, sha1=None):
445
"""See VersionedFile.add_lines for the non deprecated api."""
446
return self._add(name, text, map(self._maybe_lookup, parents), sha1)
448
def _add(self, version_id, lines, parents, sha1=None):
449
"""Add a single text on top of the weave.
451
Returns the index number of the newly added version.
454
Symbolic name for this version.
455
(Typically the revision-id of the revision that added it.)
458
List or set of direct parent version numbers.
461
Sequence of lines to be added in the new version.
464
assert isinstance(version_id, basestring)
466
sha1 = sha_strings(lines)
467
if version_id in self._name_map:
468
return self._check_repeated_add(version_id, parents, lines, sha1)
470
self._check_versions(parents)
471
## self._check_lines(lines)
472
new_version = len(self._parents)
474
# if we abort after here the (in-memory) weave will be corrupt because only
475
# some fields are updated
476
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
477
# - Robert Collins 20060226
478
self._parents.append(parents[:])
479
self._sha1s.append(sha1)
480
self._names.append(version_id)
481
self._name_map[version_id] = new_version
485
# special case; adding with no parents revision; can do
486
# this more quickly by just appending unconditionally.
487
# even more specially, if we're adding an empty text we
488
# need do nothing at all.
490
self._weave.append(('{', new_version))
491
self._weave.extend(lines)
492
self._weave.append(('}', None))
495
if len(parents) == 1:
496
pv = list(parents)[0]
497
if sha1 == self._sha1s[pv]:
498
# special case: same as the single parent
502
ancestors = self._inclusions(parents)
506
# basis a list of (origin, lineno, line)
509
for origin, lineno, line in self._extract(ancestors):
510
basis_lineno.append(lineno)
511
basis_lines.append(line)
513
# another small special case: a merge, producing the same text
515
if lines == basis_lines:
518
# add a sentinal, because we can also match against the final line
519
basis_lineno.append(len(self._weave))
521
# XXX: which line of the weave should we really consider
522
# matches the end of the file? the current code says it's the
523
# last line of the weave?
525
#print 'basis_lines:', basis_lines
526
#print 'new_lines: ', lines
528
s = SequenceMatcher(None, basis_lines, lines)
530
# offset gives the number of lines that have been inserted
531
# into the weave up to the current point; if the original edit instruction
532
# says to change line A then we actually change (A+offset)
535
for tag, i1, i2, j1, j2 in s.get_opcodes():
536
# i1,i2 are given in offsets within basis_lines; we need to map them
537
# back to offsets within the entire weave
538
#print 'raw match', tag, i1, i2, j1, j2
542
i1 = basis_lineno[i1]
543
i2 = basis_lineno[i2]
545
assert 0 <= j1 <= j2 <= len(lines)
547
#print tag, i1, i2, j1, j2
549
# the deletion and insertion are handled separately.
550
# first delete the region.
552
self._weave.insert(i1+offset, ('[', new_version))
553
self._weave.insert(i2+offset+1, (']', new_version))
557
# there may have been a deletion spanning up to
558
# i2; we want to insert after this region to make sure
559
# we don't destroy ourselves
561
self._weave[i:i] = ([('{', new_version)]
564
offset += 2 + (j2 - j1)
567
def _clone_text(self, new_version_id, old_version_id, parents):
568
"""See VersionedFile.clone_text."""
569
old_lines = self.get_text(old_version_id)
570
self.add_lines(new_version_id, parents, old_lines)
572
def _inclusions(self, versions):
573
"""Return set of all ancestors of given version(s)."""
574
if not len(versions):
577
for v in xrange(max(versions), 0, -1):
579
# include all its parents
580
i.update(self._parents[v])
582
## except IndexError:
583
## raise ValueError("version %d not present in weave" % v)
585
@deprecated_method(zero_eight)
586
def inclusions(self, version_ids):
587
"""Deprecated - see VersionedFile.get_ancestry for the replacement."""
590
if isinstance(version_ids[0], int):
591
return [self._idx_to_name(v) for v in self._inclusions(version_ids)]
593
return self.get_ancestry(version_ids)
595
def get_ancestry(self, version_ids):
596
"""See VersionedFile.get_ancestry."""
