1
# Copyright (C) 2005 by 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
# Remaing to do is to figure out if get_graph should return a simple
21
# map, or a graph object of some kind.
24
"""Versioned text file storage api."""
27
from copy import deepcopy
28
from unittest import TestSuite
31
import bzrlib.errors as errors
32
from bzrlib.inter import InterObject
33
from bzrlib.symbol_versioning import *
34
from bzrlib.textmerge import TextMerge
35
from bzrlib.transport.memory import MemoryTransport
36
from bzrlib.tsort import topo_sort
40
class VersionedFile(object):
41
"""Versioned text file storage.
43
A versioned file manages versions of line-based text files,
44
keeping track of the originating version for each line.
46
To clients the "lines" of the file are represented as a list of
47
strings. These strings will typically have terminal newline
48
characters, but this is not required. In particular files commonly
49
do not have a newline at the end of the file.
51
Texts are identified by a version-id string.
54
def __init__(self, access_mode):
56
self._access_mode = access_mode
58
def copy_to(self, name, transport):
59
"""Copy this versioned file to name on transport."""
60
raise NotImplementedError(self.copy_to)
62
@deprecated_method(zero_eight)
64
"""Return a list of all the versions in this versioned file.
66
Please use versionedfile.versions() now.
68
return self.versions()
71
"""Return a unsorted list of versions."""
72
raise NotImplementedError(self.versions)
74
def has_ghost(self, version_id):
75
"""Returns whether version is present as a ghost."""
76
raise NotImplementedError(self.has_ghost)
78
def has_version(self, version_id):
79
"""Returns whether version is present."""
80
raise NotImplementedError(self.has_version)
82
def add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
83
"""Add a text to the versioned file via a pregenerated delta.
85
:param version_id: The version id being added.
86
:param parents: The parents of the version_id.
87
:param delta_parent: The parent this delta was created against.
88
:param sha1: The sha1 of the full text.
89
:param delta: The delta instructions. See get_delta for details.
91
self._check_write_ok()
92
if self.has_version(version_id):
93
raise errors.RevisionAlreadyPresent(version_id, self)
94
return self._add_delta(version_id, parents, delta_parent, sha1, noeol, delta)
96
def _add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
97
"""Class specific routine to add a delta.
99
This generic version simply applies the delta to the delta_parent and
102
# strip annotation from delta
104
for start, stop, delta_len, delta_lines in delta:
105
new_delta.append((start, stop, delta_len, [text for origin, text in delta_lines]))
106
if delta_parent is not None:
107
parent_full = self.get_lines(delta_parent)
110
new_full = self._apply_delta(parent_full, new_delta)
111
# its impossible to have noeol on an empty file
112
if noeol and new_full[-1][-1] == '\n':
113
new_full[-1] = new_full[-1][:-1]
114
self.add_lines(version_id, parents, new_full)
116
def add_lines(self, version_id, parents, lines, parent_texts=None):
117
"""Add a single text on top of the versioned file.
119
Must raise RevisionAlreadyPresent if the new version is
120
already present in file history.
122
Must raise RevisionNotPresent if any of the given parents are
123
not present in file history.
124
:param parent_texts: An optional dictionary containing the opaque
125
representations of some or all of the parents of
126
version_id to allow delta optimisations.
127
VERY IMPORTANT: the texts must be those returned
128
by add_lines or data corruption can be caused.
129
:return: An opaque representation of the inserted version which can be
130
provided back to future add_lines calls in the parent_texts
133
self._check_write_ok()
134
return self._add_lines(version_id, parents, lines, parent_texts)
136
def _add_lines(self, version_id, parents, lines, parent_texts):
137
"""Helper to do the class specific add_lines."""
138
raise NotImplementedError(self.add_lines)
140
def add_lines_with_ghosts(self, version_id, parents, lines,
142
"""Add lines to the versioned file, allowing ghosts to be present.
144
This takes the same parameters as add_lines.
146
self._check_write_ok()
147
return self._add_lines_with_ghosts(version_id, parents, lines,
150
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
151
"""Helper to do class specific add_lines_with_ghosts."""
152
raise NotImplementedError(self.add_lines_with_ghosts)
154
def check(self, progress_bar=None):
155
"""Check the versioned file for integrity."""
156
raise NotImplementedError(self.check)
158
def _check_lines_not_unicode(self, lines):
159
"""Check that lines being added to a versioned file are not unicode."""
