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
51
# TODO: Perhaps have copy method for Weave instances?
53
# XXX: If we do weaves this way, will a merge still behave the same
54
# way if it's done in a different order? That's a pretty desirable
57
# TODO: Nothing here so far assumes the lines are really \n newlines,
58
# rather than being split up in some other way. We could accomodate
59
# binaries, perhaps by naively splitting on \n or perhaps using
60
# something like a rolling checksum.
62
# TODO: Track version names as well as indexes.
64
# TODO: End marker for each version so we can stop reading?
66
# TODO: Check that no insertion occurs inside a deletion that was
67
# active in the version of the insertion.
69
# TODO: In addition to the SHA-1 check, perhaps have some code that
70
# checks structural constraints of the weave: ie that insertions are
71
# properly nested, that there is no text outside of an insertion, that
72
# insertions or deletions are not repeated, etc.
74
# TODO: Parallel-extract that passes back each line along with a
75
# description of which revisions include it. Nice for checking all
81
class WeaveError(Exception):
82
"""Exception in processing weave"""
85
class WeaveFormatError(WeaveError):
86
"""Weave invariant violated"""
90
"""weave - versioned text file storage.
92
A Weave manages versions of line-based text files, keeping track
93
of the originating version for each line.
95
To clients the "lines" of the file are represented as a list of strings.
96
These strings will typically have terminal newline characters, but
97
this is not required. In particular files commonly do not have a newline
98
at the end of the file.
100
Texts can be identified in either of two ways:
102
* a nonnegative index number.
104
* a version-id string. (not implemented yet)
106
Typically the index number will be valid only inside this weave and
107
the version-id is used to reference it in the larger world.
109
The weave is represented as a list mixing edit instructions and
110
literal text. Each entry in _weave can be either a string (or
111
unicode), or a tuple. If a string, it means that the given line
112
should be output in the currently active revisions.
114
If a tuple, it gives a processing instruction saying in which
115
revisions the enclosed lines are active. The tuple has the form
116
(instruction, version).
118
The instruction can be '{' or '}' for an insertion block, and '['
119
and ']' for a deletion block respectively. The version is the
120
integer version index. There is no replace operator, only deletes
121
and inserts. For '}', the end of an insertion, there is no
122
version parameter because it always closes the most recently
127
* A later version can delete lines that were introduced by any
128
number of ancestor versions; this implies that deletion
129
instructions can span insertion blocks without regard to the
130
insertion block's nesting.
132
* Similarly, deletions need not be properly nested with regard to
133
each other, because they might have been generated by
134
independent revisions.
136
* Insertions are always made by inserting a new bracketed block
137
into a single point in the previous weave. This implies they
138
can nest but not overlap, and the nesting must always have later
139
insertions on the inside.
141
* It doesn't seem very useful to have an active insertion
142
inside an inactive insertion, but it might happen.
144
* Therefore, all instructions are always"considered"; that
145
is passed onto and off the stack. An outer inactive block
146
doesn't disable an inner block.
148
* Lines are enabled if the most recent enclosing insertion is
149
active and none of the enclosing deletions are active.
151
* There is no point having a deletion directly inside its own
152
insertion; you might as well just not write it. And there
153
should be no way to get an earlier version deleting a later
157
Text of the weave; list of control instruction tuples and strings.
160
List of parents, indexed by version number.
161
It is only necessary to store the minimal set of parents for
162
each version; the parent's parents are implied.
165
List of hex SHA-1 of each version, or None if not recorded.
168
__slots__ = ['_weave', '_parents', '_sha1s']
176
def __eq__(self, other):
177
if not isinstance(other, Weave):
179
return self._parents == other._parents \
180
and self._weave == other._weave
183
def __ne__(self, other):
184
return not self.__eq__(other)
187
def add(self, parents, text):
188
"""Add a single text on top of the weave.
190
Returns the index number of the newly added version.
193
List or set of direct parent version numbers.
196
Sequence of lines to be added in the new version."""