597
if isinstance(version_ids, basestring):
598
version_ids = [version_ids]
599
i = self._inclusions([self._lookup(v) for v in version_ids])
600
return [self._idx_to_name(v) for v in i]
602
def _check_lines(self, text):
603
if not isinstance(text, list):
604
raise ValueError("text should be a list, not %s" % type(text))
607
if not isinstance(l, basestring):
608
raise ValueError("text line should be a string or unicode, not %s"
613
def _check_versions(self, indexes):
614
"""Check everything in the sequence of indexes is valid"""
619
raise IndexError("invalid version number %r" % i)
621
def _compatible_parents(self, my_parents, other_parents):
622
"""During join check that other_parents are joinable with my_parents.
624
Joinable is defined as 'is a subset of' - supersets may require
625
regeneration of diffs, but subsets do not.
627
return len(other_parents.difference(my_parents)) == 0
629
def annotate(self, version_id):
630
if isinstance(version_id, int):
631
warn('Weave.annotate(int) is deprecated. Please use version names'
632
' in all circumstances as of 0.8',
637
for origin, lineno, text in self._extract([version_id]):
638
result.append((origin, text))
641
return super(Weave, self).annotate(version_id)
643
def annotate_iter(self, version_id):
644
"""Yield list of (version-id, line) pairs for the specified version.
646
The index indicates when the line originated in the weave."""
647
incls = [self._lookup(version_id)]
648
for origin, lineno, text in self._extract(incls):
649
yield self._idx_to_name(origin), text
651
@deprecated_method(zero_eight)
653
"""_walk has become visit, a supported api."""
654
return self._walk_internal()
656
def iter_lines_added_or_present_in_versions(self, version_ids=None):
657
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
658
if version_ids is None:
659
version_ids = self.versions()
660
version_ids = set(version_ids)
661
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
662
# if inserted not in version_ids then it was inserted before the
663
# versions we care about, but because weaves cannot represent ghosts
664
# properly, we do not filter down to that
665
# if inserted not in version_ids: continue
671
#@deprecated_method(zero_eight)
672
def walk(self, version_ids=None):
673
"""See VersionedFile.walk."""
674
return self._walk_internal(version_ids)
676
def _walk_internal(self, version_ids=None):
677
"""Helper method for weave actions."""
682
lineno = 0 # line of weave, 0-based
684
for l in self._weave:
685
if l.__class__ == tuple:
689
istack.append(self._names[v])
693
assert self._names[v] not in dset
694
dset.add(self._names[v])
696
dset.remove(self._names[v])
698
raise WeaveFormatError('unexpected instruction %r' % v)
700
assert l.__class__ in (str, unicode)
702
yield lineno, istack[-1], frozenset(dset), l
706
raise WeaveFormatError("unclosed insertion blocks "
707
"at end of weave: %s" % istack)
709
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
712
def _extract(self, versions):
713
"""Yield annotation of lines in included set.
715
Yields a sequence of tuples (origin, lineno, text), where
716
origin is the origin version, lineno the index in the weave,
717
and text the text of the line.
719
The set typically but not necessarily corresponds to a version.
722
if not isinstance(i, int):
725
included = self._inclusions(versions)
731
lineno = 0 # line of weave, 0-based
737
WFE = WeaveFormatError
740
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
741
# +285282 0 1676.8040 1676.8040 +<isinstance>
742
# 1.6 seconds in 'isinstance'.
743
# changing the first isinstance:
744
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
745
# +140414 0 762.8050 762.8050 +<isinstance>
746
# note that the inline time actually dropped (less function calls)
747
# and total processing time was halved.
748
# we're still spending ~1/4 of the method in isinstance though.
749
# so lets hard code the acceptable string classes we expect:
750
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
751
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
753
# yay, down to ~1/4 the initial extract time, and our inline time
754
# has shrunk again, with isinstance no longer dominating.
755
# tweaking the stack inclusion test to use a set gives:
756
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
757
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
759
# - a 5% win, or possibly just noise. However with large istacks that
760
# 'in' test could dominate, so I'm leaving this change in place -
761
# when its fast enough to consider profiling big datasets we can review.
766
for l in self._weave:
767
if l.__class__ == tuple:
775
iset.remove(istack.pop())
786
assert l.__class__ in (str, unicode)
788
isactive = (not dset) and istack and (istack[-1] in included)
790
result.append((istack[-1], lineno, l))
793
raise WeaveFormatError("unclosed insertion blocks "
794
"at end of weave: %s" % istack)
796
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
800
@deprecated_method(zero_eight)
801
def get_iter(self, name_or_index):
802
"""Deprecated, please do not use. Lookups are not not needed.
