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?
90
from cStringIO import StringIO
93
class WeaveError(Exception):
94
"""Exception in processing weave"""
97
class WeaveFormatError(WeaveError):
98
"""Weave invariant violated"""
102
"""weave - versioned text file storage.
104
A Weave manages versions of line-based text files, keeping track
105
of the originating version for each line.
107
To clients the "lines" of the file are represented as a list of strings.
108
These strings will typically have terminal newline characters, but
109
this is not required. In particular files commonly do not have a newline
110
at the end of the file.
112
Texts can be identified in either of two ways:
114
* a nonnegative index number.
116
* a version-id string. (not implemented yet)
118
Typically the index number will be valid only inside this weave and
119
the version-id is used to reference it in the larger world.
121
The weave is represented as a list mixing edit instructions and
122
literal text. Each entry in _weave can be either a string (or
123
unicode), or a tuple. If a string, it means that the given line
124
should be output in the currently active revisions.
126
If a tuple, it gives a processing instruction saying in which
127
revisions the enclosed lines are active. The tuple has the form
128
(instruction, version).
130
The instruction can be '{' or '}' for an insertion block, and '['
131
and ']' for a deletion block respectively. The version is the
132
integer version index. There is no replace operator, only deletes
133
and inserts. For '}', the end of an insertion, there is no
134
version parameter because it always closes the most recently
139
* A later version can delete lines that were introduced by any
140
number of ancestor versions; this implies that deletion
141
instructions can span insertion blocks without regard to the
142
insertion block's nesting.
144
* Similarly, deletions need not be properly nested with regard to
145
each other, because they might have been generated by
146
independent revisions.
148
* Insertions are always made by inserting a new bracketed block
149
into a single point in the previous weave. This implies they
150
can nest but not overlap, and the nesting must always have later
151
insertions on the inside.
153
* It doesn't seem very useful to have an active insertion
154
inside an inactive insertion, but it might happen.
156
* Therefore, all instructions are always"considered"; that
157
is passed onto and off the stack. An outer inactive block
158
doesn't disable an inner block.
160
* Lines are enabled if the most recent enclosing insertion is
161
active and none of the enclosing deletions are active.
163
* There is no point having a deletion directly inside its own
164
insertion; you might as well just not write it. And there
165
should be no way to get an earlier version deleting a later
169
Text of the weave; list of control instruction tuples and strings.
172
List of parents, indexed by version number.
173
It is only necessary to store the minimal set of parents for
174
each version; the parent's parents are implied.
177
List of hex SHA-1 of each version.
180
List of symbolic names for each version. Each should be unique.
183
For each name, the version number.
186
Descriptive name of this weave; typically the filename if known.
190
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
193
def __init__(self, weave_name=None):
199
self._weave_name = weave_name
202
def __eq__(self, other):
203
if not isinstance(other, Weave):
205
return self._parents == other._parents \
206
and self._weave == other._weave \
207
and self._sha1s == other._sha1s
210
def __ne__(self, other):
211
return not self.__eq__(other)
214
def lookup(self, name):
216
return self._name_map[name]
218
raise WeaveError("name %s not present in weave %s" %
219
(name, self._weave_name))
222
def idx_to_name(self, version):
223
return self._names[version]
226
def add(self, name, parents, text):
227
"""Add a single text on top of the weave.
229
Returns the index number of the newly added version.
232
Symbolic name for this version.
233
(Typically the revision-id of the revision that added it.)
236
List or set of direct parent version numbers.
239
Sequence of lines to be added in the new version."""
241
assert isinstance(name, basestring)
242
if name in self._name_map:
243
raise WeaveError("name %r already present in weave" % name)
245
self._check_versions(parents)
246
## self._check_lines(text)
247
new_version = len(self._parents)
254
# if we abort after here the (in-memory) weave will be corrupt because only
255
# some fields are updated
256
self._parents.append(parents[:])
257
self._sha1s.append(sha1)
258
self._names.append(name)
