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
34
from bzrlib.transport.memory import MemoryTransport
37
from bzrlib.inter import InterObject
38
from bzrlib.textmerge import TextMerge
39
from bzrlib.symbol_versioning import (deprecated_function,
45
class VersionedFile(object):
46
"""Versioned text file storage.
48
A versioned file manages versions of line-based text files,
49
keeping track of the originating version for each line.
51
To clients the "lines" of the file are represented as a list of
52
strings. These strings will typically have terminal newline
53
characters, but this is not required. In particular files commonly
54
do not have a newline at the end of the file.
56
Texts are identified by a version-id string.
59
def __init__(self, access_mode):
61
self._access_mode = access_mode
64
def check_not_reserved_id(version_id):
65
revision.check_not_reserved_id(version_id)
67
def copy_to(self, name, transport):
68
"""Copy this versioned file to name on transport."""
69
raise NotImplementedError(self.copy_to)
71
@deprecated_method(zero_eight)
73
"""Return a list of all the versions in this versioned file.
75
Please use versionedfile.versions() now.
77
return self.versions()
80
"""Return a unsorted list of versions."""
81
raise NotImplementedError(self.versions)
83
def has_ghost(self, version_id):
84
"""Returns whether version is present as a ghost."""
85
raise NotImplementedError(self.has_ghost)
87
def has_version(self, version_id):
88
"""Returns whether version is present."""
89
raise NotImplementedError(self.has_version)
91
def add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
92
"""Add a text to the versioned file via a pregenerated delta.
94
:param version_id: The version id being added.
95
:param parents: The parents of the version_id.
96
:param delta_parent: The parent this delta was created against.
97
:param sha1: The sha1 of the full text.
98
:param delta: The delta instructions. See get_delta for details.
100
version_id = osutils.safe_revision_id(version_id)
101
parents = [osutils.safe_revision_id(v) for v in parents]
102
self._check_write_ok()
103
if self.has_version(version_id):
104
raise errors.RevisionAlreadyPresent(version_id, self)
105
return self._add_delta(version_id, parents, delta_parent, sha1, noeol, delta)
107
def _add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
108
"""Class specific routine to add a delta.
110
This generic version simply applies the delta to the delta_parent and
113
# strip annotation from delta
115
for start, stop, delta_len, delta_lines in delta:
116
new_delta.append((start, stop, delta_len, [text for origin, text in delta_lines]))
117
if delta_parent is not None:
118
parent_full = self.get_lines(delta_parent)
121
new_full = self._apply_delta(parent_full, new_delta)
122
# its impossible to have noeol on an empty file
123
if noeol and new_full[-1][-1] == '\n':
124
new_full[-1] = new_full[-1][:-1]
125
self.add_lines(version_id, parents, new_full)
127
def add_lines(self, version_id, parents, lines, parent_texts=None):
128
"""Add a single text on top of the versioned file.
130
Must raise RevisionAlreadyPresent if the new version is
131
already present in file history.
133
Must raise RevisionNotPresent if any of the given parents are
134
not present in file history.
135
:param parent_texts: An optional dictionary containing the opaque
136
representations of some or all of the parents of
137
version_id to allow delta optimisations.
138
VERY IMPORTANT: the texts must be those returned
139
by add_lines or data corruption can be caused.
140
:return: An opaque representation of the inserted version which can be
141
provided back to future add_lines calls in the parent_texts
144
version_id = osutils.safe_revision_id(version_id)
145
parents = [osutils.safe_revision_id(v) for v in parents]
146
self._check_write_ok()
147
return self._add_lines(version_id, parents, lines, parent_texts)
149
def _add_lines(self, version_id, parents, lines, parent_texts):
150
"""Helper to do the class specific add_lines."""
151
raise NotImplementedError(self.add_lines)
153
def add_lines_with_ghosts(self, version_id, parents, lines,
155
"""Add lines to the versioned file, allowing ghosts to be present.
157
This takes the same parameters as add_lines.
