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
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
import bzrlib.patiencediff
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',
183
'_weave_name', '_matcher']
185
def __init__(self, weave_name=None, access_mode='w', matcher=None):
186
super(Weave, self).__init__(access_mode)
192
self._weave_name = weave_name
194
self._matcher = bzrlib.patiencediff.PatienceSequenceMatcher
196
self._matcher = matcher
199
return "Weave(%r)" % self._weave_name
202
"""Return a deep copy of self.
204
The copy can be modified without affecting the original weave."""
206
other._weave = self._weave[:]
207
other._parents = self._parents[:]
208
other._sha1s = self._sha1s[:]
209
other._names = self._names[:]
210
other._name_map = self._name_map.copy()
211
other._weave_name = self._weave_name
214
def __eq__(self, other):
215
if not isinstance(other, Weave):
217
return self._parents == other._parents \
218
and self._weave == other._weave \
219
and self._sha1s == other._sha1s
221
def __ne__(self, other):
222
return not self.__eq__(other)
224
@deprecated_method(zero_eight)
225
def idx_to_name(self, index):
226
"""Old public interface, the public interface is all names now."""
229
def _idx_to_name(self, version):
230
return self._names[version]
232
@deprecated_method(zero_eight)
233
def lookup(self, name):
234
"""Backwards compatability thunk:
236
Return name, as name is valid in the api now, and spew deprecation
241
def _lookup(self, name):
242
"""Convert symbolic version name to index."""
244
return self._name_map[name]
246
raise RevisionNotPresent(name, self._weave_name)
248
@deprecated_method(zero_eight)
249
def iter_names(self):
250
"""Deprecated convenience function, please see VersionedFile.names()."""
251
return iter(self.names())
253
@deprecated_method(zero_eight)
255
"""See Weave.versions for the current api."""
256
return self.versions()
259
"""See VersionedFile.versions."""
260
return self._names[:]
262
def has_version(self, version_id):
263
"""See VersionedFile.has_version."""
264
return self._name_map.has_key(version_id)
266
__contains__ = has_version
268
def get_delta(self, version_id):
269
"""See VersionedFile.get_delta."""
270
return self.get_deltas([version_id])[version_id]
272
def get_deltas(self, version_ids):
273
"""See VersionedFile.get_deltas."""
274
version_ids = self.get_ancestry(version_ids)
275
for version_id in version_ids:
276
if not self.has_version(version_id):
277
raise RevisionNotPresent(version_id, self)
278
# try extracting all versions; parallel extraction is used
279
nv = self.num_versions()
285
last_parent_lines = {}
287
parent_inclusions = {}
292
# its simplest to generate a full set of prepared variables.
294
name = self._names[i]
295
sha1s[name] = self.get_sha1(name)
296
parents_list = self.get_parents(name)
298
parent = parents_list[0]
299
parents[name] = parent
300
parent_inclusions[name] = inclusions[parent]
303
parent_inclusions[name] = set()
304
# we want to emit start, finish, replacement_length, replacement_lines tuples.
305
diff_hunks[name] = []
306
current_hunks[name] = [0, 0, 0, []] # #start, finish, repl_length, repl_tuples
307
parent_linenums[name] = 0
309
parent_noeols[name] = False
310
last_parent_lines[name] = None
311
new_inc = set([name])
312
for p in self._parents[i]:
313
new_inc.update(inclusions[self._idx_to_name(p)])
314
# debug only, known good so far.
315
#assert set(new_inc) == set(self.get_ancestry(name)), \
316
# 'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
317
inclusions[name] = new_inc
319
nlines = len(self._weave)
321
for lineno, inserted, deletes, line in self._walk_internal():
322
# a line is active in a version if:
323
# insert is in the versions inclusions
325
# deleteset & the versions inclusions is an empty set.
326
# so - if we have a included by mapping - version is included by
327
# children, we get a list of children to examine for deletes affect
328
# ing them, which is less than the entire set of children.
329
for version_id in version_ids:
