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"""
24
# before intset (r923) 2000 versions in 41.5s
25
# with intset (r926) 2000 versions in 93s !!!
26
# better to just use plain sets.
28
# making _extract build and return a list, rather than being a generator
31
# with python -O, r923 does 2000 versions in 36.87s
33
# with optimizations to avoid mutating lists - 35.75! I guess copying
34
# all the elements every time costs more than the small manipulations.
35
# a surprisingly small change.
37
# r931, which avoids using a generator for extract, does 36.98s
39
# with memoized inclusions, takes 41.49s; not very good
41
# with slots, takes 37.35s; without takes 39.16, a bit surprising
43
# with the delta calculation mixed in with the add method, rather than
44
# separated, takes 36.78s
46
# with delta folded in and mutation of the list, 36.13s
48
# with all this and simplification of add code, 33s
54
# TODO: Perhaps have copy method for Weave instances?
56
# XXX: If we do weaves this way, will a merge still behave the same
57
# way if it's done in a different order? That's a pretty desirable
60
# TODO: Nothing here so far assumes the lines are really \n newlines,
61
# rather than being split up in some other way. We could accomodate
62
# binaries, perhaps by naively splitting on \n or perhaps using
63
# something like a rolling checksum.
65
# TODO: End marker for each version so we can stop reading?
67
# TODO: Check that no insertion occurs inside a deletion that was
68
# active in the version of the insertion.
70
# TODO: In addition to the SHA-1 check, perhaps have some code that
71
# checks structural constraints of the weave: ie that insertions are
72
# properly nested, that there is no text outside of an insertion, that
73
# insertions or deletions are not repeated, etc.
75
# TODO: Parallel-extract that passes back each line along with a
76
# description of which revisions include it. Nice for checking all
79
# TODO: Using a single _extract routine and then processing the output
80
# is probably inefficient. It's simple enough that we can afford to
81
# have slight specializations for different ways its used: annotate,
82
# basis for add, get, etc.
84
# TODO: Perhaps the API should work only in names to hide the integer
85
# indexes from the user?
87
# TODO: Is there any potential performance win by having an add()
88
# variant that is passed a pre-cooked version of the single basis
94
from difflib import SequenceMatcher
96
from cStringIO import StringIO
98
from bzrlib.osutils import sha_strings
101
class WeaveError(Exception):
102
"""Exception in processing weave"""
105
class WeaveFormatError(WeaveError):
106
"""Weave invariant violated"""
110
"""weave - versioned text file storage.
112
A Weave manages versions of line-based text files, keeping track
113
of the originating version for each line.
115
To clients the "lines" of the file are represented as a list of strings.
116
These strings will typically have terminal newline characters, but
117
this is not required. In particular files commonly do not have a newline
118
at the end of the file.
120
Texts can be identified in either of two ways:
122
* a nonnegative index number.
124
* a version-id string.
126
Typically the index number will be valid only inside this weave and
127
the version-id is used to reference it in the larger world.
129
The weave is represented as a list mixing edit instructions and
130
literal text. Each entry in _weave can be either a string (or
131
unicode), or a tuple. If a string, it means that the given line
132
should be output in the currently active revisions.
134
If a tuple, it gives a processing instruction saying in which
135
revisions the enclosed lines are active. The tuple has the form
136
(instruction, version).
138
The instruction can be '{' or '}' for an insertion block, and '['
139
and ']' for a deletion block respectively. The version is the
140
integer version index. There is no replace operator, only deletes
141
and inserts. For '}', the end of an insertion, there is no
142
version parameter because it always closes the most recently
147
* A later version can delete lines that were introduced by any
148
number of ancestor versions; this implies that deletion
149
instructions can span insertion blocks without regard to the
150
insertion block's nesting.
152
* Similarly, deletions need not be properly nested with regard to
153
each other, because they might have been generated by
154
independent revisions.
156
* Insertions are always made by inserting a new bracketed block
157
into a single point in the previous weave. This implies they
158
can nest but not overlap, and the nesting must always have later
159
insertions on the inside.
161
* It doesn't seem very useful to have an active insertion
162
inside an inactive insertion, but it might happen.
164
* Therefore, all instructions are always"considered"; that
165
is passed onto and off the stack. An outer inactive block
166
doesn't disable an inner block.
