1
# Copyright (C) 2005, 2006 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."""
22
from bzrlib.lazy_import import lazy_import
23
lazy_import(globals(), """
24
from copy import deepcopy
35
from bzrlib.transport.memory import MemoryTransport
38
from bzrlib.inter import InterObject
39
from bzrlib.textmerge import TextMerge
40
from bzrlib.symbol_versioning import (deprecated_function,
46
class VersionedFile(object):
47
"""Versioned text file storage.
49
A versioned file manages versions of line-based text files,
50
keeping track of the originating version for each line.
52
To clients the "lines" of the file are represented as a list of
53
strings. These strings will typically have terminal newline
54
characters, but this is not required. In particular files commonly
55
do not have a newline at the end of the file.
57
Texts are identified by a version-id string.
60
def __init__(self, access_mode):
62
self._access_mode = access_mode
65
def check_not_reserved_id(version_id):
66
revision.check_not_reserved_id(version_id)
68
def copy_to(self, name, transport):
69
"""Copy this versioned file to name on transport."""
70
raise NotImplementedError(self.copy_to)
72
@deprecated_method(zero_eight)
74
"""Return a list of all the versions in this versioned file.
76
Please use versionedfile.versions() now.
78
return self.versions()
81
"""Return a unsorted list of versions."""
82
raise NotImplementedError(self.versions)
84
def has_ghost(self, version_id):
85
"""Returns whether version is present as a ghost."""
86
raise NotImplementedError(self.has_ghost)
88
def has_version(self, version_id):
89
"""Returns whether version is present."""
90
raise NotImplementedError(self.has_version)
92
def add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
93
"""Add a text to the versioned file via a pregenerated delta.
95
:param version_id: The version id being added.
96
:param parents: The parents of the version_id.
97
:param delta_parent: The parent this delta was created against.
98
:param sha1: The sha1 of the full text.
99
:param delta: The delta instructions. See get_delta for details.
101
version_id = osutils.safe_revision_id(version_id)
102
parents = [osutils.safe_revision_id(v) for v in parents]
103
self._check_write_ok()
104
if self.has_version(version_id):
105
raise errors.RevisionAlreadyPresent(version_id, self)
106
return self._add_delta(version_id, parents, delta_parent, sha1, noeol, delta)
108
def _add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
109
"""Class specific routine to add a delta.
111
This generic version simply applies the delta to the delta_parent and
114
# strip annotation from delta
116
for start, stop, delta_len, delta_lines in delta:
117
new_delta.append((start, stop, delta_len, [text for origin, text in delta_lines]))
118
if delta_parent is not None:
119
parent_full = self.get_lines(delta_parent)
122
new_full = self._apply_delta(parent_full, new_delta)
123
# its impossible to have noeol on an empty file
124
if noeol and new_full[-1][-1] == '\n':
125
new_full[-1] = new_full[-1][:-1]
126
self.add_lines(version_id, parents, new_full)
128
def add_lines(self, version_id, parents, lines, parent_texts=None):
129
"""Add a single text on top of the versioned file.
131
Must raise RevisionAlreadyPresent if the new version is
132
already present in file history.
134
Must raise RevisionNotPresent if any of the given parents are
135
not present in file history.
136
:param parent_texts: An optional dictionary containing the opaque
137
representations of some or all of the parents of
138
version_id to allow delta optimisations.
139
VERY IMPORTANT: the texts must be those returned
140
by add_lines or data corruption can be caused.
141
:return: An opaque representation of the inserted version which can be
142
provided back to future add_lines calls in the parent_texts
145
version_id = osutils.safe_revision_id(version_id)
146
parents = [osutils.safe_revision_id(v) for v in parents]
147
self._check_write_ok()
148
return self._add_lines(version_id, parents, lines, parent_texts)
150
def _add_lines(self, version_id, parents, lines, parent_texts):
151
"""Helper to do the class specific add_lines."""
152
raise NotImplementedError(self.add_lines)
154
def add_lines_with_ghosts(self, version_id, parents, lines,
156
"""Add lines to the versioned file, allowing ghosts to be present.
158
This takes the same parameters as add_lines.
