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.transport.memory import MemoryTransport
35
from bzrlib.tsort import topo_sort
39
class VersionedFile(object):
40
"""Versioned text file storage.
42
A versioned file manages versions of line-based text files,
43
keeping track of the originating version for each line.
45
To clients the "lines" of the file are represented as a list of
46
strings. These strings will typically have terminal newline
47
characters, but this is not required. In particular files commonly
48
do not have a newline at the end of the file.
50
Texts are identified by a version-id string.
53
def __init__(self, access_mode):
55
self._access_mode = access_mode
57
def copy_to(self, name, transport):
58
"""Copy this versioned file to name on transport."""
59
raise NotImplementedError(self.copy_to)
61
@deprecated_method(zero_eight)
63
"""Return a list of all the versions in this versioned file.
65
Please use versionedfile.versions() now.
67
return self.versions()
70
"""Return a unsorted list of versions."""
71
raise NotImplementedError(self.versions)
73
def has_ghost(self, version_id):
74
"""Returns whether version is present as a ghost."""
75
raise NotImplementedError(self.has_ghost)
77
def has_version(self, version_id):
78
"""Returns whether version is present."""
79
raise NotImplementedError(self.has_version)
81
def add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
82
"""Add a text to the versioned file via a pregenerated delta.
84
:param version_id: The version id being added.
85
:param parents: The parents of the version_id.
86
:param delta_parent: The parent this delta was created against.
87
:param sha1: The sha1 of the full text.
88
:param delta: The delta instructions. See get_delta for details.
90
self._check_write_ok()
91
if self.has_version(version_id):
92
raise errors.RevisionAlreadyPresent(version_id, self)
93
return self._add_delta(version_id, parents, delta_parent, sha1, noeol, delta)
95
def _add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
96
"""Class specific routine to add a delta.
98
This generic version simply applies the delta to the delta_parent and
101
# strip annotation from delta
103
for start, stop, delta_len, delta_lines in delta:
104
new_delta.append((start, stop, delta_len, [text for origin, text in delta_lines]))
105
if delta_parent is not None:
106
parent_full = self.get_lines(delta_parent)
109
new_full = self._apply_delta(parent_full, new_delta)
110
# its impossible to have noeol on an empty file
111
if noeol and new_full[-1][-1] == '\n':
112
new_full[-1] = new_full[-1][:-1]
113
self.add_lines(version_id, parents, new_full)
115
def add_lines(self, version_id, parents, lines, parent_texts=None):
116
"""Add a single text on top of the versioned file.
118
Must raise RevisionAlreadyPresent if the new version is
119
already present in file history.
121
Must raise RevisionNotPresent if any of the given parents are
122
not present in file history.
123
:param parent_texts: An optional dictionary containing the opaque
124
representations of some or all of the parents of
125
version_id to allow delta optimisations.
126
VERY IMPORTANT: the texts must be those returned
127
by add_lines or data corruption can be caused.
128
:return: An opaque representation of the inserted version which can be
129
provided back to future add_lines calls in the parent_texts
132
self._check_write_ok()
133
return self._add_lines(version_id, parents, lines, parent_texts)
135
def _add_lines(self, version_id, parents, lines, parent_texts):
136
"""Helper to do the class specific add_lines."""
137
raise NotImplementedError(self.add_lines)
139
def add_lines_with_ghosts(self, version_id, parents, lines,
141
"""Add lines to the versioned file, allowing ghosts to be present.
143
This takes the same parameters as add_lines.
145
self._check_write_ok()
146
return self._add_lines_with_ghosts(version_id, parents, lines,
149
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
150
"""Helper to do class specific add_lines_with_ghosts."""
151
raise NotImplementedError(self.add_lines_with_ghosts)
153
def check(self, progress_bar=None):
154
"""Check the versioned file for integrity."""
155
raise NotImplementedError(self.check)
157
def _check_write_ok(self):
158
"""Is the versioned file marked as 'finished' ? Raise if it is."""
160
raise errors.OutSideTransaction()
161
if self._access_mode != 'w':
162
raise errors.ReadOnlyObjectDirtiedError(self)
164
def clear_cache(self):
165
"""Remove any data cached in the versioned file object."""
167
def clone_text(self, new_version_id, old_version_id, parents):
168
"""Add an identical text to old_version_id as new_version_id.