330
# The active inclusion must be an ancestor,
331
# and no ancestors must have deleted this line,
332
# because we don't support resurrection.
333
parent_inclusion = parent_inclusions[version_id]
334
inclusion = inclusions[version_id]
335
parent_active = inserted in parent_inclusion and not (deletes & parent_inclusion)
336
version_active = inserted in inclusion and not (deletes & inclusion)
337
if not parent_active and not version_active:
338
# unrelated line of ancestry
340
elif parent_active and version_active:
342
parent_linenum = parent_linenums[version_id]
343
if current_hunks[version_id] != [parent_linenum, parent_linenum, 0, []]:
344
diff_hunks[version_id].append(tuple(current_hunks[version_id]))
346
current_hunks[version_id] = [parent_linenum, parent_linenum, 0, []]
347
parent_linenums[version_id] = parent_linenum
350
noeols[version_id] = True
353
elif parent_active and not version_active:
355
current_hunks[version_id][1] += 1
356
parent_linenums[version_id] += 1
357
last_parent_lines[version_id] = line
358
elif not parent_active and version_active:
360
# noeol only occurs at the end of a file because we
361
# diff linewise. We want to show noeol changes as a
362
# empty diff unless the actual eol-less content changed.
365
if last_parent_lines[version_id][-1] != '\n':
366
parent_noeols[version_id] = True
367
except (TypeError, IndexError):
370
if theline[-1] != '\n':
371
noeols[version_id] = True
375
parent_should_go = False
377
if parent_noeols[version_id] == noeols[version_id]:
378
# no noeol toggle, so trust the weaves statement
379
# that this line is changed.
381
if parent_noeols[version_id]:
382
theline = theline + '\n'
383
elif parent_noeols[version_id]:
384
# parent has no eol, we do:
385
# our line is new, report as such..
387
elif noeols[version_id]:
388
# append a eol so that it looks like
390
theline = theline + '\n'
391
if parents[version_id] is not None:
392
#if last_parent_lines[version_id] is not None:
393
parent_should_go = True
394
if last_parent_lines[version_id] != theline:
397
#parent_should_go = False
399
current_hunks[version_id][2] += 1
400
current_hunks[version_id][3].append((inserted, theline))
402
# last hunk last parent line is not eaten
403
current_hunks[version_id][1] -= 1
404
if current_hunks[version_id][1] < 0:
405
current_hunks[version_id][1] = 0
406
# import pdb;pdb.set_trace()
407
# assert current_hunks[version_id][1] >= 0
411
version = self._idx_to_name(i)
412
if current_hunks[version] != [0, 0, 0, []]:
413
diff_hunks[version].append(tuple(current_hunks[version]))
415
for version_id in version_ids:
416
result[version_id] = (
420
diff_hunks[version_id],
424
def get_parents(self, version_id):
425
"""See VersionedFile.get_parent."""
426
return map(self._idx_to_name, self._parents[self._lookup(version_id)])
428
def _check_repeated_add(self, name, parents, text, sha1):
429
"""Check that a duplicated add is OK.
431
If it is, return the (old) index; otherwise raise an exception.
433
idx = self._lookup(name)
434
if sorted(self._parents[idx]) != sorted(parents) \
435
or sha1 != self._sha1s[idx]:
436
raise RevisionAlreadyPresent(name, self._weave_name)
439
@deprecated_method(zero_eight)
440
def add_identical(self, old_rev_id, new_rev_id, parents):
441
"""Please use Weave.clone_text now."""
442
return self.clone_text(new_rev_id, old_rev_id, parents)
444
def _add_lines(self, version_id, parents, lines, parent_texts):
445
"""See VersionedFile.add_lines."""
446
return self._add(version_id, lines, map(self._lookup, parents))
448
@deprecated_method(zero_eight)
449
def add(self, name, parents, text, sha1=None):
450
"""See VersionedFile.add_lines for the non deprecated api."""
451
return self._add(name, text, map(self._maybe_lookup, parents), sha1)
453
def _add(self, version_id, lines, parents, sha1=None):
454
"""Add a single text on top of the weave.
456
Returns the index number of the newly added version.
459
Symbolic name for this version.
460
(Typically the revision-id of the revision that added it.)
463
List or set of direct parent version numbers.
466
Sequence of lines to be added in the new version.
469
assert isinstance(version_id, basestring)
470
self._check_lines_not_unicode(lines)
471
self._check_lines_are_lines(lines)
473
sha1 = sha_strings(lines)
474
if version_id in self._name_map:
475
return self._check_repeated_add(version_id, parents, lines, sha1)
477
self._check_versions(parents)
478
## self._check_lines(lines)
479
new_version = len(self._parents)
481
# if we abort after here the (in-memory) weave will be corrupt because only
482
# some fields are updated
483
# XXX: FIXME implement a succeed-or-fail of the rest of this routine.
484
# - Robert Collins 20060226
485
self._parents.append(parents[:])
486
self._sha1s.append(sha1)
487
self._names.append(version_id)