159
version_id = osutils.safe_revision_id(version_id)
160
parents = [osutils.safe_revision_id(v) for v in parents]
161
self._check_write_ok()
162
return self._add_lines_with_ghosts(version_id, parents, lines,
165
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
166
"""Helper to do class specific add_lines_with_ghosts."""
167
raise NotImplementedError(self.add_lines_with_ghosts)
169
def check(self, progress_bar=None):
170
"""Check the versioned file for integrity."""
171
raise NotImplementedError(self.check)
173
def _check_lines_not_unicode(self, lines):
174
"""Check that lines being added to a versioned file are not unicode."""
176
if line.__class__ is not str:
177
raise errors.BzrBadParameterUnicode("lines")
179
def _check_lines_are_lines(self, lines):
180
"""Check that the lines really are full lines without inline EOL."""
182
if '\n' in line[:-1]:
183
raise errors.BzrBadParameterContainsNewline("lines")
185
def _check_write_ok(self):
186
"""Is the versioned file marked as 'finished' ? Raise if it is."""
188
raise errors.OutSideTransaction()
189
if self._access_mode != 'w':
190
raise errors.ReadOnlyObjectDirtiedError(self)
192
def enable_cache(self):
193
"""Tell this versioned file that it should cache any data it reads.
195
This is advisory, implementations do not have to support caching.
199
def clear_cache(self):
200
"""Remove any data cached in the versioned file object.
202
This only needs to be supported if caches are supported
206
def clone_text(self, new_version_id, old_version_id, parents):
207
"""Add an identical text to old_version_id as new_version_id.
209
Must raise RevisionNotPresent if the old version or any of the
210
parents are not present in file history.
212
Must raise RevisionAlreadyPresent if the new version is
213
already present in file history."""
214
new_version_id = osutils.safe_revision_id(new_version_id)
215
old_version_id = osutils.safe_revision_id(old_version_id)
216
self._check_write_ok()
217
return self._clone_text(new_version_id, old_version_id, parents)
219
def _clone_text(self, new_version_id, old_version_id, parents):
220
"""Helper function to do the _clone_text work."""
221
raise NotImplementedError(self.clone_text)
223
def create_empty(self, name, transport, mode=None):
224
"""Create a new versioned file of this exact type.
226
:param name: the file name
227
:param transport: the transport
228
:param mode: optional file mode.
230
raise NotImplementedError(self.create_empty)
232
def fix_parents(self, version_id, new_parents):
233
"""Fix the parents list for version.
235
This is done by appending a new version to the index
236
with identical data except for the parents list.
237
the parents list must be a superset of the current
240
version_id = osutils.safe_revision_id(version_id)
241
new_parents = [osutils.safe_revision_id(p) for p in new_parents]
242
self._check_write_ok()
243
return self._fix_parents(version_id, new_parents)
245
def _fix_parents(self, version_id, new_parents):
246
"""Helper for fix_parents."""
247
raise NotImplementedError(self.fix_parents)
249
def get_delta(self, version):
250
"""Get a delta for constructing version from some other version.
252
:return: (delta_parent, sha1, noeol, delta)
253
Where delta_parent is a version id or None to indicate no parent.
255
raise NotImplementedError(self.get_delta)
257
def get_deltas(self, version_ids):
258
"""Get multiple deltas at once for constructing versions.
260
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
261
Where delta_parent is a version id or None to indicate no parent, and
262
version_id is the version_id created by that delta.
265
for version_id in version_ids:
266
result[version_id] = self.get_delta(version_id)
269
def get_sha1(self, version_id):
270
"""Get the stored sha1 sum for the given revision.
272
:param name: The name of the version to lookup
274
raise NotImplementedError(self.get_sha1)
276
def get_suffixes(self):
277
"""Return the file suffixes associated with this versioned file."""
278
raise NotImplementedError(self.get_suffixes)
280
def get_text(self, version_id):
281
"""Return version contents as a text string.
283
Raises RevisionNotPresent if version is not present in
286
return ''.join(self.get_lines(version_id))
287
get_string = get_text
289
def get_texts(self, version_ids):
290
"""Return the texts of listed versions as a list of strings.
