« 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:
- bzrlib/mdiff.py
- plugins/changeset
- files removed:
- HACKING
- bzrlib/export
- bzrlib/store
- bzrlib/tests/blackbox
- bzrlib/transport
- bzrlib/util
- bzrlib/util/configobj
- bzrlib/util/configobj/docs
- contrib/emacs
- files renamed:
- bzr_man.py => bzr-man.py
- contrib/newinventory.py => bzrlib/newinventory.py
- bzrlib/tests/ => bzrlib/selftest/
- bzrlib/tests/blackbox/__init__.py => bzrlib/selftest/blackbox.py
- bzrlib/tests/test_plugins.py => bzrlib/selftest/plugins.py
- bzrlib/tests/test_hashcache.py => bzrlib/selftest/testhashcache.py
- bzrlib/tests/test_merge3.py => bzrlib/selftest/testmerge3.py
- bzrlib/tests/blackbox/versioning.py => bzrlib/selftest/versioning.py
- bzrlib/tests/test_whitebox.py => bzrlib/selftest/whitebox.py
- bzrlib/store/__init__.py => bzrlib/store.py
- bzrlib/util/effbot/ => effbot/
- bzrlib/util/elementtree/ => elementtree/
- bzrlib/plugins/ => plugins/
- bzrlib/tests/test_weave.py => tools/testweave.py
- bzrlib/util/urlgrabber/ => urlgrabber/
- files modified:
- .bzrignore
- bzr *
- setup.py *
added
removed
bzrlib/weave.py
21
21
22
22
"""Weave - storage of related text file versions"""
23
23
24
# TODO: Perhaps have copy method for Weave instances?
24
25
25
26
# XXX: If we do weaves this way, will a merge still behave the same
26
27
# way if it's done in a different order? That's a pretty desirable
27
28
# property.
28
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
29
33
# TODO: Nothing here so far assumes the lines are really \n newlines,
30
34
# rather than being split up in some other way. We could accomodate
31
35
# binaries, perhaps by naively splitting on \n or perhaps using
32
36
# something like a rolling checksum.
33
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
34
48
# TODO: End marker for each version so we can stop reading?
35
49
36
50
# TODO: Check that no insertion occurs inside a deletion that was
37
51
# active in the version of the insertion.
38
52
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.
43
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.
47
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.
52
53
# TODO: Probably the API should work only in names to hide the integer
54
# indexes from the user.
55
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
58
# version?
59
60
# TODO: Reweave can possibly be made faster by remembering diffs
61
# where the basis and destination are unchanged.
62
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.
68
69
70
import sha
71
from difflib import SequenceMatcher
72
73
from bzrlib.trace import mutter
74
from bzrlib.errors import WeaveError, WeaveFormatError, WeaveParentMismatch, \
75
WeaveRevisionNotPresent, WeaveRevisionAlreadyPresent
76
from bzrlib.tsort import topo_sort
77
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
78
75
79
76
class Weave(object):
80
77
"""weave - versioned text file storage.
97
94
the version-id is used to reference it in the larger world.
98
95
99
96
The weave is represented as a list mixing edit instructions and
100
literal text. Each entry in _weave can be either a string (or
97
literal text. Each entry in _l can be either a string (or
101
98
unicode), or a tuple. If a string, it means that the given line
102
99
should be output in the currently active revisions.
103
100
108
105
The instruction can be '{' or '}' for an insertion block, and '['
109
106
and ']' for a deletion block respectively. The version is the
110
107
integer version index. There is no replace operator, only deletes
111
and inserts. For '}', the end of an insertion, there is no
112
version parameter because it always closes the most recently
113
opened insertion.
108
and inserts.
114
109
115
110
Constraints/notes:
116
111
143
138
should be no way to get an earlier version deleting a later
144
139
version.
145
140
146
_weave
147
Text of the weave; list of control instruction tuples and strings.
148
149
_parents
150
List of parents, indexed by version number.
151
It is only necessary to store the minimal set of parents for
152
each version; the parent's parents are implied.
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.
153
152
154
153
_sha1s
155
List of hex SHA-1 of each version.
156
157
_names
158
List of symbolic names for each version. Each should be unique.
159
160
_name_map
161
For each name, the version number.
162
163
_weave_name
164
Descriptive name of this weave; typically the filename if known.
165
Set by read_weave.
154
List of hex SHA-1 of each version, or None if not recorded.