488
self._name_map[version_id] = new_version
492
# special case; adding with no parents revision; can do
493
# this more quickly by just appending unconditionally.
494
# even more specially, if we're adding an empty text we
495
# need do nothing at all.
497
self._weave.append(('{', new_version))
498
self._weave.extend(lines)
499
self._weave.append(('}', None))
502
if len(parents) == 1:
503
pv = list(parents)[0]
504
if sha1 == self._sha1s[pv]:
505
# special case: same as the single parent
509
ancestors = self._inclusions(parents)
513
# basis a list of (origin, lineno, line)
516
for origin, lineno, line in self._extract(ancestors):
517
basis_lineno.append(lineno)
518
basis_lines.append(line)
520
# another small special case: a merge, producing the same text
522
if lines == basis_lines:
525
# add a sentinal, because we can also match against the final line
526
basis_lineno.append(len(self._weave))
528
# XXX: which line of the weave should we really consider
529
# matches the end of the file? the current code says it's the
530
# last line of the weave?
532
#print 'basis_lines:', basis_lines
533
#print 'new_lines: ', lines
535
s = self._matcher(None, basis_lines, lines)
537
# offset gives the number of lines that have been inserted
538
# into the weave up to the current point; if the original edit instruction
539
# says to change line A then we actually change (A+offset)
542
for tag, i1, i2, j1, j2 in s.get_opcodes():
543
# i1,i2 are given in offsets within basis_lines; we need to map them
544
# back to offsets within the entire weave
545
#print 'raw match', tag, i1, i2, j1, j2
549
i1 = basis_lineno[i1]
550
i2 = basis_lineno[i2]
552
assert 0 <= j1 <= j2 <= len(lines)
554
#print tag, i1, i2, j1, j2
556
# the deletion and insertion are handled separately.
557
# first delete the region.
559
self._weave.insert(i1+offset, ('[', new_version))
560
self._weave.insert(i2+offset+1, (']', new_version))
564
# there may have been a deletion spanning up to
565
# i2; we want to insert after this region to make sure
566
# we don't destroy ourselves
568
self._weave[i:i] = ([('{', new_version)]
571
offset += 2 + (j2 - j1)
574
def _clone_text(self, new_version_id, old_version_id, parents):
575
"""See VersionedFile.clone_text."""
576
old_lines = self.get_text(old_version_id)
577
self.add_lines(new_version_id, parents, old_lines)
579
def _inclusions(self, versions):
580
"""Return set of all ancestors of given version(s)."""
581
if not len(versions):
584
for v in xrange(max(versions), 0, -1):
586
# include all its parents
587
i.update(self._parents[v])
589
## except IndexError:
590
## raise ValueError("version %d not present in weave" % v)
592
@deprecated_method(zero_eight)
593
def inclusions(self, version_ids):
594
"""Deprecated - see VersionedFile.get_ancestry for the replacement."""
597
if isinstance(version_ids[0], int):
598
return [self._idx_to_name(v) for v in self._inclusions(version_ids)]
600
return self.get_ancestry(version_ids)
602
def get_ancestry(self, version_ids):
603
"""See VersionedFile.get_ancestry."""
604
if isinstance(version_ids, basestring):
605
version_ids = [version_ids]
606
i = self._inclusions([self._lookup(v) for v in version_ids])
607
return [self._idx_to_name(v) for v in i]
609
def _check_lines(self, text):
610
if not isinstance(text, list):
611
raise ValueError("text should be a list, not %s" % type(text))
614
if not isinstance(l, basestring):
615
raise ValueError("text line should be a string or unicode, not %s"
620
def _check_versions(self, indexes):
621
"""Check everything in the sequence of indexes is valid"""
626
raise IndexError("invalid version number %r" % i)
628
def _compatible_parents(self, my_parents, other_parents):
629
"""During join check that other_parents are joinable with my_parents.
631
Joinable is defined as 'is a subset of' - supersets may require
632
regeneration of diffs, but subsets do not.
634
return len(other_parents.difference(my_parents)) == 0
636
def annotate(self, version_id):
637
if isinstance(version_id, int):
638
warn('Weave.annotate(int) is deprecated. Please use version names'
639
' in all circumstances as of 0.8',
644
for origin, lineno, text in self._extract([version_id]):
645
result.append((origin, text))
648
return super(Weave, self).annotate(version_id)
650
def annotate_iter(self, version_id):
651
"""Yield list of (version-id, line) pairs for the specified version.
653
The index indicates when the line originated in the weave."""
654
incls = [self._lookup(version_id)]
655
for origin, lineno, text in self._extract(incls):
656
yield self._idx_to_name(origin), text
658
@deprecated_method(zero_eight)
660
"""_walk has become visit, a supported api."""
