« back to all changes in this revision
Viewing changes to bzrlib/weave.py
- browse files at revision 854
- compare with another revision
- download diff
- download tarball
- view history from revision 854
- Committer: Martin Pool
- Date: 2005-07-07 10:22:02 UTC
- Revision ID: mbp@sourcefrog.net-20050707102201-2d2a13a25098b101
- rearrange and clear up merged weave
- files added:
- bzr-man.py
- bzrlib/selftest
- effbot
- effbot/org
- patches
- plugins
- plugins/changeset
- tools
- files removed:
- bzrlib/tests.py
- files renamed:
- contrib/bash/bzr => contrib/bash/bzr.simple
- files modified:
- .bzrignore
added
removed
bzrlib/weave.py
1
#! /usr/bin/python
2
3
# Copyright (C) 2005 Canonical Ltd
4
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.
9
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.
14
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
18
19
# Author: Martin Pool <mbp@canonical.com>
20
21
22
"""Weave - storage of related text file versions"""
23
24
# TODO: Perhaps have copy method for Weave instances?
25
26
# XXX: If we do weaves this way, will a merge still behave the same
27
# way if it's done in a different order? That's a pretty desirable
28
# property.
29
30
# TODO: How to write these to disk? One option is cPickle, which
31
# would be fast but less friendly to C, and perhaps not portable. Another is
32
33
# TODO: Nothing here so far assumes the lines are really \n newlines,
34
# rather than being split up in some other way. We could accomodate
35
# binaries, perhaps by naively splitting on \n or perhaps using
36
# something like a rolling checksum.
37
38
# TODO: Perhaps track SHA-1 in the header for protection? This would
39
# be redundant with it being stored in the inventory, but perhaps
40
# usefully so?
41
42
# TODO: Track version names as well as indexes.
43
44
# TODO: Probably do transitive expansion when specifying parents?
45
46
# TODO: Separate out some code to read and write weaves.
47
48
# TODO: End marker for each version so we can stop reading?
49
50
# TODO: Check that no insertion occurs inside a deletion that was
51
# active in the version of the insertion.
52
53
# TODO: Perhaps a special slower check() method that verifies more
54
# nesting constraints and the MD5 of each version?
55
56
57
58
try:
59
set
60
frozenset
61
except NameError:
62
from sets import Set, ImmutableSet
63
set = Set
64
frozenset = ImmutableSet
65
del Set, ImmutableSet
66
67
68
class WeaveError(Exception):
69
"""Exception in processing weave"""
70
71
72
class WeaveFormatError(WeaveError):
73
"""Weave invariant violated"""
74
75
76
class Weave(object):
77
"""weave - versioned text file storage.
78
79
A Weave manages versions of line-based text files, keeping track
80
of the originating version for each line.
81
82
To clients the "lines" of the file are represented as a list of strings.
83
These strings will typically have terminal newline characters, but
84
this is not required. In particular files commonly do not have a newline
85
at the end of the file.
86
87
Texts can be identified in either of two ways:
88
89
* a nonnegative index number.
90
91
* a version-id string.
92
93
Typically the index number will be valid only inside this weave and
94
the version-id is used to reference it in the larger world.
95
96
The weave is represented as a list mixing edit instructions and
97
literal text. Each entry in _l can be either a string (or
98
unicode), or a tuple. If a string, it means that the given line
99
should be output in the currently active revisions.
100
101
If a tuple, it gives a processing instruction saying in which
102
revisions the enclosed lines are active. The tuple has the form
103
(instruction, version).
104
105
The instruction can be '{' or '}' for an insertion block, and '['
106
and ']' for a deletion block respectively. The version is the
107
integer version index. There is no replace operator, only deletes
108
and inserts.
109
110
Constraints/notes:
111
112
* A later version can delete lines that were introduced by any
113
number of ancestor versions; this implies that deletion
114
instructions can span insertion blocks without regard to the
115
insertion block's nesting.
116
117
* Similarly, deletions need not be properly nested with regard to
118
each other, because they might have been generated by
119
independent revisions.
120
121
* Insertions are always made by inserting a new bracketed block
122
into a single point in the previous weave. This implies they
123
can nest but not overlap, and the nesting must always have later
124
insertions on the inside.
125
126
* It doesn't seem very useful to have an active insertion
127
inside an inactive insertion, but it might happen.
128
129
* Therefore, all instructions are always"considered"; that
130
is passed onto and off the stack. An outer inactive block
131
doesn't disable an inner block.