166
155
"""
167
168
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
169
'_weave_name']
170
171
def __init__(self, weave_name=None):
172
self._weave = []
173
self._parents = []
156
def __init__(self):
157
self._l = []
158
self._v = []
174
159
self._sha1s = []
175
self._names = []
176
self._name_map = {}
177
self._weave_name = weave_name
178
179
def __repr__(self):
180
return "Weave(%r)" % self._weave_name
181
182
183
def copy(self):
184
"""Return a deep copy of self.
185
186
The copy can be modified without affecting the original weave."""
187
other = Weave()
188
other._weave = self._weave[:]
189
other._parents = self._parents[:]
190
other._sha1s = self._sha1s[:]
191
other._names = self._names[:]
192
other._name_map = self._name_map.copy()
193
other._weave_name = self._weave_name
194
return other
160
195
161
196
162
def __eq__(self, other):
197
163
if not isinstance(other, Weave):
198
164
return False
199
return self._parents == other._parents \
200
and self._weave == other._weave \
201
and self._sha1s == other._sha1s
165
return self._v == other._v \
166
and self._l == other._l
167
202
168
203
204
169
def __ne__(self, other):
205
170
return not self.__eq__(other)
206
171
207
def __contains__(self, name):
208
return self._name_map.has_key(name)
209
210
def maybe_lookup(self, name_or_index):
211
"""Convert possible symbolic name to index, or pass through indexes."""
212
if isinstance(name_or_index, (int, long)):
213
return name_or_index
214
else:
215
return self.lookup(name_or_index)
216
217
218
def lookup(self, name):
219
"""Convert symbolic version name to index."""
220
try:
221
return self._name_map[name]
222
except KeyError:
223
raise WeaveRevisionNotPresent(name, self)
224
225
def names(self):
226
return self._names[:]
227
228
def iter_names(self):
229
"""Yield a list of all names in this weave."""
230
return iter(self._names)
231
232
def idx_to_name(self, version):
233
return self._names[version]
234
235
def _check_repeated_add(self, name, parents, text, sha1):
236
"""Check that a duplicated add is OK.
237
238
If it is, return the (old) index; otherwise raise an exception.
239
"""
240
idx = self.lookup(name)
241
if sorted(self._parents[idx]) != sorted(parents) \
242
or sha1 != self._sha1s[idx]:
243
raise WeaveRevisionAlreadyPresent(name, self)
244
return idx
245
246
def add(self, name, parents, text, sha1=None):
172
173
def add(self, parents, text):
247
174
"""Add a single text on top of the weave.
248
175
249
176
Returns the index number of the newly added version.
250
177
251
name
252
Symbolic name for this version.
253
(Typically the revision-id of the revision that added it.)
254
255
178
parents
256
List or set of direct parent version numbers.
257
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
258
182
text
259
Sequence of lines to be added in the new version.
260
261
sha -- SHA-1 of the file, if known. This is trusted to be
262
correct if supplied.
263
"""
264
from bzrlib.osutils import sha_strings
265
266
assert isinstance(name, basestring)
267
if sha1 is None:
268
sha1 = sha_strings(text)
269
if name in self._name_map:
270
return self._check_repeated_add(name, parents, text, sha1)
271
272
parents = map(self.maybe_lookup, parents)
273
self._check_versions(parents)
183
Sequence of lines to be added in the new version."""
184
## self._check_versions(parents)
274
185
## self._check_lines(text)
275
new_version = len(self._parents)
276
277
278
# if we abort after here the (in-memory) weave will be corrupt because only
279
# some fields are updated
280
self._parents.append(parents[:])
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
281
236
self._sha1s.append(sha1)
282
self._names.append(name)
283
self._name_map[name] = new_version
284
285
286
if not parents:
287
# special case; adding with no parents revision; can do
288
# this more quickly by just appending unconditionally.
289
# even more specially, if we're adding an empty text we
290
# need do nothing at all.
291
if text:
292
self._weave.append(('{', new_version))
293
self._weave.extend(text)
294
self._weave.append(('}', None))
295
296
return new_version
297
298
if len(parents) == 1:
299
pv = list(parents)[0]
300
if sha1 == self._sha1s[pv]:
301
# special case: same as the single parent
302
return new_version
303
304
305
ancestors = self.inclusions(parents)
306
307
l = self._weave
308
309
# basis a list of (origin, lineno, line)
310
basis_lineno = []
311
basis_lines = []
312
for origin, lineno, line in self._extract(ancestors):
313
basis_lineno.append(lineno)
314
basis_lines.append(line)
315
316
# another small special case: a merge, producing the same text
317
# as auto-merge
318
if text == basis_lines:
319
return new_version
320
321
# add a sentinal, because we can also match against the final line
322
basis_lineno.append(len(self._weave))