160
version_id = osutils.safe_revision_id(version_id)
161
parents = [osutils.safe_revision_id(v) for v in parents]
162
self._check_write_ok()
163
return self._add_lines_with_ghosts(version_id, parents, lines,
166
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
167
"""Helper to do class specific add_lines_with_ghosts."""
168
raise NotImplementedError(self.add_lines_with_ghosts)
170
def check(self, progress_bar=None):
171
"""Check the versioned file for integrity."""
172
raise NotImplementedError(self.check)
174
def _check_lines_not_unicode(self, lines):
175
"""Check that lines being added to a versioned file are not unicode."""
177
if line.__class__ is not str:
178
raise errors.BzrBadParameterUnicode("lines")
180
def _check_lines_are_lines(self, lines):
181
"""Check that the lines really are full lines without inline EOL."""
183
if '\n' in line[:-1]:
184
raise errors.BzrBadParameterContainsNewline("lines")
186
def _check_write_ok(self):
187
"""Is the versioned file marked as 'finished' ? Raise if it is."""
189
raise errors.OutSideTransaction()
190
if self._access_mode != 'w':
191
raise errors.ReadOnlyObjectDirtiedError(self)
193
def enable_cache(self):
194
"""Tell this versioned file that it should cache any data it reads.
196
This is advisory, implementations do not have to support caching.
200
def clear_cache(self):
201
"""Remove any data cached in the versioned file object.
203
This only needs to be supported if caches are supported
207
def clone_text(self, new_version_id, old_version_id, parents):
208
"""Add an identical text to old_version_id as new_version_id.
210
Must raise RevisionNotPresent if the old version or any of the
211
parents are not present in file history.
213
Must raise RevisionAlreadyPresent if the new version is
214
already present in file history."""
215
new_version_id = osutils.safe_revision_id(new_version_id)
216
old_version_id = osutils.safe_revision_id(old_version_id)
217
self._check_write_ok()
218
return self._clone_text(new_version_id, old_version_id, parents)
220
def _clone_text(self, new_version_id, old_version_id, parents):
221
"""Helper function to do the _clone_text work."""
222
raise NotImplementedError(self.clone_text)
224
def create_empty(self, name, transport, mode=None):
225
"""Create a new versioned file of this exact type.
227
:param name: the file name
228
:param transport: the transport
229
:param mode: optional file mode.
231
raise NotImplementedError(self.create_empty)
233
def fix_parents(self, version_id, new_parents):
234
"""Fix the parents list for version.
236
This is done by appending a new version to the index
237
with identical data except for the parents list.
238
the parents list must be a superset of the current
241
version_id = osutils.safe_revision_id(version_id)
242
new_parents = [osutils.safe_revision_id(p) for p in new_parents]
243
self._check_write_ok()
244
return self._fix_parents(version_id, new_parents)
246
def _fix_parents(self, version_id, new_parents):
247
"""Helper for fix_parents."""
248
raise NotImplementedError(self.fix_parents)
250
def get_delta(self, version):
251
"""Get a delta for constructing version from some other version.
253
:return: (delta_parent, sha1, noeol, delta)
254
Where delta_parent is a version id or None to indicate no parent.
256
raise NotImplementedError(self.get_delta)
258
def get_deltas(self, version_ids):
259
"""Get multiple deltas at once for constructing versions.
261
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
262
Where delta_parent is a version id or None to indicate no parent, and
263
version_id is the version_id created by that delta.
266
for version_id in version_ids:
267
result[version_id] = self.get_delta(version_id)
270
def make_mpdiff(self, version_id):
271
lines = self._get_line_list([version_id]+self.get_parents(version_id))
274
return multiparent.MultiParent.from_lines(target, parents)
276
def add_mpdiff(self, version, parents, mpdiff):
277
parent_lines = self._get_line_list(parents)
278
mpvf = multiparent.MultiMemoryVersionedFile()
279
for parent_id, lines in zip(parents, parent_lines):
280
mpvf.add_version(lines, parent_id, [])
281
mpvf.add_diff(mpdiff, version, parents)
282
lines = mpvf.get_line_list([version])[0]
283
self.add_lines(version, parents, lines)
285
def get_sha1(self, version_id):
286
"""Get the stored sha1 sum for the given revision.
