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.
69
# FIXME: the conflict markers should be *7* characters
72
from cStringIO import StringIO
73
from difflib import SequenceMatcher
78
from bzrlib.trace import mutter
79
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
80
RevisionAlreadyPresent,
82
WeaveRevisionAlreadyPresent,
83
WeaveRevisionNotPresent,
85
import bzrlib.errors as errors
86
from bzrlib.osutils import sha_strings
87
from bzrlib.symbol_versioning import *
88
from bzrlib.tsort import topo_sort
89
from bzrlib.versionedfile import VersionedFile, InterVersionedFile
90
from bzrlib.weavefile import _read_weave_v5, write_weave_v5
93
class Weave(VersionedFile):
94
"""weave - versioned text file storage.
96
A Weave manages versions of line-based text files, keeping track
97
of the originating version for each line.
99
To clients the "lines" of the file are represented as a list of strings.
100
These strings will typically have terminal newline characters, but
101
this is not required. In particular files commonly do not have a newline
102
at the end of the file.
104
Texts can be identified in either of two ways:
106
* a nonnegative index number.
108
* a version-id string.
110
Typically the index number will be valid only inside this weave and
111
the version-id is used to reference it in the larger world.
113
The weave is represented as a list mixing edit instructions and
114
literal text. Each entry in _weave can be either a string (or
115
unicode), or a tuple. If a string, it means that the given line
116
should be output in the currently active revisions.
118
If a tuple, it gives a processing instruction saying in which
119
revisions the enclosed lines are active. The tuple has the form
120
(instruction, version).
122
The instruction can be '{' or '}' for an insertion block, and '['
123
and ']' for a deletion block respectively. The version is the
124
integer version index. There is no replace operator, only deletes
125
and inserts. For '}', the end of an insertion, there is no
126
version parameter because it always closes the most recently
131
* A later version can delete lines that were introduced by any
132
number of ancestor versions; this implies that deletion
133
instructions can span insertion blocks without regard to the
134
insertion block's nesting.
136
* Similarly, deletions need not be properly nested with regard to
137
each other, because they might have been generated by
138
independent revisions.
140
* Insertions are always made by inserting a new bracketed block
141
into a single point in the previous weave. This implies they
142
can nest but not overlap, and the nesting must always have later
143
insertions on the inside.
145
* It doesn't seem very useful to have an active insertion
146
inside an inactive insertion, but it might happen.
148
* Therefore, all instructions are always"considered"; that
149
is passed onto and off the stack. An outer inactive block
150
doesn't disable an inner block.
152
* Lines are enabled if the most recent enclosing insertion is
153
active and none of the enclosing deletions are active.
155
* There is no point having a deletion directly inside its own
156
insertion; you might as well just not write it. And there
157
should be no way to get an earlier version deleting a later
161
Text of the weave; list of control instruction tuples and strings.
164
List of parents, indexed by version number.
165
It is only necessary to store the minimal set of parents for
166
each version; the parent's parents are implied.
169
List of hex SHA-1 of each version.
172
List of symbolic names for each version. Each should be unique.
175
For each name, the version number.
178
Descriptive name of this weave; typically the filename if known.
182
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
185
def __init__(self, weave_name=None, access_mode='w'):
186
super(Weave, self).__init__(access_mode)
192
self._weave_name = weave_name
195
return "Weave(%r)" % self._weave_name
198
"""Return a deep copy of self.
200
The copy can be modified without affecting the original weave."""
202
other._weave = self._weave[:]
203
other._parents = self._parents[:]
204
other._sha1s = self._sha1s[:]
205
other._names = self._names[:]
206
other._name_map = self._name_map.copy()
207
other._weave_name = self._weave_name
210
def __eq__(self, other):
211
if not isinstance(other, Weave):
213
return self._parents == other._parents \
214
and self._weave == other._weave \
215
and self._sha1s == other._sha1s
217
def __ne__(self, other):
218
return not self.__eq__(other)
220
@deprecated_method(zero_eight)
221
def idx_to_name(self, index):
222
"""Old public interface, the public interface is all names now."""
225
def _idx_to_name(self, version):
226
return self._names[version]
228
@deprecated_method(zero_eight)
229
def lookup(self, name):
230
"""Backwards compatability thunk:
232
Return name, as name is valid in the api now, and spew deprecation
237
def _lookup(self, name):
238
"""Convert symbolic version name to index."""
240
return self._name_map[name]
242
raise RevisionNotPresent(name, self._weave_name)
244
@deprecated_method(zero_eight)
245
def iter_names(self):
246
"""Deprecated convenience function, please see VersionedFile.names()."""
