« back to all changes in this revision
Viewing changes to bzrlib/weave.py
- browse files at revision 886
- compare with another revision
- download diff
- download tarball
- view history from revision 886
- Committer: Martin Pool
- Date: 2005-07-11 03:42:53 UTC
- Revision ID: mbp@sourcefrog.net-20050711034253-412281abeb9f56ad
doc
- files added:
- bzrlib/mdiff.py
- plugins/changeset
- files removed:
- BRANCH.TODO
- bzrlib/export
- bzrlib/store
- bzrlib/tests/blackbox
- bzrlib/transport
- bzrlib/ui
- 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/test_too_much.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/test_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 os
71
import sha
72
from difflib import SequenceMatcher
73
74
from bzrlib.trace import mutter
75
from bzrlib.errors import (WeaveError, WeaveFormatError, WeaveParentMismatch,
76
WeaveRevisionNotPresent, WeaveRevisionAlreadyPresent)
77
import bzrlib.errors as errors
78
from bzrlib.tsort import topo_sort
79
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
80
75
81
76
class Weave(object):
82
77
"""weave - versioned text file storage.
99
94
the version-id is used to reference it in the larger world.
100
95
101
96
The weave is represented as a list mixing edit instructions and
102
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
103
98
unicode), or a tuple. If a string, it means that the given line
104
99
should be output in the currently active revisions.
105
100
110
105
The instruction can be '{' or '}' for an insertion block, and '['
111
106
and ']' for a deletion block respectively. The version is the
112
107
integer version index. There is no replace operator, only deletes
113
and inserts. For '}', the end of an insertion, there is no
114
version parameter because it always closes the most recently
115
opened insertion.
108
and inserts.
116
109
117
110
Constraints/notes:
118
111
145
138
should be no way to get an earlier version deleting a later
146
139
version.
147
140
148
_weave
149
Text of the weave; list of control instruction tuples and strings.
150
151
_parents
152
List of parents, indexed by version number.
153
It is only necessary to store the minimal set of parents for
154
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.
155
152
156
153
_sha1s
157
List of hex SHA-1 of each version.
158
159
_names
160
List of symbolic names for each version. Each should be unique.
161
162
_name_map
163
For each name, the version number.
164
165
_weave_name
166
Descriptive name of this weave; typically the filename if known.
167
Set by read_weave.
154
List of hex SHA-1 of each version, or None if not recorded.
168
155
"""
169
170
__slots__ = ['_weave', '_parents', '_sha1s', '_names', '_name_map',
171
'_weave_name']
172
173
def __init__(self, weave_name=None):
174
self._weave = []
175
self._parents = []
156
def __init__(self):
157
self._l = []
158
self._v = []
176
159
self._sha1s = []
177
self._names = []
178
self._name_map = {}
179
self._weave_name = weave_name
180
181
def __repr__(self):
182
return "Weave(%r)" % self._weave_name
183
184
185
def copy(self):
186
"""Return a deep copy of self.
187
188
The copy can be modified without affecting the original weave."""
189
other = Weave()
190
other._weave = self._weave[:]
191
other._parents = self._parents[:]
192
other._sha1s = self._sha1s[:]
193
other._names = self._names[:]
194
other._name_map = self._name_map.copy()
195
other._weave_name = self._weave_name
196
return other
160
197
161
198
162
def __eq__(self, other):
199
163
if not isinstance(other, Weave):
200
164
return False
201
return self._parents == other._parents \
202
and self._weave == other._weave \
203
and self._sha1s == other._sha1s
165
return self._v == other._v \
166
and self._l == other._l
167
204
168
205
206
169
def __ne__(self, other):
207
170
return not self.__eq__(other)
208
171
209
def __contains__(self, name):
210
return self._name_map.has_key(name)
211
212
def maybe_lookup(self, name_or_index):
213
"""Convert possible symbolic name to index, or pass through indexes."""
214
if isinstance(name_or_index, (int, long)):
215
return name_or_index
216
else:
217
return self.lookup(name_or_index)
218
219
220
def lookup(self, name):
221
"""Convert symbolic version name to index."""
222
try:
223
return self._name_map[name]
224
except KeyError:
225
raise WeaveRevisionNotPresent(name, self)
226
227
def names(self):
228
return self._names[:]
229
230
def iter_names(self):
231
"""Yield a list of all names in this weave."""
232
return iter(self._names)
233
234
def idx_to_name(self, version):
235
return self._names[version]
236
237
def _check_repeated_add(self, name, parents, text, sha1):
238
"""Check that a duplicated add is OK.
239
240
If it is, return the (old) index; otherwise raise an exception.
241
"""
242
idx = self.lookup(name)
243
if sorted(self._parents[idx]) != sorted(parents) \
244
or sha1 != self._sha1s[idx]:
245
raise WeaveRevisionAlreadyPresent(name, self)
246
return idx
247
248
def add(self, name, parents, text, sha1=None):
172
173
def add(self, parents, text):
249
174
"""Add a single text on top of the weave.