168
* Lines are enabled if the most recent enclosing insertion is
169
active and none of the enclosing deletions are active.
171
* There is no point having a deletion directly inside its own
172
insertion; you might as well just not write it. And there
173
should be no way to get an earlier version deleting a later
177
Text of the weave; list of control instruction tuples and strings.
180
List of parents, indexed by version number.
181
It is only necessary to store the minimal set of parents for
182
each version; the parent's parents are implied.
185
List of hex SHA-1 of each version.
188
List of symbolic names for each version. Each should be unique.
191
For each name, the version number.
194
Descriptive name of this weave; typically the filename if known.
198
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
201
def __init__(self, weave_name=None):
207
self._weave_name = 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
218
def __ne__(self, other):
219
return not self.__eq__(other)
222
def maybe_lookup(self, name_or_index):
223
"""Convert possible symbolic name to index, or pass through indexes."""
224
if isinstance(name_or_index, (int, long)):
227
return self.lookup(name_or_index)
230
def lookup(self, name):
231
"""Convert symbolic version name to index."""
233
return self._name_map[name]
235
raise WeaveError("name %r not present in weave %r" %
236
(name, self._weave_name))
239
def idx_to_name(self, version):
240
return self._names[version]
243
def _check_repeated_add(self, name, parents, text, sha1):
244
"""Check that a duplicated add is OK.
246
If it is, return the (old) index; otherwise raise an exception.
248
idx = self.lookup(name)
249
if sorted(self._parents[idx]) != sorted(parents):
250
raise WeaveError("name \"%s\" already present in weave "
251
"with different parents" % name)
252
if sha1 != self._sha1s[idx]:
253
raise WeaveError("name \"%s\" already present in weave "
254
"with different text" % name)
259
def add(self, name, parents, text, sha1=None):
260
"""Add a single text on top of the weave.
262
Returns the index number of the newly added version.
265
Symbolic name for this version.
266
(Typically the revision-id of the revision that added it.)
269
List or set of direct parent version numbers.
272
Sequence of lines to be added in the new version.
274
sha -- SHA-1 of the file, if known. This is trusted to be
278
assert isinstance(name, basestring)
280
sha1 = sha_strings(text)
281
if name in self._name_map:
282
return self._check_repeated_add(name, parents, text, sha1)
284
parents = map(self.maybe_lookup, parents)
285
self._check_versions(parents)
286
## self._check_lines(text)
287
new_version = len(self._parents)
290
# if we abort after here the (in-memory) weave will be corrupt because only
291
# some fields are updated
292
self._parents.append(parents[:])
293
self._sha1s.append(sha1)
294
self._names.append(name)
295
self._name_map[name] = new_version
299
# special case; adding with no parents revision; can do
300
# this more quickly by just appending unconditionally.
301
# even more specially, if we're adding an empty text we
302
# need do nothing at all.
304
self._weave.append(('{', new_version))
305
self._weave.extend(text)
306
self._weave.append(('}', None))
310
if len(parents) == 1:
311
pv = list(parents)[0]
312
if sha1 == self._sha1s[pv]:
313
# special case: same as the single parent
317
ancestors = self.inclusions(parents)
321
# basis a list of (origin, lineno, line)
324
for origin, lineno, line in self._extract(ancestors):
325
basis_lineno.append(lineno)
326
basis_lines.append(line)
328
# another small special case: a merge, producing the same text
330
if text == basis_lines:
333
# add a sentinal, because we can also match against the final line
334
basis_lineno.append(len(self._weave))
336
# XXX: which line of the weave should we really consider
337
# matches the end of the file? the current code says it's the
338
# last line of the weave?
340
#print 'basis_lines:', basis_lines
341
#print 'new_lines: ', lines
343
s = SequenceMatcher(None, basis_lines, text)
345
# offset gives the number of lines that have been inserted
346
# into the weave up to the current point; if the original edit instruction
347
# says to change line A then we actually change (A+offset)
350
for tag, i1, i2, j1, j2 in s.get_opcodes():
351
# i1,i2 are given in offsets within basis_lines; we need to map them
352
# back to offsets within the entire weave
353
#print 'raw match', tag, i1, i2, j1, j2
357
i1 = basis_lineno[i1]
358
i2 = basis_lineno[i2]
360
assert 0 <= j1 <= j2 <= len(text)