247
return iter(self.names())
249
@deprecated_method(zero_eight)
251
"""See Weave.versions for the current api."""
252
return self.versions()
255
"""See VersionedFile.versions."""
256
return self._names[:]
258
def has_version(self, version_id):
259
"""See VersionedFile.has_version."""
260
return self._name_map.has_key(version_id)
262
__contains__ = has_version
264
def get_delta(self, version_id):
265
"""See VersionedFile.get_delta."""
266
return self.get_deltas([version_id])[version_id]
268
def get_deltas(self, version_ids):
269
"""See VersionedFile.get_deltas."""
270
version_ids = self.get_ancestry(version_ids)
271
for version_id in version_ids:
272
if not self.has_version(version_id):
273
raise RevisionNotPresent(version_id, self)
274
# try extracting all versions; parallel extraction is used
275
nv = self.num_versions()
281
last_parent_lines = {}
283
parent_inclusions = {}
288
# its simplest to generate a full set of prepared variables.
290
name = self._names[i]
291
sha1s[name] = self.get_sha1(name)
292
parents_list = self.get_parents(name)
294
parent = parents_list[0]
295
parents[name] = parent
296
parent_inclusions[name] = inclusions[parent]
299
parent_inclusions[name] = set()
300
# we want to emit start, finish, replacement_length, replacement_lines tuples.
301
diff_hunks[name] = []
302
current_hunks[name] = [0, 0, 0, []] # #start, finish, repl_length, repl_tuples
303
parent_linenums[name] = 0
305
parent_noeols[name] = False
306
last_parent_lines[name] = None
307
new_inc = set([name])
308
for p in self._parents[i]:
309
new_inc.update(inclusions[self._idx_to_name(p)])
310
# debug only, known good so far.
311
#assert set(new_inc) == set(self.get_ancestry(name)), \
312
# 'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
313
inclusions[name] = new_inc
315
nlines = len(self._weave)
317
for lineno, inserted, deletes, line in self._walk_internal():
318
# a line is active in a version if:
319
# insert is in the versions inclusions
321
# deleteset & the versions inclusions is an empty set.
322
# so - if we have a included by mapping - version is included by
323
# children, we get a list of children to examine for deletes affect
324
# ing them, which is less than the entire set of children.
325
for version_id in version_ids:
326
# The active inclusion must be an ancestor,
327
# and no ancestors must have deleted this line,
328
# because we don't support resurrection.
329
parent_inclusion = parent_inclusions[version_id]
330
inclusion = inclusions[version_id]
331
parent_active = inserted in parent_inclusion and not (deletes & parent_inclusion)
332
version_active = inserted in inclusion and not (deletes & inclusion)
333
if not parent_active and not version_active:
334
# unrelated line of ancestry
336
elif parent_active and version_active:
338
parent_linenum = parent_linenums[version_id]
339
if current_hunks[version_id] != [parent_linenum, parent_linenum, 0, []]:
340
diff_hunks[version_id].append(tuple(current_hunks[version_id]))
342
current_hunks[version_id] = [parent_linenum, parent_linenum, 0, []]
343
parent_linenums[version_id] = parent_linenum
346
noeols[version_id] = True
349
elif parent_active and not version_active:
351
current_hunks[version_id][1] += 1
352
parent_linenums[version_id] += 1
353
last_parent_lines[version_id] = line
354
elif not parent_active and version_active:
356
# noeol only occurs at the end of a file because we
357
# diff linewise. We want to show noeol changes as a
358
# empty diff unless the actual eol-less content changed.
361
if last_parent_lines[version_id][-1] != '\n':
362
parent_noeols[version_id] = True
363
except (TypeError, IndexError):
366
if theline[-1] != '\n':
367
noeols[version_id] = True
371
parent_should_go = False
373
if parent_noeols[version_id] == noeols[version_id]:
374
# no noeol toggle, so trust the weaves statement
375
# that this line is changed.
377
if parent_noeols[version_id]:
378
theline = theline + '\n'
379
elif parent_noeols[version_id]:
380
# parent has no eol, we do:
381
# our line is new, report as such..
383
elif noeols[version_id]:
384
# append a eol so that it looks like
386
theline = theline + '\n'
387
if parents[version_id] is not None:
388
#if last_parent_lines[version_id] is not None:
389
parent_should_go = True
390
if last_parent_lines[version_id] != theline:
393
#parent_should_go = False
395
current_hunks[version_id][2] += 1
396
current_hunks[version_id][3].append((inserted, theline))
398
# last hunk last parent line is not eaten
399
current_hunks[version_id][1] -= 1
400
if current_hunks[version_id][1] < 0:
401
current_hunks[version_id][1] = 0
402
# import pdb;pdb.set_trace()
403
# assert current_hunks[version_id][1] >= 0
407
version = self._idx_to_name(i)
408
if current_hunks[version] != [0, 0, 0, []]:
409
diff_hunks[version].append(tuple(current_hunks[version]))
411
for version_id in version_ids:
412
result[version_id] = (
416
diff_hunks[version_id],
420
def get_parents(self, version_id):
421
"""See VersionedFile.get_parent."""