250
175
251
176
Returns the index number of the newly added version.
252
177
253
name
254
Symbolic name for this version.
255
(Typically the revision-id of the revision that added it.)
256
257
178
parents
258
List or set of direct parent version numbers.
259
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
260
182
text
261
Sequence of lines to be added in the new version.
262
263
sha -- SHA-1 of the file, if known. This is trusted to be
264
correct if supplied.
265
"""
266
from bzrlib.osutils import sha_strings
267
268
assert isinstance(name, basestring)
269
if sha1 is None:
270
sha1 = sha_strings(text)
271
if name in self._name_map:
272
return self._check_repeated_add(name, parents, text, sha1)
273
274
parents = map(self.maybe_lookup, parents)
275
self._check_versions(parents)
183
Sequence of lines to be added in the new version."""
184
## self._check_versions(parents)
276
185
## self._check_lines(text)
277
new_version = len(self._parents)
278
279
280
# if we abort after here the (in-memory) weave will be corrupt because only
281
# some fields are updated
282
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
283
236
self._sha1s.append(sha1)
284
self._names.append(name)
285
self._name_map[name] = new_version
286
287
288
if not parents:
289
# special case; adding with no parents revision; can do
290
# this more quickly by just appending unconditionally.
291
# even more specially, if we're adding an empty text we
292
# need do nothing at all.
293
if text:
294
self._weave.append(('{', new_version))
295
self._weave.extend(text)
296
self._weave.append(('}', None))
297
298
return new_version
299
300
if len(parents) == 1:
301
pv = list(parents)[0]
302
if sha1 == self._sha1s[pv]:
303
# special case: same as the single parent
304
return new_version
305
306
307
ancestors = self.inclusions(parents)
308
309
l = self._weave
310
311
# basis a list of (origin, lineno, line)
312
basis_lineno = []
313
basis_lines = []
314
for origin, lineno, line in self._extract(ancestors):
315
basis_lineno.append(lineno)
316
basis_lines.append(line)
317
318
# another small special case: a merge, producing the same text
319
# as auto-merge
320
if text == basis_lines:
321
return new_version
322
323
# add a sentinal, because we can also match against the final line
324
basis_lineno.append(len(self._weave))
325
326
# XXX: which line of the weave should we really consider
327
# matches the end of the file? the current code says it's the
328
# last line of the weave?
329
330
#print 'basis_lines:', basis_lines
331
#print 'new_lines: ', lines
332
333
s = SequenceMatcher(None, basis_lines, text)
334
335
# offset gives the number of lines that have been inserted
336
# into the weave up to the current point; if the original edit instruction
337
# says to change line A then we actually change (A+offset)
338
offset = 0
339
340
for tag, i1, i2, j1, j2 in s.get_opcodes():
341
# i1,i2 are given in offsets within basis_lines; we need to map them
342
# back to offsets within the entire weave
343
#print 'raw match', tag, i1, i2, j1, j2
344
if tag == 'equal':
345
continue
346
347
i1 = basis_lineno[i1]
348
i2 = basis_lineno[i2]
349
350
assert 0 <= j1 <= j2 <= len(text)
351
352
#print tag, i1, i2, j1, j2
353
354
# the deletion and insertion are handled separately.
355
# first delete the region.
356
if i1 != i2:
357
self._weave.insert(i1+offset, ('[', new_version))
358
self._weave.insert(i2+offset+1, (']', new_version))
359
offset += 2
360
361
if j1 != j2:
362
# there may have been a deletion spanning up to
363
# i2; we want to insert after this region to make sure
364
# we don't destroy ourselves
365
i = i2 + offset
366
self._weave[i:i] = ([('{', new_version)]
367
+ text[j1:j2]
368
+ [('}', None)])
369
offset += 2 + (j2 - j1)
370
371
return new_version
372
373
def add_identical(self, old_rev_id, new_rev_id, parents):
374
"""Add an identical text to old_rev_id as new_rev_id."""
375
old_lines = self.get(self.lookup(old_rev_id))
376
self.add(new_rev_id, parents, old_lines)
237
238
return idx
239
377
240
378
241
def inclusions(self, versions):
379
"""Return set of all ancestors of given version(s)."""
242
"""Expand out everything included by versions."""
380
243
i = set(versions)
381
for v in xrange(max(versions), 0, -1):
382
if v in i:
383
# include all its parents
384
i.update(self._parents[v])
244
for v in versions:
245
try:
246
i.update(self._v[v])
247
except IndexError:
248
raise ValueError("version %d not present in weave" % v)
385
249
return i
386
## except IndexError:
387
## raise ValueError("version %d not present in weave" % v)
388
389
390
def parents(self, version):
391
return self._parents[version]
392
393
394
def parent_names(self, version):
395
"""Return version names for parents of a version."""