661
return self._walk_internal()
663
def iter_lines_added_or_present_in_versions(self, version_ids=None):
664
"""See VersionedFile.iter_lines_added_or_present_in_versions()."""
665
if version_ids is None:
666
version_ids = self.versions()
667
version_ids = set(version_ids)
668
for lineno, inserted, deletes, line in self._walk_internal(version_ids):
669
# if inserted not in version_ids then it was inserted before the
670
# versions we care about, but because weaves cannot represent ghosts
671
# properly, we do not filter down to that
672
# if inserted not in version_ids: continue
678
#@deprecated_method(zero_eight)
679
def walk(self, version_ids=None):
680
"""See VersionedFile.walk."""
681
return self._walk_internal(version_ids)
683
def _walk_internal(self, version_ids=None):
684
"""Helper method for weave actions."""
689
lineno = 0 # line of weave, 0-based
691
for l in self._weave:
692
if l.__class__ == tuple:
696
istack.append(self._names[v])
700
assert self._names[v] not in dset
701
dset.add(self._names[v])
703
dset.remove(self._names[v])
705
raise WeaveFormatError('unexpected instruction %r' % v)
707
assert l.__class__ in (str, unicode)
709
yield lineno, istack[-1], frozenset(dset), l
713
raise WeaveFormatError("unclosed insertion blocks "
714
"at end of weave: %s" % istack)
716
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
719
def plan_merge(self, ver_a, ver_b):
720
"""Return pseudo-annotation indicating how the two versions merge.
722
This is computed between versions a and b and their common
725
Weave lines present in none of them are skipped entirely.
727
inc_a = set(self.get_ancestry([ver_a]))
728
inc_b = set(self.get_ancestry([ver_b]))
729
inc_c = inc_a & inc_b
731
for lineno, insert, deleteset, line in\
732
self.walk([ver_a, ver_b]):
733
if deleteset & inc_c:
734
# killed in parent; can't be in either a or b
735
# not relevant to our work
736
yield 'killed-base', line
737
elif insert in inc_c:
738
# was inserted in base
739
killed_a = bool(deleteset & inc_a)
740
killed_b = bool(deleteset & inc_b)
741
if killed_a and killed_b:
742
yield 'killed-both', line
744
yield 'killed-a', line
746
yield 'killed-b', line
748
yield 'unchanged', line
749
elif insert in inc_a:
750
if deleteset & inc_a:
751
yield 'ghost-a', line
755
elif insert in inc_b:
756
if deleteset & inc_b:
757
yield 'ghost-b', line
761
# not in either revision
762
yield 'irrelevant', line
764
yield 'unchanged', '' # terminator
766
def _extract(self, versions):
767
"""Yield annotation of lines in included set.
769
Yields a sequence of tuples (origin, lineno, text), where
770
origin is the origin version, lineno the index in the weave,
771
and text the text of the line.
773
The set typically but not necessarily corresponds to a version.
776
if not isinstance(i, int):
779
included = self._inclusions(versions)
785
lineno = 0 # line of weave, 0-based
791
WFE = WeaveFormatError
794
# 449 0 4474.6820 2356.5590 bzrlib.weave:556(_extract)
795
# +285282 0 1676.8040 1676.8040 +<isinstance>
796
# 1.6 seconds in 'isinstance'.
797
# changing the first isinstance:
798
# 449 0 2814.2660 1577.1760 bzrlib.weave:556(_extract)
799
# +140414 0 762.8050 762.8050 +<isinstance>
800
# note that the inline time actually dropped (less function calls)
801
# and total processing time was halved.
802
# we're still spending ~1/4 of the method in isinstance though.
803
# so lets hard code the acceptable string classes we expect:
804
# 449 0 1202.9420 786.2930 bzrlib.weave:556(_extract)
805
# +71352 0 377.5560 377.5560 +<method 'append' of 'list'
807
# yay, down to ~1/4 the initial extract time, and our inline time
808
# has shrunk again, with isinstance no longer dominating.
809
# tweaking the stack inclusion test to use a set gives:
810
# 449 0 1122.8030 713.0080 bzrlib.weave:556(_extract)
811
# +71352 0 354.9980 354.9980 +<method 'append' of 'list'
813
# - a 5% win, or possibly just noise. However with large istacks that
814
# 'in' test could dominate, so I'm leaving this change in place -
815
# when its fast enough to consider profiling big datasets we can review.
820
for l in self._weave:
821
if l.__class__ == tuple:
829
iset.remove(istack.pop())
840
assert l.__class__ in (str, unicode)
842
isactive = (not dset) and istack and (istack[-1] in included)
844
result.append((istack[-1], lineno, l))
847
raise WeaveFormatError("unclosed insertion blocks "
848
"at end of weave: %s" % istack)
850
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
854
@deprecated_method(zero_eight)
855
def get_iter(self, name_or_index):
856
"""Deprecated, please do not use. Lookups are not not needed.
