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(), """
32
from bzrlib.transport.memory import MemoryTransport
35
from bzrlib.inter import InterObject
36
from bzrlib.textmerge import TextMerge
37
from bzrlib.symbol_versioning import (deprecated_function,
43
class VersionedFile(object):
44
"""Versioned text file storage.
46
A versioned file manages versions of line-based text files,
47
keeping track of the originating version for each line.
49
To clients the "lines" of the file are represented as a list of
50
strings. These strings will typically have terminal newline
51
characters, but this is not required. In particular files commonly
52
do not have a newline at the end of the file.
54
Texts are identified by a version-id string.
57
def __init__(self, access_mode):
59
self._access_mode = access_mode
62
def check_not_reserved_id(version_id):
63
revision.check_not_reserved_id(version_id)
65
def copy_to(self, name, transport):
66
"""Copy this versioned file to name on transport."""
67
raise NotImplementedError(self.copy_to)
69
@deprecated_method(zero_eight)
71
"""Return a list of all the versions in this versioned file.
73
Please use versionedfile.versions() now.
75
return self.versions()
78
"""Return a unsorted list of versions."""
79
raise NotImplementedError(self.versions)
81
def has_ghost(self, version_id):
82
"""Returns whether version is present as a ghost."""
83
raise NotImplementedError(self.has_ghost)
85
def has_version(self, version_id):
86
"""Returns whether version is present."""
87
raise NotImplementedError(self.has_version)
89
def add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
90
"""Add a text to the versioned file via a pregenerated delta.
92
:param version_id: The version id being added.
93
:param parents: The parents of the version_id.
94
:param delta_parent: The parent this delta was created against.
95
:param sha1: The sha1 of the full text.
96
:param delta: The delta instructions. See get_delta for details.
98
version_id = osutils.safe_revision_id(version_id)
99
parents = [osutils.safe_revision_id(v) for v in parents]
100
self._check_write_ok()
101
if self.has_version(version_id):
102
raise errors.RevisionAlreadyPresent(version_id, self)
103
return self._add_delta(version_id, parents, delta_parent, sha1, noeol, delta)
105
def _add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
106
"""Class specific routine to add a delta.
108
This generic version simply applies the delta to the delta_parent and
111
# strip annotation from delta
113
for start, stop, delta_len, delta_lines in delta:
114
new_delta.append((start, stop, delta_len, [text for origin, text in delta_lines]))
115
if delta_parent is not None:
116
parent_full = self.get_lines(delta_parent)
119
new_full = self._apply_delta(parent_full, new_delta)
120
# its impossible to have noeol on an empty file
121
if noeol and new_full[-1][-1] == '\n':
122
new_full[-1] = new_full[-1][:-1]
123
self.add_lines(version_id, parents, new_full)
125
def add_lines(self, version_id, parents, lines, parent_texts=None):
126
"""Add a single text on top of the versioned file.
128
Must raise RevisionAlreadyPresent if the new version is
129
already present in file history.
131
Must raise RevisionNotPresent if any of the given parents are
132
not present in file history.
133
:param parent_texts: An optional dictionary containing the opaque
134
representations of some or all of the parents of
135
version_id to allow delta optimisations.
136
VERY IMPORTANT: the texts must be those returned
137
by add_lines or data corruption can be caused.
138
:return: An opaque representation of the inserted version which can be
139
provided back to future add_lines calls in the parent_texts
142
version_id = osutils.safe_revision_id(version_id)
143
parents = [osutils.safe_revision_id(v) for v in parents]
144
self._check_write_ok()
145
return self._add_lines(version_id, parents, lines, parent_texts)
147
def _add_lines(self, version_id, parents, lines, parent_texts):
148
"""Helper to do the class specific add_lines."""
149
raise NotImplementedError(self.add_lines)
151
def add_lines_with_ghosts(self, version_id, parents, lines,
153
"""Add lines to the versioned file, allowing ghosts to be present.
155
This takes the same parameters as add_lines.
157
version_id = osutils.safe_revision_id(version_id)
158
parents = [osutils.safe_revision_id(v) for v in parents]
159
self._check_write_ok()
160
return self._add_lines_with_ghosts(version_id, parents, lines,
163
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
164
"""Helper to do class specific add_lines_with_ghosts."""
165
raise NotImplementedError(self.add_lines_with_ghosts)
167
def check(self, progress_bar=None):
168
"""Check the versioned file for integrity."""
169
raise NotImplementedError(self.check)
171
def _check_lines_not_unicode(self, lines):
172
"""Check that lines being added to a versioned file are not unicode."""
174
if line.__class__ is not str:
175
raise errors.BzrBadParameterUnicode("lines")
177
def _check_lines_are_lines(self, lines):
178
"""Check that the lines really are full lines without inline EOL."""