804
Please use get_lines now.
806
return iter(self.get_lines(self._maybe_lookup(name_or_index)))
808
@deprecated_method(zero_eight)
809
def maybe_lookup(self, name_or_index):
810
"""Deprecated, please do not use. Lookups are not not needed."""
811
return self._maybe_lookup(name_or_index)
813
def _maybe_lookup(self, name_or_index):
814
"""Convert possible symbolic name to index, or pass through indexes.
818
if isinstance(name_or_index, (int, long)):
821
return self._lookup(name_or_index)
823
@deprecated_method(zero_eight)
824
def get(self, version_id):
825
"""Please use either Weave.get_text or Weave.get_lines as desired."""
826
return self.get_lines(version_id)
828
def get_lines(self, version_id):
829
"""See VersionedFile.get_lines()."""
830
int_index = self._maybe_lookup(version_id)
831
result = [line for (origin, lineno, line) in self._extract([int_index])]
832
expected_sha1 = self._sha1s[int_index]
833
measured_sha1 = sha_strings(result)
834
if measured_sha1 != expected_sha1:
835
raise errors.WeaveInvalidChecksum(
836
'file %s, revision %s, expected: %s, measured %s'
837
% (self._weave_name, version_id,
838
expected_sha1, measured_sha1))
841
def get_sha1(self, name):
842
"""Get the stored sha1 sum for the given revision.
844
:param name: The name of the version to lookup
846
return self._sha1s[self._lookup(name)]
848
@deprecated_method(zero_eight)
849
def numversions(self):
850
"""How many versions are in this weave?
852
Deprecated in favour of num_versions.
854
return self.num_versions()
856
def num_versions(self):
857
"""How many versions are in this weave?"""
858
l = len(self._parents)
859
assert l == len(self._sha1s)
862
__len__ = num_versions
864
def check(self, progress_bar=None):
865
# TODO evaluate performance hit of using string sets in this routine.
866
# TODO: check no circular inclusions
867
# TODO: create a nested progress bar
868
for version in range(self.num_versions()):
869
inclusions = list(self._parents[version])
872
if inclusions[-1] >= version:
873
raise WeaveFormatError("invalid included version %d for index %d"
874
% (inclusions[-1], version))
876
# try extracting all versions; parallel extraction is used
877
nv = self.num_versions()
882
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
883
# The problem is that set membership is much more expensive
884
name = self._idx_to_name(i)
885
sha1s[name] = sha.new()
887
new_inc = set([name])
888
for p in self._parents[i]:
889
new_inc.update(inclusions[self._idx_to_name(p)])
891
assert set(new_inc) == set(self.get_ancestry(name)), \
892
'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
893
inclusions[name] = new_inc
895
nlines = len(self._weave)
897
update_text = 'checking weave'
899
short_name = os.path.basename(self._weave_name)
900
update_text = 'checking %s' % (short_name,)
901
update_text = update_text[:25]
903
for lineno, insert, deleteset, line in self._walk_internal():
905
progress_bar.update(update_text, lineno, nlines)
907
for name, name_inclusions in inclusions.items():
908
# The active inclusion must be an ancestor,
909
# and no ancestors must have deleted this line,
910
# because we don't support resurrection.
911
if (insert in name_inclusions) and not (deleteset & name_inclusions):
912
sha1s[name].update(line)
915
version = self._idx_to_name(i)
916
hd = sha1s[version].hexdigest()
917
expected = self._sha1s[i]
919
raise errors.WeaveInvalidChecksum(
920
"mismatched sha1 for version %s: "
921
"got %s, expected %s"
922
% (version, hd, expected))
924
# TODO: check insertions are properly nested, that there are
925
# no lines outside of insertion blocks, that deletions are
926
# properly paired, etc.
928
def _join(self, other, pb, msg, version_ids, ignore_missing):
929
"""Worker routine for join()."""