858
Please use get_lines now.
860
return iter(self.get_lines(self._maybe_lookup(name_or_index)))
862
@deprecated_method(zero_eight)
863
def maybe_lookup(self, name_or_index):
864
"""Deprecated, please do not use. Lookups are not not needed."""
865
return self._maybe_lookup(name_or_index)
867
def _maybe_lookup(self, name_or_index):
868
"""Convert possible symbolic name to index, or pass through indexes.
872
if isinstance(name_or_index, (int, long)):
875
return self._lookup(name_or_index)
877
@deprecated_method(zero_eight)
878
def get(self, version_id):
879
"""Please use either Weave.get_text or Weave.get_lines as desired."""
880
return self.get_lines(version_id)
882
def get_lines(self, version_id):
883
"""See VersionedFile.get_lines()."""
884
int_index = self._maybe_lookup(version_id)
885
result = [line for (origin, lineno, line) in self._extract([int_index])]
886
expected_sha1 = self._sha1s[int_index]
887
measured_sha1 = sha_strings(result)
888
if measured_sha1 != expected_sha1:
889
raise errors.WeaveInvalidChecksum(
890
'file %s, revision %s, expected: %s, measured %s'
891
% (self._weave_name, version_id,
892
expected_sha1, measured_sha1))
895
def get_sha1(self, version_id):
896
"""See VersionedFile.get_sha1()."""
897
return self._sha1s[self._lookup(version_id)]
899
@deprecated_method(zero_eight)
900
def numversions(self):
901
"""How many versions are in this weave?
903
Deprecated in favour of num_versions.
905
return self.num_versions()
907
def num_versions(self):
908
"""How many versions are in this weave?"""
909
l = len(self._parents)
910
assert l == len(self._sha1s)
913
__len__ = num_versions
915
def check(self, progress_bar=None):
916
# TODO evaluate performance hit of using string sets in this routine.
917
# TODO: check no circular inclusions
918
# TODO: create a nested progress bar
919
for version in range(self.num_versions()):
920
inclusions = list(self._parents[version])
923
if inclusions[-1] >= version:
924
raise WeaveFormatError("invalid included version %d for index %d"
925
% (inclusions[-1], version))
927
# try extracting all versions; parallel extraction is used
928
nv = self.num_versions()
933
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
934
# The problem is that set membership is much more expensive
935
name = self._idx_to_name(i)
936
sha1s[name] = sha.new()
938
new_inc = set([name])
939
for p in self._parents[i]:
940
new_inc.update(inclusions[self._idx_to_name(p)])
942
assert set(new_inc) == set(self.get_ancestry(name)), \
943
'failed %s != %s' % (set(new_inc), set(self.get_ancestry(name)))
944
inclusions[name] = new_inc
946
nlines = len(self._weave)
948
update_text = 'checking weave'
950
short_name = os.path.basename(self._weave_name)
951
update_text = 'checking %s' % (short_name,)
952
update_text = update_text[:25]
954
for lineno, insert, deleteset, line in self._walk_internal():
956
progress_bar.update(update_text, lineno, nlines)
958
for name, name_inclusions in inclusions.items():
959
# The active inclusion must be an ancestor,
960
# and no ancestors must have deleted this line,
961
# because we don't support resurrection.
962
if (insert in name_inclusions) and not (deleteset & name_inclusions):
963
sha1s[name].update(line)
966
version = self._idx_to_name(i)
967
hd = sha1s[version].hexdigest()
968
expected = self._sha1s[i]
970
raise errors.WeaveInvalidChecksum(
971
"mismatched sha1 for version %s: "
972
"got %s, expected %s"
973
% (version, hd, expected))
975
# TODO: check insertions are properly nested, that there are
976
# no lines outside of insertion blocks, that deletions are
977
# properly paired, etc.
979
def _join(self, other, pb, msg, version_ids, ignore_missing):
980
"""Worker routine for join()."""
