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_mpdiffs(self, version_ids):
271
knit_versions = set()
272
for version_id in version_ids:
273
knit_versions.add(version_id)
274
knit_versions.update(self.get_parents(version_id))
275
lines = dict(zip(knit_versions, self._get_line_list(knit_versions)))
277
for version_id in version_ids:
278
target = lines[version_id]
279
parents = [lines[p] for p in self.get_parents(version_id)]
281
left_parent_blocks = self._extract_blocks(version_id,
284
left_parent_blocks = None
285
diffs.append(multiparent.MultiParent.from_lines(target, parents,
289
def _extract_blocks(self, version_id, source, target):
292
def add_mpdiffs(self, records):
294
for version, parents, expected_sha1, mpdiff in records:
295
mpvf = multiparent.MultiMemoryVersionedFile()
296
needed_parents = [p for p in parents if not mpvf.has_version(p)]
297
parent_lines = self._get_line_list(needed_parents)
298
for parent_id, lines in zip(needed_parents, parent_lines):
299
mpvf.add_version(lines, parent_id, [])
300
mpvf.add_diff(mpdiff, version, parents)
301
lines = mpvf.get_line_list([version])[0]
302
version_text = self.add_lines(version, parents, lines, vf_parents)
303
vf_parents[version] = version_text
304
assert expected_sha1 == self.get_sha1(version)
306
def get_sha1(self, version_id):
307
"""Get the stored sha1 sum for the given revision.
309
:param name: The name of the version to lookup
311
raise NotImplementedError(self.get_sha1)
313
def get_suffixes(self):
314
"""Return the file suffixes associated with this versioned file."""
315
raise NotImplementedError(self.get_suffixes)
317
def get_text(self, version_id):
318
"""Return version contents as a text string.
320
Raises RevisionNotPresent if version is not present in
323
return ''.join(self.get_lines(version_id))
324
get_string = get_text
326
def get_texts(self, version_ids):
327
"""Return the texts of listed versions as a list of strings.
329
Raises RevisionNotPresent if version is not present in
332
return [''.join(self.get_lines(v)) for v in version_ids]
334
def get_lines(self, version_id):
335
"""Return version contents as a sequence of lines.
337
Raises RevisionNotPresent if version is not present in
340
raise NotImplementedError(self.get_lines)
342
def _get_line_list(self, version_ids):
343
return [t.splitlines(True) for t in self.get_texts(version_ids)]
345
def get_ancestry(self, version_ids):
346
"""Return a list of all ancestors of given version(s). This
347
will not include the null revision.
349
Must raise RevisionNotPresent if any of the given versions are
350
not present in file history."""
351
if isinstance(version_ids, basestring):
352
version_ids = [version_ids]
353
raise NotImplementedError(self.get_ancestry)
355
def get_ancestry_with_ghosts(self, version_ids):
356
"""Return a list of all ancestors of given version(s). This
357
will not include the null revision.
359
Must raise RevisionNotPresent if any of the given versions are
360
not present in file history.
362
Ghosts that are known about will be included in ancestry list,
363
but are not explicitly marked.
365
raise NotImplementedError(self.get_ancestry_with_ghosts)
367
def get_graph(self, version_ids=None):
368
"""Return a graph from the versioned file.
370
Ghosts are not listed or referenced in the graph.
371
:param version_ids: Versions to select.
372
None means retrieve all versions.
375
if version_ids is None:
376
for version in self.versions():
377
result[version] = self.get_parents(version)
379
pending = set(osutils.safe_revision_id(v) for v in version_ids)
381
version = pending.pop()
382
if version in result:
384
parents = self.get_parents(version)
385
for parent in parents:
389
result[version] = parents
392
def get_graph_with_ghosts(self):
393
"""Return a graph for the entire versioned file.
395
Ghosts are referenced in parents list but are not
398
raise NotImplementedError(self.get_graph_with_ghosts)
400
@deprecated_method(zero_eight)
401
def parent_names(self, version):
402
"""Return version names for parents of a version.
404
See get_parents for the current api.
406
return self.get_parents(version)
408
def get_parents(self, version_id):
409
"""Return version names for parents of a version.
411
Must raise RevisionNotPresent if version is not present in
414
raise NotImplementedError(self.get_parents)
416
def get_parents_with_ghosts(self, version_id):
417
"""Return version names for parents of version_id.
419
Will raise RevisionNotPresent if version_id is not present
422
Ghosts that are known about will be included in the parent list,
423
but are not explicitly marked.
425
raise NotImplementedError(self.get_parents_with_ghosts)
427
def annotate_iter(self, version_id):
428
"""Yield list of (version-id, line) pairs for the specified
431
Must raise RevisionNotPresent if any of the given versions are
432
not present in file history.