170
Must raise RevisionNotPresent if the old version or any of the
171
parents are not present in file history.
173
Must raise RevisionAlreadyPresent if the new version is
174
already present in file history."""
175
self._check_write_ok()
176
return self._clone_text(new_version_id, old_version_id, parents)
178
def _clone_text(self, new_version_id, old_version_id, parents):
179
"""Helper function to do the _clone_text work."""
180
raise NotImplementedError(self.clone_text)
182
def create_empty(self, name, transport, mode=None):
183
"""Create a new versioned file of this exact type.
185
:param name: the file name
186
:param transport: the transport
187
:param mode: optional file mode.
189
raise NotImplementedError(self.create_empty)
191
def fix_parents(self, version, new_parents):
192
"""Fix the parents list for version.
194
This is done by appending a new version to the index
195
with identical data except for the parents list.
196
the parents list must be a superset of the current
199
self._check_write_ok()
200
return self._fix_parents(version, new_parents)
202
def _fix_parents(self, version, new_parents):
203
"""Helper for fix_parents."""
204
raise NotImplementedError(self.fix_parents)
206
def get_delta(self, version):
207
"""Get a delta for constructing version from some other version.
209
:return: (delta_parent, sha1, noeol, delta)
210
Where delta_parent is a version id or None to indicate no parent.
212
raise NotImplementedError(self.get_delta)
214
def get_deltas(self, versions):
215
"""Get multiple deltas at once for constructing versions.
217
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
218
Where delta_parent is a version id or None to indicate no parent, and
219
version_id is the version_id created by that delta.
222
for version in versions:
223
result[version] = self.get_delta(version)
226
def get_suffixes(self):
227
"""Return the file suffixes associated with this versioned file."""
228
raise NotImplementedError(self.get_suffixes)
230
def get_text(self, version_id):
231
"""Return version contents as a text string.
233
Raises RevisionNotPresent if version is not present in
236
return ''.join(self.get_lines(version_id))
237
get_string = get_text
239
def get_lines(self, version_id):
240
"""Return version contents as a sequence of lines.
242
Raises RevisionNotPresent if version is not present in
245
raise NotImplementedError(self.get_lines)
247
def get_ancestry(self, version_ids):
248
"""Return a list of all ancestors of given version(s). This
249
will not include the null revision.
251
Must raise RevisionNotPresent if any of the given versions are
252
not present in file history."""
253
if isinstance(version_ids, basestring):
254
version_ids = [version_ids]
255
raise NotImplementedError(self.get_ancestry)
257
def get_ancestry_with_ghosts(self, version_ids):
258
"""Return a list of all ancestors of given version(s). This
259
will not include the null revision.
261
Must raise RevisionNotPresent if any of the given versions are
262
not present in file history.
264
Ghosts that are known about will be included in ancestry list,
265
but are not explicitly marked.
267
raise NotImplementedError(self.get_ancestry_with_ghosts)
270
"""Return a graph for the entire versioned file.
272
Ghosts are not listed or referenced in the graph.
275
for version in self.versions():
276
result[version] = self.get_parents(version)
279
def get_graph_with_ghosts(self):
280
"""Return a graph for the entire versioned file.
282
Ghosts are referenced in parents list but are not
285
raise NotImplementedError(self.get_graph_with_ghosts)
287
@deprecated_method(zero_eight)
288
def parent_names(self, version):
289
"""Return version names for parents of a version.
291
See get_parents for the current api.
293
return self.get_parents(version)
295
def get_parents(self, version_id):
296
"""Return version names for parents of a version.
298
Must raise RevisionNotPresent if version is not present in
301
raise NotImplementedError(self.get_parents)
303
def get_parents_with_ghosts(self, version_id):
304
"""Return version names for parents of version_id.
306
Will raise RevisionNotPresent if version_id is not present
309
Ghosts that are known about will be included in the parent list,
310
but are not explicitly marked.
312
raise NotImplementedError(self.get_parents_with_ghosts)
314
def annotate_iter(self, version_id):
315
"""Yield list of (version-id, line) pairs for the specified
318
Must raise RevisionNotPresent if any of the given versions are
319
not present in file history.
321
raise NotImplementedError(self.annotate_iter)
323
def annotate(self, version_id):
324
return list(self.annotate_iter(version_id))
326
def _apply_delta(self, lines, delta):
327
"""Apply delta to lines."""