259
self._name_map[name] = new_version
263
# special case; adding with no parents revision; can do
264
# this more quickly by just appending unconditionally.
265
# even more specially, if we're adding an empty text we
266
# need do nothing at all.
268
self._weave.append(('{', new_version))
269
self._weave.extend(text)
270
self._weave.append(('}', None))
274
if len(parents) == 1:
275
pv = list(parents)[0]
276
if sha1 == self._sha1s[pv]:
277
# special case: same as the single parent
281
ancestors = self.inclusions(parents)
285
# basis a list of (origin, lineno, line)
288
for origin, lineno, line in self._extract(ancestors):
289
basis_lineno.append(lineno)
290
basis_lines.append(line)
292
# another small special case: a merge, producing the same text
294
if text == basis_lines:
297
# add a sentinal, because we can also match against the final line
298
basis_lineno.append(len(self._weave))
300
# XXX: which line of the weave should we really consider
301
# matches the end of the file? the current code says it's the
302
# last line of the weave?
304
#print 'basis_lines:', basis_lines
305
#print 'new_lines: ', lines
307
from difflib import SequenceMatcher
308
s = SequenceMatcher(None, basis_lines, text)
310
# offset gives the number of lines that have been inserted
311
# into the weave up to the current point; if the original edit instruction
312
# says to change line A then we actually change (A+offset)
315
for tag, i1, i2, j1, j2 in s.get_opcodes():
316
# i1,i2 are given in offsets within basis_lines; we need to map them
317
# back to offsets within the entire weave
318
#print 'raw match', tag, i1, i2, j1, j2
322
i1 = basis_lineno[i1]
323
i2 = basis_lineno[i2]
325
assert 0 <= j1 <= j2 <= len(text)
327
#print tag, i1, i2, j1, j2
329
# the deletion and insertion are handled separately.
330
# first delete the region.
332
self._weave.insert(i1+offset, ('[', new_version))
333
self._weave.insert(i2+offset+1, (']', new_version))
337
# there may have been a deletion spanning up to
338
# i2; we want to insert after this region to make sure
339
# we don't destroy ourselves
341
self._weave[i:i] = ([('{', new_version)]
344
offset += 2 + (j2 - j1)
349
def inclusions(self, versions):
350
"""Return set of all ancestors of given version(s)."""
356
# include all its parents
357
i.update(self._parents[v])
361
raise ValueError("version %d not present in weave" % v)
364
def parents(self, version):
365
return self._parents[version]
368
def minimal_parents(self, version):
369
"""Find the minimal set of parents for the version."""
370
included = self._parents[version]
375
li.sort(reverse=True)
383
gotit.update(self.inclusions(pv))
385
assert mininc[0] >= 0
386
assert mininc[-1] < version
391
def _check_lines(self, text):
392
if not isinstance(text, list):
393
raise ValueError("text should be a list, not %s" % type(text))
396
if not isinstance(l, basestring):
397
raise ValueError("text line should be a string or unicode, not %s"
402
def _check_versions(self, indexes):
403
"""Check everything in the sequence of indexes is valid"""
408
raise IndexError("invalid version number %r" % i)
411
def annotate(self, index):
412
return list(self.annotate_iter(index))
415
def annotate_iter(self, version):
416
"""Yield list of (index-id, line) pairs for the specified version.
418
The index indicates when the line originated in the weave."""
419
for origin, lineno, text in self._extract([version]):
427
(lineno, insert, deletes, text)
428
for each literal line.
434
lineno = 0 # line of weave, 0-based
436
for l in self._weave:
437
if isinstance(l, tuple):
450
raise WeaveFormatError('unexpected instruction %r'
453
assert isinstance(l, basestring)
455
yield lineno, istack[-1], dset, l
460
def _extract(self, versions):
461
"""Yield annotation of lines in included set.
463
Yields a sequence of tuples (origin, lineno, text), where
464
origin is the origin version, lineno the index in the weave,
465
and text the text of the line.
467
The set typically but not necessarily corresponds to a version.
470
if not isinstance(i, int):
473
included = self.inclusions(versions)
478
lineno = 0 # line of weave, 0-based
484
WFE = WeaveFormatError
486
for l in self._weave:
487
if isinstance(l, tuple):
491
assert v not in istack
505
assert isinstance(l, basestring)
507
isactive = (not dset) and istack and (istack[-1] in included)
509
result.append((istack[-1], lineno, l))
513
raise WFE("unclosed insertion blocks at end of weave",
516
raise WFE("unclosed deletion blocks at end of weave",
523
def get_iter(self, version):
524
"""Yield lines for the specified version."""
525
for origin, lineno, line in self._extract([version]):
529
def get_text(self, version):
530
assert isinstance(version, int)
532
s.writelines(self.get_iter(version))
536
def get(self, index):
537
return list(self.get_iter(index))
540
def mash_iter(self, included):
541
"""Return composed version of multiple included versions."""