132
133
* Lines are enabled if the most recent enclosing insertion is
134
active and none of the enclosing deletions are active.
135
136
* There is no point having a deletion directly inside its own
137
insertion; you might as well just not write it. And there
138
should be no way to get an earlier version deleting a later
139
version.
140
141
_l
142
Text of the weave.
143
144
_v
145
List of versions, indexed by index number.
146
147
For each version we store the set (included_versions), which
148
lists the previous versions also considered active; the
149
versions included in those versions are included transitively.
150
So new versions created from nothing list []; most versions
151
have a single entry; some have more.
152
153
_sha1s
154
List of hex SHA-1 of each version, or None if not recorded.
155
"""
156
def __init__(self):
157
self._l = []
158
self._v = []
159
self._sha1s = []
160
161
162
def __eq__(self, other):
163
if not isinstance(other, Weave):
164
return False
165
return self._v == other._v \
166
and self._l == other._l
167
168
169
def __ne__(self, other):
170
return not self.__eq__(other)
171
172
173
def add(self, parents, text):
174
"""Add a single text on top of the weave.
175
176
Returns the index number of the newly added version.
177
178
parents
179
List or set of parent version numbers. This must normally include
180
the parents and the parent's parents, or wierd things might happen.
181
182
text
183
Sequence of lines to be added in the new version."""
184
## self._check_versions(parents)
185
## self._check_lines(text)
186
idx = len(self._v)
187
188
import sha
189
s = sha.new()
190
for l in text:
191
s.update(l)
192
sha1 = s.hexdigest()
193
del s
194
195
if parents:
196
delta = self._delta(self.inclusions(parents), text)
197
198
# offset gives the number of lines that have been inserted
199
# into the weave up to the current point; if the original edit instruction
200
# says to change line A then we actually change (A+offset)
201
offset = 0
202
203
for i1, i2, newlines in delta:
204
assert 0 <= i1
205
assert i1 <= i2
206
assert i2 <= len(self._l)
207
208
# the deletion and insertion are handled separately.
209
# first delete the region.
210
if i1 != i2:
211
self._l.insert(i1+offset, ('[', idx))
212
self._l.insert(i2+offset+1, (']', idx))
213
offset += 2
214
# is this OK???
215
216
if newlines:
217
# there may have been a deletion spanning up to
218
# i2; we want to insert after this region to make sure
219
# we don't destroy ourselves
220
i = i2 + offset
221
self._l[i:i] = [('{', idx)] \
222
+ newlines \
223
+ [('}', idx)]
224
offset += 2 + len(newlines)
225
226
self._addversion(parents)
227
else:
228
# special case; adding with no parents revision; can do this
229
# more quickly by just appending unconditionally
230
self._l.append(('{', idx))
231
self._l += text
232
self._l.append(('}', idx))
233
234
self._addversion(None)
235
236
self._sha1s.append(sha1)
237
238
return idx
239
240
241
def inclusions(self, versions):
242
"""Expand out everything included by versions."""
243
i = set(versions)
244
for v in versions:
245
i.update(self._v[v])
246
return i
247
248
249
def _addversion(self, parents):
250
if parents:
251
self._v.append(frozenset(parents))
252
else:
253
self._v.append(frozenset())
254
255
256
def _check_lines(self, text):
257
if not isinstance(text, list):
258
raise ValueError("text should be a list, not %s" % type(text))
259
260
for l in text:
261
if not isinstance(l, basestring):
262
raise ValueError("text line should be a string or unicode, not %s" % type(l))
263
264
265
266
def _check_versions(self, indexes):
267
"""Check everything in the sequence of indexes is valid"""
268
for i in indexes:
269
try:
270
self._v[i]
271
except IndexError:
272
raise IndexError("invalid version number %r" % i)
273
274
275
def annotate(self, index):
276
return list(self.annotate_iter(index))
277
278
279
def annotate_iter(self, version):
280
"""Yield list of (index-id, line) pairs for the specified version.
281
282
The index indicates when the line originated in the weave."""
283
included = self.inclusions([version])
284
for origin, lineno, text in self._extract(included):
285
yield origin, text
286
287
288
def _extract(self, included):
289
"""Yield annotation of lines in included set.
290
291
Yields a sequence of tuples (origin, lineno, text), where
292
origin is the origin version, lineno the index in the weave,
293
and text the text of the line.
294
295
The set typically but not necessarily corresponds to a version.