930
if not other.versions():
931
return # nothing to update, easy
934
for version_id in version_ids:
935
if not other.has_version(version_id) and not ignore_missing:
936
raise RevisionNotPresent(version_id, self._weave_name)
938
version_ids = other.versions()
940
# two loops so that we do not change ourselves before verifying it
942
# work through in index order to make sure we get all dependencies
945
# get the selected versions only that are in other.versions.
946
version_ids = set(other.versions()).intersection(set(version_ids))
947
# pull in the referenced graph.
948
version_ids = other.get_ancestry(version_ids)
949
pending_graph = [(version, other.get_parents(version)) for
950
version in version_ids]
951
for name in topo_sort(pending_graph):
952
other_idx = other._name_map[name]
953
# returns True if we have it, False if we need it.
954
if not self._check_version_consistent(other, other_idx, name):
955
names_to_join.append((other_idx, name))
964
for other_idx, name in names_to_join:
965
# TODO: If all the parents of the other version are already
966
# present then we can avoid some work by just taking the delta
967
# and adjusting the offsets.
968
new_parents = self._imported_parents(other, other_idx)
969
sha1 = other._sha1s[other_idx]
974
pb.update(msg, merged, len(names_to_join))
976
lines = other.get_lines(other_idx)
977
self._add(name, lines, new_parents, sha1)
979
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
980
merged, processed, self._weave_name, time.time()-time0))
982
def _imported_parents(self, other, other_idx):
983
"""Return list of parents in self corresponding to indexes in other."""
985
for parent_idx in other._parents[other_idx]:
986
parent_name = other._names[parent_idx]
987
if parent_name not in self._name_map:
988
# should not be possible
989
raise WeaveError("missing parent {%s} of {%s} in %r"
990
% (parent_name, other._name_map[other_idx], self))
991
new_parents.append(self._name_map[parent_name])
994
def _check_version_consistent(self, other, other_idx, name):
995
"""Check if a version in consistent in this and other.
997
To be consistent it must have:
1000
* the same direct parents (by name, not index, and disregarding
1003
If present & correct return True;
1004
if not present in self return False;
1005
if inconsistent raise error."""
1006
this_idx = self._name_map.get(name, -1)
1008
if self._sha1s[this_idx] != other._sha1s[other_idx]:
1009
raise errors.WeaveTextDiffers(name, self, other)
1010
self_parents = self._parents[this_idx]
1011
other_parents = other._parents[other_idx]
1012
n1 = set([self._names[i] for i in self_parents])
1013
n2 = set([other._names[i] for i in other_parents])
1014
if not self._compatible_parents(n1, n2):
1015
raise WeaveParentMismatch("inconsistent parents "
1016
"for version {%s}: %s vs %s" % (name, n1, n2))
1022
@deprecated_method(zero_eight)
1023
def reweave(self, other, pb=None, msg=None):
1024
"""reweave has been superceded by plain use of join."""
1025
return self.join(other, pb, msg)
1027
def _reweave(self, other, pb, msg):
1028
"""Reweave self with other - internal helper for join().
1030
:param other: The other weave to merge
1031
:param pb: An optional progress bar, indicating how far done we are
1032
:param msg: An optional message for the progress
1034
new_weave = _reweave(self, other, pb=pb, msg=msg)
1035
self._copy_weave_content(new_weave)
1037
def _copy_weave_content(self, otherweave):
1038
"""adsorb the content from otherweave."""
1039
for attr in self.__slots__:
1040
if attr != '_weave_name':
1041
setattr(self, attr, copy(getattr(otherweave, attr)))
1044
class WeaveFile(Weave):
1045
"""A WeaveFile represents a Weave on disk and writes on change."""
1047
WEAVE_SUFFIX = '.weave'
1049
def __init__(self, name, transport, filemode=None, create=False, access_mode='w'):
1050
"""Create a WeaveFile.
1052
:param create: If not True, only open an existing knit.
1054
super(WeaveFile, self).__init__(name, access_mode)
1055
self._transport = transport
1056
self._filemode = filemode
1058
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
1059
except errors.NoSuchFile:
1065
def _add_lines(self, version_id, parents, lines, parent_texts):
1066
"""Add a version and save the weave."""
1067
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
1072
def _clone_text(self, new_version_id, old_version_id, parents):
1073
"""See VersionedFile.clone_text."""
1074
super(WeaveFile, self)._clone_text(new_version_id, old_version_id, parents)
1077
def copy_to(self, name, transport):
1078
"""See VersionedFile.copy_to()."""