288
:param name: The name of the version to lookup
290
raise NotImplementedError(self.get_sha1)
292
def get_suffixes(self):
293
"""Return the file suffixes associated with this versioned file."""
294
raise NotImplementedError(self.get_suffixes)
296
def get_text(self, version_id):
297
"""Return version contents as a text string.
299
Raises RevisionNotPresent if version is not present in
302
return ''.join(self.get_lines(version_id))
303
get_string = get_text
305
def get_texts(self, version_ids):
306
"""Return the texts of listed versions as a list of strings.
308
Raises RevisionNotPresent if version is not present in
311
return [''.join(self.get_lines(v)) for v in version_ids]
313
def get_lines(self, version_id):
314
"""Return version contents as a sequence of lines.
316
Raises RevisionNotPresent if version is not present in
319
raise NotImplementedError(self.get_lines)
321
def _get_line_list(self, version_ids):
322
return [t.splitlines(True) for t in self.get_texts(version_ids)]
324
def get_ancestry(self, version_ids):
325
"""Return a list of all ancestors of given version(s). This
326
will not include the null revision.
328
Must raise RevisionNotPresent if any of the given versions are
329
not present in file history."""
330
if isinstance(version_ids, basestring):
331
version_ids = [version_ids]
332
raise NotImplementedError(self.get_ancestry)
334
def get_ancestry_with_ghosts(self, version_ids):
335
"""Return a list of all ancestors of given version(s). This
336
will not include the null revision.
338
Must raise RevisionNotPresent if any of the given versions are
339
not present in file history.
341
Ghosts that are known about will be included in ancestry list,
342
but are not explicitly marked.
344
raise NotImplementedError(self.get_ancestry_with_ghosts)
346
def get_graph(self, version_ids=None):
347
"""Return a graph from the versioned file.
349
Ghosts are not listed or referenced in the graph.
350
:param version_ids: Versions to select.
351
None means retrieve all versions.
354
if version_ids is None:
355
for version in self.versions():
356
result[version] = self.get_parents(version)
358
pending = set(osutils.safe_revision_id(v) for v in version_ids)
360
version = pending.pop()
361
if version in result:
363
parents = self.get_parents(version)
364
for parent in parents:
368
result[version] = parents
371
def get_graph_with_ghosts(self):
372
"""Return a graph for the entire versioned file.
374
Ghosts are referenced in parents list but are not
377
raise NotImplementedError(self.get_graph_with_ghosts)
379
@deprecated_method(zero_eight)
380
def parent_names(self, version):
381
"""Return version names for parents of a version.
383
See get_parents for the current api.
385
return self.get_parents(version)
387
def get_parents(self, version_id):
388
"""Return version names for parents of a version.
390
Must raise RevisionNotPresent if version is not present in
393
raise NotImplementedError(self.get_parents)
395
def get_parents_with_ghosts(self, version_id):
396
"""Return version names for parents of version_id.
398
Will raise RevisionNotPresent if version_id is not present
401
Ghosts that are known about will be included in the parent list,
402
but are not explicitly marked.
404
raise NotImplementedError(self.get_parents_with_ghosts)
406
def annotate_iter(self, version_id):
407
"""Yield list of (version-id, line) pairs for the specified
410
Must raise RevisionNotPresent if any of the given versions are
411
not present in file history.
413
raise NotImplementedError(self.annotate_iter)
415
def annotate(self, version_id):
416
return list(self.annotate_iter(version_id))
418
def _apply_delta(self, lines, delta):
419
"""Apply delta to lines."""
422
for start, end, count, delta_lines in delta:
423
lines[offset+start:offset+end] = delta_lines
424
offset = offset + (start - end) + count
427
def join(self, other, pb=None, msg=None, version_ids=None,
428
ignore_missing=False):
429
"""Integrate versions from other into this versioned file.
431
If version_ids is None all versions from other should be
432
incorporated into this versioned file.
434
Must raise RevisionNotPresent if any of the specified versions
435
are not present in the other files history unless ignore_missing
436
is supplied when they are silently skipped.
438
self._check_write_ok()
439
return InterVersionedFile.get(other, self).join(
445
def iter_lines_added_or_present_in_versions(self, version_ids=None,
447
"""Iterate over the lines in the versioned file from version_ids.
