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
"""Versioned text file storage api."""
23
from copy import deepcopy
24
from unittest import TestSuite
27
import bzrlib.errors as errors
28
from bzrlib.inter import InterObject
29
from bzrlib.symbol_versioning import *
30
from bzrlib.textmerge import TextMerge
31
from bzrlib.transport.memory import MemoryTransport
32
from bzrlib.tsort import topo_sort
36
class VersionedFile(object):
37
"""Versioned text file storage.
39
A versioned file manages versions of line-based text files,
40
keeping track of the originating version for each line.
42
To clients the "lines" of the file are represented as a list of
43
strings. These strings will typically have terminal newline
44
characters, but this is not required. In particular files commonly
45
do not have a newline at the end of the file.
47
Texts are identified by a version-id string.
50
def __init__(self, access_mode):
52
self._access_mode = access_mode
54
def copy_to(self, name, transport):
55
"""Copy this versioned file to name on transport."""
56
raise NotImplementedError(self.copy_to)
58
@deprecated_method(zero_eight)
60
"""Return a list of all the versions in this versioned file.
62
Please use versionedfile.versions() now.
64
return self.versions()
67
"""Return a unsorted list of versions."""
68
raise NotImplementedError(self.versions)
70
def has_ghost(self, version_id):
71
"""Returns whether version is present as a ghost."""
72
raise NotImplementedError(self.has_ghost)
74
def has_version(self, version_id):
75
"""Returns whether version is present."""
76
raise NotImplementedError(self.has_version)
78
def add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
79
"""Add a text to the versioned file via a pregenerated delta.
81
:param version_id: The version id being added.
82
:param parents: The parents of the version_id.
83
:param delta_parent: The parent this delta was created against.
84
:param sha1: The sha1 of the full text.
85
:param delta: The delta instructions. See get_delta for details.
87
self._check_write_ok()
88
if self.has_version(version_id):
89
raise errors.RevisionAlreadyPresent(version_id, self)
90
return self._add_delta(version_id, parents, delta_parent, sha1, noeol, delta)
92
def _add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
93
"""Class specific routine to add a delta.
95
This generic version simply applies the delta to the delta_parent and
98
# strip annotation from delta
100
for start, stop, delta_len, delta_lines in delta:
101
new_delta.append((start, stop, delta_len, [text for origin, text in delta_lines]))
102
if delta_parent is not None:
103
parent_full = self.get_lines(delta_parent)
106
new_full = self._apply_delta(parent_full, new_delta)
107
# its impossible to have noeol on an empty file
108
if noeol and new_full[-1][-1] == '\n':
109
new_full[-1] = new_full[-1][:-1]
110
self.add_lines(version_id, parents, new_full)
112
def add_lines(self, version_id, parents, lines, parent_texts=None):
113
"""Add a single text on top of the versioned file.
115
Must raise RevisionAlreadyPresent if the new version is
116
already present in file history.
118
Must raise RevisionNotPresent if any of the given parents are
119
not present in file history.
120
:param parent_texts: An optional dictionary containing the opaque
121
representations of some or all of the parents of
122
version_id to allow delta optimisations.
123
VERY IMPORTANT: the texts must be those returned
124
by add_lines or data corruption can be caused.
125
:return: An opaque representation of the inserted version which can be
126
provided back to future add_lines calls in the parent_texts
129
self._check_write_ok()
130
return self._add_lines(version_id, parents, lines, parent_texts)
132
def _add_lines(self, version_id, parents, lines, parent_texts):
133
"""Helper to do the class specific add_lines."""
134
raise NotImplementedError(self.add_lines)
136
def add_lines_with_ghosts(self, version_id, parents, lines,
138
"""Add lines to the versioned file, allowing ghosts to be present.
140
This takes the same parameters as add_lines.
142
self._check_write_ok()
143
return self._add_lines_with_ghosts(version_id, parents, lines,
146
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
147
"""Helper to do class specific add_lines_with_ghosts."""
148
raise NotImplementedError(self.add_lines_with_ghosts)
150
def check(self, progress_bar=None):
151
"""Check the versioned file for integrity."""
152
raise NotImplementedError(self.check)
154
def _check_lines_not_unicode(self, lines):
155
"""Check that lines being added to a versioned file are not unicode."""
157
if line.__class__ is not str:
158
raise errors.BzrBadParameterUnicode("lines")
160
def _check_lines_are_lines(self, lines):
161
"""Check that the lines really are full lines without inline EOL."""
163
if '\n' in line[:-1]:
164
raise errors.BzrBadParameterContainsNewline("lines")
166
def _check_write_ok(self):
167
"""Is the versioned file marked as 'finished' ? Raise if it is."""