422
return map(self._idx_to_name, self._parents[self._lookup(version_id)])
424
def _check_repeated_add(self, name, parents, text, sha1):
425
"""Check that a duplicated add is OK.
427
If it is, return the (old) index; otherwise raise an exception.
429
idx = self._lookup(name)
430
if sorted(self._parents[idx]) != sorted(parents) \
431
or sha1 != self._sha1s[idx]:
432
raise RevisionAlreadyPresent(name, self._weave_name)
435
@deprecated_method(zero_eight)
436
def add_identical(self, old_rev_id, new_rev_id, parents):
437
"""Please use Weave.clone_text now."""
438
return self.clone_text(new_rev_id, old_rev_id, parents)
440
def _add_lines(self, version_id, parents, lines, parent_texts):
441
"""See VersionedFile.add_lines."""
442
return self._add(version_id, lines, map(self._lookup, parents))
444
@deprecated_method(zero_eight)
445
def add(self, name, parents, text, sha1=None):
446
"""See VersionedFile.add_lines for the non deprecated api."""
447
return self._add(name, text, map(self._maybe_lookup, parents), sha1)
449
def _add(self, version_id, lines, parents, sha1=None):
450
"""Add a single text on top of the weave.
452
Returns the index number of the newly added version.
455
Symbolic name for this version.
456
(Typically the revision-id of the revision that added it.)
459
List or set of direct parent version numbers.
462
Sequence of lines to be added in the new version.
465
assert isinstance(version_id, basestring)
467
sha1 = sha_strings(lines)
468
if version_id in self._name_map:
469
return self._check_repeated_add(version_id, parents, lines, sha1)
471
self._check_versions(parents)
472
## self._check_lines(lines)
473
new_version = len(self._parents)
475
# if we abort after here the (in-memory) weave will be corrupt because only
476
# some fields are updated
477
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
478
# - Robert Collins 20060226
479
self._parents.append(parents[:])
480
self._sha1s.append(sha1)
481
self._names.append(version_id)
482
self._name_map[version_id] = new_version
486
# special case; adding with no parents revision; can do
487
# this more quickly by just appending unconditionally.
488
# even more specially, if we're adding an empty text we
489
# need do nothing at all.
491
self._weave.append(('{', new_version))
492
self._weave.extend(lines)
493
self._weave.append(('}', None))
496
if len(parents) == 1:
497
pv = list(parents)[0]
498
if sha1 == self._sha1s[pv]:
499
# special case: same as the single parent
503
ancestors = self._inclusions(parents)
507
# basis a list of (origin, lineno, line)
510
for origin, lineno, line in self._extract(ancestors):
511
basis_lineno.append(lineno)
512
basis_lines.append(line)
514
# another small special case: a merge, producing the same text
516
if lines == basis_lines:
519
# add a sentinal, because we can also match against the final line
520
basis_lineno.append(len(self._weave))
522
# XXX: which line of the weave should we really consider
523
# matches the end of the file? the current code says it's the
524
# last line of the weave?
526
#print 'basis_lines:', basis_lines
527
#print 'new_lines: ', lines
529
s = SequenceMatcher(None, basis_lines, lines)
531
# offset gives the number of lines that have been inserted
532
# into the weave up to the current point; if the original edit instruction
533
# says to change line A then we actually change (A+offset)
536
for tag, i1, i2, j1, j2 in s.get_opcodes():
537
# i1,i2 are given in offsets within basis_lines; we need to map them
538
# back to offsets within the entire weave
539
#print 'raw match', tag, i1, i2, j1, j2
543
i1 = basis_lineno[i1]
544
i2 = basis_lineno[i2]
546
assert 0 <= j1 <= j2 <= len(lines)
548
#print tag, i1, i2, j1, j2
550
# the deletion and insertion are handled separately.
551
# first delete the region.
553
self._weave.insert(i1+offset, ('[', new_version))
554
self._weave.insert(i2+offset+1, (']', new_version))
558
# there may have been a deletion spanning up to
559
# i2; we want to insert after this region to make sure
560
# we don't destroy ourselves
562
self._weave[i:i] = ([('{', new_version)]
565
offset += 2 + (j2 - j1)
568
def _clone_text(self, new_version_id, old_version_id, parents):
569
"""See VersionedFile.clone_text."""