1079
# as we are all in memory always, just serialise to the new place.
1081
write_weave_v5(self, sio)
1083
transport.put(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
1085
def create_empty(self, name, transport, filemode=None):
1086
return WeaveFile(name, transport, filemode, create=True)
1089
"""Save the weave."""
1090
self._check_write_ok()
1092
write_weave_v5(self, sio)
1094
self._transport.put(self._weave_name + WeaveFile.WEAVE_SUFFIX,
1100
"""See VersionedFile.get_suffixes()."""
1101
return [WeaveFile.WEAVE_SUFFIX]
1103
def join(self, other, pb=None, msg=None, version_ids=None,
1104
ignore_missing=False):
1105
"""Join other into self and save."""
1106
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1110
@deprecated_function(zero_eight)
1111
def reweave(wa, wb, pb=None, msg=None):
1112
"""reweaving is deprecation, please just use weave.join()."""
1113
_reweave(wa, wb, pb, msg)
1115
def _reweave(wa, wb, pb=None, msg=None):
1116
"""Combine two weaves and return the result.
1118
This works even if a revision R has different parents in
1119
wa and wb. In the resulting weave all the parents are given.
1121
This is done by just building up a new weave, maintaining ordering
1122
of the versions in the two inputs. More efficient approaches
1123
might be possible but it should only be necessary to do
1124
this operation rarely, when a new previously ghost version is
1127
:param pb: An optional progress bar, indicating how far done we are
1128
:param msg: An optional message for the progress
1132
queue_a = range(wa.num_versions())
1133
queue_b = range(wb.num_versions())
1134
# first determine combined parents of all versions
1135
# map from version name -> all parent names
1136
combined_parents = _reweave_parent_graphs(wa, wb)
1137
mutter("combined parents: %r", combined_parents)
1138
order = topo_sort(combined_parents.iteritems())
1139
mutter("order to reweave: %r", order)
1144
for idx, name in enumerate(order):
1146
pb.update(msg, idx, len(order))
1147
if name in wa._name_map:
1148
lines = wa.get_lines(name)
1149
if name in wb._name_map:
1150
lines_b = wb.get_lines(name)
1151
if lines != lines_b:
1152
mutter('Weaves differ on content. rev_id {%s}', name)
1153
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1155
lines = list(difflib.unified_diff(lines, lines_b,
1156
wa._weave_name, wb._weave_name))
1157
mutter('lines:\n%s', ''.join(lines))
1158
raise errors.WeaveTextDiffers(name, wa, wb)
1160
lines = wb.get_lines(name)
1161
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1164
def _reweave_parent_graphs(wa, wb):
1165
"""Return combined parent ancestry for two weaves.
1167
Returned as a list of (version_name, set(parent_names))"""
1169
for weave in [wa, wb]:
1170
for idx, name in enumerate(weave._names):
1171
p = combined.setdefault(name, set())
1172
p.update(map(weave._idx_to_name, weave._parents[idx]))
1177
"""Show the weave's table-of-contents"""
1178
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1179
for i in (6, 50, 10, 10):
1182
for i in range(w.num_versions()):
1185
parent_str = ' '.join(map(str, w._parents[i]))
1186
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1190
def weave_stats(weave_file, pb):
1191
from bzrlib.weavefile import read_weave
1193
wf = file(weave_file, 'rb')
1194
w = read_weave(wf, WeaveVersionedFile)
1195
# FIXME: doesn't work on pipes
1196
weave_size = wf.tell()
1200
for i in range(vers):
1201
pb.update('checking sizes', i, vers)
1202
for origin, lineno, line in w._extract([i]):
1207
print 'versions %9d' % vers
1208
print 'weave file %9d bytes' % weave_size
1209
print 'total contents %9d bytes' % total
1210
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1213
print 'average size %9d bytes' % avg
1214
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1218
print """bzr weave tool
1220
Experimental tool for weave algorithm.
1223
weave init WEAVEFILE
1224
Create an empty weave file
1225
weave get WEAVEFILE VERSION
1226
Write out specified version.
1227
weave check WEAVEFILE
1228
Check consistency of all versions.
1230
Display table of contents.
1231
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1232
Add NEWTEXT, with specified parent versions.
1233
weave annotate WEAVEFILE VERSION
1234
Display origin of each line.
1235
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1236
Auto-merge two versions and display conflicts.