323
324
# XXX: which line of the weave should we really consider
325
# matches the end of the file? the current code says it's the
326
# last line of the weave?
327
328
#print 'basis_lines:', basis_lines
329
#print 'new_lines: ', lines
330
331
s = SequenceMatcher(None, basis_lines, text)
332
333
# offset gives the number of lines that have been inserted
334
# into the weave up to the current point; if the original edit instruction
335
# says to change line A then we actually change (A+offset)
336
offset = 0
337
338
for tag, i1, i2, j1, j2 in s.get_opcodes():
339
# i1,i2 are given in offsets within basis_lines; we need to map them
340
# back to offsets within the entire weave
341
#print 'raw match', tag, i1, i2, j1, j2
342
if tag == 'equal':
343
continue
344
345
i1 = basis_lineno[i1]
346
i2 = basis_lineno[i2]
347
348
assert 0 <= j1 <= j2 <= len(text)
349
350
#print tag, i1, i2, j1, j2
351
352
# the deletion and insertion are handled separately.
353
# first delete the region.
354
if i1 != i2:
355
self._weave.insert(i1+offset, ('[', new_version))
356
self._weave.insert(i2+offset+1, (']', new_version))
357
offset += 2
358
359
if j1 != j2:
360
# there may have been a deletion spanning up to
361
# i2; we want to insert after this region to make sure
362
# we don't destroy ourselves
363
i = i2 + offset
364
self._weave[i:i] = ([('{', new_version)]
365
+ text[j1:j2]
366
+ [('}', None)])
367
offset += 2 + (j2 - j1)
368
369
return new_version
370
371
def add_identical(self, old_rev_id, new_rev_id, parents):
372
"""Add an identical text to old_rev_id as new_rev_id."""
373
old_lines = self.get(self.lookup(old_rev_id))
374
self.add(new_rev_id, parents, old_lines)
237
238
return idx
239
375
240
376
241
def inclusions(self, versions):
377
"""Return set of all ancestors of given version(s)."""
242
"""Expand out everything included by versions."""
378
243
i = set(versions)
379
for v in xrange(max(versions), 0, -1):
380
if v in i:
381
# include all its parents
382
i.update(self._parents[v])
244
for v in versions:
245
i.update(self._v[v])
383
246
return i
384
## except IndexError:
385
## raise ValueError("version %d not present in weave" % v)
386
387
388
def parents(self, version):
389
return self._parents[version]
390
391
392
def parent_names(self, version):
393
"""Return version names for parents of a version."""
394
return map(self.idx_to_name, self._parents[self.lookup(version)])
395
396
397
def minimal_parents(self, version):
398
"""Find the minimal set of parents for the version."""
399
included = self._parents[version]
400
if not included:
401
return []
402
403
li = list(included)
404
li.sort(reverse=True)
405
406
mininc = []
407
gotit = set()
408
409
for pv in li:
410
if pv not in gotit:
411
mininc.append(pv)
412
gotit.update(self.inclusions(pv))
413
414
assert mininc[0] >= 0
415
assert mininc[-1] < version
416
return mininc
417
247
248
249
def _addversion(self, parents):
250
if parents:
251
self._v.append(frozenset(parents))
252
else:
253
self._v.append(frozenset())
418
254
419
255
420
256
def _check_lines(self, text):
423
259
424
260
for l in text:
425
261
if not isinstance(l, basestring):
426
raise ValueError("text line should be a string or unicode, not %s"
427
% type(l))
262
raise ValueError("text line should be a string or unicode, not %s" % type(l))
428
263
429
264
430
265
432
267
"""Check everything in the sequence of indexes is valid"""
433
268
for i in indexes:
434
269
try:
435
self._parents[i]
270
self._v[i]
436
271
except IndexError:
437
272
raise IndexError("invalid version number %r" % i)
438
273
439
274
440
def annotate(self, name_or_index):
441
return list(self.annotate_iter(name_or_index))
442
443
444
def annotate_iter(self, name_or_index):
275
def annotate(self, index):
276
return list(self.annotate_iter(index))
277
278
279
def annotate_iter(self, version):
445
280
"""Yield list of (index-id, line) pairs for the specified version.