449
This may return lines from other versions, and does not return the
450
specific version marker at this point. The api may be changed
451
during development to include the version that the versioned file
452
thinks is relevant, but given that such hints are just guesses,
453
its better not to have it if we don't need it.
455
If a progress bar is supplied, it may be used to indicate progress.
456
The caller is responsible for cleaning up progress bars (because this
459
NOTES: Lines are normalised: they will all have \n terminators.
460
Lines are returned in arbitrary order.
462
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
464
def transaction_finished(self):
465
"""The transaction that this file was opened in has finished.
467
This records self.finished = True and should cause all mutating
472
@deprecated_method(zero_eight)
473
def walk(self, version_ids=None):
474
"""Walk the versioned file as a weave-like structure, for
475
versions relative to version_ids. Yields sequence of (lineno,
476
insert, deletes, text) for each relevant line.
478
Must raise RevisionNotPresent if any of the specified versions
479
are not present in the file history.
481
:param version_ids: the version_ids to walk with respect to. If not
482
supplied the entire weave-like structure is walked.
484
walk is deprecated in favour of iter_lines_added_or_present_in_versions
486
raise NotImplementedError(self.walk)
488
@deprecated_method(zero_eight)
489
def iter_names(self):
490
"""Walk the names list."""
491
return iter(self.versions())
493
def plan_merge(self, ver_a, ver_b):
494
"""Return pseudo-annotation indicating how the two versions merge.
496
This is computed between versions a and b and their common
499
Weave lines present in none of them are skipped entirely.
502
killed-base Dead in base revision
503
killed-both Killed in each revision
506
unchanged Alive in both a and b (possibly created in both)
509
ghost-a Killed in a, unborn in b
510
ghost-b Killed in b, unborn in a
511
irrelevant Not in either revision
513
raise NotImplementedError(VersionedFile.plan_merge)
515
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
516
b_marker=TextMerge.B_MARKER):
517
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
520
class PlanWeaveMerge(TextMerge):
521
"""Weave merge that takes a plan as its input.
523
This exists so that VersionedFile.plan_merge is implementable.
524
Most callers will want to use WeaveMerge instead.
527
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
528
b_marker=TextMerge.B_MARKER):
529
TextMerge.__init__(self, a_marker, b_marker)
532
def _merge_struct(self):
537
def outstanding_struct():
538
if not lines_a and not lines_b:
540
elif ch_a and not ch_b:
543
elif ch_b and not ch_a:
545
elif lines_a == lines_b:
548
yield (lines_a, lines_b)
550
# We previously considered either 'unchanged' or 'killed-both' lines
551
# to be possible places to resynchronize. However, assuming agreement
552
# on killed-both lines may be too aggressive. -- mbp 20060324
553
for state, line in self.plan:
554
if state == 'unchanged':
555
# resync and flush queued conflicts changes if any
556
for struct in outstanding_struct():
562
if state == 'unchanged':
565
elif state == 'killed-a':
568
elif state == 'killed-b':
571
elif state == 'new-a':
574
elif state == 'new-b':
578
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
579
'killed-base', 'killed-both'), state
580
for struct in outstanding_struct():
584
class WeaveMerge(PlanWeaveMerge):
585
"""Weave merge that takes a VersionedFile and two versions as its input"""
587
def __init__(self, versionedfile, ver_a, ver_b,
588
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
589
plan = versionedfile.plan_merge(ver_a, ver_b)
590
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
593
class InterVersionedFile(InterObject):
594
"""This class represents operations taking place between two versionedfiles..
596
Its instances have methods like join, and contain
597
references to the source and target versionedfiles these operations can be
600
Often we will provide convenience methods on 'versionedfile' which carry out
601
operations with another versionedfile - they will always forward to
602
InterVersionedFile.get(other).method_name(parameters).
606
"""The available optimised InterVersionedFile types."""
608
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
609
"""Integrate versions from self.source into self.target.
611
If version_ids is None all versions from source should be
612
incorporated into this versioned file.
614
Must raise RevisionNotPresent if any of the specified versions
615
are not present in the other files history unless ignore_missing is
616
supplied when they are silently skipped.
