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(), """
33
from bzrlib.transport.memory import MemoryTransport
36
from cStringIO import StringIO
38
from bzrlib.inter import InterObject
39
from bzrlib.textmerge import TextMerge
42
class VersionedFile(object):
43
"""Versioned text file storage.
45
A versioned file manages versions of line-based text files,
46
keeping track of the originating version for each line.
48
To clients the "lines" of the file are represented as a list of
49
strings. These strings will typically have terminal newline
50
characters, but this is not required. In particular files commonly
51
do not have a newline at the end of the file.
53
Texts are identified by a version-id string.
56
def __init__(self, access_mode):
58
self._access_mode = access_mode
61
def check_not_reserved_id(version_id):
62
revision.check_not_reserved_id(version_id)
64
def copy_to(self, name, transport):
65
"""Copy this versioned file to name on transport."""
66
raise NotImplementedError(self.copy_to)
69
"""Return a unsorted list of versions."""
70
raise NotImplementedError(self.versions)
72
def has_ghost(self, version_id):
73
"""Returns whether version is present as a ghost."""
74
raise NotImplementedError(self.has_ghost)
76
def has_version(self, version_id):
77
"""Returns whether version is present."""
78
raise NotImplementedError(self.has_version)
80
def add_lines(self, version_id, parents, lines, parent_texts=None,
81
left_matching_blocks=None, nostore_sha=None, random_id=False,
83
"""Add a single text on top of the versioned file.
85
Must raise RevisionAlreadyPresent if the new version is
86
already present in file history.
88
Must raise RevisionNotPresent if any of the given parents are
89
not present in file history.
91
:param lines: A list of lines. Each line must be a bytestring. And all
92
of them except the last must be terminated with \n and contain no
93
other \n's. The last line may either contain no \n's or a single
94
terminated \n. If the lines list does meet this constraint the add
95
routine may error or may succeed - but you will be unable to read
96
the data back accurately. (Checking the lines have been split
97
correctly is expensive and extremely unlikely to catch bugs so it
98
is not done at runtime unless check_content is True.)
99
:param parent_texts: An optional dictionary containing the opaque
100
representations of some or all of the parents of version_id to
101
allow delta optimisations. VERY IMPORTANT: the texts must be those
102
returned by add_lines or data corruption can be caused.
103
:param left_matching_blocks: a hint about which areas are common
104
between the text and its left-hand-parent. The format is
105
the SequenceMatcher.get_matching_blocks format.
106
:param nostore_sha: Raise ExistingContent and do not add the lines to
107
the versioned file if the digest of the lines matches this.
108
:param random_id: If True a random id has been selected rather than
109
an id determined by some deterministic process such as a converter
110
from a foreign VCS. When True the backend may choose not to check
111
for uniqueness of the resulting key within the versioned file, so
112
this should only be done when the result is expected to be unique
114
:param check_content: If True, the lines supplied are verified to be
115
bytestrings that are correctly formed lines.
116
:return: The text sha1, the number of bytes in the text, and an opaque
117
representation of the inserted version which can be provided
118
back to future add_lines calls in the parent_texts dictionary.
120
version_id = osutils.safe_revision_id(version_id)
121
parents = [osutils.safe_revision_id(v) for v in parents]
122
self._check_write_ok()
123
return self._add_lines(version_id, parents, lines, parent_texts,
124
left_matching_blocks, nostore_sha, random_id, check_content)
126
def _add_lines(self, version_id, parents, lines, parent_texts,
127
left_matching_blocks, nostore_sha, random_id, check_content):
128
"""Helper to do the class specific add_lines."""
129
raise NotImplementedError(self.add_lines)
131
def add_lines_with_ghosts(self, version_id, parents, lines,
132
parent_texts=None, nostore_sha=None, random_id=False,
134
"""Add lines to the versioned file, allowing ghosts to be present.
136
This takes the same parameters as add_lines and returns the same.
138
version_id = osutils.safe_revision_id(version_id)
139
parents = [osutils.safe_revision_id(v) for v in parents]
140
self._check_write_ok()
141
return self._add_lines_with_ghosts(version_id, parents, lines,
142
parent_texts, nostore_sha, random_id, check_content)
144
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts,
145
nostore_sha, random_id, check_content):
146
"""Helper to do class specific add_lines_with_ghosts."""