396
return map(self.idx_to_name, self._parents[self.lookup(version)])
397
398
399
def minimal_parents(self, version):
400
"""Find the minimal set of parents for the version."""
401
included = self._parents[version]
402
if not included:
403
return []
404
405
li = list(included)
406
li.sort(reverse=True)
407
408
mininc = []
409
gotit = set()
410
411
for pv in li:
412
if pv not in gotit:
413
mininc.append(pv)
414
gotit.update(self.inclusions(pv))
415
416
assert mininc[0] >= 0
417
assert mininc[-1] < version
418
return mininc
419
250
251
252
def _addversion(self, parents):
253
if parents:
254
self._v.append(frozenset(parents))
255
else:
256
self._v.append(frozenset())
420
257
421
258
422
259
def _check_lines(self, text):
434
271
"""Check everything in the sequence of indexes is valid"""
435
272
for i in indexes:
436
273
try:
437
self._parents[i]
274
self._v[i]
438
275
except IndexError:
439
276
raise IndexError("invalid version number %r" % i)
440
277
441
278
442
def annotate(self, name_or_index):
443
return list(self.annotate_iter(name_or_index))
444
445
446
def annotate_iter(self, name_or_index):
279
def annotate(self, index):
280
return list(self.annotate_iter(index))
281
282
283
def annotate_iter(self, version):
447
284
"""Yield list of (index-id, line) pairs for the specified version.
448
285
449
286
The index indicates when the line originated in the weave."""
450
incls = [self.maybe_lookup(name_or_index)]
451
for origin, lineno, text in self._extract(incls):
287
included = self.inclusions([version])
288
for origin, lineno, text in self._extract(included):
452
289
yield origin, text
453
290
454
def _walk(self):
455
"""Walk the weave.
456
457
Yields sequence of
458
(lineno, insert, deletes, text)
459
for each literal line.
460
"""
461
462
istack = []
463
dset = set()
464
465
lineno = 0 # line of weave, 0-based
466
467
for l in self._weave:
468
if isinstance(l, tuple):
469
c, v = l
470
isactive = None
471
if c == '{':
472
istack.append(v)
473
elif c == '}':
474
istack.pop()
475
elif c == '[':
476
assert v not in dset
477
dset.add(v)
478
elif c == ']':
479
dset.remove(v)
480
else:
481
raise WeaveFormatError('unexpected instruction %r' % v)
482
else:
483
assert isinstance(l, basestring)
484
assert istack
485
yield lineno, istack[-1], dset, l
486
lineno += 1
487
488
if istack:
489
raise WeaveFormatError("unclosed insertion blocks "
490
"at end of weave: %s" % istack)
491
if dset:
492
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
493
% dset)
494
495
def _extract(self, versions):
291
292
def _extract(self, included):
496
293
"""Yield annotation of lines in included set.
497
294
498
295
Yields a sequence of tuples (origin, lineno, text), where
501
298
502
299
The set typically but not necessarily corresponds to a version.
503
300
"""
504
for i in versions:
505
if not isinstance(i, int):
506
raise ValueError(i)
507
508
included = self.inclusions(versions)
509
301
510
302
istack = []
511
303
dset = set()
512
304
513
305
lineno = 0 # line of weave, 0-based
514
515
isactive = None
516
517
result = []
306
isactive = False
518
307
519
308
WFE = WeaveFormatError
520
309
521
for l in self._weave:
310
for l in self._l:
522
311
if isinstance(l, tuple):
523
312
c, v = l
524
isactive = None
525
if c == '{':
526
assert v not in istack
527
istack.append(v)
528
elif c == '}':
529
istack.pop()
530
elif c == '[':
531
if v in included:
313
if v in included: # only active blocks are interesting
314
if c == '{':
315
assert v not in istack
316
istack.append(v)
317
isactive = not dset
318
elif c == '}':
319
oldv = istack.pop()
320
assert oldv == v
321
isactive = istack and not dset
322
elif c == '[':
532
323
assert v not in dset
533
324
dset.add(v)
534
else:
535
assert c == ']'
536
if v in included:
325
isactive = False
326
else:
327
assert c == ']'
537
328
assert v in dset
538
329
dset.remove(v)
330
isactive = istack and not dset
539
331
else:
540
332
assert isinstance(l, basestring)
541
if isactive is None:
542
isactive = (not dset) and istack and (istack[-1] in included)
543
333
if isactive:
544
result.append((istack[-1], lineno, l))
334
yield origin, lineno, l
545
335
lineno += 1
336
546
337
if istack:
547
raise WeaveFormatError("unclosed insertion blocks "
548
"at end of weave: %s" % istack)
338
raise WFE("unclosed insertion blocks at end of weave",
339
istack)
549
340
if dset:
550
raise WeaveFormatError("unclosed deletion blocks at end of weave: %s"
551
% dset)
552
return result
553
554
555
def get_iter(self, name_or_index):
341
raise WFE("unclosed deletion blocks at end of weave",
342
dset)
343
344
345
def get_iter(self, version):
556
346
"""Yield lines for the specified version."""
