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 cStringIO import StringIO
40
from bzrlib.inter import InterObject
41
from bzrlib.textmerge import TextMerge
42
from bzrlib.symbol_versioning import (deprecated_function,
48
class VersionedFile(object):
49
"""Versioned text file storage.
51
A versioned file manages versions of line-based text files,
52
keeping track of the originating version for each line.
54
To clients the "lines" of the file are represented as a list of
55
strings. These strings will typically have terminal newline
56
characters, but this is not required. In particular files commonly
57
do not have a newline at the end of the file.
59
Texts are identified by a version-id string.
62
def __init__(self, access_mode):
64
self._access_mode = access_mode
67
def check_not_reserved_id(version_id):
68
revision.check_not_reserved_id(version_id)
70
def copy_to(self, name, transport):
71
"""Copy this versioned file to name on transport."""
72
raise NotImplementedError(self.copy_to)
74
@deprecated_method(zero_eight)
76
"""Return a list of all the versions in this versioned file.
78
Please use versionedfile.versions() now.
80
return self.versions()
83
"""Return a unsorted list of versions."""
84
raise NotImplementedError(self.versions)
86
def has_ghost(self, version_id):
87
"""Returns whether version is present as a ghost."""
88
raise NotImplementedError(self.has_ghost)
90
def has_version(self, version_id):
91
"""Returns whether version is present."""
92
raise NotImplementedError(self.has_version)
94
def add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
95
"""Add a text to the versioned file via a pregenerated delta.
97
:param version_id: The version id being added.
98
:param parents: The parents of the version_id.
99
:param delta_parent: The parent this delta was created against.
100
:param sha1: The sha1 of the full text.
101
:param delta: The delta instructions. See get_delta for details.
103
version_id = osutils.safe_revision_id(version_id)
104
parents = [osutils.safe_revision_id(v) for v in parents]
105
self._check_write_ok()
106
if self.has_version(version_id):
107
raise errors.RevisionAlreadyPresent(version_id, self)
108
return self._add_delta(version_id, parents, delta_parent, sha1, noeol, delta)
110
def _add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
111
"""Class specific routine to add a delta.
113
This generic version simply applies the delta to the delta_parent and
116
# strip annotation from delta
118
for start, stop, delta_len, delta_lines in delta:
119
new_delta.append((start, stop, delta_len, [text for origin, text in delta_lines]))
120
if delta_parent is not None:
121
parent_full = self.get_lines(delta_parent)
124
new_full = self._apply_delta(parent_full, new_delta)
125
# its impossible to have noeol on an empty file
126
if noeol and new_full[-1][-1] == '\n':
127
new_full[-1] = new_full[-1][:-1]
128
self.add_lines(version_id, parents, new_full)
130
def add_lines(self, version_id, parents, lines, parent_texts=None):
131
"""Add a single text on top of the versioned file.
133
Must raise RevisionAlreadyPresent if the new version is
134
already present in file history.
136
Must raise RevisionNotPresent if any of the given parents are
137
not present in file history.
138
:param parent_texts: An optional dictionary containing the opaque
139
representations of some or all of the parents of
140
version_id to allow delta optimisations.
141
VERY IMPORTANT: the texts must be those returned
142
by add_lines or data corruption can be caused.
143
:return: An opaque representation of the inserted version which can be
144
provided back to future add_lines calls in the parent_texts
147
version_id = osutils.safe_revision_id(version_id)
148
parents = [osutils.safe_revision_id(v) for v in parents]
149
self._check_write_ok()
150
return self._add_lines(version_id, parents, lines, parent_texts)
152
def _add_lines(self, version_id, parents, lines, parent_texts):
153
"""Helper to do the class specific add_lines."""
154
raise NotImplementedError(self.add_lines)
156
def add_lines_with_ghosts(self, version_id, parents, lines,
158
"""Add lines to the versioned file, allowing ghosts to be present.
160
This takes the same parameters as add_lines.
162
version_id = osutils.safe_revision_id(version_id)
163
parents = [osutils.safe_revision_id(v) for v in parents]
164
self._check_write_ok()
165
return self._add_lines_with_ghosts(version_id, parents, lines,
168
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
169
"""Helper to do class specific add_lines_with_ghosts."""
170
raise NotImplementedError(self.add_lines_with_ghosts)
172
def check(self, progress_bar=None):
173
"""Check the versioned file for integrity."""