330
for start, end, count, delta_lines in delta:
331
lines[offset+start:offset+end] = delta_lines
332
offset = offset + (start - end) + count
335
def join(self, other, pb=None, msg=None, version_ids=None,
336
ignore_missing=False):
337
"""Integrate versions from other into this versioned file.
339
If version_ids is None all versions from other should be
340
incorporated into this versioned file.
342
Must raise RevisionNotPresent if any of the specified versions
343
are not present in the other files history unless ignore_missing
344
is supplied when they are silently skipped.
346
self._check_write_ok()
347
return InterVersionedFile.get(other, self).join(
353
def iter_lines_added_or_present_in_versions(self, version_ids=None):
354
"""Iterate over the lines in the versioned file from version_ids.
356
This may return lines from other versions, and does not return the
357
specific version marker at this point. The api may be changed
358
during development to include the version that the versioned file
359
thinks is relevant, but given that such hints are just guesses,
360
its better not to have it if we dont need it.
362
NOTES: Lines are normalised: they will all have \n terminators.
363
Lines are returned in arbitrary order.
365
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
367
def transaction_finished(self):
368
"""The transaction that this file was opened in has finished.
370
This records self.finished = True and should cause all mutating
375
@deprecated_method(zero_eight)
376
def walk(self, version_ids=None):
377
"""Walk the versioned file as a weave-like structure, for
378
versions relative to version_ids. Yields sequence of (lineno,
379
insert, deletes, text) for each relevant line.
381
Must raise RevisionNotPresent if any of the specified versions
382
are not present in the file history.
384
:param version_ids: the version_ids to walk with respect to. If not
385
supplied the entire weave-like structure is walked.
387
walk is deprecated in favour of iter_lines_added_or_present_in_versions
389
raise NotImplementedError(self.walk)
391
@deprecated_method(zero_eight)
392
def iter_names(self):
393
"""Walk the names list."""
394
return iter(self.versions())
396
def plan_merge(self, ver_a, ver_b):
397
"""Return pseudo-annotation indicating how the two versions merge.
399
This is computed between versions a and b and their common
402
Weave lines present in none of them are skipped entirely.
404
inc_a = set(self.get_ancestry([ver_a]))
405
inc_b = set(self.get_ancestry([ver_b]))
406
inc_c = inc_a & inc_b
408
for lineno, insert, deleteset, line in self.walk([ver_a, ver_b]):
409
if deleteset & inc_c:
410
# killed in parent; can't be in either a or b
411
# not relevant to our work
412
yield 'killed-base', line
413
elif insert in inc_c:
414
# was inserted in base
415
killed_a = bool(deleteset & inc_a)
416
killed_b = bool(deleteset & inc_b)
417
if killed_a and killed_b:
418
yield 'killed-both', line
420
yield 'killed-a', line
422
yield 'killed-b', line
424
yield 'unchanged', line
425
elif insert in inc_a:
426
if deleteset & inc_a:
427
yield 'ghost-a', line
431
elif insert in inc_b:
432
if deleteset & inc_b:
433
yield 'ghost-b', line
437
# not in either revision
438
yield 'irrelevant', line
440
yield 'unchanged', '' # terminator
442
def weave_merge(self, plan, a_marker='<<<<<<< \n', b_marker='>>>>>>> \n'):
446
# TODO: Return a structured form of the conflicts (e.g. 2-tuples for
447
# conflicted regions), rather than just inserting the markers.
449
# TODO: Show some version information (e.g. author, date) on
450
# conflicted regions.
452
# We previously considered either 'unchanged' or 'killed-both' lines
453
# to be possible places to resynchronize. However, assuming agreement
454
# on killed-both lines may be too agressive. -- mbp 20060324
455
for state, line in plan:
456
if state == 'unchanged':
457
# resync and flush queued conflicts changes if any
458
if not lines_a and not lines_b:
460
elif ch_a and not ch_b:
462
for l in lines_a: yield l
463
elif ch_b and not ch_a:
464
for l in lines_b: yield l
465
elif lines_a == lines_b:
466
for l in lines_a: yield l
469
for l in lines_a: yield l
471
for l in lines_b: yield l
478
if state == 'unchanged':
481
elif state == 'killed-a':
484
elif state == 'killed-b':
487
elif state == 'new-a':
490
elif state == 'new-b':
494
assert state in ('irrelevant', 'ghost-a', 'ghost-b', 'killed-base',
499
class InterVersionedFile(InterObject):
500
"""This class represents operations taking place between two versionedfiles..