557
incls = [self.maybe_lookup(name_or_index)]
558
if len(incls) == 1:
559
index = incls[0]
560
cur_sha = sha.new()
561
else:
562
# We don't have sha1 sums for multiple entries
563
cur_sha = None
564
for origin, lineno, line in self._extract(incls):
565
if cur_sha:
566
cur_sha.update(line)
347
for origin, lineno, line in self._extract(self.inclusions([version])):
567
348
yield line
568
if cur_sha:
569
expected_sha1 = self._sha1s[index]
570
measured_sha1 = cur_sha.hexdigest()
571
if measured_sha1 != expected_sha1:
572
raise errors.WeaveInvalidChecksum(
573
'file %s, revision %s, expected: %s, measured %s'
574
% (self._weave_name, self._names[index],
575
expected_sha1, measured_sha1))
576
577
578
def get_text(self, name_or_index):
579
return ''.join(self.get_iter(name_or_index))
580
assert isinstance(version, int)
581
582
583
def get_lines(self, name_or_index):
584
return list(self.get_iter(name_or_index))
585
586
587
get = get_lines
588
589
590
def get_sha1(self, name):
591
"""Get the stored sha1 sum for the given revision.
592
593
:param name: The name of the version to lookup
594
"""
595
return self._sha1s[self.lookup(name)]
349
350
351
def get(self, index):
352
return list(self.get_iter(index))
353
596
354
597
355
def mash_iter(self, included):
598
356
"""Return composed version of multiple included versions."""
599
included = map(self.maybe_lookup, included)
357
included = frozenset(included)
600
358
for origin, lineno, text in self._extract(included):
601
359
yield text
602
360
603
361
604
362
def dump(self, to_file):
605
363
from pprint import pprint
606
print >>to_file, "Weave._weave = ",
607
pprint(self._weave, to_file)
608
print >>to_file, "Weave._parents = ",
609
pprint(self._parents, to_file)
364
print >>to_file, "Weave._l = ",
365
pprint(self._l, to_file)
366
print >>to_file, "Weave._v = ",
367
pprint(self._v, to_file)
610
368
611
369
612
370
613
371
def numversions(self):
614
l = len(self._parents)
372
l = len(self._v)
615
373
assert l == len(self._sha1s)
616
374
return l
617
375
618
376
619
def __len__(self):
620
return self.numversions()
621
622
def check(self, progress_bar=None):
377
def check(self):
623
378
# check no circular inclusions
624
379
for version in range(self.numversions()):
625
inclusions = list(self._parents[version])
380
inclusions = list(self._v[version])
626
381
if inclusions:
627
382
inclusions.sort()
628
383
if inclusions[-1] >= version:
629
384
raise WeaveFormatError("invalid included version %d for index %d"
630
385
% (inclusions[-1], version))
631
386
632
# try extracting all versions; parallel extraction is used
633
nv = self.numversions()
634
sha1s = [sha.new() for i in range(nv)]
635
texts = [[] for i in range(nv)]
636
inclusions = []
637
for i in range(nv):
638
# For creating the ancestry, IntSet is much faster (3.7s vs 0.17s)
639
# The problem is that set membership is much more expensive
640
new_inc = set([i])
641
for p in self._parents[i]:
642
new_inc.update(inclusions[p])
643
644
#assert set(new_inc) == self.inclusions([i]), 'failed %s != %s' % (new_inc, self.inclusions([i]))
645
inclusions.append(new_inc)
646
647
nlines = len(self._weave)
648
649
update_text = 'checking weave'
650
if self._weave_name:
651
short_name = os.path.basename(self._weave_name)
652
update_text = 'checking %s' % (short_name,)
653
update_text = update_text[:25]
654
655
for lineno, insert, deleteset, line in self._walk():
656
if progress_bar:
657
progress_bar.update(update_text, lineno, nlines)
658
659
for j, j_inc in enumerate(inclusions):
660
# The active inclusion must be an ancestor,
661
# and no ancestors must have deleted this line,
662
# because we don't support resurrection.
663
if (insert in j_inc) and not (deleteset & j_inc):
664
sha1s[j].update(line)
665
666
for version in range(nv):
667
hd = sha1s[version].hexdigest()
387
# try extracting all versions; this is a bit slow and parallel
388
# extraction could be used
389
import sha
390
for version in range(self.numversions()):
391
s = sha.new()
392
for l in self.get_iter(version):
393
s.update(l)
394
hd = s.hexdigest()
668
395
expected = self._sha1s[version]
669
396
if hd != expected:
670
raise errors.WeaveInvalidChecksum(
671
"mismatched sha1 for version %s: "
672
"got %s, expected %s"
673
% (self._names[version], hd, expected))
397
raise WeaveError("mismatched sha1 for version %d; "
398
"got %s, expected %s"
399
% (version, hd, expected))
674
400
675
401
# TODO: check insertions are properly nested, that there are
676
402
# no lines outside of insertion blocks, that deletions are
677
403
# properly paired, etc.
678
404
405
406
407
def merge(self, merge_versions):
408
"""Automerge and mark conflicts between versions.