198
self._check_versions(parents)
199
## self._check_lines(text)
200
new_version = len(self._parents)
208
# if we abort after here the weave will be corrupt
209
self._parents.append(frozenset(parents))
210
self._sha1s.append(sha1)
214
# special case; adding with no parents revision; can do
215
# this more quickly by just appending unconditionally.
216
# even more specially, if we're adding an empty text we
217
# need do nothing at all.
219
self._weave.append(('{', new_version))
220
self._weave.extend(text)
221
self._weave.append(('}', None))
225
if len(parents) == 1:
226
pv = list(parents)[0]
227
if sha1 == self._sha1s[pv]:
228
# special case: same as the single parent
232
ancestors = self.inclusions(parents)
236
# basis a list of (origin, lineno, line)
239
for origin, lineno, line in self._extract(ancestors):
240
basis_lineno.append(lineno)
241
basis_lines.append(line)
243
# another small special case: a merge, producing the same text
245
if text == basis_lines:
248
# add a sentinal, because we can also match against the final line
249
basis_lineno.append(len(self._weave))
251
# XXX: which line of the weave should we really consider
252
# matches the end of the file? the current code says it's the
253
# last line of the weave?
255
#print 'basis_lines:', basis_lines
256
#print 'new_lines: ', lines
258
from difflib import SequenceMatcher
259
s = SequenceMatcher(None, basis_lines, text)
261
# offset gives the number of lines that have been inserted
262
# into the weave up to the current point; if the original edit instruction
263
# says to change line A then we actually change (A+offset)
266
for tag, i1, i2, j1, j2 in s.get_opcodes():
267
# i1,i2 are given in offsets within basis_lines; we need to map them
268
# back to offsets within the entire weave
269
#print 'raw match', tag, i1, i2, j1, j2
273
i1 = basis_lineno[i1]
274
i2 = basis_lineno[i2]
276
assert 0 <= j1 <= j2 <= len(text)
278
#print tag, i1, i2, j1, j2
280
# the deletion and insertion are handled separately.
281
# first delete the region.
283
self._weave.insert(i1+offset, ('[', new_version))
284
self._weave.insert(i2+offset+1, (']', new_version))
288
# there may have been a deletion spanning up to
289
# i2; we want to insert after this region to make sure
290
# we don't destroy ourselves
292
self._weave[i:i] = ([('{', new_version)]
295
offset += 2 + (j2 - j1)
300
def inclusions(self, versions):
301
"""Return set of all ancestors of given version(s)."""
307
# include all its parents
308
i.update(self._parents[v])
312
raise ValueError("version %d not present in weave" % v)
315
def minimal_parents(self, version):
316
"""Find the minimal set of parents for the version."""
317
included = self._parents[version]
322
li.sort(reverse=True)
330
gotit.update(self.inclusions(pv))
332
assert mininc[0] >= 0
333
assert mininc[-1] < version
338
def _check_lines(self, text):
339
if not isinstance(text, list):
340
raise ValueError("text should be a list, not %s" % type(text))
343
if not isinstance(l, basestring):
344
raise ValueError("text line should be a string or unicode, not %s"
349
def _check_versions(self, indexes):
350
"""Check everything in the sequence of indexes is valid"""
355
raise IndexError("invalid version number %r" % i)
358
def annotate(self, index):
359
return list(self.annotate_iter(index))
362
def annotate_iter(self, version):
363
"""Yield list of (index-id, line) pairs for the specified version.
365
The index indicates when the line originated in the weave."""
366
for origin, lineno, text in self._extract([version]):
374
(lineno, insert, deletes, text)
375
for each literal line.
381
lineno = 0 # line of weave, 0-based
383
for l in self._weave:
384
if isinstance(l, tuple):
397
raise WeaveFormatError('unexpected instruction %r'
400
assert isinstance(l, basestring)
402
yield lineno, istack[-1], dset, l
407
def _extract(self, versions):
408
"""Yield annotation of lines in included set.
410
Yields a sequence of tuples (origin, lineno, text), where
411
origin is the origin version, lineno the index in the weave,
412
and text the text of the line.
414
The set typically but not necessarily corresponds to a version.
416
included = self.inclusions(versions)
421
lineno = 0 # line of weave, 0-based
427
WFE = WeaveFormatError
429
for l in self._weave:
430
if isinstance(l, tuple):
434
assert v not in istack
448
assert isinstance(l, basestring)
450
isactive = (not dset) and istack and (istack[-1] in included)
452
result.append((istack[-1], lineno, l))
456
raise WFE("unclosed insertion blocks at end of weave",
459
raise WFE("unclosed deletion blocks at end of weave",
466
def get_iter(self, version):
467
"""Yield lines for the specified version."""
468
for origin, lineno, line in self._extract([version]):
472
def get(self, index):
473
return list(self.get_iter(index))
476
def mash_iter(self, included):
477
"""Return composed version of multiple included versions."""