981
if not other.versions():
982
return # nothing to update, easy
985
# versions is never none, InterWeave checks this.
988
# two loops so that we do not change ourselves before verifying it
990
# work through in index order to make sure we get all dependencies
993
# get the selected versions only that are in other.versions.
994
version_ids = set(other.versions()).intersection(set(version_ids))
995
# pull in the referenced graph.
996
version_ids = other.get_ancestry(version_ids)
997
pending_graph = [(version, other.get_parents(version)) for
998
version in version_ids]
999
for name in topo_sort(pending_graph):
1000
other_idx = other._name_map[name]
1001
# returns True if we have it, False if we need it.
1002
if not self._check_version_consistent(other, other_idx, name):
1003
names_to_join.append((other_idx, name))
1012
for other_idx, name in names_to_join:
1013
# TODO: If all the parents of the other version are already
1014
# present then we can avoid some work by just taking the delta
1015
# and adjusting the offsets.
1016
new_parents = self._imported_parents(other, other_idx)
1017
sha1 = other._sha1s[other_idx]
1022
pb.update(msg, merged, len(names_to_join))
1024
lines = other.get_lines(other_idx)
1025
self._add(name, lines, new_parents, sha1)
1027
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
1028
merged, processed, self._weave_name, time.time()-time0))
1030
def _imported_parents(self, other, other_idx):
1031
"""Return list of parents in self corresponding to indexes in other."""
1033
for parent_idx in other._parents[other_idx]:
1034
parent_name = other._names[parent_idx]
1035
if parent_name not in self._name_map:
1036
# should not be possible
1037
raise WeaveError("missing parent {%s} of {%s} in %r"
1038
% (parent_name, other._name_map[other_idx], self))
1039
new_parents.append(self._name_map[parent_name])
1042
def _check_version_consistent(self, other, other_idx, name):
1043
"""Check if a version in consistent in this and other.
1045
To be consistent it must have:
1048
* the same direct parents (by name, not index, and disregarding
1051
If present & correct return True;
1052
if not present in self return False;
1053
if inconsistent raise error."""
1054
this_idx = self._name_map.get(name, -1)
1056
if self._sha1s[this_idx] != other._sha1s[other_idx]:
1057
raise errors.WeaveTextDiffers(name, self, other)
1058
self_parents = self._parents[this_idx]
1059
other_parents = other._parents[other_idx]
1060
n1 = set([self._names[i] for i in self_parents])
1061
n2 = set([other._names[i] for i in other_parents])
1062
if not self._compatible_parents(n1, n2):
1063
raise WeaveParentMismatch("inconsistent parents "
1064
"for version {%s}: %s vs %s" % (name, n1, n2))
1070
@deprecated_method(zero_eight)
1071
def reweave(self, other, pb=None, msg=None):
1072
"""reweave has been superceded by plain use of join."""
1073
return self.join(other, pb, msg)
1075
def _reweave(self, other, pb, msg):
1076
"""Reweave self with other - internal helper for join().
1078
:param other: The other weave to merge
1079
:param pb: An optional progress bar, indicating how far done we are
1080
:param msg: An optional message for the progress
1082
new_weave = _reweave(self, other, pb=pb, msg=msg)
1083
self._copy_weave_content(new_weave)
1085
def _copy_weave_content(self, otherweave):
1086
"""adsorb the content from otherweave."""
1087
for attr in self.__slots__:
1088
if attr != '_weave_name':
1089
setattr(self, attr, copy(getattr(otherweave, attr)))
1092
class WeaveFile(Weave):
1093
"""A WeaveFile represents a Weave on disk and writes on change."""
1095
WEAVE_SUFFIX = '.weave'
1097
def __init__(self, name, transport, filemode=None, create=False, access_mode='w'):
1098
"""Create a WeaveFile.
1100
:param create: If not True, only open an existing knit.
1102
super(WeaveFile, self).__init__(name, access_mode)
1103
self._transport = transport
1104
self._filemode = filemode
1106
_read_weave_v5(self._transport.get(name + WeaveFile.WEAVE_SUFFIX), self)
1107
except errors.NoSuchFile:
1113
def _add_lines(self, version_id, parents, lines, parent_texts):
1114
"""Add a version and save the weave."""
1115
result = super(WeaveFile, self)._add_lines(version_id, parents, lines,
1120
def _clone_text(self, new_version_id, old_version_id, parents):
1121
"""See VersionedFile.clone_text."""
1122
super(WeaveFile, self)._clone_text(new_version_id, old_version_id, parents)
1125
def copy_to(self, name, transport):
1126
"""See VersionedFile.copy_to()."""
1127
# as we are all in memory always, just serialise to the new place.
1129
write_weave_v5(self, sio)
1131
transport.put(name + WeaveFile.WEAVE_SUFFIX, sio, self._filemode)
1133
def create_empty(self, name, transport, filemode=None):
1134
return WeaveFile(name, transport, filemode, create=True)
1137
"""Save the weave."""