362
#print tag, i1, i2, j1, j2
364
# the deletion and insertion are handled separately.
365
# first delete the region.
367
self._weave.insert(i1+offset, ('[', new_version))
368
self._weave.insert(i2+offset+1, (']', new_version))
372
# there may have been a deletion spanning up to
373
# i2; we want to insert after this region to make sure
374
# we don't destroy ourselves
376
self._weave[i:i] = ([('{', new_version)]
379
offset += 2 + (j2 - j1)
384
def inclusions(self, versions):
385
"""Return set of all ancestors of given version(s)."""
387
for v in xrange(max(versions), 0, -1):
389
# include all its parents
390
i.update(self._parents[v])
392
## except IndexError:
393
## raise ValueError("version %d not present in weave" % v)
396
def parents(self, version):
397
return self._parents[version]
400
def minimal_parents(self, version):
401
"""Find the minimal set of parents for the version."""
402
included = self._parents[version]
407
li.sort(reverse=True)
415
gotit.update(self.inclusions(pv))
417
assert mininc[0] >= 0
418
assert mininc[-1] < version
423
def _check_lines(self, text):
424
if not isinstance(text, list):
425
raise ValueError("text should be a list, not %s" % type(text))
428
if not isinstance(l, basestring):
429
raise ValueError("text line should be a string or unicode, not %s"
434
def _check_versions(self, indexes):
435
"""Check everything in the sequence of indexes is valid"""
440
raise IndexError("invalid version number %r" % i)
443
def annotate(self, name_or_index):
444
return list(self.annotate_iter(name_or_index))
447
def annotate_iter(self, name_or_index):
448
"""Yield list of (index-id, line) pairs for the specified version.
450
The index indicates when the line originated in the weave."""
451
incls = [self.maybe_lookup(name_or_index)]
452
for origin, lineno, text in self._extract(incls):
460
(lineno, insert, deletes, text)
461
for each literal line.
467
lineno = 0 # line of weave, 0-based
469
for l in self._weave:
470
if isinstance(l, tuple):
483
raise WeaveFormatError('unexpected instruction %r'
486
assert isinstance(l, basestring)
488
yield lineno, istack[-1], dset, l
493
def _extract(self, versions):
494
"""Yield annotation of lines in included set.
496
Yields a sequence of tuples (origin, lineno, text), where
497
origin is the origin version, lineno the index in the weave,
498
and text the text of the line.
500
The set typically but not necessarily corresponds to a version.
503
if not isinstance(i, int):
506
included = self.inclusions(versions)
511
lineno = 0 # line of weave, 0-based
517
WFE = WeaveFormatError
519
for l in self._weave:
520
if isinstance(l, tuple):
524
assert v not in istack
538
assert isinstance(l, basestring)
540
isactive = (not dset) and istack and (istack[-1] in included)
542
result.append((istack[-1], lineno, l))
546
raise WFE("unclosed insertion blocks at end of weave",
549
raise WFE("unclosed deletion blocks at end of weave",
556
def get_iter(self, name_or_index):
557
"""Yield lines for the specified version."""
558
incls = [self.maybe_lookup(name_or_index)]
559
for origin, lineno, line in self._extract(incls):
563
def get_text(self, version):
564
assert isinstance(version, int)
566
s.writelines(self.get_iter(version))
570
def get(self, name_or_index):
571
return list(self.get_iter(name_or_index))
574
def mash_iter(self, included):
575
"""Return composed version of multiple included versions."""
576
for origin, lineno, text in self._extract(included):
580
def dump(self, to_file):
581
from pprint import pprint
582
print >>to_file, "Weave._weave = ",
583
pprint(self._weave, to_file)
584
print >>to_file, "Weave._parents = ",
585
pprint(self._parents, to_file)
589
def numversions(self):
590
l = len(self._parents)
591
assert l == len(self._sha1s)
596
return self.numversions()
599
def check(self, progress_bar=None):
600
# check no circular inclusions
601
for version in range(self.numversions()):
602
inclusions = list(self._parents[version])
605
if inclusions[-1] >= version:
606
raise WeaveFormatError("invalid included version %d for index %d"
607
% (inclusions[-1], version))
609
# try extracting all versions; this is a bit slow and parallel
610
# extraction could be used
611
nv = self.numversions()
612
for version in range(nv):
614
progress_bar.update('checking text', version, nv)
616
for l in self.get_iter(version):
619
expected = self._sha1s[version]
621
raise WeaveError("mismatched sha1 for version %d; "
622
"got %s, expected %s"
623
% (version, hd, expected))
625
# TODO: check insertions are properly nested, that there are
626
# no lines outside of insertion blocks, that deletions are
627
# properly paired, etc.
631
def merge(self, merge_versions):
632
"""Automerge and mark conflicts between versions.