174
raise NotImplementedError(self.check)
176
def _check_lines_not_unicode(self, lines):
177
"""Check that lines being added to a versioned file are not unicode."""
179
if line.__class__ is not str:
180
raise errors.BzrBadParameterUnicode("lines")
182
def _check_lines_are_lines(self, lines):
183
"""Check that the lines really are full lines without inline EOL."""
185
if '\n' in line[:-1]:
186
raise errors.BzrBadParameterContainsNewline("lines")
188
def _check_write_ok(self):
189
"""Is the versioned file marked as 'finished' ? Raise if it is."""
191
raise errors.OutSideTransaction()
192
if self._access_mode != 'w':
193
raise errors.ReadOnlyObjectDirtiedError(self)
195
def enable_cache(self):
196
"""Tell this versioned file that it should cache any data it reads.
198
This is advisory, implementations do not have to support caching.
202
def clear_cache(self):
203
"""Remove any data cached in the versioned file object.
205
This only needs to be supported if caches are supported
209
def clone_text(self, new_version_id, old_version_id, parents):
210
"""Add an identical text to old_version_id as new_version_id.
212
Must raise RevisionNotPresent if the old version or any of the
213
parents are not present in file history.
215
Must raise RevisionAlreadyPresent if the new version is
216
already present in file history."""
217
new_version_id = osutils.safe_revision_id(new_version_id)
218
old_version_id = osutils.safe_revision_id(old_version_id)
219
self._check_write_ok()
220
return self._clone_text(new_version_id, old_version_id, parents)
222
def _clone_text(self, new_version_id, old_version_id, parents):
223
"""Helper function to do the _clone_text work."""
224
raise NotImplementedError(self.clone_text)
226
def create_empty(self, name, transport, mode=None):
227
"""Create a new versioned file of this exact type.
229
:param name: the file name
230
:param transport: the transport
231
:param mode: optional file mode.
233
raise NotImplementedError(self.create_empty)
235
def fix_parents(self, version_id, new_parents):
236
"""Fix the parents list for version.
238
This is done by appending a new version to the index
239
with identical data except for the parents list.
240
the parents list must be a superset of the current
243
version_id = osutils.safe_revision_id(version_id)
244
new_parents = [osutils.safe_revision_id(p) for p in new_parents]
245
self._check_write_ok()
246
return self._fix_parents(version_id, new_parents)
248
def _fix_parents(self, version_id, new_parents):
249
"""Helper for fix_parents."""
250
raise NotImplementedError(self.fix_parents)
252
def get_delta(self, version):
253
"""Get a delta for constructing version from some other version.
255
:return: (delta_parent, sha1, noeol, delta)
256
Where delta_parent is a version id or None to indicate no parent.
258
raise NotImplementedError(self.get_delta)
260
def get_deltas(self, version_ids):
261
"""Get multiple deltas at once for constructing versions.
263
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
264
Where delta_parent is a version id or None to indicate no parent, and
265
version_id is the version_id created by that delta.
268
for version_id in version_ids:
269
result[version_id] = self.get_delta(version_id)
272
def make_mpdiffs(self, version_ids):
273
knit_versions = set()
274
for version_id in version_ids:
275
knit_versions.add(version_id)
276
knit_versions.update(self.get_parents(version_id))
277
lines = dict(zip(knit_versions,
278
self._get_lf_split_line_list(knit_versions)))
280
for version_id in version_ids:
281
target = lines[version_id]
282
parents = [lines[p] for p in self.get_parents(version_id)]
284
left_parent_blocks = self._extract_blocks(version_id,
287
left_parent_blocks = None
288
diffs.append(multiparent.MultiParent.from_lines(target, parents,
292
def _extract_blocks(self, version_id, source, target):
295
def add_mpdiffs(self, records):
297
for version, parents, expected_sha1, mpdiff in records:
298
mpvf = multiparent.MultiMemoryVersionedFile()
299
needed_parents = [p for p in parents if not mpvf.has_version(p)]
300
parent_lines = self._get_lf_split_line_list(needed_parents)
301
for parent_id, lines in zip(needed_parents, parent_lines):
302
mpvf.add_version(lines, parent_id, [])
303
mpvf.add_diff(mpdiff, version, parents)
304
lines = mpvf.get_line_list([version])[0]
305
version_text = self.add_lines(version, parents, lines, vf_parents)
306
vf_parents[version] = version_text
307
if expected_sha1 != self.get_sha1(version):
308
raise errors.VersionedFileInvalidChecksum(version)
310
def get_sha1(self, version_id):
311
"""Get the stored sha1 sum for the given revision.