570
old_lines = self.get_text(old_version_id)
571
self.add_lines(new_version_id, parents, old_lines)
573
def _inclusions(self, versions):
574
"""Return set of all ancestors of given version(s)."""
575
if not len(versions):
578
for v in xrange(max(versions), 0, -1):
580
# include all its parents
581
i.update(self._parents[v])
583
## except IndexError:
584
## raise ValueError("version %d not present in weave" % v)
586
@deprecated_method(zero_eight)
587
def inclusions(self, version_ids):
588
"""Deprecated - see VersionedFile.get_ancestry for the replacement."""
591
if isinstance(version_ids[0], int):
592
return [self._idx_to_name(v) for v in self._inclusions(version_ids)]
594
return self.get_ancestry(version_ids)
596
def get_ancestry(self, version_ids):
597
"""See VersionedFile.get_ancestry."""
598
if isinstance(version_ids, basestring):
599
version_ids = [version_ids]
600
i = self._inclusions([self._lookup(v) for v in version_ids])
601
return [self._idx_to_name(v) for v in i]
603
def _check_lines(self, text):
604
if not isinstance(text, list):
605
raise ValueError("text should be a list, not %s" % type(text))
608
if not isinstance(l, basestring):
609
raise ValueError("text line should be a string or unicode, not %s"
614
def _check_versions(self, indexes):
615
"""Check everything in the sequence of indexes is valid"""
620
raise IndexError("invalid version number %r" % i)
622
def _compatible_parents(self, my_parents, other_parents):
623
"""During join check that other_parents are joinable with my_parents.
625
Joinable is defined as 'is a subset of' - supersets may require
626
regeneration of diffs, but subsets do not.
628
return len(other_parents.difference(my_parents)) == 0
630
def annotate(self, version_id):
631
if isinstance(version_id, int):
632
warn('Weave.annotate(int) is deprecated. Please use version names'
633
' in all circumstances as of 0.8',
638
for origin, lineno, text in self._extract([version_id]):
639
result.append((origin, text))
642
return super(Weave, self).annotate(version_id)
644
def annotate_iter(self, version_id):
645
"""Yield list of (version-id, line) pairs for the specified version.
647
The index indicates when the line originated in the weave."""
648
incls = [self._lookup(version_id)]
649
for origin, lineno, text in self._extract(incls):
650
yield self._idx_to_name(origin), text
652
@deprecated_method(zero_eight)
654
"""_walk has become visit, a supported api."""
655
return self._walk_internal()
657
def iter_lines_added_or_present_in_versions(self, version_ids=None):
658
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
659
if version_ids is None:
660
version_ids = self.versions()
661
version_ids = set(version_ids)
662
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
663
# if inserted not in version_ids then it was inserted before the
664
# versions we care about, but because weaves cannot represent ghosts
665
# properly, we do not filter down to that
666
# if inserted not in version_ids: continue
672
#@deprecated_method(zero_eight)
673
def walk(self, version_ids=None):
674
"""See VersionedFile.walk."""
675
return self._walk_internal(version_ids)
677
def _walk_internal(self, version_ids=None):
678
"""Helper method for weave actions."""
683
lineno = 0 # line of weave, 0-based
685
for l in self._weave:
686
if l.__class__ == tuple:
690
istack.append(self._names[v])
694
assert self._names[v] not in dset
695
dset.add(self._names[v])
697
dset.remove(self._names[v])
699
raise WeaveFormatError('unexpected instruction %r' % v)
701
assert l.__class__ in (str, unicode)
703
yield lineno, istack[-1], frozenset(dset), l
707
raise WeaveFormatError("unclosed insertion blocks "
708
"at end of weave: %s" % istack)
710
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
713
def _extract(self, versions):
714
"""Yield annotation of lines in included set.
716
Yields a sequence of tuples (origin, lineno, text), where
717
origin is the origin version, lineno the index in the weave,
718
and text the text of the line.
720
The set typically but not necessarily corresponds to a version.
