1
# Copyright (C) 2005 by Canonical Ltd
4
# Johan Rydberg <jrydberg@gnu.org>
6
# This program is free software; you can redistribute it and/or modify
7
# it under the terms of the GNU General Public License as published by
8
# the Free Software Foundation; either version 2 of the License, or
9
# (at your option) any later version.
11
# This program is distributed in the hope that it will be useful,
12
# but WITHOUT ANY WARRANTY; without even the implied warranty of
13
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14
# GNU General Public License for more details.
16
# You should have received a copy of the GNU General Public License
17
# along with this program; if not, write to the Free Software
18
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20
# Remaing to do is to figure out if get_graph should return a simple
21
# map, or a graph object of some kind.
24
"""Versioned text file storage api."""
27
from copy import deepcopy
28
from unittest import TestSuite
31
import bzrlib.errors as errors
32
from bzrlib.inter import InterObject
33
from bzrlib.symbol_versioning import *
34
from bzrlib.textmerge import TextMerge
35
from bzrlib.transport.memory import MemoryTransport
36
from bzrlib.tsort import topo_sort
40
class VersionedFile(object):
41
"""Versioned text file storage.
43
A versioned file manages versions of line-based text files,
44
keeping track of the originating version for each line.
46
To clients the "lines" of the file are represented as a list of
47
strings. These strings will typically have terminal newline
48
characters, but this is not required. In particular files commonly
49
do not have a newline at the end of the file.
51
Texts are identified by a version-id string.
54
def __init__(self, access_mode):
56
self._access_mode = access_mode
58
def copy_to(self, name, transport):
59
"""Copy this versioned file to name on transport."""
60
raise NotImplementedError(self.copy_to)
62
@deprecated_method(zero_eight)
64
"""Return a list of all the versions in this versioned file.
66
Please use versionedfile.versions() now.
68
return self.versions()
71
"""Return a unsorted list of versions."""
72
raise NotImplementedError(self.versions)
74
def has_ghost(self, version_id):
75
"""Returns whether version is present as a ghost."""
76
raise NotImplementedError(self.has_ghost)
78
def has_version(self, version_id):
79
"""Returns whether version is present."""
80
raise NotImplementedError(self.has_version)
82
def add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
83
"""Add a text to the versioned file via a pregenerated delta.
85
:param version_id: The version id being added.
86
:param parents: The parents of the version_id.
87
:param delta_parent: The parent this delta was created against.
88
:param sha1: The sha1 of the full text.
89
:param delta: The delta instructions. See get_delta for details.
91
self._check_write_ok()
92
if self.has_version(version_id):
93
raise errors.RevisionAlreadyPresent(version_id, self)
94
return self._add_delta(version_id, parents, delta_parent, sha1, noeol, delta)
96
def _add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
97
"""Class specific routine to add a delta.
99
This generic version simply applies the delta to the delta_parent and
102
# strip annotation from delta
104
for start, stop, delta_len, delta_lines in delta:
105
new_delta.append((start, stop, delta_len, [text for origin, text in delta_lines]))
106
if delta_parent is not None:
107
parent_full = self.get_lines(delta_parent)
110
new_full = self._apply_delta(parent_full, new_delta)
111
# its impossible to have noeol on an empty file
112
if noeol and new_full[-1][-1] == '\n':
113
new_full[-1] = new_full[-1][:-1]
114
self.add_lines(version_id, parents, new_full)
116
def add_lines(self, version_id, parents, lines, parent_texts=None):
117
"""Add a single text on top of the versioned file.
119
Must raise RevisionAlreadyPresent if the new version is
120
already present in file history.
122
Must raise RevisionNotPresent if any of the given parents are
123
not present in file history.
124
:param parent_texts: An optional dictionary containing the opaque
125
representations of some or all of the parents of
126
version_id to allow delta optimisations.
127
VERY IMPORTANT: the texts must be those returned
128
by add_lines or data corruption can be caused.
129
:return: An opaque representation of the inserted version which can be
130
provided back to future add_lines calls in the parent_texts
133
self._check_write_ok()
134
return self._add_lines(version_id, parents, lines, parent_texts)
136
def _add_lines(self, version_id, parents, lines, parent_texts):
137
"""Helper to do the class specific add_lines."""
138
raise NotImplementedError(self.add_lines)
140
def add_lines_with_ghosts(self, version_id, parents, lines,
142
"""Add lines to the versioned file, allowing ghosts to be present.
144
This takes the same parameters as add_lines.
146
self._check_write_ok()
147
return self._add_lines_with_ghosts(version_id, parents, lines,
150
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
151
"""Helper to do class specific add_lines_with_ghosts."""