409
410
This returns a sequence, each entry describing alternatives
411
for a chunk of the file. Each of the alternatives is given as
412
a list of lines.
413
414
If there is a chunk of the file where there's no diagreement,
415
only one alternative is given.
416
"""
417
418
# approach: find the included versions common to all the
419
# merged versions
420
raise NotImplementedError()
421
422
423
679
424
def _delta(self, included, lines):
680
425
"""Return changes from basis to new revision.
681
426
695
440
If line1=line2, this is a pure insert; if newlines=[] this is a
696
441
pure delete. (Similar to difflib.)
697
442
"""
698
raise NotImplementedError()
699
700
701
def plan_merge(self, ver_a, ver_b):
702
"""Return pseudo-annotation indicating how the two versions merge.
703
704
This is computed between versions a and b and their common
705
base.
706
707
Weave lines present in none of them are skipped entirely.
708
"""
709
inc_a = self.inclusions([ver_a])
710
inc_b = self.inclusions([ver_b])
711
inc_c = inc_a & inc_b
712
713
for lineno, insert, deleteset, line in self._walk():
714
if deleteset & inc_c:
715
# killed in parent; can't be in either a or b
716
# not relevant to our work
717
yield 'killed-base', line
718
elif insert in inc_c:
719
# was inserted in base
720
killed_a = bool(deleteset & inc_a)
721
killed_b = bool(deleteset & inc_b)
722
if killed_a and killed_b:
723
yield 'killed-both', line
724
elif killed_a:
725
yield 'killed-a', line
726
elif killed_b:
727
yield 'killed-b', line
728
else:
729
yield 'unchanged', line
730
elif insert in inc_a:
731
if deleteset & inc_a:
732
yield 'ghost-a', line
733
else:
734
# new in A; not in B
735
yield 'new-a', line
736
elif insert in inc_b:
737
if deleteset & inc_b:
738
yield 'ghost-b', line
739
else:
740
yield 'new-b', line
741
else:
742
# not in either revision
743
yield 'irrelevant', line
744
745
yield 'unchanged', '' # terminator
746
747
748
749
def weave_merge(self, plan):
750
lines_a = []
751
lines_b = []
752
ch_a = ch_b = False
753
# TODO: Return a structured form of the conflicts (e.g. 2-tuples for
754
# conflicted regions), rather than just inserting the markers.
755
#
756
# TODO: Show some version information (e.g. author, date) on
757
# conflicted regions.
758
for state, line in plan:
759
if state == 'unchanged' or state == 'killed-both':
760
# resync and flush queued conflicts changes if any
761
if not lines_a and not lines_b:
762
pass
763
elif ch_a and not ch_b:
764
# one-sided change:
765
for l in lines_a: yield l
766
elif ch_b and not ch_a:
767
for l in lines_b: yield l
768
elif lines_a == lines_b:
769
for l in lines_a: yield l
770
else:
771
yield '<<<<<<<\n'
772
for l in lines_a: yield l
773
yield '=======\n'
774
for l in lines_b: yield l
775
yield '>>>>>>>\n'
776
777
del lines_a[:]
778
del lines_b[:]
779
ch_a = ch_b = False
780
781
if state == 'unchanged':
782
if line:
783
yield line
784
elif state == 'killed-a':
785
ch_a = True
786
lines_b.append(line)
787
elif state == 'killed-b':
788
ch_b = True
789
lines_a.append(line)
790
elif state == 'new-a':
791
ch_a = True
792
lines_a.append(line)
793
elif state == 'new-b':
794
ch_b = True
795
lines_b.append(line)
796
else:
797
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
798
'killed-both'), \
799
state
800
801
802
def join(self, other):
803
import sys, time
804
"""Integrate versions from other into this weave.
805
806
The resulting weave contains all the history of both weaves;
807
any version you could retrieve from either self or other can be
808
retrieved from self after this call.
809
810
It is illegal for the two weaves to contain different values
811
or different parents for any version. See also reweave().