478
for origin, lineno, text in self._extract(included):
482
def dump(self, to_file):
483
from pprint import pprint
484
print >>to_file, "Weave._weave = ",
485
pprint(self._weave, to_file)
486
print >>to_file, "Weave._parents = ",
487
pprint(self._parents, to_file)
491
def numversions(self):
492
l = len(self._parents)
493
assert l == len(self._sha1s)
498
return self.numversions()
501
def check(self, progress_bar=None):
502
# check no circular inclusions
503
for version in range(self.numversions()):
504
inclusions = list(self._parents[version])
507
if inclusions[-1] >= version:
508
raise WeaveFormatError("invalid included version %d for index %d"
509
% (inclusions[-1], version))
511
# try extracting all versions; this is a bit slow and parallel
512
# extraction could be used
514
nv = self.numversions()
515
for version in range(nv):
517
progress_bar.update('checking text', version, nv)
519
for l in self.get_iter(version):
522
expected = self._sha1s[version]
524
raise WeaveError("mismatched sha1 for version %d; "
525
"got %s, expected %s"
526
% (version, hd, expected))
528
# TODO: check insertions are properly nested, that there are
529
# no lines outside of insertion blocks, that deletions are
530
# properly paired, etc.
534
def merge(self, merge_versions):
535
"""Automerge and mark conflicts between versions.
537
This returns a sequence, each entry describing alternatives
538
for a chunk of the file. Each of the alternatives is given as
541
If there is a chunk of the file where there's no diagreement,
542
only one alternative is given.
545
# approach: find the included versions common to all the
547
raise NotImplementedError()
551
def _delta(self, included, lines):
552
"""Return changes from basis to new revision.
554
The old text for comparison is the union of included revisions.
556
This is used in inserting a new text.
558
Delta is returned as a sequence of
559
(weave1, weave2, newlines).
561
This indicates that weave1:weave2 of the old weave should be
562
replaced by the sequence of lines in newlines. Note that
563
these line numbers are positions in the total weave and don't
564
correspond to the lines in any extracted version, or even the
565
extracted union of included versions.
567
If line1=line2, this is a pure insert; if newlines=[] this is a
568
pure delete. (Similar to difflib.)
573
def plan_merge(self, ver_a, ver_b):
574
"""Return pseudo-annotation indicating how the two versions merge.
576
This is computed between versions a and b and their common
579
Weave lines present in none of them are skipped entirely.
581
inc_a = self.inclusions([ver_a])
582
inc_b = self.inclusions([ver_b])
583
inc_c = inc_a & inc_b
585
for lineno, insert, deleteset, line in self._walk():
586
if deleteset & inc_c:
587
# killed in parent; can't be in either a or b
588
# not relevant to our work
589
yield 'killed-base', line
590
elif insert in inc_c:
591
# was inserted in base
592
killed_a = bool(deleteset & inc_a)
593
killed_b = bool(deleteset & inc_b)
594
if killed_a and killed_b:
595
yield 'killed-both', line
597
yield 'killed-a', line
599
yield 'killed-b', line
601
yield 'unchanged', line
602
elif insert in inc_a:
603
if deleteset & inc_a:
604
yield 'ghost-a', line
608
elif insert in inc_b:
609
if deleteset & inc_b:
610
yield 'ghost-b', line
614
# not in either revision
615
yield 'irrelevant', line
617
yield 'unchanged', '' # terminator
621
def weave_merge(self, plan):
626
for state, line in plan:
627
if state == 'unchanged' or state == 'killed-both':
628
# resync and flush queued conflicts changes if any
629
if not lines_a and not lines_b:
631
elif ch_a and not ch_b:
633
for l in lines_a: yield l
634
elif ch_b and not ch_a:
635
for l in lines_b: yield l
636
elif lines_a == lines_b:
637
for l in lines_a: yield l
640
for l in lines_a: yield l
642
for l in lines_b: yield l
649
if state == 'unchanged':
652
elif state == 'killed-a':
655
elif state == 'killed-b':
658
elif state == 'new-a':
661
elif state == 'new-b':
665
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
676
"""Show some text information about the weave."""