180
if '\n' in line[:-1]:
181
raise errors.BzrBadParameterContainsNewline("lines")
183
def _check_write_ok(self):
184
"""Is the versioned file marked as 'finished' ? Raise if it is."""
186
raise errors.OutSideTransaction()
187
if self._access_mode != 'w':
188
raise errors.ReadOnlyObjectDirtiedError(self)
190
def enable_cache(self):
191
"""Tell this versioned file that it should cache any data it reads.
193
This is advisory, implementations do not have to support caching.
197
def clear_cache(self):
198
"""Remove any data cached in the versioned file object.
200
This only needs to be supported if caches are supported
204
def clone_text(self, new_version_id, old_version_id, parents):
205
"""Add an identical text to old_version_id as new_version_id.
207
Must raise RevisionNotPresent if the old version or any of the
208
parents are not present in file history.
210
Must raise RevisionAlreadyPresent if the new version is
211
already present in file history."""
212
new_version_id = osutils.safe_revision_id(new_version_id)
213
old_version_id = osutils.safe_revision_id(old_version_id)
214
self._check_write_ok()
215
return self._clone_text(new_version_id, old_version_id, parents)
217
def _clone_text(self, new_version_id, old_version_id, parents):
218
"""Helper function to do the _clone_text work."""
219
raise NotImplementedError(self.clone_text)
221
def create_empty(self, name, transport, mode=None):
222
"""Create a new versioned file of this exact type.
224
:param name: the file name
225
:param transport: the transport
226
:param mode: optional file mode.
228
raise NotImplementedError(self.create_empty)
230
def fix_parents(self, version_id, new_parents):
231
"""Fix the parents list for version.
233
This is done by appending a new version to the index
234
with identical data except for the parents list.
235
the parents list must be a superset of the current
238
version_id = osutils.safe_revision_id(version_id)
239
new_parents = [osutils.safe_revision_id(p) for p in new_parents]
240
self._check_write_ok()
241
return self._fix_parents(version_id, new_parents)
243
def _fix_parents(self, version_id, new_parents):
244
"""Helper for fix_parents."""
245
raise NotImplementedError(self.fix_parents)
247
def get_delta(self, version):
248
"""Get a delta for constructing version from some other version.
250
:return: (delta_parent, sha1, noeol, delta)
251
Where delta_parent is a version id or None to indicate no parent.
253
raise NotImplementedError(self.get_delta)
255
def get_deltas(self, version_ids):
256
"""Get multiple deltas at once for constructing versions.
258
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
259
Where delta_parent is a version id or None to indicate no parent, and
260
version_id is the version_id created by that delta.
263
for version_id in version_ids:
264
result[version_id] = self.get_delta(version_id)
267
def get_sha1(self, version_id):
268
"""Get the stored sha1 sum for the given revision.
270
:param name: The name of the version to lookup
272
raise NotImplementedError(self.get_sha1)
274
def get_suffixes(self):
275
"""Return the file suffixes associated with this versioned file."""
276
raise NotImplementedError(self.get_suffixes)
278
def get_text(self, version_id):
279
"""Return version contents as a text string.
281
Raises RevisionNotPresent if version is not present in
284
return ''.join(self.get_lines(version_id))
285
get_string = get_text
287
def get_texts(self, version_ids):
288
"""Return the texts of listed versions as a list of strings.
290
Raises RevisionNotPresent if version is not present in
293
return [''.join(self.get_lines(v)) for v in version_ids]
295
def get_lines(self, version_id):
296
"""Return version contents as a sequence of lines.
298
Raises RevisionNotPresent if version is not present in
301
raise NotImplementedError(self.get_lines)
303
def get_ancestry(self, version_ids, topo_sorted=True):
304
"""Return a list of all ancestors of given version(s). This
305
will not include the null revision.
307
This list will not be topologically sorted if topo_sorted=False is
310
Must raise RevisionNotPresent if any of the given versions are
311
not present in file history."""
312
if isinstance(version_ids, basestring):
313
version_ids = [version_ids]
314
raise NotImplementedError(self.get_ancestry)
316
def get_ancestry_with_ghosts(self, version_ids):
317
"""Return a list of all ancestors of given version(s). This
318
will not include the null revision.
320
Must raise RevisionNotPresent if any of the given versions are
321
not present in file history.
323
Ghosts that are known about will be included in ancestry list,
324
but are not explicitly marked.
326
raise NotImplementedError(self.get_ancestry_with_ghosts)
328
def get_graph(self, version_ids=None):
329
"""Return a graph from the versioned file.
331
Ghosts are not listed or referenced in the graph.
332
:param version_ids: Versions to select.
333
None means retrieve all versions.