812
"""
813
if other.numversions() == 0:
814
return # nothing to update, easy
815
# two loops so that we do not change ourselves before verifying it
816
# will be ok
817
# work through in index order to make sure we get all dependencies
818
for other_idx, name in enumerate(other._names):
819
self._check_version_consistent(other, other_idx, name)
820
821
merged = 0
822
processed = 0
823
time0 = time.time( )
824
for other_idx, name in enumerate(other._names):
825
# TODO: If all the parents of the other version are already
826
# present then we can avoid some work by just taking the delta
827
# and adjusting the offsets.
828
new_parents = self._imported_parents(other, other_idx)
829
sha1 = other._sha1s[other_idx]
830
831
processed += 1
832
833
if name in self._names:
834
idx = self.lookup(name)
835
n1 = map(other.idx_to_name, other._parents[other_idx] )
836
n2 = map(self.idx_to_name, self._parents[other_idx] )
837
if sha1 == self._sha1s[idx] and n1 == n2:
838
continue
839
840
merged += 1
841
lines = other.get_lines(other_idx)
842
self.add(name, new_parents, lines, sha1)
843
844
mutter("merged = %d, processed = %d, file_id=%s; deltat=%d"%(
845
merged,processed,self._weave_name, time.time( )-time0))
846
847
848
849
850
def _imported_parents(self, other, other_idx):
851
"""Return list of parents in self corresponding to indexes in other."""
852
new_parents = []
853
for parent_idx in other._parents[other_idx]:
854
parent_name = other._names[parent_idx]
855
if parent_name not in self._names:
856
# should not be possible
857
raise WeaveError("missing parent {%s} of {%s} in %r"
858
% (parent_name, other._name_map[other_idx], self))
859
new_parents.append(self._name_map[parent_name])
860
return new_parents
861
862
def _check_version_consistent(self, other, other_idx, name):
863
"""Check if a version in consistent in this and other.
864
865
To be consistent it must have:
866
867
* the same text
868
* the same direct parents (by name, not index, and disregarding
869
order)
870
871
If present & correct return True;
872
if not present in self return False;
873
if inconsistent raise error."""
874
this_idx = self._name_map.get(name, -1)
875
if this_idx != -1:
876
if self._sha1s[this_idx] != other._sha1s[other_idx]:
877
raise WeaveError("inconsistent texts for version {%s} "
878
"when joining weaves"
879
% (name))
880
self_parents = self._parents[this_idx]
881
other_parents = other._parents[other_idx]
882
n1 = [self._names[i] for i in self_parents]
883
n2 = [other._names[i] for i in other_parents]
884
n1.sort()
885
n2.sort()
886
if n1 != n2:
887
raise WeaveParentMismatch("inconsistent parents "
888
"for version {%s}: %s vs %s" % (name, n1, n2))
889
else:
890
return True # ok!
891
else:
892
return False
893
894
def reweave(self, other):
895
"""Reweave self with other."""
896
new_weave = reweave(self, other)
897
for attr in self.__slots__:
898
setattr(self, attr, getattr(new_weave, attr))
899
900
901
def reweave(wa, wb):
902
"""Combine two weaves and return the result.
903
904
This works even if a revision R has different parents in
905
wa and wb. In the resulting weave all the parents are given.
906
907
This is done by just building up a new weave, maintaining ordering
908
of the versions in the two inputs. More efficient approaches
909
might be possible but it should only be necessary to do
910
this operation rarely, when a new previously ghost version is
911
inserted.
912
"""
913
wr = Weave()
914
ia = ib = 0
915
queue_a = range(wa.numversions())
916
queue_b = range(wb.numversions())
917
# first determine combined parents of all versions
918
# map from version name -> all parent names
919
combined_parents = _reweave_parent_graphs(wa, wb)
920
mutter("combined parents: %r", combined_parents)
921
order = topo_sort(combined_parents.iteritems())
922
mutter("order to reweave: %r", order)
923
for name in order:
924
if name in wa._name_map:
925
lines = wa.get_lines(name)
926
if name in wb._name_map:
927
assert lines == wb.get_lines(name)
928
else:
929
lines = wb.get_lines(name)
930
wr.add(name, combined_parents[name], lines)
931
return wr
932
933
934
def _reweave_parent_graphs(wa, wb):
935
"""Return combined parent ancestry for two weaves.
936
937
Returned as a list of (version_name, set(parent_names))"""
938
combined = {}
939
for weave in [wa, wb]:
940
for idx, name in enumerate(weave._names):
941
p = combined.setdefault(name, set())
942
p.update(map(weave.idx_to_name, weave._parents[idx]))
943
return combined
944
945
946
def weave_toc(w):
947
"""Show the weave's table-of-contents"""
948
print '%6s %50s %10s %10s' % ('ver', 'name', 'sha1', 'parents')
949
for i in (6, 50, 10, 10):
950
print '-' * i,
951
print
952
for i in range(w.numversions()):
953
sha1 = w._sha1s[i]
954
name = w._names[i]
955
parent_str = ' '.join(map(str, w._parents[i]))
956
print '%6d %-50.50s %10.10s %s' % (i, name, sha1, parent_str)
957
958
959
960
def weave_stats(weave_file, pb):
961
from bzrlib.weavefile import read_weave
962
963
wf = file(weave_file, 'rb')
443
# basis a list of (origin, lineno, line)
444
basis_lineno = []
445
basis_lines = []
446
for origin, lineno, line in self._extract(included):
447
basis_lineno.append(lineno)
448
basis_lines.append(line)
449
450
# add a sentinal, because we can also match against the final line
451
basis_lineno.append(len(self._l))