161
if line.__class__ is not str:
162
raise errors.BzrBadParameterUnicode("lines")
164
def _check_lines_are_lines(self, lines):
165
"""Check that the lines really are full lines without inline EOL."""
167
if '\n' in line[:-1]:
168
raise errors.BzrBadParameterContainsNewline("lines")
170
def _check_write_ok(self):
171
"""Is the versioned file marked as 'finished' ? Raise if it is."""
173
raise errors.OutSideTransaction()
174
if self._access_mode != 'w':
175
raise errors.ReadOnlyObjectDirtiedError(self)
177
def clear_cache(self):
178
"""Remove any data cached in the versioned file object."""
180
def clone_text(self, new_version_id, old_version_id, parents):
181
"""Add an identical text to old_version_id as new_version_id.
183
Must raise RevisionNotPresent if the old version or any of the
184
parents are not present in file history.
186
Must raise RevisionAlreadyPresent if the new version is
187
already present in file history."""
188
self._check_write_ok()
189
return self._clone_text(new_version_id, old_version_id, parents)
191
def _clone_text(self, new_version_id, old_version_id, parents):
192
"""Helper function to do the _clone_text work."""
193
raise NotImplementedError(self.clone_text)
195
def create_empty(self, name, transport, mode=None):
196
"""Create a new versioned file of this exact type.
198
:param name: the file name
199
:param transport: the transport
200
:param mode: optional file mode.
202
raise NotImplementedError(self.create_empty)
204
def fix_parents(self, version, new_parents):
205
"""Fix the parents list for version.
207
This is done by appending a new version to the index
208
with identical data except for the parents list.
209
the parents list must be a superset of the current
212
self._check_write_ok()
213
return self._fix_parents(version, new_parents)
215
def _fix_parents(self, version, new_parents):
216
"""Helper for fix_parents."""
217
raise NotImplementedError(self.fix_parents)
219
def get_delta(self, version):
220
"""Get a delta for constructing version from some other version.
222
:return: (delta_parent, sha1, noeol, delta)
223
Where delta_parent is a version id or None to indicate no parent.
225
raise NotImplementedError(self.get_delta)
227
def get_deltas(self, versions):
228
"""Get multiple deltas at once for constructing versions.
230
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
231
Where delta_parent is a version id or None to indicate no parent, and
232
version_id is the version_id created by that delta.
235
for version in versions:
236
result[version] = self.get_delta(version)
239
def get_sha1(self, version_id):
240
"""Get the stored sha1 sum for the given revision.
242
:param name: The name of the version to lookup
244
raise NotImplementedError(self.get_sha1)
246
def get_suffixes(self):
247
"""Return the file suffixes associated with this versioned file."""
248
raise NotImplementedError(self.get_suffixes)
250
def get_text(self, version_id):
251
"""Return version contents as a text string.
253
Raises RevisionNotPresent if version is not present in
256
return ''.join(self.get_lines(version_id))
257
get_string = get_text
259
def get_lines(self, version_id):
260
"""Return version contents as a sequence of lines.
262
Raises RevisionNotPresent if version is not present in
265
raise NotImplementedError(self.get_lines)
267
def get_ancestry(self, version_ids):
268
"""Return a list of all ancestors of given version(s). This
269
will not include the null revision.
271
Must raise RevisionNotPresent if any of the given versions are
272
not present in file history."""
273
if isinstance(version_ids, basestring):
274
version_ids = [version_ids]
275
raise NotImplementedError(self.get_ancestry)
277
def get_ancestry_with_ghosts(self, version_ids):
278
"""Return a list of all ancestors of given version(s). This
279
will not include the null revision.
281
Must raise RevisionNotPresent if any of the given versions are
282
not present in file history.
284
Ghosts that are known about will be included in ancestry list,
285
but are not explicitly marked.
287
raise NotImplementedError(self.get_ancestry_with_ghosts)
290
"""Return a graph for the entire versioned file.
292
Ghosts are not listed or referenced in the graph.
295
for version in self.versions():
296
result[version] = self.get_parents(version)
299
def get_graph_with_ghosts(self):
300
"""Return a graph for the entire versioned file.
302
Ghosts are referenced in parents list but are not
305
raise NotImplementedError(self.get_graph_with_ghosts)
307
@deprecated_method(zero_eight)
308
def parent_names(self, version):
309
"""Return version names for parents of a version.
311
See get_parents for the current api.
313
return self.get_parents(version)
315
def get_parents(self, version_id):
316
"""Return version names for parents of a version.