313
:param name: The name of the version to lookup
315
raise NotImplementedError(self.get_sha1)
317
def get_sha1s(self, version_ids):
318
"""Get the stored sha1 sums for the given revisions.
320
:param version_ids: The names of the versions to lookup
321
:return: a list of sha1s in order according to the version_ids
323
raise NotImplementedError(self.get_sha1)
325
def get_suffixes(self):
326
"""Return the file suffixes associated with this versioned file."""
327
raise NotImplementedError(self.get_suffixes)
329
def get_text(self, version_id):
330
"""Return version contents as a text string.
332
Raises RevisionNotPresent if version is not present in
335
return ''.join(self.get_lines(version_id))
336
get_string = get_text
338
def get_texts(self, version_ids):
339
"""Return the texts of listed versions as a list of strings.
341
Raises RevisionNotPresent if version is not present in
344
return [''.join(self.get_lines(v)) for v in version_ids]
346
def get_lines(self, version_id):
347
"""Return version contents as a sequence of lines.
349
Raises RevisionNotPresent if version is not present in
352
raise NotImplementedError(self.get_lines)
354
def _get_lf_split_line_list(self, version_ids):
355
return [StringIO(t).readlines() for t in self.get_texts(version_ids)]
357
def get_ancestry(self, version_ids, topo_sorted=True):
358
"""Return a list of all ancestors of given version(s). This
359
will not include the null revision.
361
This list will not be topologically sorted if topo_sorted=False is
364
Must raise RevisionNotPresent if any of the given versions are
365
not present in file history."""
366
if isinstance(version_ids, basestring):
367
version_ids = [version_ids]
368
raise NotImplementedError(self.get_ancestry)
370
def get_ancestry_with_ghosts(self, version_ids):
371
"""Return a list of all ancestors of given version(s). This
372
will not include the null revision.
374
Must raise RevisionNotPresent if any of the given versions are
375
not present in file history.
377
Ghosts that are known about will be included in ancestry list,
378
but are not explicitly marked.
380
raise NotImplementedError(self.get_ancestry_with_ghosts)
382
def get_graph(self, version_ids=None):
383
"""Return a graph from the versioned file.
385
Ghosts are not listed or referenced in the graph.
386
:param version_ids: Versions to select.
387
None means retrieve all versions.
390
if version_ids is None:
391
for version in self.versions():
392
result[version] = self.get_parents(version)
394
pending = set(osutils.safe_revision_id(v) for v in version_ids)
396
version = pending.pop()
397
if version in result:
399
parents = self.get_parents(version)
400
for parent in parents:
404
result[version] = parents
407
def get_graph_with_ghosts(self):
408
"""Return a graph for the entire versioned file.
410
Ghosts are referenced in parents list but are not
413
raise NotImplementedError(self.get_graph_with_ghosts)
415
@deprecated_method(zero_eight)
416
def parent_names(self, version):
417
"""Return version names for parents of a version.
419
See get_parents for the current api.
421
return self.get_parents(version)
423
def get_parents(self, version_id):
424
"""Return version names for parents of a version.
426
Must raise RevisionNotPresent if version is not present in
429
raise NotImplementedError(self.get_parents)
431
def get_parents_with_ghosts(self, version_id):
432
"""Return version names for parents of version_id.
434
Will raise RevisionNotPresent if version_id is not present
437
Ghosts that are known about will be included in the parent list,
438
but are not explicitly marked.
440
raise NotImplementedError(self.get_parents_with_ghosts)
442
def annotate_iter(self, version_id):
443
"""Yield list of (version-id, line) pairs for the specified
446
Must raise RevisionNotPresent if any of the given versions are
447
not present in file history.
449
raise NotImplementedError(self.annotate_iter)
451
def annotate(self, version_id):
452
return list(self.annotate_iter(version_id))
454
def _apply_delta(self, lines, delta):
455
"""Apply delta to lines."""
458
for start, end, count, delta_lines in delta:
459
lines[offset+start:offset+end] = delta_lines
460
offset = offset + (start - end) + count
463
def join(self, other, pb=None, msg=None, version_ids=None,
464
ignore_missing=False):
465
"""Integrate versions from other into this versioned file.
467
If version_ids is None all versions from other should be
468
incorporated into this versioned file.