1138
self._check_write_ok()
1140
write_weave_v5(self, sio)
1142
self._transport.put(self._weave_name + WeaveFile.WEAVE_SUFFIX,
1148
"""See VersionedFile.get_suffixes()."""
1149
return [WeaveFile.WEAVE_SUFFIX]
1151
def join(self, other, pb=None, msg=None, version_ids=None,
1152
ignore_missing=False):
1153
"""Join other into self and save."""
1154
super(WeaveFile, self).join(other, pb, msg, version_ids, ignore_missing)
1158
@deprecated_function(zero_eight)
1159
def reweave(wa, wb, pb=None, msg=None):
1160
"""reweaving is deprecation, please just use weave.join()."""
1161
_reweave(wa, wb, pb, msg)
1163
def _reweave(wa, wb, pb=None, msg=None):
1164
"""Combine two weaves and return the result.
1166
This works even if a revision R has different parents in
1167
wa and wb. In the resulting weave all the parents are given.
1169
This is done by just building up a new weave, maintaining ordering
1170
of the versions in the two inputs. More efficient approaches
1171
might be possible but it should only be necessary to do
1172
this operation rarely, when a new previously ghost version is
1175
:param pb: An optional progress bar, indicating how far done we are
1176
:param msg: An optional message for the progress
1180
queue_a = range(wa.num_versions())
1181
queue_b = range(wb.num_versions())
1182
# first determine combined parents of all versions
1183
# map from version name -> all parent names
1184
combined_parents = _reweave_parent_graphs(wa, wb)
1185
mutter("combined parents: %r", combined_parents)
1186
order = topo_sort(combined_parents.iteritems())
1187
mutter("order to reweave: %r", order)
1192
for idx, name in enumerate(order):
1194
pb.update(msg, idx, len(order))
1195
if name in wa._name_map:
1196
lines = wa.get_lines(name)
1197
if name in wb._name_map:
1198
lines_b = wb.get_lines(name)
1199
if lines != lines_b:
1200
mutter('Weaves differ on content. rev_id {%s}', name)
1201
mutter('weaves: %s, %s', wa._weave_name, wb._weave_name)
1203
lines = list(difflib.unified_diff(lines, lines_b,
1204
wa._weave_name, wb._weave_name))
1205
mutter('lines:\n%s', ''.join(lines))
1206
raise errors.WeaveTextDiffers(name, wa, wb)
1208
lines = wb.get_lines(name)
1209
wr._add(name, lines, [wr._lookup(i) for i in combined_parents[name]])
1212
def _reweave_parent_graphs(wa, wb):
1213
"""Return combined parent ancestry for two weaves.
1215
Returned as a list of (version_name, set(parent_names))"""
1217
for weave in [wa, wb]:
1218
for idx, name in enumerate(weave._names):
1219
p = combined.setdefault(name, set())
1220
p.update(map(weave._idx_to_name, weave._parents[idx]))
1225
"""Show the weave's table-of-contents"""
1226
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
1227
for i in (6, 50, 10, 10):
1230
for i in range(w.num_versions()):
1233
parent_str = ' '.join(map(str, w._parents[i]))
1234
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
1238
def weave_stats(weave_file, pb):
1239
from bzrlib.weavefile import read_weave
1241
wf = file(weave_file, 'rb')
1242
w = read_weave(wf, WeaveVersionedFile)
1243
# FIXME: doesn't work on pipes
1244
weave_size = wf.tell()
1248
for i in range(vers):
1249
pb.update('checking sizes', i, vers)
1250
for origin, lineno, line in w._extract([i]):
1255
print 'versions %9d' % vers
1256
print 'weave file %9d bytes' % weave_size
1257
print 'total contents %9d bytes' % total
1258
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
1261
print 'average size %9d bytes' % avg
1262
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
1266
print """bzr weave tool
1268
Experimental tool for weave algorithm.
1271
weave init WEAVEFILE
1272
Create an empty weave file
1273
weave get WEAVEFILE VERSION
1274
Write out specified version.
1275
weave check WEAVEFILE
1276
Check consistency of all versions.
1278
Display table of contents.
1279
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
1280
Add NEWTEXT, with specified parent versions.
1281
weave annotate WEAVEFILE VERSION
1282
Display origin of each line.
1283
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1284
Auto-merge two versions and display conflicts.
1285
weave diff WEAVEFILE VERSION1 VERSION2
1286
Show differences between two versions.