147
raise NotImplementedError(self.add_lines_with_ghosts)
149
def check(self, progress_bar=None):
150
"""Check the versioned file for integrity."""
151
raise NotImplementedError(self.check)
153
def _check_lines_not_unicode(self, lines):
154
"""Check that lines being added to a versioned file are not unicode."""
156
if line.__class__ is not str:
157
raise errors.BzrBadParameterUnicode("lines")
159
def _check_lines_are_lines(self, lines):
160
"""Check that the lines really are full lines without inline EOL."""
162
if '\n' in line[:-1]:
163
raise errors.BzrBadParameterContainsNewline("lines")
165
def _check_write_ok(self):
166
"""Is the versioned file marked as 'finished' ? Raise if it is."""
168
raise errors.OutSideTransaction()
169
if self._access_mode != 'w':
170
raise errors.ReadOnlyObjectDirtiedError(self)
172
def enable_cache(self):
173
"""Tell this versioned file that it should cache any data it reads.
175
This is advisory, implementations do not have to support caching.
179
def clear_cache(self):
180
"""Remove any data cached in the versioned file object.
182
This only needs to be supported if caches are supported
186
def clone_text(self, new_version_id, old_version_id, parents):
187
"""Add an identical text to old_version_id as new_version_id.
189
Must raise RevisionNotPresent if the old version or any of the
190
parents are not present in file history.
192
Must raise RevisionAlreadyPresent if the new version is
193
already present in file history."""
194
new_version_id = osutils.safe_revision_id(new_version_id)
195
old_version_id = osutils.safe_revision_id(old_version_id)
196
self._check_write_ok()
197
return self._clone_text(new_version_id, old_version_id, parents)
199
def _clone_text(self, new_version_id, old_version_id, parents):
200
"""Helper function to do the _clone_text work."""
201
raise NotImplementedError(self.clone_text)
203
def create_empty(self, name, transport, mode=None):
204
"""Create a new versioned file of this exact type.
206
:param name: the file name
207
:param transport: the transport
208
:param mode: optional file mode.
210
raise NotImplementedError(self.create_empty)
212
def fix_parents(self, version_id, new_parents):
213
"""Fix the parents list for version.
215
This is done by appending a new version to the index
216
with identical data except for the parents list.
217
the parents list must be a superset of the current
220
version_id = osutils.safe_revision_id(version_id)
221
new_parents = [osutils.safe_revision_id(p) for p in new_parents]
222
self._check_write_ok()
223
return self._fix_parents(version_id, new_parents)
225
def _fix_parents(self, version_id, new_parents):
226
"""Helper for fix_parents."""
227
raise NotImplementedError(self.fix_parents)
229
def get_format_signature(self):
230
"""Get a text description of the data encoding in this file.
234
raise NotImplementedError(self.get_format_signature)
236
def make_mpdiffs(self, version_ids):
237
"""Create multiparent diffs for specified versions"""
238
knit_versions = set()
239
for version_id in version_ids:
240
knit_versions.add(version_id)
241
knit_versions.update(self.get_parents(version_id))
242
lines = dict(zip(knit_versions,
243
self._get_lf_split_line_list(knit_versions)))
245
for version_id in version_ids:
246
target = lines[version_id]
247
parents = [lines[p] for p in self.get_parents(version_id)]
249
left_parent_blocks = self._extract_blocks(version_id,
252
left_parent_blocks = None
253
diffs.append(multiparent.MultiParent.from_lines(target, parents,
257
def _extract_blocks(self, version_id, source, target):
260
def add_mpdiffs(self, records):
261
"""Add mpdiffs to this versionedfile
263
Records should be iterables of version, parents, expected_sha1,
264
mpdiff. mpdiff should be a MultiParent instance.
267
mpvf = multiparent.MultiMemoryVersionedFile()
269
for version, parent_ids, expected_sha1, mpdiff in records:
270
versions.append(version)
271
mpvf.add_diff(mpdiff, version, parent_ids)
272
needed_parents = set()
273
for version, parent_ids, expected_sha1, mpdiff in records:
274
needed_parents.update(p for p in parent_ids
275
if not mpvf.has_version(p))
276
for parent_id, lines in zip(needed_parents,
277
self._get_lf_split_line_list(needed_parents)):
278
mpvf.add_version(lines, parent_id, [])
279
for (version, parent_ids, expected_sha1, mpdiff), lines in\
280
zip(records, mpvf.get_line_list(versions)):
281
if len(parent_ids) == 1:
282
left_matching_blocks = list(mpdiff.get_matching_blocks(0,
283
mpvf.get_diff(parent_ids[0]).num_lines()))
285
left_matching_blocks = None
286
_, _, version_text = self.add_lines(version, parent_ids, lines,
287
vf_parents, left_matching_blocks=left_matching_blocks)
288
vf_parents[version] = version_text
289
for (version, parent_ids, expected_sha1, mpdiff), sha1 in\
290
zip(records, self.get_sha1s(versions)):
291
if expected_sha1 != sha1:
292
raise errors.VersionedFileInvalidChecksum(version)
294
def get_sha1(self, version_id):
295
"""Get the stored sha1 sum for the given revision.