152
raise NotImplementedError(self.add_lines_with_ghosts)
154
def check(self, progress_bar=None):
155
"""Check the versioned file for integrity."""
156
raise NotImplementedError(self.check)
158
def _check_write_ok(self):
159
"""Is the versioned file marked as 'finished' ? Raise if it is."""
161
raise errors.OutSideTransaction()
162
if self._access_mode != 'w':
163
raise errors.ReadOnlyObjectDirtiedError(self)
165
def clear_cache(self):
166
"""Remove any data cached in the versioned file object."""
168
def clone_text(self, new_version_id, old_version_id, parents):
169
"""Add an identical text to old_version_id as new_version_id.
171
Must raise RevisionNotPresent if the old version or any of the
172
parents are not present in file history.
174
Must raise RevisionAlreadyPresent if the new version is
175
already present in file history."""
176
self._check_write_ok()
177
return self._clone_text(new_version_id, old_version_id, parents)
179
def _clone_text(self, new_version_id, old_version_id, parents):
180
"""Helper function to do the _clone_text work."""
181
raise NotImplementedError(self.clone_text)
183
def create_empty(self, name, transport, mode=None):
184
"""Create a new versioned file of this exact type.
186
:param name: the file name
187
:param transport: the transport
188
:param mode: optional file mode.
190
raise NotImplementedError(self.create_empty)
192
def fix_parents(self, version, new_parents):
193
"""Fix the parents list for version.
195
This is done by appending a new version to the index
196
with identical data except for the parents list.
197
the parents list must be a superset of the current
200
self._check_write_ok()
201
return self._fix_parents(version, new_parents)
203
def _fix_parents(self, version, new_parents):
204
"""Helper for fix_parents."""
205
raise NotImplementedError(self.fix_parents)
207
def get_delta(self, version):
208
"""Get a delta for constructing version from some other version.
210
:return: (delta_parent, sha1, noeol, delta)
211
Where delta_parent is a version id or None to indicate no parent.
213
raise NotImplementedError(self.get_delta)
215
def get_deltas(self, versions):
216
"""Get multiple deltas at once for constructing versions.
218
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
219
Where delta_parent is a version id or None to indicate no parent, and
220
version_id is the version_id created by that delta.
223
for version in versions:
224
result[version] = self.get_delta(version)
227
def get_suffixes(self):
228
"""Return the file suffixes associated with this versioned file."""
229
raise NotImplementedError(self.get_suffixes)
231
def get_text(self, version_id):
232
"""Return version contents as a text string.
234
Raises RevisionNotPresent if version is not present in
237
return ''.join(self.get_lines(version_id))
238
get_string = get_text
240
def get_lines(self, version_id):
241
"""Return version contents as a sequence of lines.
243
Raises RevisionNotPresent if version is not present in
246
raise NotImplementedError(self.get_lines)
248
def get_ancestry(self, version_ids):
249
"""Return a list of all ancestors of given version(s). This
250
will not include the null revision.
252
Must raise RevisionNotPresent if any of the given versions are
253
not present in file history."""
254
if isinstance(version_ids, basestring):
255
version_ids = [version_ids]
256
raise NotImplementedError(self.get_ancestry)
258
def get_ancestry_with_ghosts(self, version_ids):
259
"""Return a list of all ancestors of given version(s). This
260
will not include the null revision.
262
Must raise RevisionNotPresent if any of the given versions are
263
not present in file history.
265
Ghosts that are known about will be included in ancestry list,
266
but are not explicitly marked.
268
raise NotImplementedError(self.get_ancestry_with_ghosts)
271
"""Return a graph for the entire versioned file.
273
Ghosts are not listed or referenced in the graph.
276
for version in self.versions():
277
result[version] = self.get_parents(version)
280
def get_graph_with_ghosts(self):
281
"""Return a graph for the entire versioned file.
283
Ghosts are referenced in parents list but are not
286
raise NotImplementedError(self.get_graph_with_ghosts)
288
@deprecated_method(zero_eight)
289
def parent_names(self, version):
290
"""Return version names for parents of a version.
292
See get_parents for the current api.
294
return self.get_parents(version)
296
def get_parents(self, version_id):
297
"""Return version names for parents of a version.
299
Must raise RevisionNotPresent if version is not present in
302
raise NotImplementedError(self.get_parents)
304
def get_parents_with_ghosts(self, version_id):
305
"""Return version names for parents of version_id.
307
Will raise RevisionNotPresent if version_id is not present
310
Ghosts that are known about will be included in the parent list,
311
but are not explicitly marked.