434
raise NotImplementedError(self.annotate_iter)
436
def annotate(self, version_id):
437
return list(self.annotate_iter(version_id))
439
def _apply_delta(self, lines, delta):
440
"""Apply delta to lines."""
443
for start, end, count, delta_lines in delta:
444
lines[offset+start:offset+end] = delta_lines
445
offset = offset + (start - end) + count
448
def join(self, other, pb=None, msg=None, version_ids=None,
449
ignore_missing=False):
450
"""Integrate versions from other into this versioned file.
452
If version_ids is None all versions from other should be
453
incorporated into this versioned file.
455
Must raise RevisionNotPresent if any of the specified versions
456
are not present in the other files history unless ignore_missing
457
is supplied when they are silently skipped.
459
self._check_write_ok()
460
return InterVersionedFile.get(other, self).join(
466
def iter_lines_added_or_present_in_versions(self, version_ids=None,
468
"""Iterate over the lines in the versioned file from version_ids.
470
This may return lines from other versions, and does not return the
471
specific version marker at this point. The api may be changed
472
during development to include the version that the versioned file
473
thinks is relevant, but given that such hints are just guesses,
474
its better not to have it if we don't need it.
476
If a progress bar is supplied, it may be used to indicate progress.
477
The caller is responsible for cleaning up progress bars (because this
480
NOTES: Lines are normalised: they will all have \n terminators.
481
Lines are returned in arbitrary order.
483
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
485
def transaction_finished(self):
486
"""The transaction that this file was opened in has finished.
488
This records self.finished = True and should cause all mutating
493
@deprecated_method(zero_eight)
494
def walk(self, version_ids=None):
495
"""Walk the versioned file as a weave-like structure, for
496
versions relative to version_ids. Yields sequence of (lineno,
497
insert, deletes, text) for each relevant line.
499
Must raise RevisionNotPresent if any of the specified versions
500
are not present in the file history.
502
:param version_ids: the version_ids to walk with respect to. If not
503
supplied the entire weave-like structure is walked.
505
walk is deprecated in favour of iter_lines_added_or_present_in_versions
507
raise NotImplementedError(self.walk)
509
@deprecated_method(zero_eight)
510
def iter_names(self):
511
"""Walk the names list."""
512
return iter(self.versions())
514
def plan_merge(self, ver_a, ver_b):
515
"""Return pseudo-annotation indicating how the two versions merge.
517
This is computed between versions a and b and their common
520
Weave lines present in none of them are skipped entirely.
523
killed-base Dead in base revision
524
killed-both Killed in each revision
527
unchanged Alive in both a and b (possibly created in both)
530
ghost-a Killed in a, unborn in b
531
ghost-b Killed in b, unborn in a
532
irrelevant Not in either revision
534
raise NotImplementedError(VersionedFile.plan_merge)
536
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
537
b_marker=TextMerge.B_MARKER):
538
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
541
class PlanWeaveMerge(TextMerge):
542
"""Weave merge that takes a plan as its input.
544
This exists so that VersionedFile.plan_merge is implementable.
545
Most callers will want to use WeaveMerge instead.
548
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
549
b_marker=TextMerge.B_MARKER):
550
TextMerge.__init__(self, a_marker, b_marker)
553
def _merge_struct(self):
558
def outstanding_struct():
559
if not lines_a and not lines_b:
561
elif ch_a and not ch_b:
564
elif ch_b and not ch_a:
566
elif lines_a == lines_b:
569
yield (lines_a, lines_b)
571
# We previously considered either 'unchanged' or 'killed-both' lines
572
# to be possible places to resynchronize. However, assuming agreement
573
# on killed-both lines may be too aggressive. -- mbp 20060324
574
for state, line in self.plan:
575
if state == 'unchanged':
576
# resync and flush queued conflicts changes if any
577
for struct in outstanding_struct():
583
if state == 'unchanged':
586
elif state == 'killed-a':
589
elif state == 'killed-b':
592
elif state == 'new-a':
595
elif state == 'new-b':
599
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
600
'killed-base', 'killed-both'), state
601
for struct in outstanding_struct():
605
class WeaveMerge(PlanWeaveMerge):
606
"""Weave merge that takes a VersionedFile and two versions as its input"""
608
def __init__(self, versionedfile, ver_a, ver_b,
609
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
610
plan = versionedfile.plan_merge(ver_a, ver_b)
611
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
614
class InterVersionedFile(InterObject):
615
"""This class represents operations taking place between two versionedfiles..
617
Its instances have methods like join, and contain
618
references to the source and target versionedfiles these operations can be
621
Often we will provide convenience methods on 'versionedfile' which carry out
622
operations with another versionedfile - they will always forward to
623
InterVersionedFile.get(other).method_name(parameters).