634
This returns a sequence, each entry describing alternatives
635
for a chunk of the file. Each of the alternatives is given as
638
If there is a chunk of the file where there's no diagreement,
639
only one alternative is given.
642
# approach: find the included versions common to all the
644
raise NotImplementedError()
648
def _delta(self, included, lines):
649
"""Return changes from basis to new revision.
651
The old text for comparison is the union of included revisions.
653
This is used in inserting a new text.
655
Delta is returned as a sequence of
656
(weave1, weave2, newlines).
658
This indicates that weave1:weave2 of the old weave should be
659
replaced by the sequence of lines in newlines. Note that
660
these line numbers are positions in the total weave and don't
661
correspond to the lines in any extracted version, or even the
662
extracted union of included versions.
664
If line1=line2, this is a pure insert; if newlines=[] this is a
665
pure delete. (Similar to difflib.)
670
def plan_merge(self, ver_a, ver_b):
671
"""Return pseudo-annotation indicating how the two versions merge.
673
This is computed between versions a and b and their common
676
Weave lines present in none of them are skipped entirely.
678
inc_a = self.inclusions([ver_a])
679
inc_b = self.inclusions([ver_b])
680
inc_c = inc_a & inc_b
682
for lineno, insert, deleteset, line in self._walk():
683
if deleteset & inc_c:
684
# killed in parent; can't be in either a or b
685
# not relevant to our work
686
yield 'killed-base', line
687
elif insert in inc_c:
688
# was inserted in base
689
killed_a = bool(deleteset & inc_a)
690
killed_b = bool(deleteset & inc_b)
691
if killed_a and killed_b:
692
yield 'killed-both', line
694
yield 'killed-a', line
696
yield 'killed-b', line
698
yield 'unchanged', line
699
elif insert in inc_a:
700
if deleteset & inc_a:
701
yield 'ghost-a', line
705
elif insert in inc_b:
706
if deleteset & inc_b:
707
yield 'ghost-b', line
711
# not in either revision
712
yield 'irrelevant', line
714
yield 'unchanged', '' # terminator
718
def weave_merge(self, plan):
723
for state, line in plan:
724
if state == 'unchanged' or state == 'killed-both':
725
# resync and flush queued conflicts changes if any
726
if not lines_a and not lines_b:
728
elif ch_a and not ch_b:
730
for l in lines_a: yield l
731
elif ch_b and not ch_a:
732
for l in lines_b: yield l
733
elif lines_a == lines_b:
734
for l in lines_a: yield l
737
for l in lines_a: yield l
739
for l in lines_b: yield l
746
if state == 'unchanged':
749
elif state == 'killed-a':
752
elif state == 'killed-b':
755
elif state == 'new-a':
758
elif state == 'new-b':
762
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
773
"""Show the weave's table-of-contents"""
774
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
775
for i in (6, 50, 10, 10):
778
for i in range(w.numversions()):
781
parent_str = ' '.join(map(str, w._parents[i]))
782
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
786
def weave_stats(weave_file):
787
from bzrlib.progress import ProgressBar
788
from bzrlib.weavefile import read_weave
792
wf = file(weave_file, 'rb')
794
# FIXME: doesn't work on pipes
795
weave_size = wf.tell()
799
for i in range(vers):
800
pb.update('checking sizes', i, vers)
801
for line in w.get_iter(i):
806
print 'versions %9d' % vers
807
print 'weave file %9d bytes' % weave_size
808
print 'total contents %9d bytes' % total
809
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
812
print 'average size %9d bytes' % avg
813
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
817
print """bzr weave tool
819
Experimental tool for weave algorithm.