169
raise errors.OutSideTransaction()
170
if self._access_mode != 'w':
171
raise errors.ReadOnlyObjectDirtiedError(self)
173
def clear_cache(self):
174
"""Remove any data cached in the versioned file object."""
176
def clone_text(self, new_version_id, old_version_id, parents):
177
"""Add an identical text to old_version_id as new_version_id.
179
Must raise RevisionNotPresent if the old version or any of the
180
parents are not present in file history.
182
Must raise RevisionAlreadyPresent if the new version is
183
already present in file history."""
184
self._check_write_ok()
185
return self._clone_text(new_version_id, old_version_id, parents)
187
def _clone_text(self, new_version_id, old_version_id, parents):
188
"""Helper function to do the _clone_text work."""
189
raise NotImplementedError(self.clone_text)
191
def create_empty(self, name, transport, mode=None):
192
"""Create a new versioned file of this exact type.
194
:param name: the file name
195
:param transport: the transport
196
:param mode: optional file mode.
198
raise NotImplementedError(self.create_empty)
200
def fix_parents(self, version, new_parents):
201
"""Fix the parents list for version.
203
This is done by appending a new version to the index
204
with identical data except for the parents list.
205
the parents list must be a superset of the current
208
self._check_write_ok()
209
return self._fix_parents(version, new_parents)
211
def _fix_parents(self, version, new_parents):
212
"""Helper for fix_parents."""
213
raise NotImplementedError(self.fix_parents)
215
def get_delta(self, version):
216
"""Get a delta for constructing version from some other version.
218
:return: (delta_parent, sha1, noeol, delta)
219
Where delta_parent is a version id or None to indicate no parent.
221
raise NotImplementedError(self.get_delta)
223
def get_deltas(self, versions):
224
"""Get multiple deltas at once for constructing versions.
226
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
227
Where delta_parent is a version id or None to indicate no parent, and
228
version_id is the version_id created by that delta.
231
for version in versions:
232
result[version] = self.get_delta(version)
235
def get_sha1(self, version_id):
236
"""Get the stored sha1 sum for the given revision.
238
:param name: The name of the version to lookup
240
raise NotImplementedError(self.get_sha1)
242
def get_suffixes(self):
243
"""Return the file suffixes associated with this versioned file."""
244
raise NotImplementedError(self.get_suffixes)
246
def get_text(self, version_id):
247
"""Return version contents as a text string.
249
Raises RevisionNotPresent if version is not present in
252
return ''.join(self.get_lines(version_id))
253
get_string = get_text
255
def get_texts(self, version_ids):
256
"""Return the texts of listed versions as a list of strings.
258
Raises RevisionNotPresent if version is not present in
261
return [''.join(self.get_lines(v)) for v in version_ids]
263
def get_lines(self, version_id):
264
"""Return version contents as a sequence of lines.
266
Raises RevisionNotPresent if version is not present in
269
raise NotImplementedError(self.get_lines)
271
def get_ancestry(self, version_ids):
272
"""Return a list of all ancestors of given version(s). This
273
will not include the null revision.
275
Must raise RevisionNotPresent if any of the given versions are
276
not present in file history."""
277
if isinstance(version_ids, basestring):
278
version_ids = [version_ids]
279
raise NotImplementedError(self.get_ancestry)
281
def get_ancestry_with_ghosts(self, version_ids):
282
"""Return a list of all ancestors of given version(s). This
283
will not include the null revision.
285
Must raise RevisionNotPresent if any of the given versions are
286
not present in file history.
288
Ghosts that are known about will be included in ancestry list,
289
but are not explicitly marked.
291
raise NotImplementedError(self.get_ancestry_with_ghosts)
293
def get_graph(self, version_ids=None):
294
"""Return a graph from the versioned file.
296
Ghosts are not listed or referenced in the graph.
297
:param version_ids: Versions to select.
298
None means retrieve all versions.
301
if version_ids is None:
302
for version in self.versions():
303
result[version] = self.get_parents(version)
305
pending = set(version_ids)
307
version = pending.pop()
308
if version in result:
310
parents = self.get_parents(version)
311
for parent in parents:
315
result[version] = parents
318
def get_graph_with_ghosts(self):
319
"""Return a graph for the entire versioned file.
321
Ghosts are referenced in parents list but are not
324
raise NotImplementedError(self.get_graph_with_ghosts)
326
@deprecated_method(zero_eight)
327
def parent_names(self, version):
328
"""Return version names for parents of a version.
330
See get_parents for the current api.
332
return self.get_parents(version)
334
def get_parents(self, version_id):
335
"""Return version names for parents of a version.
337
Must raise RevisionNotPresent if version is not present in
340
raise NotImplementedError(self.get_parents)
342
def get_parents_with_ghosts(self, version_id):
343
"""Return version names for parents of version_id.
345
Will raise RevisionNotPresent if version_id is not present
348
Ghosts that are known about will be included in the parent list,
349
but are not explicitly marked.