318
Must raise RevisionNotPresent if version is not present in
321
raise NotImplementedError(self.get_parents)
323
def get_parents_with_ghosts(self, version_id):
324
"""Return version names for parents of version_id.
326
Will raise RevisionNotPresent if version_id is not present
329
Ghosts that are known about will be included in the parent list,
330
but are not explicitly marked.
332
raise NotImplementedError(self.get_parents_with_ghosts)
334
def annotate_iter(self, version_id):
335
"""Yield list of (version-id, line) pairs for the specified
338
Must raise RevisionNotPresent if any of the given versions are
339
not present in file history.
341
raise NotImplementedError(self.annotate_iter)
343
def annotate(self, version_id):
344
return list(self.annotate_iter(version_id))
346
def _apply_delta(self, lines, delta):
347
"""Apply delta to lines."""
350
for start, end, count, delta_lines in delta:
351
lines[offset+start:offset+end] = delta_lines
352
offset = offset + (start - end) + count
355
def join(self, other, pb=None, msg=None, version_ids=None,
356
ignore_missing=False):
357
"""Integrate versions from other into this versioned file.
359
If version_ids is None all versions from other should be
360
incorporated into this versioned file.
362
Must raise RevisionNotPresent if any of the specified versions
363
are not present in the other files history unless ignore_missing
364
is supplied when they are silently skipped.
366
self._check_write_ok()
367
return InterVersionedFile.get(other, self).join(
373
def iter_lines_added_or_present_in_versions(self, version_ids=None):
374
"""Iterate over the lines in the versioned file from version_ids.
376
This may return lines from other versions, and does not return the
377
specific version marker at this point. The api may be changed
378
during development to include the version that the versioned file
379
thinks is relevant, but given that such hints are just guesses,
380
its better not to have it if we dont need it.
382
NOTES: Lines are normalised: they will all have \n terminators.
383
Lines are returned in arbitrary order.
385
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
387
def transaction_finished(self):
388
"""The transaction that this file was opened in has finished.
390
This records self.finished = True and should cause all mutating
395
@deprecated_method(zero_eight)
396
def walk(self, version_ids=None):
397
"""Walk the versioned file as a weave-like structure, for
398
versions relative to version_ids. Yields sequence of (lineno,
399
insert, deletes, text) for each relevant line.
401
Must raise RevisionNotPresent if any of the specified versions
402
are not present in the file history.
404
:param version_ids: the version_ids to walk with respect to. If not
405
supplied the entire weave-like structure is walked.
407
walk is deprecated in favour of iter_lines_added_or_present_in_versions
409
raise NotImplementedError(self.walk)
411
@deprecated_method(zero_eight)
412
def iter_names(self):
413
"""Walk the names list."""
414
return iter(self.versions())
416
def plan_merge(self, ver_a, ver_b):
417
"""Return pseudo-annotation indicating how the two versions merge.
419
This is computed between versions a and b and their common
422
Weave lines present in none of them are skipped entirely.
425
killed-base Dead in base revision
426
killed-both Killed in each revision
429
unchanged Alive in both a and b (possibly created in both)
432
ghost-a Killed in a, unborn in b
433
ghost-b Killed in b, unborn in a
434
irrelevant Not in either revision
436
raise NotImplementedError(VersionedFile.plan_merge)
438
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
439
b_marker=TextMerge.B_MARKER):
440
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
443
class PlanWeaveMerge(TextMerge):
444
"""Weave merge that takes a plan as its input.
446
This exists so that VersionedFile.plan_merge is implementable.
447
Most callers will want to use WeaveMerge instead.
450
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
451
b_marker=TextMerge.B_MARKER):
452
TextMerge.__init__(self, a_marker, b_marker)
455
def _merge_struct(self):
460
def outstanding_struct():
461
if not lines_a and not lines_b:
463
elif ch_a and not ch_b:
466
elif ch_b and not ch_a:
468
elif lines_a == lines_b:
471
yield (lines_a, lines_b)
473
# We previously considered either 'unchanged' or 'killed-both' lines
474
# to be possible places to resynchronize. However, assuming agreement
475
# on killed-both lines may be too agressive. -- mbp 20060324
476
for state, line in self.plan:
477
if state == 'unchanged':
478
# resync and flush queued conflicts changes if any
479
for struct in outstanding_struct():
485
if state == 'unchanged':
488
elif state == 'killed-a':
491
elif state == 'killed-b':
494
elif state == 'new-a':
497
elif state == 'new-b':
501
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
502
'killed-base', 'killed-both'), state
503
for struct in outstanding_struct():
507
class WeaveMerge(PlanWeaveMerge):
508
"""Weave merge that takes a VersionedFile and two versions as its input"""
510
def __init__(self, versionedfile, ver_a, ver_b,
511
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
512
plan = versionedfile.plan_merge(ver_a, ver_b)
513
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
516
class InterVersionedFile(InterObject):
517
"""This class represents operations taking place between two versionedfiles..