470
Must raise RevisionNotPresent if any of the specified versions
471
are not present in the other files history unless ignore_missing
472
is supplied when they are silently skipped.
474
self._check_write_ok()
475
return InterVersionedFile.get(other, self).join(
481
def iter_lines_added_or_present_in_versions(self, version_ids=None,
483
"""Iterate over the lines in the versioned file from version_ids.
485
This may return lines from other versions, and does not return the
486
specific version marker at this point. The api may be changed
487
during development to include the version that the versioned file
488
thinks is relevant, but given that such hints are just guesses,
489
its better not to have it if we don't need it.
491
If a progress bar is supplied, it may be used to indicate progress.
492
The caller is responsible for cleaning up progress bars (because this
495
NOTES: Lines are normalised: they will all have \n terminators.
496
Lines are returned in arbitrary order.
498
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
500
def transaction_finished(self):
501
"""The transaction that this file was opened in has finished.
503
This records self.finished = True and should cause all mutating
508
@deprecated_method(zero_eight)
509
def walk(self, version_ids=None):
510
"""Walk the versioned file as a weave-like structure, for
511
versions relative to version_ids. Yields sequence of (lineno,
512
insert, deletes, text) for each relevant line.
514
Must raise RevisionNotPresent if any of the specified versions
515
are not present in the file history.
517
:param version_ids: the version_ids to walk with respect to. If not
518
supplied the entire weave-like structure is walked.
520
walk is deprecated in favour of iter_lines_added_or_present_in_versions
522
raise NotImplementedError(self.walk)
524
@deprecated_method(zero_eight)
525
def iter_names(self):
526
"""Walk the names list."""
527
return iter(self.versions())
529
def plan_merge(self, ver_a, ver_b):
530
"""Return pseudo-annotation indicating how the two versions merge.
532
This is computed between versions a and b and their common
535
Weave lines present in none of them are skipped entirely.
538
killed-base Dead in base revision
539
killed-both Killed in each revision
542
unchanged Alive in both a and b (possibly created in both)
545
ghost-a Killed in a, unborn in b
546
ghost-b Killed in b, unborn in a
547
irrelevant Not in either revision
549
raise NotImplementedError(VersionedFile.plan_merge)
551
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
552
b_marker=TextMerge.B_MARKER):
553
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
556
class PlanWeaveMerge(TextMerge):
557
"""Weave merge that takes a plan as its input.
559
This exists so that VersionedFile.plan_merge is implementable.
560
Most callers will want to use WeaveMerge instead.
563
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
564
b_marker=TextMerge.B_MARKER):
565
TextMerge.__init__(self, a_marker, b_marker)
568
def _merge_struct(self):
573
def outstanding_struct():
574
if not lines_a and not lines_b:
576
elif ch_a and not ch_b:
579
elif ch_b and not ch_a:
581
elif lines_a == lines_b:
584
yield (lines_a, lines_b)
586
# We previously considered either 'unchanged' or 'killed-both' lines
587
# to be possible places to resynchronize. However, assuming agreement
588
# on killed-both lines may be too aggressive. -- mbp 20060324
589
for state, line in self.plan:
590
if state == 'unchanged':
591
# resync and flush queued conflicts changes if any
592
for struct in outstanding_struct():
598
if state == 'unchanged':
601
elif state == 'killed-a':
604
elif state == 'killed-b':
607
elif state == 'new-a':
610
elif state == 'new-b':
614
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
615
'killed-base', 'killed-both'), state
616
for struct in outstanding_struct():
620
class WeaveMerge(PlanWeaveMerge):
621
"""Weave merge that takes a VersionedFile and two versions as its input"""
623
def __init__(self, versionedfile, ver_a, ver_b,
624
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
625
plan = versionedfile.plan_merge(ver_a, ver_b)
626
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
629
class InterVersionedFile(InterObject):
630
"""This class represents operations taking place between two versionedfiles..
632
Its instances have methods like join, and contain
633
references to the source and target versionedfiles these operations can be
636
Often we will provide convenience methods on 'versionedfile' which carry out
637
operations with another versionedfile - they will always forward to
638
InterVersionedFile.get(other).method_name(parameters).
642
"""The available optimised InterVersionedFile types."""
644
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
645
"""Integrate versions from self.source into self.target.
647
If version_ids is None all versions from source should be
648
incorporated into this versioned file.
650
Must raise RevisionNotPresent if any of the specified versions
651
are not present in the other files history unless ignore_missing is
652
supplied when they are silently skipped.