292
Raises RevisionNotPresent if version is not present in
295
return [''.join(self.get_lines(v)) for v in version_ids]
297
def get_lines(self, version_id):
298
"""Return version contents as a sequence of lines.
300
Raises RevisionNotPresent if version is not present in
303
raise NotImplementedError(self.get_lines)
305
def get_ancestry(self, version_ids):
306
"""Return a list of all ancestors of given version(s). This
307
will not include the null revision.
309
Must raise RevisionNotPresent if any of the given versions are
310
not present in file history."""
311
if isinstance(version_ids, basestring):
312
version_ids = [version_ids]
313
raise NotImplementedError(self.get_ancestry)
315
def get_ancestry_with_ghosts(self, version_ids):
316
"""Return a list of all ancestors of given version(s). This
317
will not include the null revision.
319
Must raise RevisionNotPresent if any of the given versions are
320
not present in file history.
322
Ghosts that are known about will be included in ancestry list,
323
but are not explicitly marked.
325
raise NotImplementedError(self.get_ancestry_with_ghosts)
327
def get_graph(self, version_ids=None):
328
"""Return a graph from the versioned file.
330
Ghosts are not listed or referenced in the graph.
331
:param version_ids: Versions to select.
332
None means retrieve all versions.
335
if version_ids is None:
336
for version in self.versions():
337
result[version] = self.get_parents(version)
339
pending = set(osutils.safe_revision_id(v) for v in version_ids)
341
version = pending.pop()
342
if version in result:
344
parents = self.get_parents(version)
345
for parent in parents:
349
result[version] = parents
352
def get_graph_with_ghosts(self):
353
"""Return a graph for the entire versioned file.
355
Ghosts are referenced in parents list but are not
358
raise NotImplementedError(self.get_graph_with_ghosts)
360
@deprecated_method(zero_eight)
361
def parent_names(self, version):
362
"""Return version names for parents of a version.
364
See get_parents for the current api.
366
return self.get_parents(version)
368
def get_parents(self, version_id):
369
"""Return version names for parents of a version.
371
Must raise RevisionNotPresent if version is not present in
374
raise NotImplementedError(self.get_parents)
376
def get_parents_with_ghosts(self, version_id):
377
"""Return version names for parents of version_id.
379
Will raise RevisionNotPresent if version_id is not present
382
Ghosts that are known about will be included in the parent list,
383
but are not explicitly marked.
385
raise NotImplementedError(self.get_parents_with_ghosts)
387
def annotate_iter(self, version_id):
388
"""Yield list of (version-id, line) pairs for the specified
391
Must raise RevisionNotPresent if any of the given versions are
392
not present in file history.
394
raise NotImplementedError(self.annotate_iter)
396
def annotate(self, version_id):
397
return list(self.annotate_iter(version_id))
399
def _apply_delta(self, lines, delta):
400
"""Apply delta to lines."""
403
for start, end, count, delta_lines in delta:
404
lines[offset+start:offset+end] = delta_lines
405
offset = offset + (start - end) + count
408
def join(self, other, pb=None, msg=None, version_ids=None,
409
ignore_missing=False):
410
"""Integrate versions from other into this versioned file.
412
If version_ids is None all versions from other should be
413
incorporated into this versioned file.
415
Must raise RevisionNotPresent if any of the specified versions
416
are not present in the other files history unless ignore_missing
417
is supplied when they are silently skipped.
419
self._check_write_ok()
420
return InterVersionedFile.get(other, self).join(
426
def iter_lines_added_or_present_in_versions(self, version_ids=None,
428
"""Iterate over the lines in the versioned file from version_ids.
430
This may return lines from other versions, and does not return the
431
specific version marker at this point. The api may be changed
432
during development to include the version that the versioned file
433
thinks is relevant, but given that such hints are just guesses,
434
its better not to have it if we don't need it.
436
If a progress bar is supplied, it may be used to indicate progress.
437
The caller is responsible for cleaning up progress bars (because this
440
NOTES: Lines are normalised: they will all have \n terminators.