519
Its instances have methods like join, and contain
520
references to the source and target versionedfiles these operations can be
523
Often we will provide convenience methods on 'versionedfile' which carry out
524
operations with another versionedfile - they will always forward to
525
InterVersionedFile.get(other).method_name(parameters).
529
"""The available optimised InterVersionedFile types."""
531
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
532
"""Integrate versions from self.source into self.target.
534
If version_ids is None all versions from source should be
535
incorporated into this versioned file.
537
Must raise RevisionNotPresent if any of the specified versions
538
are not present in the other files history unless ignore_missing is
539
supplied when they are silently skipped.
542
# - if the target is empty, just add all the versions from
543
# source to target, otherwise:
544
# - make a temporary versioned file of type target
545
# - insert the source content into it one at a time
547
if not self.target.versions():
550
# Make a new target-format versioned file.
551
temp_source = self.target.create_empty("temp", MemoryTransport())
553
graph = self.source.get_graph()
554
order = topo_sort(graph.items())
555
pb = ui.ui_factory.nested_progress_bar()
558
# TODO for incremental cross-format work:
559
# make a versioned file with the following content:
560
# all revisions we have been asked to join
561
# all their ancestors that are *not* in target already.
562
# the immediate parents of the above two sets, with
563
# empty parent lists - these versions are in target already
564
# and the incorrect version data will be ignored.
565
# TODO: for all ancestors that are present in target already,
566
# check them for consistent data, this requires moving sha1 from
568
# TODO: remove parent texts when they are not relevant any more for
569
# memory pressure reduction. RBC 20060313
570
# pb.update('Converting versioned data', 0, len(order))
571
# deltas = self.source.get_deltas(order)
572
for index, version in enumerate(order):
573
pb.update('Converting versioned data', index, len(order))
574
parent_text = target.add_lines(version,
575
self.source.get_parents(version),
576
self.source.get_lines(version),
577
parent_texts=parent_texts)
578
parent_texts[version] = parent_text
579
#delta_parent, sha1, noeol, delta = deltas[version]
580
#target.add_delta(version,
581
# self.source.get_parents(version),
586
#target.get_lines(version)
588
# this should hit the native code path for target
589
if target is not self.target:
590
return self.target.join(temp_source,
599
class InterVersionedFileTestProviderAdapter(object):
600
"""A tool to generate a suite testing multiple inter versioned-file classes.
602
This is done by copying the test once for each interversionedfile provider
603
and injecting the transport_server, transport_readonly_server,
604
versionedfile_factory and versionedfile_factory_to classes into each copy.
605
Each copy is also given a new id() to make it easy to identify.
608
def __init__(self, transport_server, transport_readonly_server, formats):
609
self._transport_server = transport_server
610
self._transport_readonly_server = transport_readonly_server
611
self._formats = formats
613
def adapt(self, test):
615
for (interversionedfile_class,
616
versionedfile_factory,
617
versionedfile_factory_to) in self._formats:
618
new_test = deepcopy(test)
619
new_test.transport_server = self._transport_server
620
new_test.transport_readonly_server = self._transport_readonly_server
621
new_test.interversionedfile_class = interversionedfile_class
622
new_test.versionedfile_factory = versionedfile_factory
623
new_test.versionedfile_factory_to = versionedfile_factory_to
624
def make_new_test_id():
625
new_id = "%s(%s)" % (new_test.id(), interversionedfile_class.__name__)
626
return lambda: new_id
627
new_test.id = make_new_test_id()
628
result.addTest(new_test)
632
def default_test_list():
633
"""Generate the default list of interversionedfile permutations to test."""
634
from bzrlib.weave import WeaveFile
635
from bzrlib.knit import KnitVersionedFile
637
# test the fallback InterVersionedFile from weave to annotated knits
638
result.append((InterVersionedFile,
641
for optimiser in InterVersionedFile._optimisers:
642
result.append((optimiser,
643
optimiser._matching_file_factory,
644
optimiser._matching_file_factory
646
# if there are specific combinations we want to use, we can add them
added
removed
bzrlib/versionedfile.py