452
453
# XXX: which line of the weave should we really consider
454
# matches the end of the file? the current code says it's the
455
# last line of the weave?
456
457
from difflib import SequenceMatcher
458
s = SequenceMatcher(None, basis_lines, lines)
459
460
# TODO: Perhaps return line numbers from composed weave as well?
461
462
for tag, i1, i2, j1, j2 in s.get_opcodes():
463
##print tag, i1, i2, j1, j2
464
465
if tag == 'equal':
466
continue
467
468
# i1,i2 are given in offsets within basis_lines; we need to map them
469
# back to offsets within the entire weave
470
real_i1 = basis_lineno[i1]
471
real_i2 = basis_lineno[i2]
472
473
assert 0 <= j1
474
assert j1 <= j2
475
assert j2 <= len(lines)
476
477
yield real_i1, real_i2, lines[j1:j2]
478
479
480
481
def weave_info(filename, out):
482
"""Show some text information about the weave."""
483
from weavefile import read_weave
484
wf = file(filename, 'rb')
964
485
w = read_weave(wf)
965
486
# FIXME: doesn't work on pipes
966
487
weave_size = wf.tell()
488
print >>out, "weave file size %d bytes" % weave_size
489
print >>out, "weave contains %d versions" % len(w._v)
967
490
968
491
total = 0
969
vers = len(w)
970
for i in range(vers):
971
pb.update('checking sizes', i, vers)
972
for origin, lineno, line in w._extract([i]):
973
total += len(line)
974
975
pb.clear()
976
977
print 'versions %9d' % vers
978
print 'weave file %9d bytes' % weave_size
979
print 'total contents %9d bytes' % total
980
print 'compression ratio %9.2fx' % (float(total) / float(weave_size))
981
if vers:
982
avg = total/vers
983
print 'average size %9d bytes' % avg
984
print 'relative size %9.2fx' % (float(weave_size) / float(avg))
492
print '%6s %6s %8s %40s %20s' % ('ver', 'lines', 'bytes', 'sha1', 'parents')
493
for i in (6, 6, 8, 40, 20):
494
print '-' * i,
495
print
496
for i in range(len(w._v)):
497
text = w.get(i)
498
lines = len(text)
499
bytes = sum((len(a) for a in text))
500
sha1 = w._sha1s[i]
501
print '%6d %6d %8d %40s' % (i, lines, bytes, sha1),
502
print ', '.join(map(str, w._v[i]))
503
total += bytes
504
505
print >>out, "versions total %d bytes" % total
506
print >>out, "compression ratio %.3f" % (float(total)/float(weave_size))
985
507
986
508
987
509
def usage():
996
518
Write out specified version.
997
519
weave check WEAVEFILE
998
520
Check consistency of all versions.
999
weave toc WEAVEFILE
521
weave info WEAVEFILE
1000
522
Display table of contents.
1001
weave add WEAVEFILE NAME [BASE...] < NEWTEXT
523
weave add WEAVEFILE [BASE...] < NEWTEXT
1002
524
Add NEWTEXT, with specified parent versions.
1003
525
weave annotate WEAVEFILE VERSION
1004
526
Display origin of each line.
1006
528
Display composite of all selected versions.
1007
529
weave merge WEAVEFILE VERSION1 VERSION2 > OUT
1008
530
Auto-merge two versions and display conflicts.
1009
weave diff WEAVEFILE VERSION1 VERSION2
1010
Show differences between two versions.
1011
531
1012
532
example:
1013
533
1014
534
% weave init foo.weave
1015
535
% vi foo.txt
1016
% weave add foo.weave ver0 < foo.txt
536
% weave add foo.weave < foo.txt
1017
537
added version 0
1018
538
1019
539
(create updated version)
1020
540
% vi foo.txt
1021
541
% weave get foo.weave 0 | diff -u - foo.txt
1022
% weave add foo.weave ver1 0 < foo.txt
542
% weave add foo.weave 0 < foo.txt
1023
543
added version 1
1024
544
1025
545
% weave get foo.weave 0 > foo.txt (create forked version)
1026
546
% vi foo.txt
1027
% weave add foo.weave ver2 0 < foo.txt
547
% weave add foo.weave 0 < foo.txt
1028
548
added version 2
1029
549
1030
550
% weave merge foo.weave 1 2 > foo.txt (merge them)
1031
551
% vi foo.txt (resolve conflicts)
1032
% weave add foo.weave merged 1 2 < foo.txt (commit merged version)
552
% weave add foo.weave 1 2 < foo.txt (commit merged version)
1033
553
1034
554
"""
1035
555
1038
558
def main(argv):
1039
559
import sys
1040
560
import os
1041
try:
1042
import bzrlib
1043
except ImportError:
1044
# in case we're run directly from the subdirectory
1045
sys.path.append('..')