723
if not isinstance(i, int):
726
included = self._inclusions(versions)
732
lineno = 0 # line of weave, 0-based
738
WFE = WeaveFormatError
741
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
742
# +285282 0 1676.8040 1676.8040 +<isinstance>
743
# 1.6 seconds in 'isinstance'.
744
# changing the first isinstance:
745
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
746
# +140414 0 762.8050 762.8050 +<isinstance>
747
# note that the inline time actually dropped (less function calls)
748
# and total processing time was halved.
749
# we're still spending ~1/4 of the method in isinstance though.
750
# so lets hard code the acceptable string classes we expect:
751
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
752
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
754
# yay, down to ~1/4 the initial extract time, and our inline time
755
# has shrunk again, with isinstance no longer dominating.
756
# tweaking the stack inclusion test to use a set gives:
757
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
758
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
760
# - a 5% win, or possibly just noise. However with large istacks that
761
# 'in' test could dominate, so I'm leaving this change in place -
762
# when its fast enough to consider profiling big datasets we can review.
767
for l in self._weave:
768
if l.__class__ == tuple:
776
iset.remove(istack.pop())
787
assert l.__class__ in (str, unicode)
789
isactive = (not dset) and istack and (istack[-1] in included)
791
result.append((istack[-1], lineno, l))
794
raise WeaveFormatError("unclosed insertion blocks "
795
"at end of weave: %s" % istack)
797
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
801
@deprecated_method(zero_eight)
802
def get_iter(self, name_or_index):
803
"""Deprecated, please do not use. Lookups are not not needed.
805
Please use get_lines now.
807
return iter(self.get_lines(self._maybe_lookup(name_or_index)))
809
@deprecated_method(zero_eight)
810
def maybe_lookup(self, name_or_index):
811
"""Deprecated, please do not use. Lookups are not not needed."""
812
return self._maybe_lookup(name_or_index)
814
def _maybe_lookup(self, name_or_index):
815
"""Convert possible symbolic name to index, or pass through indexes.
819
if isinstance(name_or_index, (int, long)):
822
return self._lookup(name_or_index)
824
@deprecated_method(zero_eight)
825
def get(self, version_id):
826
"""Please use either Weave.get_text or Weave.get_lines as desired."""
827
return self.get_lines(version_id)
829
def get_lines(self, version_id):
830
"""See VersionedFile.get_lines()."""
831
int_index = self._maybe_lookup(version_id)
832
result = [line for (origin, lineno, line) in self._extract([int_index])]
833
expected_sha1 = self._sha1s[int_index]
834
measured_sha1 = sha_strings(result)
835
if measured_sha1 != expected_sha1:
836
raise errors.WeaveInvalidChecksum(
837
'file %s, revision %s, expected: %s, measured %s'
838
% (self._weave_name, version_id,
839
expected_sha1, measured_sha1))
842
def get_sha1(self, name):
843
"""Get the stored sha1 sum for the given revision.
845
:param name: The name of the version to lookup
847
return self._sha1s[self._lookup(name)]
849
@deprecated_method(zero_eight)
850
def numversions(self):
851
"""How many versions are in this weave?
853
Deprecated in favour of num_versions.
855
return self.num_versions()
857
def num_versions(self):
858
"""How many versions are in this weave?"""
859
l = len(self._parents)
860
assert l == len(self._sha1s)
863
__len__ = num_versions
865
def check(self, progress_bar=None):
866
# TODO evaluate performance hit of using string sets in this routine.
867
# TODO: check no circular inclusions
868
# TODO: create a nested progress bar
869
for version in range(self.num_versions()):
870
inclusions = list(self._parents[version])
873
if inclusions[-1] >= version:
874
raise WeaveFormatError("invalid included version %d for index %d"
875
% (inclusions[-1], version))
877
# try extracting all versions; parallel extraction is used
878
nv = self.num_versions()
883
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
884
# The problem is that set membership is much more expensive
885
name = self._idx_to_name(i)
886
sha1s[name] = sha.new()
888
new_inc = set([name])
889
for p in self._parents[i]:
890
new_inc.update(inclusions[self._idx_to_name(p)])
892
assert set(new_inc) == set(self.get_ancestry(name)), \
893
'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
894
inclusions[name] = new_inc
896
nlines = len(self._weave)
898
update_text = 'checking weave'
900
short_name = os.path.basename(self._weave_name)
901
update_text = 'checking %s' % (short_name,)
902
update_text = update_text[:25]
904
for lineno, insert, deleteset, line in self._walk_internal():
906
progress_bar.update(update_text, lineno, nlines)
908
for name, name_inclusions in inclusions.items():
909
# The active inclusion must be an ancestor,
910
# and no ancestors must have deleted this line,
911
# because we don't support resurrection.
912
if (insert in name_inclusions) and not (deleteset & name_inclusions):
913
sha1s[name].update(line)
916
version = self._idx_to_name(i)
917
hd = sha1s[version].hexdigest()
918
expected = self._sha1s[i]
920
raise errors.WeaveInvalidChecksum(
921
"mismatched sha1 for version %s: "
922
"got %s, expected %s"
923
% (version, hd, expected))
925
# TODO: check insertions are properly nested, that there are
926
# no lines outside of insertion blocks, that deletions are
927
# properly paired, etc.
929
def _join(self, other, pb, msg, version_ids, ignore_missing):
930
"""Worker routine for join()."""