335
if version_ids is None:
336
return dict(self.iter_parents(self.versions()))
338
pending = set(osutils.safe_revision_id(v) for v in version_ids)
340
this_iteration = pending
342
for version, parents in self.iter_parents(this_iteration):
343
result[version] = parents
344
pending.update(parents)
345
pending.difference_update(result)
348
def get_graph_with_ghosts(self):
349
"""Return a graph for the entire versioned file.
351
Ghosts are referenced in parents list but are not
354
raise NotImplementedError(self.get_graph_with_ghosts)
356
@deprecated_method(zero_eight)
357
def parent_names(self, version):
358
"""Return version names for parents of a version.
360
See get_parents for the current api.
362
return self.get_parents(version)
364
def get_parents(self, version_id):
365
"""Return version names for parents of a version.
367
Must raise RevisionNotPresent if version is not present in
370
raise NotImplementedError(self.get_parents)
372
def get_parents_with_ghosts(self, version_id):
373
"""Return version names for parents of version_id.
375
Will raise RevisionNotPresent if version_id is not present
378
Ghosts that are known about will be included in the parent list,
379
but are not explicitly marked.
381
raise NotImplementedError(self.get_parents_with_ghosts)
383
def annotate_iter(self, version_id):
384
"""Yield list of (version-id, line) pairs for the specified
387
Must raise RevisionNotPresent if any of the given versions are
388
not present in file history.
390
raise NotImplementedError(self.annotate_iter)
392
def annotate(self, version_id):
393
return list(self.annotate_iter(version_id))
395
def _apply_delta(self, lines, delta):
396
"""Apply delta to lines."""
399
for start, end, count, delta_lines in delta:
400
lines[offset+start:offset+end] = delta_lines
401
offset = offset + (start - end) + count
404
def join(self, other, pb=None, msg=None, version_ids=None,
405
ignore_missing=False):
406
"""Integrate versions from other into this versioned file.
408
If version_ids is None all versions from other should be
409
incorporated into this versioned file.
411
Must raise RevisionNotPresent if any of the specified versions
412
are not present in the other files history unless ignore_missing
413
is supplied when they are silently skipped.
415
self._check_write_ok()
416
return InterVersionedFile.get(other, self).join(
422
def iter_lines_added_or_present_in_versions(self, version_ids=None,
424
"""Iterate over the lines in the versioned file from version_ids.
426
This may return lines from other versions, and does not return the
427
specific version marker at this point. The api may be changed
428
during development to include the version that the versioned file
429
thinks is relevant, but given that such hints are just guesses,
430
its better not to have it if we don't need it.
432
If a progress bar is supplied, it may be used to indicate progress.
433
The caller is responsible for cleaning up progress bars (because this
436
NOTES: Lines are normalised: they will all have \n terminators.
437
Lines are returned in arbitrary order.
439
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
441
def iter_parents(self, version_ids):
442
"""Iterate through the parents for many version ids.
444
:param version_ids: An iterable yielding version_ids.
445
:return: An iterator that yields (version_id, parents). Requested
446
version_ids not present in the versioned file are simply skipped.
447
The order is undefined, allowing for different optimisations in
448
the underlying implementation.
450
for version_id in version_ids:
452
yield version_id, tuple(self.get_parents(version_id))
453
except errors.RevisionNotPresent:
456
def transaction_finished(self):
457
"""The transaction that this file was opened in has finished.
459
This records self.finished = True and should cause all mutating
464
@deprecated_method(zero_eight)
465
def walk(self, version_ids=None):
466
"""Walk the versioned file as a weave-like structure, for
467
versions relative to version_ids. Yields sequence of (lineno,
468
insert, deletes, text) for each relevant line.
470
Must raise RevisionNotPresent if any of the specified versions
471
are not present in the file history.
473
:param version_ids: the version_ids to walk with respect to. If not
474
supplied the entire weave-like structure is walked.
476
walk is deprecated in favour of iter_lines_added_or_present_in_versions
478
raise NotImplementedError(self.walk)
480
@deprecated_method(zero_eight)
481
def iter_names(self):
482
"""Walk the names list."""
483
return iter(self.versions())
485
def plan_merge(self, ver_a, ver_b):
486
"""Return pseudo-annotation indicating how the two versions merge.
488
This is computed between versions a and b and their common
491
Weave lines present in none of them are skipped entirely.
494
killed-base Dead in base revision
495
killed-both Killed in each revision
498
unchanged Alive in both a and b (possibly created in both)
501
ghost-a Killed in a, unborn in b
502
ghost-b Killed in b, unborn in a
503
irrelevant Not in either revision
505
raise NotImplementedError(VersionedFile.plan_merge)
507
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
508
b_marker=TextMerge.B_MARKER):
509
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
512
class PlanWeaveMerge(TextMerge):
513
"""Weave merge that takes a plan as its input.