446
281
447
282
The index indicates when the line originated in the weave."""
448
incls = [self.maybe_lookup(name_or_index)]
449
for origin, lineno, text in self._extract(incls):
283
included = self.inclusions([version])
284
for origin, lineno, text in self._extract(included):
450
285
yield origin, text
451
286
452
def _walk(self):
453
"""Walk the weave.
454
455
Yields sequence of
456
(lineno, insert, deletes, text)
457
for each literal line.
458
"""
459
460
istack = []
461
dset = set()
462
463
lineno = 0 # line of weave, 0-based
464
465
for l in self._weave:
466
if isinstance(l, tuple):
467
c, v = l
468
isactive = None
469
if c == '{':
470
istack.append(v)
471
elif c == '}':
472
istack.pop()
473
elif c == '[':
474
assert v not in dset
475
dset.add(v)
476
elif c == ']':
477
dset.remove(v)
478
else:
479
raise WeaveFormatError('unexpected instruction %r' % v)
480
else:
481
assert isinstance(l, basestring)
482
assert istack
483
yield lineno, istack[-1], dset, l
484
lineno += 1
485
486
487
488
def _extract(self, versions):
287
288
def _extract(self, included):
489
289
"""Yield annotation of lines in included set.
490
290
491
291
Yields a sequence of tuples (origin, lineno, text), where
494
294
495
295
The set typically but not necessarily corresponds to a version.
496
296
"""
497
for i in versions:
498
if not isinstance(i, int):
499
raise ValueError(i)
500
501
included = self.inclusions(versions)
502
503
istack = []
504
dset = set()
505
506
lineno = 0 # line of weave, 0-based
297
istack = [] # versions for which an insertion block is current
298
299
dset = set() # versions for which a deletion block is current
507
300
508
301
isactive = None
509
302
510
result = []
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.
511
314
512
315
WFE = WeaveFormatError
513
316
514
for l in self._weave:
317
for l in self._l:
515
318
if isinstance(l, tuple):
319
isactive = None # recalculate
516
320
c, v = l
517
isactive = None
518
321
if c == '{':
519
assert v not in istack
322
if istack and (istack[-1] >= v):
323
raise WFE("improperly nested insertions %d>=%d on line %d"
324
% (istack[-1], v, lineno))
520
325
istack.append(v)
521
326
elif c == '}':
522
istack.pop()
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))
523
335
elif c == '[':
524
if v in included:
525
assert v not in dset
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
526
345
dset.add(v)
527
else:
528
assert c == ']'
529
if v in included:
530
assert v in dset
346
elif c == ']':
347
if v in dset:
531
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))
532
355
else:
533
356
assert isinstance(l, basestring)
534
if isactive is None:
535
isactive = (not dset) and istack and (istack[-1] in included)
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)
536
363
if isactive:
537
result.append((istack[-1], lineno, l))
364
origin = istack[-1]
365
yield origin, lineno, l
538
366
lineno += 1
367
539
368
if istack:
540
raise WeaveFormatError("unclosed insertion blocks "
541
"at end of weave: %s" % istack)
369
raise WFE("unclosed insertion blocks at end of weave",
370
istack)
542
371
if dset:
543
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
544
% dset)
545
return result
546
547
548
def get_iter(self, name_or_index):
372
raise WFE("unclosed deletion blocks at end of weave",
373
dset)
374
375
376
def get_iter(self, version):
549
377
"""Yield lines for the specified version."""
550
incls = [self.maybe_lookup(name_or_index)]
551
for origin, lineno, line in self._extract(incls):
378
for origin, lineno, line in self._extract(self.inclusions([version])):
552
379
yield line
553
380
554
381
555
def get_text(self, name_or_index):
556
return ''.join(self.get_iter(name_or_index))
557
assert isinstance(version, int)
558
559
560
def get_lines(self, name_or_index):
561
return list(self.get_iter(name_or_index))
562
563
564
get = get_lines
382
def get(self, index):
383
return list(self.get_iter(index))
565
384
566
385
567
386
def mash_iter(self, included):
568
387
"""Return composed version of multiple included versions."""