619
# - if the target is empty, just add all the versions from
620
# source to target, otherwise:
621
# - make a temporary versioned file of type target
622
# - insert the source content into it one at a time
624
if not self.target.versions():
627
# Make a new target-format versioned file.
628
temp_source = self.target.create_empty("temp", MemoryTransport())
630
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
631
graph = self.source.get_graph(version_ids)
632
order = tsort.topo_sort(graph.items())
633
pb = ui.ui_factory.nested_progress_bar()
636
# TODO for incremental cross-format work:
637
# make a versioned file with the following content:
638
# all revisions we have been asked to join
639
# all their ancestors that are *not* in target already.
640
# the immediate parents of the above two sets, with
641
# empty parent lists - these versions are in target already
642
# and the incorrect version data will be ignored.
643
# TODO: for all ancestors that are present in target already,
644
# check them for consistent data, this requires moving sha1 from
646
# TODO: remove parent texts when they are not relevant any more for
647
# memory pressure reduction. RBC 20060313
648
# pb.update('Converting versioned data', 0, len(order))
649
# deltas = self.source.get_deltas(order)
650
for index, version in enumerate(order):
651
pb.update('Converting versioned data', index, len(order))
652
parent_text = target.add_lines(version,
653
self.source.get_parents(version),
654
self.source.get_lines(version),
655
parent_texts=parent_texts)
656
parent_texts[version] = parent_text
657
#delta_parent, sha1, noeol, delta = deltas[version]
658
#target.add_delta(version,
659
# self.source.get_parents(version),
664
#target.get_lines(version)
666
# this should hit the native code path for target
667
if target is not self.target:
668
return self.target.join(temp_source,
676
def _get_source_version_ids(self, version_ids, ignore_missing):
677
"""Determine the version ids to be used from self.source.
679
:param version_ids: The caller-supplied version ids to check. (None
680
for all). If None is in version_ids, it is stripped.
681
:param ignore_missing: if True, remove missing ids from the version
682
list. If False, raise RevisionNotPresent on
683
a missing version id.
684
:return: A set of version ids.
686
if version_ids is None:
687
# None cannot be in source.versions
688
return set(self.source.versions())
690
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
692
return set(self.source.versions()).intersection(set(version_ids))
694
new_version_ids = set()
695
for version in version_ids:
698
if not self.source.has_version(version):
699
raise errors.RevisionNotPresent(version, str(self.source))
701
new_version_ids.add(version)
702
return new_version_ids
705
class InterVersionedFileTestProviderAdapter(object):
706
"""A tool to generate a suite testing multiple inter versioned-file classes.
708
This is done by copying the test once for each InterVersionedFile provider
709
and injecting the transport_server, transport_readonly_server,
710
versionedfile_factory and versionedfile_factory_to classes into each copy.
711
Each copy is also given a new id() to make it easy to identify.
714
def __init__(self, transport_server, transport_readonly_server, formats):
715
self._transport_server = transport_server
716
self._transport_readonly_server = transport_readonly_server
717
self._formats = formats
719
def adapt(self, test):
720
result = unittest.TestSuite()
721
for (interversionedfile_class,
722
versionedfile_factory,
723
versionedfile_factory_to) in self._formats:
724
new_test = deepcopy(test)
725
new_test.transport_server = self._transport_server
726
new_test.transport_readonly_server = self._transport_readonly_server
727
new_test.interversionedfile_class = interversionedfile_class
728
new_test.versionedfile_factory = versionedfile_factory
729
new_test.versionedfile_factory_to = versionedfile_factory_to
730
def make_new_test_id():
731
new_id = "%s(%s)" % (new_test.id(), interversionedfile_class.__name__)
732
return lambda: new_id
733
new_test.id = make_new_test_id()
734
result.addTest(new_test)
738
def default_test_list():
739
"""Generate the default list of interversionedfile permutations to test."""
740
from bzrlib.weave import WeaveFile
741
from bzrlib.knit import KnitVersionedFile
743
# test the fallback InterVersionedFile from annotated knits to weave
744
result.append((InterVersionedFile,
747
for optimiser in InterVersionedFile._optimisers:
748
result.append((optimiser,
749
optimiser._matching_file_from_factory,
750
optimiser._matching_file_to_factory
752
# if there are specific combinations we want to use, we can add them
added
removed
bzrlib/versionedfile.py