515
This exists so that VersionedFile.plan_merge is implementable.
516
Most callers will want to use WeaveMerge instead.
519
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
520
b_marker=TextMerge.B_MARKER):
521
TextMerge.__init__(self, a_marker, b_marker)
524
def _merge_struct(self):
529
def outstanding_struct():
530
if not lines_a and not lines_b:
532
elif ch_a and not ch_b:
535
elif ch_b and not ch_a:
537
elif lines_a == lines_b:
540
yield (lines_a, lines_b)
542
# We previously considered either 'unchanged' or 'killed-both' lines
543
# to be possible places to resynchronize. However, assuming agreement
544
# on killed-both lines may be too aggressive. -- mbp 20060324
545
for state, line in self.plan:
546
if state == 'unchanged':
547
# resync and flush queued conflicts changes if any
548
for struct in outstanding_struct():
554
if state == 'unchanged':
557
elif state == 'killed-a':
560
elif state == 'killed-b':
563
elif state == 'new-a':
566
elif state == 'new-b':
570
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
571
'killed-base', 'killed-both'), state
572
for struct in outstanding_struct():
576
class WeaveMerge(PlanWeaveMerge):
577
"""Weave merge that takes a VersionedFile and two versions as its input"""
579
def __init__(self, versionedfile, ver_a, ver_b,
580
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
581
plan = versionedfile.plan_merge(ver_a, ver_b)
582
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
585
class InterVersionedFile(InterObject):
586
"""This class represents operations taking place between two versionedfiles..
588
Its instances have methods like join, and contain
589
references to the source and target versionedfiles these operations can be
592
Often we will provide convenience methods on 'versionedfile' which carry out
593
operations with another versionedfile - they will always forward to
594
InterVersionedFile.get(other).method_name(parameters).
598
"""The available optimised InterVersionedFile types."""
600
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
601
"""Integrate versions from self.source into self.target.
603
If version_ids is None all versions from source should be
604
incorporated into this versioned file.
606
Must raise RevisionNotPresent if any of the specified versions
607
are not present in the other files history unless ignore_missing is
608
supplied when they are silently skipped.
611
# - if the target is empty, just add all the versions from
612
# source to target, otherwise:
613
# - make a temporary versioned file of type target
614
# - insert the source content into it one at a time
616
if not self.target.versions():
619
# Make a new target-format versioned file.
620
temp_source = self.target.create_empty("temp", MemoryTransport())
622
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
623
graph = self.source.get_graph(version_ids)
624
order = tsort.topo_sort(graph.items())
625
pb = ui.ui_factory.nested_progress_bar()
628
# TODO for incremental cross-format work:
629
# make a versioned file with the following content:
630
# all revisions we have been asked to join
631
# all their ancestors that are *not* in target already.
632
# the immediate parents of the above two sets, with
633
# empty parent lists - these versions are in target already
634
# and the incorrect version data will be ignored.
635
# TODO: for all ancestors that are present in target already,
636
# check them for consistent data, this requires moving sha1 from
638
# TODO: remove parent texts when they are not relevant any more for
639
# memory pressure reduction. RBC 20060313
640
# pb.update('Converting versioned data', 0, len(order))
641
# deltas = self.source.get_deltas(order)
642
for index, version in enumerate(order):
643
pb.update('Converting versioned data', index, len(order))
644
parent_text = target.add_lines(version,
645
self.source.get_parents(version),
646
self.source.get_lines(version),
647
parent_texts=parent_texts)
648
parent_texts[version] = parent_text
649
#delta_parent, sha1, noeol, delta = deltas[version]
650
#target.add_delta(version,
651
# self.source.get_parents(version),
656
#target.get_lines(version)
658
# this should hit the native code path for target
659
if target is not self.target:
660
return self.target.join(temp_source,
668
def _get_source_version_ids(self, version_ids, ignore_missing):
669
"""Determine the version ids to be used from self.source.
671
:param version_ids: The caller-supplied version ids to check. (None
672
for all). If None is in version_ids, it is stripped.
673
:param ignore_missing: if True, remove missing ids from the version
674
list. If False, raise RevisionNotPresent on
675
a missing version id.
676
:return: A set of version ids.
678
if version_ids is None:
679
# None cannot be in source.versions
680
return set(self.source.versions())
682
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
684
return set(self.source.versions()).intersection(set(version_ids))
686
new_version_ids = set()
687
for version in version_ids:
690
if not self.source.has_version(version):
691
raise errors.RevisionNotPresent(version, str(self.source))
693
new_version_ids.add(version)
694
return new_version_ids
added
removed
bzrlib/versionedfile.py