627
"""The available optimised InterVersionedFile types."""
629
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
630
"""Integrate versions from self.source into self.target.
632
If version_ids is None all versions from source should be
633
incorporated into this versioned file.
635
Must raise RevisionNotPresent if any of the specified versions
636
are not present in the other files history unless ignore_missing is
637
supplied when they are silently skipped.
640
# - if the target is empty, just add all the versions from
641
# source to target, otherwise:
642
# - make a temporary versioned file of type target
643
# - insert the source content into it one at a time
645
if not self.target.versions():
648
# Make a new target-format versioned file.
649
temp_source = self.target.create_empty("temp", MemoryTransport())
651
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
652
graph = self.source.get_graph(version_ids)
653
order = tsort.topo_sort(graph.items())
654
pb = ui.ui_factory.nested_progress_bar()
657
# TODO for incremental cross-format work:
658
# make a versioned file with the following content:
659
# all revisions we have been asked to join
660
# all their ancestors that are *not* in target already.
661
# the immediate parents of the above two sets, with
662
# empty parent lists - these versions are in target already
663
# and the incorrect version data will be ignored.
664
# TODO: for all ancestors that are present in target already,
665
# check them for consistent data, this requires moving sha1 from
667
# TODO: remove parent texts when they are not relevant any more for
668
# memory pressure reduction. RBC 20060313
669
# pb.update('Converting versioned data', 0, len(order))
670
# deltas = self.source.get_deltas(order)
671
for index, version in enumerate(order):
672
pb.update('Converting versioned data', index, len(order))
673
parent_text = target.add_lines(version,
674
self.source.get_parents(version),
675
self.source.get_lines(version),
676
parent_texts=parent_texts)
677
parent_texts[version] = parent_text
678
#delta_parent, sha1, noeol, delta = deltas[version]
679
#target.add_delta(version,
680
# self.source.get_parents(version),
685
#target.get_lines(version)
687
# this should hit the native code path for target
688
if target is not self.target:
689
return self.target.join(temp_source,
697
def _get_source_version_ids(self, version_ids, ignore_missing):
698
"""Determine the version ids to be used from self.source.
700
:param version_ids: The caller-supplied version ids to check. (None
701
for all). If None is in version_ids, it is stripped.
702
:param ignore_missing: if True, remove missing ids from the version
703
list. If False, raise RevisionNotPresent on
704
a missing version id.
705
:return: A set of version ids.
707
if version_ids is None:
708
# None cannot be in source.versions
709
return set(self.source.versions())
711
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
713
return set(self.source.versions()).intersection(set(version_ids))
715
new_version_ids = set()
716
for version in version_ids:
719
if not self.source.has_version(version):
720
raise errors.RevisionNotPresent(version, str(self.source))
722
new_version_ids.add(version)
723
return new_version_ids
726
class InterVersionedFileTestProviderAdapter(object):
727
"""A tool to generate a suite testing multiple inter versioned-file classes.
729
This is done by copying the test once for each InterVersionedFile provider
730
and injecting the transport_server, transport_readonly_server,
731
versionedfile_factory and versionedfile_factory_to classes into each copy.
732
Each copy is also given a new id() to make it easy to identify.
735
def __init__(self, transport_server, transport_readonly_server, formats):
736
self._transport_server = transport_server
737
self._transport_readonly_server = transport_readonly_server
738
self._formats = formats
740
def adapt(self, test):
741
result = unittest.TestSuite()
742
for (interversionedfile_class,
743
versionedfile_factory,
744
versionedfile_factory_to) in self._formats:
745
new_test = deepcopy(test)
746
new_test.transport_server = self._transport_server
747
new_test.transport_readonly_server = self._transport_readonly_server
748
new_test.interversionedfile_class = interversionedfile_class
749
new_test.versionedfile_factory = versionedfile_factory
750
new_test.versionedfile_factory_to = versionedfile_factory_to
751
def make_new_test_id():
752
new_id = "%s(%s)" % (new_test.id(), interversionedfile_class.__name__)
753
return lambda: new_id
754
new_test.id = make_new_test_id()
755
result.addTest(new_test)
759
def default_test_list():
760
"""Generate the default list of interversionedfile permutations to test."""
761
from bzrlib.weave import WeaveFile
762
from bzrlib.knit import KnitVersionedFile
764
# test the fallback InterVersionedFile from annotated knits to weave
765
result.append((InterVersionedFile,
768
for optimiser in InterVersionedFile._optimisers:
769
result.append((optimiser,
770
optimiser._matching_file_from_factory,
771
optimiser._matching_file_to_factory
773
# if there are specific combinations we want to use, we can add them
added
removed
bzrlib/versionedfile.py