569
included = map(self.maybe_lookup, included)
388
included = frozenset(included)
570
389
for origin, lineno, text in self._extract(included):
571
390
yield text
572
391
573
392
574
393
def dump(self, to_file):
575
394
from pprint import pprint
576
print >>to_file, "Weave._weave = ",
577
pprint(self._weave, to_file)
578
print >>to_file, "Weave._parents = ",
579
pprint(self._parents, to_file)
395
print >>to_file, "Weave._l = ",
396
pprint(self._l, to_file)
397
print >>to_file, "Weave._v = ",
398
pprint(self._v, to_file)
580
399
581
400
582
401
583
402
def numversions(self):
584
l = len(self._parents)
403
l = len(self._v)
585
404
assert l == len(self._sha1s)
586
405
return l
587
406
588
407
589
def __len__(self):
590
return self.numversions()
591
592
593
def check(self, progress_bar=None):
408
def check(self):
594
409
# check no circular inclusions
595
410
for version in range(self.numversions()):
596
inclusions = list(self._parents[version])
411
inclusions = list(self._v[version])
597
412
if inclusions:
598
413
inclusions.sort()
599
414
if inclusions[-1] >= version:
602
417
603
418
# try extracting all versions; this is a bit slow and parallel
604
419
# extraction could be used
605
nv = self.numversions()
606
for version in range(nv):
607
if progress_bar:
608
progress_bar.update('checking text', version, nv)
420
import sha
421
for version in range(self.numversions()):
609
422
s = sha.new()
610
423
for l in self.get_iter(version):
611
424
s.update(l)
616
429
"got %s, expected %s"
617
430
% (version, hd, expected))
618
431
619
# TODO: check insertions are properly nested, that there are
620
# no lines outside of insertion blocks, that deletions are
621
# properly paired, etc.
622
623
432
624
433
625
434
def merge(self, merge_versions):
632
441
If there is a chunk of the file where there's no diagreement,
633
442
only one alternative is given.
634
443
"""
444
635
445
# approach: find the included versions common to all the
636
446
# merged versions
637
447
raise NotImplementedError()
657
467
If line1=line2, this is a pure insert; if newlines=[] this is a
658
468
pure delete. (Similar to difflib.)
659
469
"""
660
raise NotImplementedError()
661
662
663
def plan_merge(self, ver_a, ver_b):
664
"""Return pseudo-annotation indicating how the two versions merge.
665
666
This is computed between versions a and b and their common
667
base.
668
669
Weave lines present in none of them are skipped entirely.
670
"""
671
inc_a = self.inclusions([ver_a])
672
inc_b = self.inclusions([ver_b])
673
inc_c = inc_a & inc_b
674
675
for lineno, insert, deleteset, line in self._walk():
676
if deleteset & inc_c:
677
# killed in parent; can't be in either a or b
678
# not relevant to our work
679
yield 'killed-base', line
680
elif insert in inc_c:
681
# was inserted in base
682
killed_a = bool(deleteset & inc_a)
683
killed_b = bool(deleteset & inc_b)
684
if killed_a and killed_b:
685
yield 'killed-both', line
686
elif killed_a:
687
yield 'killed-a', line
688
elif killed_b:
689
yield 'killed-b', line
690
else:
691
yield 'unchanged', line
692
elif insert in inc_a:
693
if deleteset & inc_a:
694
yield 'ghost-a', line
695
else:
696
# new in A; not in B
697
yield 'new-a', line
698
elif insert in inc_b:
699
if deleteset & inc_b:
700
yield 'ghost-b', line
701
else:
702
yield 'new-b', line
703
else:
704
# not in either revision
705
yield 'irrelevant', line
706
707
yield 'unchanged', '' # terminator
708
709
710
711
def weave_merge(self, plan):
712
lines_a = []
713
lines_b = []
714
ch_a = ch_b = False
715
716
for state, line in plan:
717
if state == 'unchanged' or state == 'killed-both':
718
# resync and flush queued conflicts changes if any
719
if not lines_a and not lines_b:
720
pass
721
elif ch_a and not ch_b:
722
# one-sided change:
723
for l in lines_a: yield l
724
elif ch_b and not ch_a:
725
for l in lines_b: yield l
726
elif lines_a == lines_b:
727
for l in lines_a: yield l
728
else:
729
yield '<<<<<<<\n'
730
for l in lines_a: yield l
731
yield '=======\n'
732
for l in lines_b: yield l
733
yield '>>>>>>>\n'
734
735
del lines_a[:]
736
del lines_b[:]
737
ch_a = ch_b = False
738
739
if state == 'unchanged':
740
if line:
741
yield line
742
elif state == 'killed-a':
743
ch_a = True
744
lines_b.append(line)
745
elif state == 'killed-b':
746
ch_b = True
747
lines_a.append(line)
748
elif state == 'new-a':
749
ch_a = True
750
lines_a.append(line)
751
elif state == 'new-b':
752
ch_b = True
753
lines_b.append(line)
754
else:
755
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
756
'killed-both'), \
757
state
758
759
760
def join(self, other):
761
"""Integrate versions from other into this weave.