502
Its instances have methods like join, and contain
503
references to the source and target versionedfiles these operations can be
506
Often we will provide convenience methods on 'versionedfile' which carry out
507
operations with another versionedfile - they will always forward to
508
InterVersionedFile.get(other).method_name(parameters).
512
"""The available optimised InterVersionedFile types."""
514
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
515
"""Integrate versions from self.source into self.target.
517
If version_ids is None all versions from source should be
518
incorporated into this versioned file.
520
Must raise RevisionNotPresent if any of the specified versions
521
are not present in the other files history unless ignore_missing is
522
supplied when they are silently skipped.
525
# - if the target is empty, just add all the versions from
526
# source to target, otherwise:
527
# - make a temporary versioned file of type target
528
# - insert the source content into it one at a time
530
if not self.target.versions():
533
# Make a new target-format versioned file.
534
temp_source = self.target.create_empty("temp", MemoryTransport())
536
graph = self.source.get_graph()
537
order = topo_sort(graph.items())
538
pb = ui.ui_factory.nested_progress_bar()
541
# TODO for incremental cross-format work:
542
# make a versioned file with the following content:
543
# all revisions we have been asked to join
544
# all their ancestors that are *not* in target already.
545
# the immediate parents of the above two sets, with
546
# empty parent lists - these versions are in target already
547
# and the incorrect version data will be ignored.
548
# TODO: for all ancestors that are present in target already,
549
# check them for consistent data, this requires moving sha1 from
551
# TODO: remove parent texts when they are not relevant any more for
552
# memory pressure reduction. RBC 20060313
553
# pb.update('Converting versioned data', 0, len(order))
554
# deltas = self.source.get_deltas(order)
555
for index, version in enumerate(order):
556
pb.update('Converting versioned data', index, len(order))
557
parent_text = target.add_lines(version,
558
self.source.get_parents(version),
559
self.source.get_lines(version),
560
parent_texts=parent_texts)
561
parent_texts[version] = parent_text
562
#delta_parent, sha1, noeol, delta = deltas[version]
563
#target.add_delta(version,
564
# self.source.get_parents(version),
569
#target.get_lines(version)
571
# this should hit the native code path for target
572
if target is not self.target:
573
return self.target.join(temp_source,
582
class InterVersionedFileTestProviderAdapter(object):
583
"""A tool to generate a suite testing multiple inter versioned-file classes.
585
This is done by copying the test once for each interversionedfile provider
586
and injecting the transport_server, transport_readonly_server,
587
versionedfile_factory and versionedfile_factory_to classes into each copy.
588
Each copy is also given a new id() to make it easy to identify.
591
def __init__(self, transport_server, transport_readonly_server, formats):
592
self._transport_server = transport_server
593
self._transport_readonly_server = transport_readonly_server
594
self._formats = formats
596
def adapt(self, test):
598
for (interversionedfile_class,
599
versionedfile_factory,
600
versionedfile_factory_to) in self._formats:
601
new_test = deepcopy(test)
602
new_test.transport_server = self._transport_server
603
new_test.transport_readonly_server = self._transport_readonly_server
604
new_test.interversionedfile_class = interversionedfile_class
605
new_test.versionedfile_factory = versionedfile_factory
606
new_test.versionedfile_factory_to = versionedfile_factory_to
607
def make_new_test_id():
608
new_id = "%s(%s)" % (new_test.id(), interversionedfile_class.__name__)
609
return lambda: new_id
610
new_test.id = make_new_test_id()
611
result.addTest(new_test)
615
def default_test_list():
616
"""Generate the default list of interversionedfile permutations to test."""
617
from bzrlib.weave import WeaveFile
618
from bzrlib.knit import KnitVersionedFile
620
# test the fallback InterVersionedFile from weave to annotated knits
621
result.append((InterVersionedFile,
624
for optimiser in InterVersionedFile._optimisers:
625
result.append((optimiser,
626
optimiser._matching_file_factory,
627
optimiser._matching_file_factory
629
# if there are specific combinations we want to use, we can add them