296
"""
297
istack = [] # versions for which an insertion block is current
298
299
dset = set() # versions for which a deletion block is current
300
301
isactive = None
302
303
lineno = 0 # line of weave, 0-based
304
305
# TODO: Probably only need to put included revisions in the istack
306
307
# TODO: Could split this into two functions, one that updates
308
# the stack and the other that processes the results -- but
309
# I'm not sure it's really needed.
310
311
# TODO: In fact, I think we only need to store the *count* of
312
# active insertions and deletions, and we can maintain that by
313
# just by just counting as we go along.
314
315
WFE = WeaveFormatError
316
317
for l in self._l:
318
if isinstance(l, tuple):
319
isactive = None # recalculate
320
c, v = l
321
if c == '{':
322
if istack and (istack[-1] >= v):
323
raise WFE("improperly nested insertions %d>=%d on line %d"
324
% (istack[-1], v, lineno))
325
istack.append(v)
326
elif c == '}':
327
try:
328
oldv = istack.pop()
329
except IndexError:
330
raise WFE("unmatched close of insertion %d on line %d"
331
% (v, lineno))
332
if oldv != v:
333
raise WFE("mismatched close of insertion %d!=%d on line %d"
334
% (oldv, v, lineno))
335
elif c == '[':
336
# block deleted in v
337
if v in dset:
338
raise WFE("repeated deletion marker for version %d on line %d"
339
% (v, lineno))
340
if istack:
341
if istack[-1] == v:
342
raise WFE("version %d deletes own text on line %d"
343
% (v, lineno))
344
# XXX
345
dset.add(v)
346
elif c == ']':
347
if v in dset:
348
dset.remove(v)
349
else:
350
raise WFE("unmatched close of deletion %d on line %d"
351
% (v, lineno))
352
else:
353
raise WFE("invalid processing instruction %r on line %d"
354
% (l, lineno))
355
else:
356
assert isinstance(l, basestring)
357
if not istack:
358
raise WFE("literal at top level on line %d"
359
% lineno)
360
if isactive == None:
361
isactive = (istack[-1] in included) \
362
and not included.intersection(dset)
363
if isactive:
364
origin = istack[-1]
365
yield origin, lineno, l
366
lineno += 1
367
368
if istack:
369
raise WFE("unclosed insertion blocks at end of weave",
370
istack)
371
if dset:
372
raise WFE("unclosed deletion blocks at end of weave",
373
dset)
374
375
376
def get_iter(self, version):
377
"""Yield lines for the specified version."""
378
for origin, lineno, line in self._extract(self.inclusions([version])):
379
yield line
380
381
382
def get(self, index):
383
return list(self.get_iter(index))
384
385
386
def mash_iter(self, included):
387
"""Return composed version of multiple included versions."""
388
included = frozenset(included)
389
for origin, lineno, text in self._extract(included):
390
yield text
391
392
393
def dump(self, to_file):
394
from pprint import pprint
395
print >>to_file, "Weave._l = ",
396
pprint(self._l, to_file)
397
print >>to_file, "Weave._v = ",
398
pprint(self._v, to_file)
399
400
401
402
def numversions(self):
403
l = len(self._v)
404
assert l == len(self._sha1s)
405
return l
406
407
408
def check(self):
409
# check no circular inclusions
410
for version in range(self.numversions()):
411
inclusions = list(self._v[version])
412
if inclusions:
413
inclusions.sort()
414
if inclusions[-1] >= version:
415
raise WeaveFormatError("invalid included version %d for index %d"
416
% (inclusions[-1], version))
417
418
# try extracting all versions; this is a bit slow and parallel
419
# extraction could be used
420
import sha
421
for version in range(self.numversions()):
422
s = sha.new()
423
for l in self.get_iter(version):
424
s.update(l)
425
hd = s.hexdigest()
426
expected = self._sha1s[version]
427
if hd != expected:
428
raise WeaveError("mismatched sha1 for version %d; "
429
"got %s, expected %s"
430
% (version, hd, expected))
431
432
433
434
def merge(self, merge_versions):
435
"""Automerge and mark conflicts between versions.
436
437
This returns a sequence, each entry describing alternatives
438
for a chunk of the file. Each of the alternatives is given as
439
a list of lines.
440
441
If there is a chunk of the file where there's no diagreement,
442
only one alternative is given.
443
"""
444
445
# approach: find the included versions common to all the
446
# merged versions
447
raise NotImplementedError()
448
449
450
451
def _delta(self, included, lines):
452
"""Return changes from basis to new revision.