762
763
The resulting weave contains all the history of both weaves;
764
any version you could retrieve from either self or other can be
765
retrieved from self after this call.
766
767
It is illegal for the two weaves to contain different values
768
or different parents for any version. See also reweave().
769
"""
770
if other.numversions() == 0:
771
return # nothing to update, easy
772
# two loops so that we do not change ourselves before verifying it
773
# will be ok
774
# work through in index order to make sure we get all dependencies
775
for other_idx, name in enumerate(other._names):
776
if self._check_version_consistent(other, other_idx, name):
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':
777
493
continue
778
for other_idx, name in enumerate(other._names):
779
# TODO: If all the parents of the other version are already
780
# present then we can avoid some work by just taking the delta
781
# and adjusting the offsets.
782
new_parents = self._imported_parents(other, other_idx)
783
lines = other.get_lines(other_idx)
784
sha1 = other._sha1s[other_idx]
785
self.add(name, new_parents, lines, sha1)
786
787
788
def _imported_parents(self, other, other_idx):
789
"""Return list of parents in self corresponding to indexes in other."""
790
new_parents = []
791
for parent_idx in other._parents[other_idx]:
792
parent_name = other._names[parent_idx]
793
if parent_name not in self._names:
794
# should not be possible
795
raise WeaveError("missing parent {%s} of {%s} in %r"
796
% (parent_name, other._name_map[other_idx], self))
797
new_parents.append(self._name_map[parent_name])
798
return new_parents
799
800
def _check_version_consistent(self, other, other_idx, name):
801
"""Check if a version in consistent in this and other.
802
803
To be consistent it must have:
804
805
* the same text
806
* the same direct parents (by name, not index, and disregarding
807
order)
808
809
If present & correct return True;
810
if not present in self return False;
811
if inconsistent raise error."""
812
this_idx = self._name_map.get(name, -1)
813
if this_idx != -1:
814
if self._sha1s[this_idx] != other._sha1s[other_idx]:
815
raise WeaveError("inconsistent texts for version {%s} "
816
"when joining weaves"
817
% (name))
818
self_parents = self._parents[this_idx]
819
other_parents = other._parents[other_idx]
820
n1 = [self._names[i] for i in self_parents]
821
n2 = [other._names[i] for i in other_parents]
822
n1.sort()
823
n2.sort()
824
if n1 != n2:
825
raise WeaveParentMismatch("inconsistent parents "
826
"for version {%s}: %s vs %s" % (name, n1, n2))
827
else:
828
return True # ok!
829
else:
830
return False
831
832
def reweave(self, other):
833
"""Reweave self with other."""
834
new_weave = reweave(self, other)
835
for attr in self.__slots__:
836
setattr(self, attr, getattr(new_weave, attr))
837
838
839
def reweave(wa, wb):
840
"""Combine two weaves and return the result.
841
842
This works even if a revision R has different parents in
843
wa and wb. In the resulting weave all the parents are given.
844
845
This is done by just building up a new weave, maintaining ordering
846
of the versions in the two inputs. More efficient approaches
847
might be possible but it should only be necessary to do
848
this operation rarely, when a new previously ghost version is
849
inserted.
850
"""
851
wr = Weave()
852
ia = ib = 0
853
queue_a = range(wa.numversions())
854
queue_b = range(wb.numversions())
855
# first determine combined parents of all versions
856
# map from version name -> all parent names
857
combined_parents = _reweave_parent_graphs(wa, wb)
858
mutter("combined parents: %r", combined_parents)
859
order = topo_sort(combined_parents.iteritems())
860
mutter("order to reweave: %r", order)
861
for name in order:
862
if name in wa._name_map:
863
lines = wa.get_lines(name)
864
if name in wb._name_map:
865
assert lines == wb.get_lines(name)
866
else:
867
lines = wb.get_lines(name)
868
wr.add(name, combined_parents[name], lines)
869
return wr
870
871
872
def _reweave_parent_graphs(wa, wb):
873
"""Return combined parent ancestry for two weaves.