931
if not other.versions():
932
return # nothing to update, easy
935
for version_id in version_ids:
936
if not other.has_version(version_id) and not ignore_missing:
937
raise RevisionNotPresent(version_id, self._weave_name)
939
version_ids = other.versions()
941
# two loops so that we do not change ourselves before verifying it
943
# work through in index order to make sure we get all dependencies
946
# get the selected versions only that are in other.versions.
947
version_ids = set(other.versions()).intersection(set(version_ids))
948
# pull in the referenced graph.
949
version_ids = other.get_ancestry(version_ids)
950
pending_graph = [(version, other.get_parents(version)) for
951
version in version_ids]
952
for name in topo_sort(pending_graph):
953
other_idx = other._name_map[name]
954
# returns True if we have it, False if we need it.
955
if not self._check_version_consistent(other, other_idx, name):
956
names_to_join.append((other_idx, name))
965
for other_idx, name in names_to_join:
966
# TODO: If all the parents of the other version are already
967
# present then we can avoid some work by just taking the delta
968
# and adjusting the offsets.
969
new_parents = self._imported_parents(other, other_idx)
970
sha1 = other._sha1s[other_idx]
975
pb.update(msg, merged, len(names_to_join))
977
lines = other.get_lines(other_idx)
978
self._add(name, lines, new_parents, sha1)
980
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
981
merged, processed, self._weave_name, time.time()-time0))
983
def _imported_parents(self, other, other_idx):
984
"""Return list of parents in self corresponding to indexes in other."""
986
for parent_idx in other._parents[other_idx]:
987
parent_name = other._names[parent_idx]
988
if parent_name not in self._name_map:
989
# should not be possible
990
raise WeaveError("missing parent {%s} of {%s} in %r"
991
% (parent_name, other._name_map[other_idx], self))
992
new_parents.append(self._name_map[parent_name])
995
def _check_version_consistent(self, other, other_idx, name):
996
"""Check if a version in consistent in this and other.
998
To be consistent it must have:
1001
* the same direct parents (by name, not index, and disregarding
1004
If present & correct return True;
1005
if not present in self return False;
1006
if inconsistent raise error."""
1007
this_idx = self._name_map.get(name, -1)
1009
if self._sha1s[this_idx] != other._sha1s[other_idx]:
1010
raise errors.WeaveTextDiffers(name, self, other)
1011
self_parents = self._parents[this_idx]
1012
other_parents = other._parents[other_idx]
1013
n1 = set([self._names[i] for i in self_parents])
1014
n2 = set([other._names[i] for i in other_parents])
1015
if not self._compatible_parents(n1, n2):
1016
raise WeaveParentMismatch("inconsistent parents "
1017
"for version {%s}: %s vs %s" % (name, n1, n2))
1023
@deprecated_method(zero_eight)
1024
def reweave(self, other, pb=None, msg=None):
1025
"""reweave has been superceded by plain use of join."""
1026
return self.join(other, pb, msg)
1028
def _reweave(self, other, pb, msg):
1029
"""Reweave self with other - internal helper for join().
1031
:param other: The other weave to merge
1032
:param pb: An optional progress bar, indicating how far done we are
1033
:param msg: An optional message for the progress
1035
new_weave = _reweave(self, other, pb=pb, msg=msg)
1036
self._copy_weave_content(new_weave)
1038
def _copy_weave_content(self, otherweave):
1039
"""adsorb the content from otherweave."""
1040
for attr in self.__slots__:
1041
if attr != '_weave_name':
1042
setattr(self, attr, copy(getattr(otherweave, attr)))
1045
class WeaveFile(Weave):
1046
"""A WeaveFile represents a Weave on disk and writes on change."""
1048
WEAVE_SUFFIX = '.weave'
1050
def __init__(self, name, transport, filemode=None, create=False, access_mode='w'):
1051
"""Create a WeaveFile.
1053
:param create: If not True, only open an existing knit.
1055
super(WeaveFile, self).__init__(name, access_mode)
1056
self._transport = transport
1057
self._filemode = filemode
1059
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
1060
except errors.NoSuchFile:
1066
def _add_lines(self, version_id, parents, lines, parent_texts):
1067
"""Add a version and save the weave."""