453
454
The old text for comparison is the union of included revisions.
455
456
This is used in inserting a new text.
457
458
Delta is returned as a sequence of
459
(weave1, weave2, newlines).
460
461
This indicates that weave1:weave2 of the old weave should be
462
replaced by the sequence of lines in newlines. Note that
463
these line numbers are positions in the total weave and don't
464
correspond to the lines in any extracted version, or even the
465
extracted union of included versions.
466
467
If line1=line2, this is a pure insert; if newlines=[] this is a
468
pure delete. (Similar to difflib.)
469
"""
470
# basis a list of (origin, lineno, line)
471
basis_lineno = []
472
basis_lines = []
473
for origin, lineno, line in self._extract(included):
474
basis_lineno.append(lineno)
475
basis_lines.append(line)
476
477
# add a sentinal, because we can also match against the final line
478
basis_lineno.append(len(self._l))
479
480
# XXX: which line of the weave should we really consider
481
# matches the end of the file? the current code says it's the
482
# last line of the weave?
483
484
from difflib import SequenceMatcher
485
s = SequenceMatcher(None, basis_lines, lines)
486
487
# TODO: Perhaps return line numbers from composed weave as well?
488
489
for tag, i1, i2, j1, j2 in s.get_opcodes():
490
##print tag, i1, i2, j1, j2
491
492
if tag == 'equal':
493
continue
494
495
# i1,i2 are given in offsets within basis_lines; we need to map them
496
# back to offsets within the entire weave
497
real_i1 = basis_lineno[i1]
498
real_i2 = basis_lineno[i2]
499
500
assert 0 <= j1
501
assert j1 <= j2
502
assert j2 <= len(lines)
503
504
yield real_i1, real_i2, lines[j1:j2]
505
506
507
508
def weave_info(filename, out):
509
"""Show some text information about the weave."""
510
from weavefile import read_weave
511
wf = file(filename, 'rb')
512
w = read_weave(wf)
513
# FIXME: doesn't work on pipes
514
weave_size = wf.tell()
515
print >>out, "weave file size %d bytes" % weave_size
516
print >>out, "weave contains %d versions" % len(w._v)
517
518
total = 0
519
print ' %8s %8s %8s %s' % ('version', 'lines', 'bytes', 'sha1')
520
print ' -------- -------- -------- ----------------------------------------'
521
for i in range(len(w._v)):
522
text = w.get(i)
523
lines = len(text)
524
bytes = sum((len(a) for a in text))
525
sha1 = w._sha1s[i]
526
print ' %8d %8d %8d %s' % (i, lines, bytes, sha1)
527
total += bytes
528
529
print >>out, "versions total %d bytes" % total
530
print >>out, "compression ratio %.3f" % (float(total)/float(weave_size))
531
532
533
534
def main(argv):
535
import sys
536
import os
537
from weavefile import write_weave_v1, read_weave
538
cmd = argv[1]
539
if cmd == 'add':
540
w = read_weave(file(argv[2], 'rb'))
541
# at the moment, based on everything in the file
542
parents = set(range(len(w._v)))
543
lines = sys.stdin.readlines()
544
ver = w.add(parents, lines)
545
write_weave_v1(w, file(argv[2], 'wb'))
546
print 'added %d' % ver
547
elif cmd == 'init':
548
fn = argv[2]
549
if os.path.exists(fn):
550
raise IOError("file exists")
551
w = Weave()
552
write_weave_v1(w, file(fn, 'wb'))
553
elif cmd == 'get':
554
w = read_weave(file(argv[2], 'rb'))
555
sys.stdout.writelines(w.get_iter(int(argv[3])))
556
elif cmd == 'annotate':
557
w = read_weave(file(argv[2], 'rb'))
558
# newline is added to all lines regardless; too hard to get
559
# reasonable formatting otherwise
560
lasto = None
561
for origin, text in w.annotate(int(argv[3])):
562
text = text.rstrip('\r\n')
563
if origin == lasto:
564
print ' | %s' % (text)
565
else:
566
print '%5d | %s' % (origin, text)
567
lasto = origin
568
elif cmd == 'info':
569
weave_info(argv[2], sys.stdout)
570
elif cmd == 'check':
571
w = read_weave(file(argv[2], 'rb'))
572
w.check()
573
else:
574
raise ValueError('unknown command %r' % cmd)
575
576
577
if __name__ == '__main__':
578
import sys
579
sys.exit(main(sys.argv))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1