1290
% weave init foo.weave
1292
% weave add foo.weave ver0 < foo.txt
1295
(create updated version)
1297
% weave get foo.weave 0 | diff -u - foo.txt
1298
% weave add foo.weave ver1 0 < foo.txt
1301
% weave get foo.weave 0 > foo.txt (create forked version)
1303
% weave add foo.weave ver2 0 < foo.txt
1306
% weave merge foo.weave 1 2 > foo.txt (merge them)
1307
% vi foo.txt (resolve conflicts)
1308
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
1320
# in case we're run directly from the subdirectory
1321
sys.path.append('..')
1323
from bzrlib.weavefile import write_weave, read_weave
1324
from bzrlib.progress import ProgressBar
1339
return read_weave(file(argv[2], 'rb'))
1345
# at the moment, based on everything in the file
1347
parents = map(int, argv[4:])
1348
lines = sys.stdin.readlines()
1349
ver = w.add(name, parents, lines)
1350
write_weave(w, file(argv[2], 'wb'))
1351
print 'added version %r %d' % (name, ver)
1354
if os.path.exists(fn):
1355
raise IOError("file exists")
1357
write_weave(w, file(fn, 'wb'))
1358
elif cmd == 'get': # get one version
1360
sys.stdout.writelines(w.get_iter(int(argv[3])))
1365
v1, v2 = map(int, argv[3:5])
1368
diff_gen = bzrlib.patiencediff.unified_diff(lines1, lines2,
1369
'%s version %d' % (fn, v1),
1370
'%s version %d' % (fn, v2))
1371
sys.stdout.writelines(diff_gen)
1373
elif cmd == 'annotate':
1375
# newline is added to all lines regardless; too hard to get
1376
# reasonable formatting otherwise
1378
for origin, text in w.annotate(int(argv[3])):
1379
text = text.rstrip('\r\n')
1381
print ' | %s' % (text)
1383
print '%5d | %s' % (origin, text)
1389
elif cmd == 'stats':
1390
weave_stats(argv[2], ProgressBar())
1392
elif cmd == 'check':
1397
print '%d versions ok' % w.num_versions()
1399
elif cmd == 'inclusions':
1401
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1403
elif cmd == 'parents':
1405
print ' '.join(map(str, w._parents[int(argv[3])]))
1407
elif cmd == 'plan-merge':
1408
# replaced by 'bzr weave-plan-merge'
1410
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1412
print '%14s | %s' % (state, line),
1413
elif cmd == 'merge':
1414
# replaced by 'bzr weave-merge-text'
1416
p = w.plan_merge(int(argv[3]), int(argv[4]))
1417
sys.stdout.writelines(w.weave_merge(p))
1419
raise ValueError('unknown command %r' % cmd)
1423
def profile_main(argv):
1424
import tempfile, hotshot, hotshot.stats
1426
prof_f = tempfile.NamedTemporaryFile()
1428
prof = hotshot.Profile(prof_f.name)
1430
ret = prof.runcall(main, argv)
1433
stats = hotshot.stats.load(prof_f.name)
1435
stats.sort_stats('cumulative')
1436
## XXX: Might like to write to stderr or the trace file instead but
1437
## print_stats seems hardcoded to stdout
1438
stats.print_stats(20)
1443
def lsprofile_main(argv):
1444
from bzrlib.lsprof import profile
1445
ret,stats = profile(main, argv)
1451
if __name__ == '__main__':
1453
if '--profile' in sys.argv:
1455
args.remove('--profile')
1456
sys.exit(profile_main(args))
1457
elif '--lsprof' in sys.argv:
1459
args.remove('--lsprof')
1460
sys.exit(lsprofile_main(args))
1462
sys.exit(main(sys.argv))
1465
class InterWeave(InterVersionedFile):
1466
"""Optimised code paths for weave to weave operations."""
1468
_matching_file_from_factory = staticmethod(WeaveFile)
1469
_matching_file_to_factory = staticmethod(WeaveFile)
1472
def is_compatible(source, target):
1473
"""Be compatible with weaves."""
1475
return (isinstance(source, Weave) and
1476
isinstance(target, Weave))
1477
except AttributeError:
1480
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
1481
"""See InterVersionedFile.join."""
1482
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
1483
if self.target.versions() == [] and version_ids is None:
1484
self.target._copy_weave_content(self.source)
1487
self.target._join(self.source, pb, msg, version_ids, ignore_missing)
1488
except errors.WeaveParentMismatch:
1489
self.target._reweave(self.source, pb, msg)
1492
InterVersionedFile.register_optimiser(InterWeave)
added
removed
bzrlib/weave.py