441
Lines are returned in arbitrary order.
443
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
445
def transaction_finished(self):
446
"""The transaction that this file was opened in has finished.
448
This records self.finished = True and should cause all mutating
453
@deprecated_method(zero_eight)
454
def walk(self, version_ids=None):
455
"""Walk the versioned file as a weave-like structure, for
456
versions relative to version_ids. Yields sequence of (lineno,
457
insert, deletes, text) for each relevant line.
459
Must raise RevisionNotPresent if any of the specified versions
460
are not present in the file history.
462
:param version_ids: the version_ids to walk with respect to. If not
463
supplied the entire weave-like structure is walked.
465
walk is deprecated in favour of iter_lines_added_or_present_in_versions
467
raise NotImplementedError(self.walk)
469
@deprecated_method(zero_eight)
470
def iter_names(self):
471
"""Walk the names list."""
472
return iter(self.versions())
474
def plan_merge(self, ver_a, ver_b):
475
"""Return pseudo-annotation indicating how the two versions merge.
477
This is computed between versions a and b and their common
480
Weave lines present in none of them are skipped entirely.
483
killed-base Dead in base revision
484
killed-both Killed in each revision
487
unchanged Alive in both a and b (possibly created in both)
490
ghost-a Killed in a, unborn in b
491
ghost-b Killed in b, unborn in a
492
irrelevant Not in either revision
494
raise NotImplementedError(VersionedFile.plan_merge)
496
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
497
b_marker=TextMerge.B_MARKER):
498
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
501
class PlanWeaveMerge(TextMerge):
502
"""Weave merge that takes a plan as its input.
504
This exists so that VersionedFile.plan_merge is implementable.
505
Most callers will want to use WeaveMerge instead.
508
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
509
b_marker=TextMerge.B_MARKER):
510
TextMerge.__init__(self, a_marker, b_marker)
513
def _merge_struct(self):
518
def outstanding_struct():
519
if not lines_a and not lines_b:
521
elif ch_a and not ch_b:
524
elif ch_b and not ch_a:
526
elif lines_a == lines_b:
529
yield (lines_a, lines_b)
531
# We previously considered either 'unchanged' or 'killed-both' lines
532
# to be possible places to resynchronize. However, assuming agreement
533
# on killed-both lines may be too aggressive. -- mbp 20060324
534
for state, line in self.plan:
535
if state == 'unchanged':
536
# resync and flush queued conflicts changes if any
537
for struct in outstanding_struct():
543
if state == 'unchanged':
546
elif state == 'killed-a':
549
elif state == 'killed-b':
552
elif state == 'new-a':
555
elif state == 'new-b':
559
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
560
'killed-base', 'killed-both'), state
561
for struct in outstanding_struct():
565
class WeaveMerge(PlanWeaveMerge):
566
"""Weave merge that takes a VersionedFile and two versions as its input"""
568
def __init__(self, versionedfile, ver_a, ver_b,
569
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
570
plan = versionedfile.plan_merge(ver_a, ver_b)
571
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
574
class InterVersionedFile(InterObject):
575
"""This class represents operations taking place between two versionedfiles..
577
Its instances have methods like join, and contain
578
references to the source and target versionedfiles these operations can be
581
Often we will provide convenience methods on 'versionedfile' which carry out
582
operations with another versionedfile - they will always forward to
583
InterVersionedFile.get(other).method_name(parameters).
587
"""The available optimised InterVersionedFile types."""
589
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
590
"""Integrate versions from self.source into self.target.
592
If version_ids is None all versions from source should be
593
incorporated into this versioned file.
595
Must raise RevisionNotPresent if any of the specified versions
596
are not present in the other files history unless ignore_missing is
597
supplied when they are silently skipped.