823
Create an empty weave file
824
weave get WEAVEFILE VERSION
825
Write out specified version.
826
weave check WEAVEFILE
827
Check consistency of all versions.
829
Display table of contents.
830
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
831
Add NEWTEXT, with specified parent versions.
832
weave annotate WEAVEFILE VERSION
833
Display origin of each line.
834
weave mash WEAVEFILE VERSION...
835
Display composite of all selected versions.
836
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
837
Auto-merge two versions and display conflicts.
838
weave diff WEAVEFILE VERSION1 VERSION2
839
Show differences between two versions.
843
% weave init foo.weave
845
% weave add foo.weave ver0 < foo.txt
848
(create updated version)
850
% weave get foo.weave 0 | diff -u - foo.txt
851
% weave add foo.weave ver1 0 < foo.txt
854
% weave get foo.weave 0 > foo.txt (create forked version)
856
% weave add foo.weave ver2 0 < foo.txt
859
% weave merge foo.weave 1 2 > foo.txt (merge them)
860
% vi foo.txt (resolve conflicts)
861
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
870
from weavefile import write_weave, read_weave
871
from bzrlib.progress import ProgressBar
886
return read_weave(file(argv[2], 'rb'))
892
# at the moment, based on everything in the file
894
parents = map(int, argv[4:])
895
lines = sys.stdin.readlines()
896
ver = w.add(name, parents, lines)
897
write_weave(w, file(argv[2], 'wb'))
898
print 'added version %r %d' % (name, ver)
901
if os.path.exists(fn):
902
raise IOError("file exists")
904
write_weave(w, file(fn, 'wb'))
905
elif cmd == 'get': # get one version
907
sys.stdout.writelines(w.get_iter(int(argv[3])))
909
elif cmd == 'mash': # get composite
911
sys.stdout.writelines(w.mash_iter(map(int, argv[3:])))
914
from difflib import unified_diff
917
v1, v2 = map(int, argv[3:5])
920
diff_gen = unified_diff(lines1, lines2,
921
'%s version %d' % (fn, v1),
922
'%s version %d' % (fn, v2))
923
sys.stdout.writelines(diff_gen)
925
elif cmd == 'annotate':
927
# newline is added to all lines regardless; too hard to get
928
# reasonable formatting otherwise
930
for origin, text in w.annotate(int(argv[3])):
931
text = text.rstrip('\r\n')
933
print ' | %s' % (text)
935
print '%5d | %s' % (origin, text)
949
print '%d versions ok' % w.numversions()
951
elif cmd == 'inclusions':
953
print ' '.join(map(str, w.inclusions([int(argv[3])])))
955
elif cmd == 'parents':
957
print ' '.join(map(str, w._parents[int(argv[3])]))
959
elif cmd == 'plan-merge':
961
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
963
print '%14s | %s' % (state, line),
967
p = w.plan_merge(int(argv[3]), int(argv[4]))
968
sys.stdout.writelines(w.weave_merge(p))
970
elif cmd == 'mash-merge':
976
v1, v2 = map(int, argv[3:5])
978
basis = w.inclusions([v1]).intersection(w.inclusions([v2]))
980
base_lines = list(w.mash_iter(basis))
981
a_lines = list(w.get(v1))
982
b_lines = list(w.get(v2))
984
from bzrlib.merge3 import Merge3
985
m3 = Merge3(base_lines, a_lines, b_lines)
987
name_a = 'version %d' % v1
988
name_b = 'version %d' % v2
989
sys.stdout.writelines(m3.merge_lines(name_a=name_a, name_b=name_b))
991
raise ValueError('unknown command %r' % cmd)
995
def profile_main(argv):
996
import tempfile, hotshot, hotshot.stats
998
prof_f = tempfile.NamedTemporaryFile()
1000
prof = hotshot.Profile(prof_f.name)
1002
ret = prof.runcall(main, argv)
1005
stats = hotshot.stats.load(prof_f.name)
1007
stats.sort_stats('cumulative')
1008
## XXX: Might like to write to stderr or the trace file instead but
1009
## print_stats seems hardcoded to stdout
1010
stats.print_stats(20)
1015
if __name__ == '__main__':
1017
if '--profile' in sys.argv:
1019
args.remove('--profile')
1020
sys.exit(profile_main(args))
1022
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