542
for origin, lineno, text in self._extract(included):
546
def dump(self, to_file):
547
from pprint import pprint
548
print >>to_file, "Weave._weave = ",
549
pprint(self._weave, to_file)
550
print >>to_file, "Weave._parents = ",
551
pprint(self._parents, to_file)
555
def numversions(self):
556
l = len(self._parents)
557
assert l == len(self._sha1s)
562
return self.numversions()
565
def check(self, progress_bar=None):
566
# check no circular inclusions
567
for version in range(self.numversions()):
568
inclusions = list(self._parents[version])
571
if inclusions[-1] >= version:
572
raise WeaveFormatError("invalid included version %d for index %d"
573
% (inclusions[-1], version))
575
# try extracting all versions; this is a bit slow and parallel
576
# extraction could be used
577
nv = self.numversions()
578
for version in range(nv):
580
progress_bar.update('checking text', version, nv)
582
for l in self.get_iter(version):
585
expected = self._sha1s[version]
587
raise WeaveError("mismatched sha1 for version %d; "
588
"got %s, expected %s"
589
% (version, hd, expected))
591
# TODO: check insertions are properly nested, that there are
592
# no lines outside of insertion blocks, that deletions are
593
# properly paired, etc.
597
def merge(self, merge_versions):
598
"""Automerge and mark conflicts between versions.
600
This returns a sequence, each entry describing alternatives
601
for a chunk of the file. Each of the alternatives is given as
604
If there is a chunk of the file where there's no diagreement,
605
only one alternative is given.
608
# approach: find the included versions common to all the
610
raise NotImplementedError()
614
def _delta(self, included, lines):
615
"""Return changes from basis to new revision.
617
The old text for comparison is the union of included revisions.
619
This is used in inserting a new text.
621
Delta is returned as a sequence of
622
(weave1, weave2, newlines).
624
This indicates that weave1:weave2 of the old weave should be
625
replaced by the sequence of lines in newlines. Note that
626
these line numbers are positions in the total weave and don't
627
correspond to the lines in any extracted version, or even the
628
extracted union of included versions.
630
If line1=line2, this is a pure insert; if newlines=[] this is a
631
pure delete. (Similar to difflib.)
636
def plan_merge(self, ver_a, ver_b):
637
"""Return pseudo-annotation indicating how the two versions merge.
639
This is computed between versions a and b and their common
642
Weave lines present in none of them are skipped entirely.
644
inc_a = self.inclusions([ver_a])
645
inc_b = self.inclusions([ver_b])
646
inc_c = inc_a & inc_b
648
for lineno, insert, deleteset, line in self._walk():
649
if deleteset & inc_c:
650
# killed in parent; can't be in either a or b
651
# not relevant to our work
652
yield 'killed-base', line
653
elif insert in inc_c:
654
# was inserted in base
655
killed_a = bool(deleteset & inc_a)
656
killed_b = bool(deleteset & inc_b)
657
if killed_a and killed_b:
658
yield 'killed-both', line
660
yield 'killed-a', line
662
yield 'killed-b', line
664
yield 'unchanged', line
665
elif insert in inc_a:
666
if deleteset & inc_a:
667
yield 'ghost-a', line
671
elif insert in inc_b:
672
if deleteset & inc_b:
673
yield 'ghost-b', line
677
# not in either revision
678
yield 'irrelevant', line
680
yield 'unchanged', '' # terminator
684
def weave_merge(self, plan):
689
for state, line in plan:
690
if state == 'unchanged' or state == 'killed-both':
691
# resync and flush queued conflicts changes if any
692
if not lines_a and not lines_b:
694
elif ch_a and not ch_b:
696
for l in lines_a: yield l
697
elif ch_b and not ch_a:
698
for l in lines_b: yield l
699
elif lines_a == lines_b:
700
for l in lines_a: yield l
703
for l in lines_a: yield l
705
for l in lines_b: yield l
712
if state == 'unchanged':
715
elif state == 'killed-a':
718
elif state == 'killed-b':
721
elif state == 'new-a':
724
elif state == 'new-b':
728
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
739
"""Show the weave's table-of-contents"""
740
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
741
for i in (6, 50, 10, 10):
744
for i in range(w.numversions()):
747
parent_str = ' '.join(map(str, w._parents[i]))
748
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
752
def weave_stats(weave_file):
753
from bzrlib.progress import ProgressBar
754
from bzrlib.weavefile import read_weave
758
wf = file(weave_file, 'rb')
760
# FIXME: doesn't work on pipes
761
weave_size = wf.tell()
765
for i in range(vers):
766
pb.update('checking sizes', i, vers)
767
for line in w.get_iter(i):
772
print 'versions %9d' % vers
773
print 'weave file %9d bytes' % weave_size
774
print 'total contents %9d bytes' % total
775
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
778
print 'average size %9d bytes' % avg
779
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
783
print """bzr weave tool
785
Experimental tool for weave algorithm.