874
875
Returned as a list of (version_name, set(parent_names))"""
876
combined = {}
877
for weave in [wa, wb]:
878
for idx, name in enumerate(weave._names):
879
p = combined.setdefault(name, set())
880
p.update(map(weave.idx_to_name, weave._parents[idx]))
881
return combined
882
883
884
def weave_toc(w):
885
"""Show the weave's table-of-contents"""
886
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
887
for i in (6, 50, 10, 10):
888
print '-' * i,
889
print
890
for i in range(w.numversions()):
891
sha1 = w._sha1s[i]
892
name = w._names[i]
893
parent_str = ' '.join(map(str, w._parents[i]))
894
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
895
896
897
898
def weave_stats(weave_file, pb):
899
from bzrlib.weavefile import read_weave
900
901
wf = file(weave_file, 'rb')
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')
902
512
w = read_weave(wf)
903
513
# FIXME: doesn't work on pipes
904
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)
905
517
906
518
total = 0
907
vers = len(w)
908
for i in range(vers):
909
pb.update('checking sizes', i, vers)
910
for origin, lineno, line in w._extract([i]):
911
total += len(line)
912
913
pb.clear()
914
915
print 'versions %9d' % vers
916
print 'weave file %9d bytes' % weave_size
917
print 'total contents %9d bytes' % total
918
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
919
if vers:
920
avg = total/vers
921
print 'average size %9d bytes' % avg
922
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
923
924
925
def usage():
926
print """bzr weave tool
927
928
Experimental tool for weave algorithm.
929
930
usage:
931
weave init WEAVEFILE
932
Create an empty weave file
933
weave get WEAVEFILE VERSION
934
Write out specified version.
935
weave check WEAVEFILE
936
Check consistency of all versions.
937
weave toc WEAVEFILE
938
Display table of contents.
939
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
940
Add NEWTEXT, with specified parent versions.
941
weave annotate WEAVEFILE VERSION
942
Display origin of each line.
943
weave mash WEAVEFILE VERSION...
944
Display composite of all selected versions.
945
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
946
Auto-merge two versions and display conflicts.
947
weave diff WEAVEFILE VERSION1 VERSION2
948
Show differences between two versions.
949
950
example:
951
952
% weave init foo.weave
953
% vi foo.txt
954
% weave add foo.weave ver0 < foo.txt
955
added version 0
956
957
(create updated version)
958
% vi foo.txt
959
% weave get foo.weave 0 | diff -u - foo.txt
960
% weave add foo.weave ver1 0 < foo.txt
961
added version 1
962
963
% weave get foo.weave 0 > foo.txt (create forked version)
964
% vi foo.txt
965
% weave add foo.weave ver2 0 < foo.txt
966
added version 2
967
968
% weave merge foo.weave 1 2 > foo.txt (merge them)
969
% vi foo.txt (resolve conflicts)
970
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
971
972
"""
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))
973
531
974
532
975
533
976
534
def main(argv):
977
535
import sys
978
536
import os
979
try:
980
import bzrlib
981
except ImportError:
982
# in case we're run directly from the subdirectory
983
sys.path.append('..')