313
raise NotImplementedError(self.get_parents_with_ghosts)
315
def annotate_iter(self, version_id):
316
"""Yield list of (version-id, line) pairs for the specified
319
Must raise RevisionNotPresent if any of the given versions are
320
not present in file history.
322
raise NotImplementedError(self.annotate_iter)
324
def annotate(self, version_id):
325
return list(self.annotate_iter(version_id))
327
def _apply_delta(self, lines, delta):
328
"""Apply delta to lines."""
331
for start, end, count, delta_lines in delta:
332
lines[offset+start:offset+end] = delta_lines
333
offset = offset + (start - end) + count
336
def join(self, other, pb=None, msg=None, version_ids=None,
337
ignore_missing=False):
338
"""Integrate versions from other into this versioned file.
340
If version_ids is None all versions from other should be
341
incorporated into this versioned file.
343
Must raise RevisionNotPresent if any of the specified versions
344
are not present in the other files history unless ignore_missing
345
is supplied when they are silently skipped.
347
self._check_write_ok()
348
return InterVersionedFile.get(other, self).join(
354
def iter_lines_added_or_present_in_versions(self, version_ids=None):
355
"""Iterate over the lines in the versioned file from version_ids.
357
This may return lines from other versions, and does not return the
358
specific version marker at this point. The api may be changed
359
during development to include the version that the versioned file
360
thinks is relevant, but given that such hints are just guesses,
361
its better not to have it if we dont need it.
363
NOTES: Lines are normalised: they will all have \n terminators.
364
Lines are returned in arbitrary order.
366
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
368
def transaction_finished(self):
369
"""The transaction that this file was opened in has finished.
371
This records self.finished = True and should cause all mutating
376
@deprecated_method(zero_eight)
377
def walk(self, version_ids=None):
378
"""Walk the versioned file as a weave-like structure, for
379
versions relative to version_ids. Yields sequence of (lineno,
380
insert, deletes, text) for each relevant line.
382
Must raise RevisionNotPresent if any of the specified versions
383
are not present in the file history.
385
:param version_ids: the version_ids to walk with respect to. If not
386
supplied the entire weave-like structure is walked.
388
walk is deprecated in favour of iter_lines_added_or_present_in_versions
390
raise NotImplementedError(self.walk)
392
@deprecated_method(zero_eight)
393
def iter_names(self):
394
"""Walk the names list."""
395
return iter(self.versions())
397
def plan_merge(self, ver_a, ver_b):
398
"""Return pseudo-annotation indicating how the two versions merge.
400
This is computed between versions a and b and their common
403
Weave lines present in none of them are skipped entirely.
405
inc_a = set(self.get_ancestry([ver_a]))
406
inc_b = set(self.get_ancestry([ver_b]))
407
inc_c = inc_a & inc_b
409
for lineno, insert, deleteset, line in self.walk([ver_a, ver_b]):
410
if deleteset & inc_c:
411
# killed in parent; can't be in either a or b
412
# not relevant to our work
413
yield 'killed-base', line
414
elif insert in inc_c:
415
# was inserted in base
416
killed_a = bool(deleteset & inc_a)
417
killed_b = bool(deleteset & inc_b)
418
if killed_a and killed_b:
419
yield 'killed-both', line
421
yield 'killed-a', line
423
yield 'killed-b', line
425
yield 'unchanged', line
426
elif insert in inc_a:
427
if deleteset & inc_a:
428
yield 'ghost-a', line
432
elif insert in inc_b:
433
if deleteset & inc_b:
434
yield 'ghost-b', line
438
# not in either revision
439
yield 'irrelevant', line
441
yield 'unchanged', '' # terminator
443
def weave_merge(self, plan, a_marker='<<<<<<< \n', b_marker='>>>>>>> \n'):
444
return WeaveMerge(plan, a_marker, b_marker).merge_lines()
446
class WeaveMerge(TextMerge):
447
def __init__(self, plan, a_marker='<<<<<<< \n', b_marker='>>>>>>> \n'):
448
TextMerge.__init__(self, a_marker, b_marker)
451
def _merge_struct(self):
455
# TODO: Show some version information (e.g. author, date) on
456
# conflicted regions.
458
# We previously considered either 'unchanged' or 'killed-both' lines
459
# to be possible places to resynchronize. However, assuming agreement
460
# on killed-both lines may be too agressive. -- mbp 20060324
461
for state, line in self.plan:
462
if state == 'unchanged':
463
# resync and flush queued conflicts changes if any
464
if not lines_a and not lines_b:
466
elif ch_a and not ch_b:
469
elif ch_b and not ch_a:
471
elif lines_a == lines_b:
474
yield (lines_a, lines_b)
480
if state == 'unchanged':
483
elif state == 'killed-a':
486
elif state == 'killed-b':
489
elif state == 'new-a':
492
elif state == 'new-b':
496
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
497
'killed-base', 'killed-both'), state
500
class InterVersionedFile(InterObject):
501
"""This class represents operations taking place between two versionedfiles..