655
# - if the target is empty, just add all the versions from
656
# source to target, otherwise:
657
# - make a temporary versioned file of type target
658
# - insert the source content into it one at a time
660
if not self.target.versions():
663
# Make a new target-format versioned file.
664
temp_source = self.target.create_empty("temp", MemoryTransport())
666
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
667
graph = self.source.get_graph(version_ids)
668
order = tsort.topo_sort(graph.items())
669
pb = ui.ui_factory.nested_progress_bar()
672
# TODO for incremental cross-format work:
673
# make a versioned file with the following content:
674
# all revisions we have been asked to join
675
# all their ancestors that are *not* in target already.
676
# the immediate parents of the above two sets, with
677
# empty parent lists - these versions are in target already
678
# and the incorrect version data will be ignored.
679
# TODO: for all ancestors that are present in target already,
680
# check them for consistent data, this requires moving sha1 from
682
# TODO: remove parent texts when they are not relevant any more for
683
# memory pressure reduction. RBC 20060313
684
# pb.update('Converting versioned data', 0, len(order))
685
# deltas = self.source.get_deltas(order)
686
for index, version in enumerate(order):
687
pb.update('Converting versioned data', index, len(order))
688
parent_text = target.add_lines(version,
689
self.source.get_parents(version),
690
self.source.get_lines(version),
691
parent_texts=parent_texts)
692
parent_texts[version] = parent_text
693
#delta_parent, sha1, noeol, delta = deltas[version]
694
#target.add_delta(version,
695
# self.source.get_parents(version),
700
#target.get_lines(version)
702
# this should hit the native code path for target
703
if target is not self.target:
704
return self.target.join(temp_source,
712
def _get_source_version_ids(self, version_ids, ignore_missing):
713
"""Determine the version ids to be used from self.source.
715
:param version_ids: The caller-supplied version ids to check. (None
716
for all). If None is in version_ids, it is stripped.
717
:param ignore_missing: if True, remove missing ids from the version
718
list. If False, raise RevisionNotPresent on
719
a missing version id.
720
:return: A set of version ids.
722
if version_ids is None:
723
# None cannot be in source.versions
724
return set(self.source.versions())
726
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
728
return set(self.source.versions()).intersection(set(version_ids))
730
new_version_ids = set()
731
for version in version_ids:
734
if not self.source.has_version(version):
735
raise errors.RevisionNotPresent(version, str(self.source))
737
new_version_ids.add(version)
738
return new_version_ids
741
class InterVersionedFileTestProviderAdapter(object):
742
"""A tool to generate a suite testing multiple inter versioned-file classes.
744
This is done by copying the test once for each InterVersionedFile provider
745
and injecting the transport_server, transport_readonly_server,
746
versionedfile_factory and versionedfile_factory_to classes into each copy.
747
Each copy is also given a new id() to make it easy to identify.
750
def __init__(self, transport_server, transport_readonly_server, formats):
751
self._transport_server = transport_server
752
self._transport_readonly_server = transport_readonly_server
753
self._formats = formats
755
def adapt(self, test):
756
result = unittest.TestSuite()
757
for (interversionedfile_class,
758
versionedfile_factory,
759
versionedfile_factory_to) in self._formats:
760
new_test = deepcopy(test)
761
new_test.transport_server = self._transport_server
762
new_test.transport_readonly_server = self._transport_readonly_server
763
new_test.interversionedfile_class = interversionedfile_class
764
new_test.versionedfile_factory = versionedfile_factory
765
new_test.versionedfile_factory_to = versionedfile_factory_to
766
def make_new_test_id():
767
new_id = "%s(%s)" % (new_test.id(), interversionedfile_class.__name__)
768
return lambda: new_id
769
new_test.id = make_new_test_id()
770
result.addTest(new_test)
774
def default_test_list():
775
"""Generate the default list of interversionedfile permutations to test."""
776
from bzrlib.weave import WeaveFile
777
from bzrlib.knit import KnitVersionedFile
779
# test the fallback InterVersionedFile from annotated knits to weave
780
result.append((InterVersionedFile,
783
for optimiser in InterVersionedFile._optimisers:
784
result.append((optimiser,
785
optimiser._matching_file_from_factory,
786
optimiser._matching_file_to_factory
788
# if there are specific combinations we want to use, we can add them
added
removed
bzrlib/versionedfile.py