351
raise NotImplementedError(self.get_parents_with_ghosts)
353
def annotate_iter(self, version_id):
354
"""Yield list of (version-id, line) pairs for the specified
357
Must raise RevisionNotPresent if any of the given versions are
358
not present in file history.
360
raise NotImplementedError(self.annotate_iter)
362
def annotate(self, version_id):
363
return list(self.annotate_iter(version_id))
365
def _apply_delta(self, lines, delta):
366
"""Apply delta to lines."""
369
for start, end, count, delta_lines in delta:
370
lines[offset+start:offset+end] = delta_lines
371
offset = offset + (start - end) + count
374
def join(self, other, pb=None, msg=None, version_ids=None,
375
ignore_missing=False):
376
"""Integrate versions from other into this versioned file.
378
If version_ids is None all versions from other should be
379
incorporated into this versioned file.
381
Must raise RevisionNotPresent if any of the specified versions
382
are not present in the other files history unless ignore_missing
383
is supplied when they are silently skipped.
385
self._check_write_ok()
386
return InterVersionedFile.get(other, self).join(
392
def iter_lines_added_or_present_in_versions(self, version_ids=None):
393
"""Iterate over the lines in the versioned file from version_ids.
395
This may return lines from other versions, and does not return the
396
specific version marker at this point. The api may be changed
397
during development to include the version that the versioned file
398
thinks is relevant, but given that such hints are just guesses,
399
its better not to have it if we don't need it.
401
NOTES: Lines are normalised: they will all have \n terminators.
402
Lines are returned in arbitrary order.
404
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
406
def transaction_finished(self):
407
"""The transaction that this file was opened in has finished.
409
This records self.finished = True and should cause all mutating
414
@deprecated_method(zero_eight)
415
def walk(self, version_ids=None):
416
"""Walk the versioned file as a weave-like structure, for
417
versions relative to version_ids. Yields sequence of (lineno,
418
insert, deletes, text) for each relevant line.
420
Must raise RevisionNotPresent if any of the specified versions
421
are not present in the file history.
423
:param version_ids: the version_ids to walk with respect to. If not
424
supplied the entire weave-like structure is walked.
426
walk is deprecated in favour of iter_lines_added_or_present_in_versions
428
raise NotImplementedError(self.walk)
430
@deprecated_method(zero_eight)
431
def iter_names(self):
432
"""Walk the names list."""
433
return iter(self.versions())
435
def plan_merge(self, ver_a, ver_b):
436
"""Return pseudo-annotation indicating how the two versions merge.
438
This is computed between versions a and b and their common
441
Weave lines present in none of them are skipped entirely.
444
killed-base Dead in base revision
445
killed-both Killed in each revision
448
unchanged Alive in both a and b (possibly created in both)
451
ghost-a Killed in a, unborn in b
452
ghost-b Killed in b, unborn in a
453
irrelevant Not in either revision
455
raise NotImplementedError(VersionedFile.plan_merge)
457
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
458
b_marker=TextMerge.B_MARKER):
459
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
462
class PlanWeaveMerge(TextMerge):
463
"""Weave merge that takes a plan as its input.
465
This exists so that VersionedFile.plan_merge is implementable.
466
Most callers will want to use WeaveMerge instead.
469
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
470
b_marker=TextMerge.B_MARKER):
471
TextMerge.__init__(self, a_marker, b_marker)
474
def _merge_struct(self):
479
def outstanding_struct():
480
if not lines_a and not lines_b:
482
elif ch_a and not ch_b:
485
elif ch_b and not ch_a:
487
elif lines_a == lines_b:
490
yield (lines_a, lines_b)
492
# We previously considered either 'unchanged' or 'killed-both' lines
493
# to be possible places to resynchronize. However, assuming agreement
494
# on killed-both lines may be too aggressive. -- mbp 20060324
495
for state, line in self.plan:
496
if state == 'unchanged':
497
# resync and flush queued conflicts changes if any
498
for struct in outstanding_struct():
504
if state == 'unchanged':
507
elif state == 'killed-a':
510
elif state == 'killed-b':
513
elif state == 'new-a':
516
elif state == 'new-b':
520
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
521
'killed-base', 'killed-both'), state
522
for struct in outstanding_struct():
526
class WeaveMerge(PlanWeaveMerge):
527
"""Weave merge that takes a VersionedFile and two versions as its input"""
529
def __init__(self, versionedfile, ver_a, ver_b,
530
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
531
plan = versionedfile.plan_merge(ver_a, ver_b)
532
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
535
class InterVersionedFile(InterObject):
536
"""This class represents operations taking place between two versionedfiles..