503
Its instances have methods like join, and contain
504
references to the source and target versionedfiles these operations can be
507
Often we will provide convenience methods on 'versionedfile' which carry out
508
operations with another versionedfile - they will always forward to
509
InterVersionedFile.get(other).method_name(parameters).
513
"""The available optimised InterVersionedFile types."""
515
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
516
"""Integrate versions from self.source into self.target.
518
If version_ids is None all versions from source should be
519
incorporated into this versioned file.
521
Must raise RevisionNotPresent if any of the specified versions
522
are not present in the other files history unless ignore_missing is
523
supplied when they are silently skipped.
526
# - if the target is empty, just add all the versions from
527
# source to target, otherwise:
528
# - make a temporary versioned file of type target
529
# - insert the source content into it one at a time
531
if not self.target.versions():
534
# Make a new target-format versioned file.
535
temp_source = self.target.create_empty("temp", MemoryTransport())
537
graph = self.source.get_graph()
538
order = topo_sort(graph.items())
539
pb = ui.ui_factory.nested_progress_bar()
542
# TODO for incremental cross-format work:
543
# make a versioned file with the following content:
544
# all revisions we have been asked to join
545
# all their ancestors that are *not* in target already.
546
# the immediate parents of the above two sets, with
547
# empty parent lists - these versions are in target already
548
# and the incorrect version data will be ignored.
549
# TODO: for all ancestors that are present in target already,
550
# check them for consistent data, this requires moving sha1 from
552
# TODO: remove parent texts when they are not relevant any more for
553
# memory pressure reduction. RBC 20060313
554
# pb.update('Converting versioned data', 0, len(order))
555
# deltas = self.source.get_deltas(order)
556
for index, version in enumerate(order):
557
pb.update('Converting versioned data', index, len(order))
558
parent_text = target.add_lines(version,
559
self.source.get_parents(version),
560
self.source.get_lines(version),
561
parent_texts=parent_texts)
562
parent_texts[version] = parent_text
563
#delta_parent, sha1, noeol, delta = deltas[version]
564
#target.add_delta(version,
565
# self.source.get_parents(version),
570
#target.get_lines(version)
572
# this should hit the native code path for target
573
if target is not self.target:
574
return self.target.join(temp_source,
583
class InterVersionedFileTestProviderAdapter(object):
584
"""A tool to generate a suite testing multiple inter versioned-file classes.
586
This is done by copying the test once for each interversionedfile provider
587
and injecting the transport_server, transport_readonly_server,
588
versionedfile_factory and versionedfile_factory_to classes into each copy.
589
Each copy is also given a new id() to make it easy to identify.
592
def __init__(self, transport_server, transport_readonly_server, formats):
593
self._transport_server = transport_server
594
self._transport_readonly_server = transport_readonly_server
595
self._formats = formats
597
def adapt(self, test):
599
for (interversionedfile_class,
600
versionedfile_factory,
601
versionedfile_factory_to) in self._formats:
602
new_test = deepcopy(test)
603
new_test.transport_server = self._transport_server
604
new_test.transport_readonly_server = self._transport_readonly_server
605
new_test.interversionedfile_class = interversionedfile_class
606
new_test.versionedfile_factory = versionedfile_factory
607
new_test.versionedfile_factory_to = versionedfile_factory_to
608
def make_new_test_id():
609
new_id = "%s(%s)" % (new_test.id(), interversionedfile_class.__name__)
610
return lambda: new_id
611
new_test.id = make_new_test_id()
612
result.addTest(new_test)
616
def default_test_list():
617
"""Generate the default list of interversionedfile permutations to test."""
618
from bzrlib.weave import WeaveFile
619
from bzrlib.knit import KnitVersionedFile
621
# test the fallback InterVersionedFile from weave to annotated knits
622
result.append((InterVersionedFile,
625
for optimiser in InterVersionedFile._optimisers:
626
result.append((optimiser,
627
optimiser._matching_file_factory,
628
optimiser._matching_file_factory
630
# if there are specific combinations we want to use, we can add them
added
removed
bzrlib/versionedfile.py