1046
import bzrlib
1047
from bzrlib.weavefile import write_weave, read_weave
1048
from bzrlib.progress import ProgressBar
1049
1050
try:
1051
import psyco
1052
psyco.full()
1053
except ImportError:
1054
pass
1055
1056
if len(argv) < 2:
1057
usage()
1058
return 0
1059
561
from weavefile import write_weave, read_weave
1060
562
cmd = argv[1]
1061
563
1062
564
def readit():
1067
569
elif cmd == 'add':
1068
570
w = readit()
1069
571
# at the moment, based on everything in the file
1070
name = argv[3]
1071
parents = map(int, argv[4:])
572
parents = map(int, argv[3:])
1072
573
lines = sys.stdin.readlines()
1073
ver = w.add(name, parents, lines)
574
ver = w.add(parents, lines)
1074
575
write_weave(w, file(argv[2], 'wb'))
1075
print 'added version %r %d' % (name, ver)
576
print 'added version %d' % ver
1076
577
elif cmd == 'init':
1077
578
fn = argv[2]
1078
579
if os.path.exists(fn):
1087
588
w = readit()
1088
589
sys.stdout.writelines(w.mash_iter(map(int, argv[3:])))
1089
590
1090
elif cmd == 'diff':
1091
from difflib import unified_diff
1092
w = readit()
1093
fn = argv[2]
1094
v1, v2 = map(int, argv[3:5])
1095
lines1 = w.get(v1)
1096
lines2 = w.get(v2)
1097
diff_gen = unified_diff(lines1, lines2,
1098
'%s version %d' % (fn, v1),
1099
'%s version %d' % (fn, v2))
1100
sys.stdout.writelines(diff_gen)
1101
1102
591
elif cmd == 'annotate':
1103
592
w = readit()
1104
593
# newline is added to all lines regardless; too hard to get
1112
601
print '%5d | %s' % (origin, text)
1113
602
lasto = origin
1114
603
1115
elif cmd == 'toc':
1116
weave_toc(readit())
1117
1118
elif cmd == 'stats':
1119
weave_stats(argv[2], ProgressBar())
604
elif cmd == 'info':
605
weave_info(argv[2], sys.stdout)
1120
606
1121
607
elif cmd == 'check':
1122
608
w = readit()
1123
pb = ProgressBar()
1124
w.check(pb)
1125
pb.clear()
1126
print '%d versions ok' % w.numversions()
1127
1128
elif cmd == 'inclusions':
1129
w = readit()
1130
print ' '.join(map(str, w.inclusions([int(argv[3])])))
1131
1132
elif cmd == 'parents':
1133
w = readit()
1134
print ' '.join(map(str, w._parents[int(argv[3])]))
1135
1136
elif cmd == 'plan-merge':
1137
w = readit()
1138
for state, line in w.plan_merge(int(argv[3]), int(argv[4])):
1139
if line:
1140
print '%14s | %s' % (state, line),
609
w.check()
1141
610
1142
611
elif cmd == 'merge':
1143
w = readit()
1144
p = w.plan_merge(int(argv[3]), int(argv[4]))
1145
sys.stdout.writelines(w.weave_merge(p))
1146
1147
elif cmd == 'mash-merge':
1148
612
if len(argv) != 5:
1149
613
usage()
1150
614
return 1
1168
632
raise ValueError('unknown command %r' % cmd)
1169
633
1170
634
1171
1172
def profile_main(argv):
1173
import tempfile, hotshot, hotshot.stats
1174
1175
prof_f = tempfile.NamedTemporaryFile()
1176
1177
prof = hotshot.Profile(prof_f.name)
1178
1179
ret = prof.runcall(main, argv)
1180
prof.close()
1181
1182
stats = hotshot.stats.load(prof_f.name)
1183
#stats.strip_dirs()
1184
stats.sort_stats('cumulative')
1185
## XXX: Might like to write to stderr or the trace file instead but
1186
## print_stats seems hardcoded to stdout
1187
stats.print_stats(20)
1188
1189
return ret
1190
1191
1192
def lsprofile_main(argv):
1193
from bzrlib.lsprof import profile
1194
ret,stats = profile(main, argv)
1195
stats.sort()
1196
stats.pprint()
1197
return ret
1198
1199
1200
635
if __name__ == '__main__':
1201
636
import sys
1202
if '--profile' in sys.argv:
1203
args = sys.argv[:]
1204
args.remove('--profile')
1205
sys.exit(profile_main(args))
1206
elif '--lsprof' in sys.argv:
1207
args = sys.argv[:]
1208
args.remove('--lsprof')
1209
sys.exit(lsprofile_main(args))
1210
else:
1211
sys.exit(main(sys.argv))
1212
637
sys.exit(main(sys.argv))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1