538
Its instances have methods like join, and contain
539
references to the source and target versionedfiles these operations can be
542
Often we will provide convenience methods on 'versionedfile' which carry out
543
operations with another versionedfile - they will always forward to
544
InterVersionedFile.get(other).method_name(parameters).
548
"""The available optimised InterVersionedFile types."""
550
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
551
"""Integrate versions from self.source into self.target.
553
If version_ids is None all versions from source should be
554
incorporated into this versioned file.
556
Must raise RevisionNotPresent if any of the specified versions
557
are not present in the other files history unless ignore_missing is
558
supplied when they are silently skipped.
561
# - if the target is empty, just add all the versions from
562
# source to target, otherwise:
563
# - make a temporary versioned file of type target
564
# - insert the source content into it one at a time
566
if not self.target.versions():
569
# Make a new target-format versioned file.
570
temp_source = self.target.create_empty("temp", MemoryTransport())
572
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
573
graph = self.source.get_graph(version_ids)
574
order = topo_sort(graph.items())
575
pb = ui.ui_factory.nested_progress_bar()
578
# TODO for incremental cross-format work:
579
# make a versioned file with the following content:
580
# all revisions we have been asked to join
581
# all their ancestors that are *not* in target already.
582
# the immediate parents of the above two sets, with
583
# empty parent lists - these versions are in target already
584
# and the incorrect version data will be ignored.
585
# TODO: for all ancestors that are present in target already,
586
# check them for consistent data, this requires moving sha1 from
588
# TODO: remove parent texts when they are not relevant any more for
589
# memory pressure reduction. RBC 20060313
590
# pb.update('Converting versioned data', 0, len(order))
591
# deltas = self.source.get_deltas(order)
592
for index, version in enumerate(order):
593
pb.update('Converting versioned data', index, len(order))
594
parent_text = target.add_lines(version,
595
self.source.get_parents(version),
596
self.source.get_lines(version),
597
parent_texts=parent_texts)
598
parent_texts[version] = parent_text
599
#delta_parent, sha1, noeol, delta = deltas[version]
600
#target.add_delta(version,
601
# self.source.get_parents(version),
606
#target.get_lines(version)
608
# this should hit the native code path for target
609
if target is not self.target:
610
return self.target.join(temp_source,
618
def _get_source_version_ids(self, version_ids, ignore_missing):
619
"""Determine the version ids to be used from self.source.
621
:param version_ids: The caller-supplied version ids to check. (None
622
for all). If None is in version_ids, it is stripped.
623
:param ignore_missing: if True, remove missing ids from the version
624
list. If False, raise RevisionNotPresent on
625
a missing version id.
626
:return: A set of version ids.
628
if version_ids is None:
629
# None cannot be in source.versions
630
return set(self.source.versions())
633
return set(self.source.versions()).intersection(set(version_ids))
635
new_version_ids = set()
636
for version in version_ids:
639
if not self.source.has_version(version):
640
raise errors.RevisionNotPresent(version, str(self.source))
642
new_version_ids.add(version)
643
return new_version_ids
646
class InterVersionedFileTestProviderAdapter(object):
647
"""A tool to generate a suite testing multiple inter versioned-file classes.
649
This is done by copying the test once for each InterVersionedFile provider
650
and injecting the transport_server, transport_readonly_server,
651
versionedfile_factory and versionedfile_factory_to classes into each copy.
652
Each copy is also given a new id() to make it easy to identify.
655
def __init__(self, transport_server, transport_readonly_server, formats):
656
self._transport_server = transport_server
657
self._transport_readonly_server = transport_readonly_server
658
self._formats = formats
660
def adapt(self, test):
662
for (interversionedfile_class,
663
versionedfile_factory,
664
versionedfile_factory_to) in self._formats:
665
new_test = deepcopy(test)
666
new_test.transport_server = self._transport_server
667
new_test.transport_readonly_server = self._transport_readonly_server
668
new_test.interversionedfile_class = interversionedfile_class
669
new_test.versionedfile_factory = versionedfile_factory
670
new_test.versionedfile_factory_to = versionedfile_factory_to
671
def make_new_test_id():
672
new_id = "%s(%s)" % (new_test.id(), interversionedfile_class.__name__)
673
return lambda: new_id
674
new_test.id = make_new_test_id()
675
result.addTest(new_test)
679
def default_test_list():
680
"""Generate the default list of interversionedfile permutations to test."""
681
from bzrlib.weave import WeaveFile
682
from bzrlib.knit import KnitVersionedFile
684
# test the fallback InterVersionedFile from annotated knits to weave
685
result.append((InterVersionedFile,
688
for optimiser in InterVersionedFile._optimisers:
689
result.append((optimiser,
690
optimiser._matching_file_from_factory,
691
optimiser._matching_file_to_factory
693
# if there are specific combinations we want to use, we can add them
added
removed
bzrlib/versionedfile.py