1068
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
1073
def _clone_text(self, new_version_id, old_version_id, parents):
1074
"""See VersionedFile.clone_text."""
1075
super(WeaveFile, self)._clone_text(new_version_id, old_version_id, parents)
1078
def copy_to(self, name, transport):
1079
"""See VersionedFile.copy_to()."""
1080
# as we are all in memory always, just serialise to the new place.
1082
write_weave_v5(self, sio)
1084
transport.put(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
1086
def create_empty(self, name, transport, filemode=None):
1087
return WeaveFile(name, transport, filemode, create=True)
1090
"""Save the weave."""
1091
self._check_write_ok()
1093
write_weave_v5(self, sio)
1095
self._transport.put(self._weave_name + WeaveFile.WEAVE_SUFFIX,
1101
"""See VersionedFile.get_suffixes()."""
1102
return [WeaveFile.WEAVE_SUFFIX]
1104
def join(self, other, pb=None, msg=None, version_ids=None,
1105
ignore_missing=False):
1106
"""Join other into self and save."""
1107
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1111
@deprecated_function(zero_eight)
1112
def reweave(wa, wb, pb=None, msg=None):
1113
"""reweaving is deprecation, please just use weave.join()."""
1114
_reweave(wa, wb, pb, msg)
1116
def _reweave(wa, wb, pb=None, msg=None):
1117
"""Combine two weaves and return the result.
1119
This works even if a revision R has different parents in
1120
wa and wb. In the resulting weave all the parents are given.
1122
This is done by just building up a new weave, maintaining ordering
1123
of the versions in the two inputs. More efficient approaches
1124
might be possible but it should only be necessary to do
1125
this operation rarely, when a new previously ghost version is
1128
:param pb: An optional progress bar, indicating how far done we are
1129
:param msg: An optional message for the progress
1133
queue_a = range(wa.num_versions())
1134
queue_b = range(wb.num_versions())
1135
# first determine combined parents of all versions
1136
# map from version name -> all parent names
1137
combined_parents = _reweave_parent_graphs(wa, wb)
1138
mutter("combined parents: %r", combined_parents)
1139
order = topo_sort(combined_parents.iteritems())
1140
mutter("order to reweave: %r", order)
1145
for idx, name in enumerate(order):
1147
pb.update(msg, idx, len(order))
1148
if name in wa._name_map:
1149
lines = wa.get_lines(name)
1150
if name in wb._name_map:
1151
lines_b = wb.get_lines(name)
1152
if lines != lines_b:
1153
mutter('Weaves differ on content. rev_id {%s}', name)
1154
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1156
lines = list(difflib.unified_diff(lines, lines_b,
1157
wa._weave_name, wb._weave_name))
1158
mutter('lines:\n%s', ''.join(lines))
1159
raise errors.WeaveTextDiffers(name, wa, wb)
1161
lines = wb.get_lines(name)
1162
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1165
def _reweave_parent_graphs(wa, wb):
1166
"""Return combined parent ancestry for two weaves.
1168
Returned as a list of (version_name, set(parent_names))"""
1170
for weave in [wa, wb]:
1171
for idx, name in enumerate(weave._names):
1172
p = combined.setdefault(name, set())
1173
p.update(map(weave._idx_to_name, weave._parents[idx]))
1178
"""Show the weave's table-of-contents"""
1179
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1180
for i in (6, 50, 10, 10):
1183
for i in range(w.num_versions()):
1186
parent_str = ' '.join(map(str, w._parents[i]))
1187
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1191
def weave_stats(weave_file, pb):
1192
from bzrlib.weavefile import read_weave
1194
wf = file(weave_file, 'rb')
1195
w = read_weave(wf, WeaveVersionedFile)
1196
# FIXME: doesn't work on pipes
1197
weave_size = wf.tell()
1201
for i in range(vers):
1202
pb.update('checking sizes', i, vers)
1203
for origin, lineno, line in w._extract([i]):
1208
print 'versions %9d' % vers
1209
print 'weave file %9d bytes' % weave_size
1210
print 'total contents %9d bytes' % total
1211
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1214
print 'average size %9d bytes' % avg
1215
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1219
print """bzr weave tool
1221
Experimental tool for weave algorithm.
1224
weave init WEAVEFILE
1225
Create an empty weave file
1226
weave get WEAVEFILE VERSION
1227
Write out specified version.
1228
weave check WEAVEFILE
1229
Check consistency of all versions.
1231
Display table of contents.
1232
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1233
Add NEWTEXT, with specified parent versions.