677
print '%6s %40s %20s' % ('ver', 'sha1', 'parents')
678
for i in (6, 40, 20):
681
for i in range(w.numversions()):
683
print '%6d %40s %s' % (i, sha1, ' '.join(map(str, w._parents[i])))
687
def weave_stats(weave_file):
688
from bzrlib.progress import ProgressBar
689
from bzrlib.weavefile import read_weave
693
wf = file(weave_file, 'rb')
695
# FIXME: doesn't work on pipes
696
weave_size = wf.tell()
700
for i in range(vers):
701
pb.update('checking sizes', i, vers)
702
for line in w.get_iter(i):
707
print 'versions %9d' % vers
708
print 'weave file %9d bytes' % weave_size
709
print 'total contents %9d bytes' % total
710
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
713
print 'average size %9d bytes' % avg
714
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
718
print """bzr weave tool
720
Experimental tool for weave algorithm.
724
Create an empty weave file
725
weave get WEAVEFILE VERSION
726
Write out specified version.
727
weave check WEAVEFILE
728
Check consistency of all versions.
730
Display table of contents.
731
weave add WEAVEFILE [BASE...] < NEWTEXT
732
Add NEWTEXT, with specified parent versions.
733
weave annotate WEAVEFILE VERSION
734
Display origin of each line.
735
weave mash WEAVEFILE VERSION...
736
Display composite of all selected versions.
737
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
738
Auto-merge two versions and display conflicts.
742
% weave init foo.weave
744
% weave add foo.weave < foo.txt
747
(create updated version)
749
% weave get foo.weave 0 | diff -u - foo.txt
750
% weave add foo.weave 0 < foo.txt
753
% weave get foo.weave 0 > foo.txt (create forked version)
755
% weave add foo.weave 0 < foo.txt
758
% weave merge foo.weave 1 2 > foo.txt (merge them)
759
% vi foo.txt (resolve conflicts)
760
% weave add foo.weave 1 2 < foo.txt (commit merged version)
769
from weavefile import write_weave, read_weave
770
from bzrlib.progress import ProgressBar
778
return read_weave(file(argv[2], 'rb'))
784
# at the moment, based on everything in the file
785
parents = map(int, argv[3:])
786
lines = sys.stdin.readlines()
787
ver = w.add(parents, lines)
788
write_weave(w, file(argv[2], 'wb'))
789
print 'added version %d' % ver
792
if os.path.exists(fn):
793
raise IOError("file exists")
795
write_weave(w, file(fn, 'wb'))
796
elif cmd == 'get': # get one version
798
sys.stdout.writelines(w.get_iter(int(argv[3])))
800
elif cmd == 'mash': # get composite
802
sys.stdout.writelines(w.mash_iter(map(int, argv[3:])))
804
elif cmd == 'annotate':
806
# newline is added to all lines regardless; too hard to get
807
# reasonable formatting otherwise
809
for origin, text in w.annotate(int(argv[3])):
810
text = text.rstrip('\r\n')
812
print ' | %s' % (text)
814
print '%5d | %s' % (origin, text)
828
print '%d versions ok' % w.numversions()
830
elif cmd == 'inclusions':
832
print ' '.join(map(str, w.inclusions([int(argv[3])])))
834
elif cmd == 'parents':
836
print ' '.join(map(str, w._parents[int(argv[3])]))
838
elif cmd == 'plan-merge':
840
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
842
print '%14s | %s' % (state, line),
846
p = w.plan_merge(int(argv[3]), int(argv[4]))
847
sys.stdout.writelines(w.weave_merge(p))
849
elif cmd == 'mash-merge':
855
v1, v2 = map(int, argv[3:5])
857
basis = w.inclusions([v1]).intersection(w.inclusions([v2]))
859
base_lines = list(w.mash_iter(basis))
860
a_lines = list(w.get(v1))
861
b_lines = list(w.get(v2))
863
from bzrlib.merge3 import Merge3
864
m3 = Merge3(base_lines, a_lines, b_lines)
866
name_a = 'version %d' % v1
867
name_b = 'version %d' % v2
868
sys.stdout.writelines(m3.merge_lines(name_a=name_a, name_b=name_b))
870
raise ValueError('unknown command %r' % cmd)
874
def profile_main(argv):
875
import tempfile, hotshot, hotshot.stats
877
prof_f = tempfile.NamedTemporaryFile()
879
prof = hotshot.Profile(prof_f.name)
881
ret = prof.runcall(main, argv)
884
stats = hotshot.stats.load(prof_f.name)
886
stats.sort_stats('time')
887
## XXX: Might like to write to stderr or the trace file instead but
888
## print_stats seems hardcoded to stdout
889
stats.print_stats(20)
894
if __name__ == '__main__':
896
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