600
# - if the target is empty, just add all the versions from
601
# source to target, otherwise:
602
# - make a temporary versioned file of type target
603
# - insert the source content into it one at a time
605
if not self.target.versions():
608
# Make a new target-format versioned file.
609
temp_source = self.target.create_empty("temp", MemoryTransport())
611
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
612
graph = self.source.get_graph(version_ids)
613
order = tsort.topo_sort(graph.items())
614
pb = ui.ui_factory.nested_progress_bar()
617
# TODO for incremental cross-format work:
618
# make a versioned file with the following content:
619
# all revisions we have been asked to join
620
# all their ancestors that are *not* in target already.
621
# the immediate parents of the above two sets, with
622
# empty parent lists - these versions are in target already
623
# and the incorrect version data will be ignored.
624
# TODO: for all ancestors that are present in target already,
625
# check them for consistent data, this requires moving sha1 from
627
# TODO: remove parent texts when they are not relevant any more for
628
# memory pressure reduction. RBC 20060313
629
# pb.update('Converting versioned data', 0, len(order))
630
# deltas = self.source.get_deltas(order)
631
for index, version in enumerate(order):
632
pb.update('Converting versioned data', index, len(order))
633
parent_text = target.add_lines(version,
634
self.source.get_parents(version),
635
self.source.get_lines(version),
636
parent_texts=parent_texts)
637
parent_texts[version] = parent_text
638
#delta_parent, sha1, noeol, delta = deltas[version]
639
#target.add_delta(version,
640
# self.source.get_parents(version),
645
#target.get_lines(version)
647
# this should hit the native code path for target
648
if target is not self.target:
649
return self.target.join(temp_source,
657
def _get_source_version_ids(self, version_ids, ignore_missing):
658
"""Determine the version ids to be used from self.source.
660
:param version_ids: The caller-supplied version ids to check. (None
661
for all). If None is in version_ids, it is stripped.
662
:param ignore_missing: if True, remove missing ids from the version
663
list. If False, raise RevisionNotPresent on
664
a missing version id.
665
:return: A set of version ids.
667
if version_ids is None:
668
# None cannot be in source.versions
669
return set(self.source.versions())
671
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
673
return set(self.source.versions()).intersection(set(version_ids))
675
new_version_ids = set()
676
for version in version_ids:
679
if not self.source.has_version(version):
680
raise errors.RevisionNotPresent(version, str(self.source))
682
new_version_ids.add(version)
683
return new_version_ids
686
class InterVersionedFileTestProviderAdapter(object):
687
"""A tool to generate a suite testing multiple inter versioned-file classes.
689
This is done by copying the test once for each InterVersionedFile provider
690
and injecting the transport_server, transport_readonly_server,
691
versionedfile_factory and versionedfile_factory_to classes into each copy.
692
Each copy is also given a new id() to make it easy to identify.
695
def __init__(self, transport_server, transport_readonly_server, formats):
696
self._transport_server = transport_server
697
self._transport_readonly_server = transport_readonly_server
698
self._formats = formats
700
def adapt(self, test):
701
result = unittest.TestSuite()
702
for (interversionedfile_class,
703
versionedfile_factory,
704
versionedfile_factory_to) in self._formats:
705
new_test = deepcopy(test)
706
new_test.transport_server = self._transport_server
707
new_test.transport_readonly_server = self._transport_readonly_server
708
new_test.interversionedfile_class = interversionedfile_class
709
new_test.versionedfile_factory = versionedfile_factory
710
new_test.versionedfile_factory_to = versionedfile_factory_to
711
def make_new_test_id():
712
new_id = "%s(%s)" % (new_test.id(), interversionedfile_class.__name__)
713
return lambda: new_id
714
new_test.id = make_new_test_id()
715
result.addTest(new_test)
719
def default_test_list():
720
"""Generate the default list of interversionedfile permutations to test."""
721
from bzrlib.weave import WeaveFile
722
from bzrlib.knit import KnitVersionedFile
724
# test the fallback InterVersionedFile from annotated knits to weave
725
result.append((InterVersionedFile,
728
for optimiser in InterVersionedFile._optimisers:
729
result.append((optimiser,
730
optimiser._matching_file_from_factory,
731
optimiser._matching_file_to_factory
733
# if there are specific combinations we want to use, we can add them
added
removed
bzrlib/versionedfile.py