789
Create an empty weave file
790
weave get WEAVEFILE VERSION
791
Write out specified version.
792
weave check WEAVEFILE
793
Check consistency of all versions.
795
Display table of contents.
796
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
797
Add NEWTEXT, with specified parent versions.
798
weave annotate WEAVEFILE VERSION
799
Display origin of each line.
800
weave mash WEAVEFILE VERSION...
801
Display composite of all selected versions.
802
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
803
Auto-merge two versions and display conflicts.
807
% weave init foo.weave
809
% weave add foo.weave ver0 < foo.txt
812
(create updated version)
814
% weave get foo.weave 0 | diff -u - foo.txt
815
% weave add foo.weave ver1 0 < foo.txt
818
% weave get foo.weave 0 > foo.txt (create forked version)
820
% weave add foo.weave ver2 0 < foo.txt
823
% weave merge foo.weave 1 2 > foo.txt (merge them)
824
% vi foo.txt (resolve conflicts)
825
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
834
from weavefile import write_weave, read_weave
835
from bzrlib.progress import ProgressBar
850
return read_weave(file(argv[2], 'rb'))
856
# at the moment, based on everything in the file
858
parents = map(int, argv[4:])
859
lines = sys.stdin.readlines()
860
ver = w.add(name, parents, lines)
861
write_weave(w, file(argv[2], 'wb'))
862
print 'added version %r %d' % (name, ver)
865
if os.path.exists(fn):
866
raise IOError("file exists")
868
write_weave(w, file(fn, 'wb'))
869
elif cmd == 'get': # get one version
871
sys.stdout.writelines(w.get_iter(int(argv[3])))
873
elif cmd == 'mash': # get composite
875
sys.stdout.writelines(w.mash_iter(map(int, argv[3:])))
877
elif cmd == 'annotate':
879
# newline is added to all lines regardless; too hard to get
880
# reasonable formatting otherwise
882
for origin, text in w.annotate(int(argv[3])):
883
text = text.rstrip('\r\n')
885
print ' | %s' % (text)
887
print '%5d | %s' % (origin, text)
901
print '%d versions ok' % w.numversions()
903
elif cmd == 'inclusions':
905
print ' '.join(map(str, w.inclusions([int(argv[3])])))
907
elif cmd == 'parents':
909
print ' '.join(map(str, w._parents[int(argv[3])]))
911
elif cmd == 'plan-merge':
913
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
915
print '%14s | %s' % (state, line),
919
p = w.plan_merge(int(argv[3]), int(argv[4]))
920
sys.stdout.writelines(w.weave_merge(p))
922
elif cmd == 'mash-merge':
928
v1, v2 = map(int, argv[3:5])
930
basis = w.inclusions([v1]).intersection(w.inclusions([v2]))
932
base_lines = list(w.mash_iter(basis))
933
a_lines = list(w.get(v1))
934
b_lines = list(w.get(v2))
936
from bzrlib.merge3 import Merge3
937
m3 = Merge3(base_lines, a_lines, b_lines)
939
name_a = 'version %d' % v1
940
name_b = 'version %d' % v2
941
sys.stdout.writelines(m3.merge_lines(name_a=name_a, name_b=name_b))
943
raise ValueError('unknown command %r' % cmd)
947
def profile_main(argv):
948
import tempfile, hotshot, hotshot.stats
950
prof_f = tempfile.NamedTemporaryFile()
952
prof = hotshot.Profile(prof_f.name)
954
ret = prof.runcall(main, argv)
957
stats = hotshot.stats.load(prof_f.name)
959
stats.sort_stats('cumulative')
960
## XXX: Might like to write to stderr or the trace file instead but
961
## print_stats seems hardcoded to stdout
962
stats.print_stats(20)
967
if __name__ == '__main__':
969
if '--profile' in sys.argv:
971
args.remove('--profile')
972
sys.exit(profile_main(args))
974
sys.exit(main(sys.argv))
added
removed
bzrlib/weave.py