984
import bzrlib
985
from bzrlib.weavefile import write_weave, read_weave
986
from bzrlib.progress import ProgressBar
987
988
try:
989
import psyco
990
psyco.full()
991
except ImportError:
992
pass
993
994
if len(argv) < 2:
995
usage()
996
return 0
997
537
from weavefile import write_weave_v1, read_weave
998
538
cmd = argv[1]
999
1000
def readit():
1001
return read_weave(file(argv[2], 'rb'))
1002
1003
if cmd == 'help':
1004
usage()
1005
elif cmd == 'add':
1006
w = readit()
539
if cmd == 'add':
540
w = read_weave(file(argv[2], 'rb'))
1007
541
# at the moment, based on everything in the file
1008
name = argv[3]
1009
parents = map(int, argv[4:])
542
parents = set(range(len(w._v)))
1010
543
lines = sys.stdin.readlines()
1011
ver = w.add(name, parents, lines)
1012
write_weave(w, file(argv[2], 'wb'))
1013
print 'added version %r %d' % (name, ver)
544
ver = w.add(parents, lines)
545
write_weave_v1(w, file(argv[2], 'wb'))
546
print 'added %d' % ver
1014
547
elif cmd == 'init':
1015
548
fn = argv[2]
1016
549
if os.path.exists(fn):
1017
550
raise IOError("file exists")
1018
551
w = Weave()
1019
write_weave(w, file(fn, 'wb'))
1020
elif cmd == 'get': # get one version
1021
w = readit()
552
write_weave_v1(w, file(fn, 'wb'))
553
elif cmd == 'get':
554
w = read_weave(file(argv[2], 'rb'))
1022
555
sys.stdout.writelines(w.get_iter(int(argv[3])))
1023
1024
elif cmd == 'mash': # get composite
1025
w = readit()
1026
sys.stdout.writelines(w.mash_iter(map(int, argv[3:])))
1027
1028
elif cmd == 'diff':
1029
from difflib import unified_diff
1030
w = readit()
1031
fn = argv[2]
1032
v1, v2 = map(int, argv[3:5])
1033
lines1 = w.get(v1)
1034
lines2 = w.get(v2)
1035
diff_gen = unified_diff(lines1, lines2,
1036
'%s version %d' % (fn, v1),
1037
'%s version %d' % (fn, v2))
1038
sys.stdout.writelines(diff_gen)
1039
1040
556
elif cmd == 'annotate':
1041
w = readit()
557
w = read_weave(file(argv[2], 'rb'))
1042
558
# newline is added to all lines regardless; too hard to get
1043
559
# reasonable formatting otherwise
1044
560
lasto = None
1049
565
else:
1050
566
print '%5d | %s' % (origin, text)
1051
567
lasto = origin
1052
1053
elif cmd == 'toc':
1054
weave_toc(readit())
1055
1056
elif cmd == 'stats':
1057
weave_stats(argv[2], ProgressBar())
1058
568
elif cmd == 'info':
569
weave_info(argv[2], sys.stdout)
1059
570
elif cmd == 'check':
1060
w = readit()
1061
pb = ProgressBar()
1062
w.check(pb)
1063
pb.clear()
1064
print '%d versions ok' % w.numversions()
1065
1066
elif cmd == 'inclusions':
1067
w = readit()
1068
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1069
1070
elif cmd == 'parents':
1071
w = readit()
1072
print ' '.join(map(str, w._parents[int(argv[3])]))
1073
1074
elif cmd == 'plan-merge':
1075
w = readit()
1076
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1077
if line:
1078
print '%14s | %s' % (state, line),
1079
1080
elif cmd == 'merge':
1081
w = readit()
1082
p = w.plan_merge(int(argv[3]), int(argv[4]))
1083
sys.stdout.writelines(w.weave_merge(p))
1084
1085
elif cmd == 'mash-merge':
1086
if len(argv) != 5:
1087
usage()
1088
return 1
1089
1090
w = readit()
1091
v1, v2 = map(int, argv[3:5])
1092
1093
basis = w.inclusions([v1]).intersection(w.inclusions([v2]))
1094
1095
base_lines = list(w.mash_iter(basis))
1096
a_lines = list(w.get(v1))
1097
b_lines = list(w.get(v2))
1098
1099
from bzrlib.merge3 import Merge3
1100
m3 = Merge3(base_lines, a_lines, b_lines)
1101
1102
name_a = 'version %d' % v1
1103
name_b = 'version %d' % v2
1104
sys.stdout.writelines(m3.merge_lines(name_a=name_a, name_b=name_b))
571
w = read_weave(file(argv[2], 'rb'))
572
w.check()
1105
573
else:
1106
574
raise ValueError('unknown command %r' % cmd)
1107
575
1108
576
1109
1110
def profile_main(argv):
1111
import tempfile, hotshot, hotshot.stats
1112
1113
prof_f = tempfile.NamedTemporaryFile()
1114
1115
prof = hotshot.Profile(prof_f.name)
1116
1117
ret = prof.runcall(main, argv)
1118
prof.close()
1119
1120
stats = hotshot.stats.load(prof_f.name)
1121
#stats.strip_dirs()
1122
stats.sort_stats('cumulative')
1123
## XXX: Might like to write to stderr or the trace file instead but
1124
## print_stats seems hardcoded to stdout
1125
stats.print_stats(20)
1126
1127
return ret
1128
1129
1130
577
if __name__ == '__main__':
1131
578
import sys
1132
if '--profile' in sys.argv:
1133
args = sys.argv[:]
1134
args.remove('--profile')
1135
sys.exit(profile_main(args))
1136
else:
1137
sys.exit(main(sys.argv))
1138
579
sys.exit(main(sys.argv))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1