1234
weave annotate WEAVEFILE VERSION
1235
Display origin of each line.
1236
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1237
Auto-merge two versions and display conflicts.
1238
weave diff WEAVEFILE VERSION1 VERSION2
1239
Show differences between two versions.
1243
% weave init foo.weave
1245
% weave add foo.weave ver0 < foo.txt
1248
(create updated version)
1250
% weave get foo.weave 0 | diff -u - foo.txt
1251
% weave add foo.weave ver1 0 < foo.txt
1254
% weave get foo.weave 0 > foo.txt (create forked version)
1256
% weave add foo.weave ver2 0 < foo.txt
1259
% weave merge foo.weave 1 2 > foo.txt (merge them)
1260
% vi foo.txt (resolve conflicts)
1261
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1273
# in case we're run directly from the subdirectory
1274
sys.path.append('..')
1276
from bzrlib.weavefile import write_weave, read_weave
1277
from bzrlib.progress import ProgressBar
1292
return read_weave(file(argv[2], 'rb'))
1298
# at the moment, based on everything in the file
1300
parents = map(int, argv[4:])
1301
lines = sys.stdin.readlines()
1302
ver = w.add(name, parents, lines)
1303
write_weave(w, file(argv[2], 'wb'))
1304
print 'added version %r %d' % (name, ver)
1307
if os.path.exists(fn):
1308
raise IOError("file exists")
1310
write_weave(w, file(fn, 'wb'))
1311
elif cmd == 'get': # get one version
1313
sys.stdout.writelines(w.get_iter(int(argv[3])))
1316
from difflib import unified_diff
1319
v1, v2 = map(int, argv[3:5])
1322
diff_gen = unified_diff(lines1, lines2,
1323
'%s version %d' % (fn, v1),
1324
'%s version %d' % (fn, v2))
1325
sys.stdout.writelines(diff_gen)
1327
elif cmd == 'annotate':
1329
# newline is added to all lines regardless; too hard to get
1330
# reasonable formatting otherwise
1332
for origin, text in w.annotate(int(argv[3])):
1333
text = text.rstrip('\r\n')
1335
print ' | %s' % (text)
1337
print '%5d | %s' % (origin, text)
1343
elif cmd == 'stats':
1344
weave_stats(argv[2], ProgressBar())
1346
elif cmd == 'check':
1351
print '%d versions ok' % w.num_versions()
1353
elif cmd == 'inclusions':
1355
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1357
elif cmd == 'parents':
1359
print ' '.join(map(str, w._parents[int(argv[3])]))
1361
elif cmd == 'plan-merge':
1362
# replaced by 'bzr weave-plan-merge'
1364
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1366
print '%14s | %s' % (state, line),
1367
elif cmd == 'merge':
1368
# replaced by 'bzr weave-merge-text'
1370
p = w.plan_merge(int(argv[3]), int(argv[4]))
1371
sys.stdout.writelines(w.weave_merge(p))
1373
raise ValueError('unknown command %r' % cmd)
1377
def profile_main(argv):
1378
import tempfile, hotshot, hotshot.stats
1380
prof_f = tempfile.NamedTemporaryFile()
1382
prof = hotshot.Profile(prof_f.name)
1384
ret = prof.runcall(main, argv)
1387
stats = hotshot.stats.load(prof_f.name)
1389
stats.sort_stats('cumulative')
1390
## XXX: Might like to write to stderr or the trace file instead but
1391
## print_stats seems hardcoded to stdout
1392
stats.print_stats(20)
1397
def lsprofile_main(argv):
1398
from bzrlib.lsprof import profile
1399
ret,stats = profile(main, argv)
1405
if __name__ == '__main__':
1407
if '--profile' in sys.argv:
1409
args.remove('--profile')
1410
sys.exit(profile_main(args))
1411
elif '--lsprof' in sys.argv:
1413
args.remove('--lsprof')
1414
sys.exit(lsprofile_main(args))
1416
sys.exit(main(sys.argv))
1419
class InterWeave(InterVersionedFile):
1420
"""Optimised code paths for weave to weave operations."""
1422
_matching_file_factory = staticmethod(WeaveFile)
1425
def is_compatible(source, target):
1426
"""Be compatible with weaves."""
1428
return (isinstance(source, Weave) and
1429
isinstance(target, Weave))
1430
except AttributeError:
1433
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1434
"""See InterVersionedFile.join."""
1435
if self.target.versions() == []:
1437
self.target._copy_weave_content(self.source)
1440
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1441
except errors.WeaveParentMismatch:
1442
self.target._reweave(self.source, pb, msg)
1445
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