1237
weave diff WEAVEFILE VERSION1 VERSION2
1238
Show differences between two versions.
1242
% weave init foo.weave
1244
% weave add foo.weave ver0 < foo.txt
1247
(create updated version)
1249
% weave get foo.weave 0 | diff -u - foo.txt
1250
% weave add foo.weave ver1 0 < foo.txt
1253
% weave get foo.weave 0 > foo.txt (create forked version)
1255
% weave add foo.weave ver2 0 < foo.txt
1258
% weave merge foo.weave 1 2 > foo.txt (merge them)
1259
% vi foo.txt (resolve conflicts)
1260
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1272
# in case we're run directly from the subdirectory
1273
sys.path.append('..')
1275
from bzrlib.weavefile import write_weave, read_weave
1276
from bzrlib.progress import ProgressBar
1291
return read_weave(file(argv[2], 'rb'))
1297
# at the moment, based on everything in the file
1299
parents = map(int, argv[4:])
1300
lines = sys.stdin.readlines()
1301
ver = w.add(name, parents, lines)
1302
write_weave(w, file(argv[2], 'wb'))
1303
print 'added version %r %d' % (name, ver)
1306
if os.path.exists(fn):
1307
raise IOError("file exists")
1309
write_weave(w, file(fn, 'wb'))
1310
elif cmd == 'get': # get one version
1312
sys.stdout.writelines(w.get_iter(int(argv[3])))
1315
from difflib import unified_diff
1318
v1, v2 = map(int, argv[3:5])
1321
diff_gen = unified_diff(lines1, lines2,
1322
'%s version %d' % (fn, v1),
1323
'%s version %d' % (fn, v2))
1324
sys.stdout.writelines(diff_gen)
1326
elif cmd == 'annotate':
1328
# newline is added to all lines regardless; too hard to get
1329
# reasonable formatting otherwise
1331
for origin, text in w.annotate(int(argv[3])):
1332
text = text.rstrip('\r\n')
1334
print ' | %s' % (text)
1336
print '%5d | %s' % (origin, text)
1342
elif cmd == 'stats':
1343
weave_stats(argv[2], ProgressBar())
1345
elif cmd == 'check':
1350
print '%d versions ok' % w.num_versions()
1352
elif cmd == 'inclusions':
1354
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1356
elif cmd == 'parents':
1358
print ' '.join(map(str, w._parents[int(argv[3])]))
1360
elif cmd == 'plan-merge':
1362
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1364
print '%14s | %s' % (state, line),
1366
elif cmd == 'merge':
1368
p = w.plan_merge(int(argv[3]), int(argv[4]))
1369
sys.stdout.writelines(w.weave_merge(p))
1371
raise ValueError('unknown command %r' % cmd)
1375
def profile_main(argv):
1376
import tempfile, hotshot, hotshot.stats
1378
prof_f = tempfile.NamedTemporaryFile()
1380
prof = hotshot.Profile(prof_f.name)
1382
ret = prof.runcall(main, argv)
1385
stats = hotshot.stats.load(prof_f.name)
1387
stats.sort_stats('cumulative')
1388
## XXX: Might like to write to stderr or the trace file instead but
1389
## print_stats seems hardcoded to stdout
1390
stats.print_stats(20)
1395
def lsprofile_main(argv):
1396
from bzrlib.lsprof import profile
1397
ret,stats = profile(main, argv)
1403
if __name__ == '__main__':
1405
if '--profile' in sys.argv:
1407
args.remove('--profile')
1408
sys.exit(profile_main(args))
1409
elif '--lsprof' in sys.argv:
1411
args.remove('--lsprof')
1412
sys.exit(lsprofile_main(args))
1414
sys.exit(main(sys.argv))
1417
class InterWeave(InterVersionedFile):
1418
"""Optimised code paths for weave to weave operations."""
1420
_matching_file_factory = staticmethod(WeaveFile)
1423
def is_compatible(source, target):
1424
"""Be compatible with weaves."""
1426
return (isinstance(source, Weave) and
1427
isinstance(target, Weave))
1428
except AttributeError:
1431
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1432
"""See InterVersionedFile.join."""
1433
if self.target.versions() == []:
1435
self.target._copy_weave_content(self.source)
1438
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1439
except errors.WeaveParentMismatch:
1440
self.target._reweave(self.source, pb, msg)
1443
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