297
:param name: The name of the version to lookup
299
raise NotImplementedError(self.get_sha1)
301
def get_sha1s(self, version_ids):
302
"""Get the stored sha1 sums for the given revisions.
304
:param version_ids: The names of the versions to lookup
305
:return: a list of sha1s in order according to the version_ids
307
raise NotImplementedError(self.get_sha1)
309
def get_suffixes(self):
310
"""Return the file suffixes associated with this versioned file."""
311
raise NotImplementedError(self.get_suffixes)
313
def get_text(self, version_id):
314
"""Return version contents as a text string.
316
Raises RevisionNotPresent if version is not present in
319
return ''.join(self.get_lines(version_id))
320
get_string = get_text
322
def get_texts(self, version_ids):
323
"""Return the texts of listed versions as a list of strings.
325
Raises RevisionNotPresent if version is not present in
328
return [''.join(self.get_lines(v)) for v in version_ids]
330
def get_lines(self, version_id):
331
"""Return version contents as a sequence of lines.
333
Raises RevisionNotPresent if version is not present in
336
raise NotImplementedError(self.get_lines)
338
def _get_lf_split_line_list(self, version_ids):
339
return [StringIO(t).readlines() for t in self.get_texts(version_ids)]
341
def get_ancestry(self, version_ids, topo_sorted=True):
342
"""Return a list of all ancestors of given version(s). This
343
will not include the null revision.
345
This list will not be topologically sorted if topo_sorted=False is
348
Must raise RevisionNotPresent if any of the given versions are
349
not present in file history."""
350
if isinstance(version_ids, basestring):
351
version_ids = [version_ids]
352
raise NotImplementedError(self.get_ancestry)
354
def get_ancestry_with_ghosts(self, version_ids):
355
"""Return a list of all ancestors of given version(s). This
356
will not include the null revision.
358
Must raise RevisionNotPresent if any of the given versions are
359
not present in file history.
361
Ghosts that are known about will be included in ancestry list,
362
but are not explicitly marked.
364
raise NotImplementedError(self.get_ancestry_with_ghosts)
366
def get_graph(self, version_ids=None):
367
"""Return a graph from the versioned file.
369
Ghosts are not listed or referenced in the graph.
370
:param version_ids: Versions to select.
371
None means retrieve all versions.
373
if version_ids is None:
374
return dict(self.iter_parents(self.versions()))
376
pending = set(osutils.safe_revision_id(v) for v in version_ids)
378
this_iteration = pending
380
for version, parents in self.iter_parents(this_iteration):
381
result[version] = parents
382
for parent in parents:
388
def get_graph_with_ghosts(self):
389
"""Return a graph for the entire versioned file.
391
Ghosts are referenced in parents list but are not
394
raise NotImplementedError(self.get_graph_with_ghosts)
396
def get_parents(self, version_id):
397
"""Return version names for parents of a version.
399
Must raise RevisionNotPresent if version is not present in
402
raise NotImplementedError(self.get_parents)
404
def get_parents_with_ghosts(self, version_id):
405
"""Return version names for parents of version_id.
407
Will raise RevisionNotPresent if version_id is not present
410
Ghosts that are known about will be included in the parent list,
411
but are not explicitly marked.
413
raise NotImplementedError(self.get_parents_with_ghosts)
415
def annotate_iter(self, version_id):
416
"""Yield list of (version-id, line) pairs for the specified
419
Must raise RevisionNotPresent if any of the given versions are
420
not present in file history.
422
raise NotImplementedError(self.annotate_iter)
424
def annotate(self, version_id):
425
return list(self.annotate_iter(version_id))
427
def _apply_delta(self, lines, delta):
428
"""Apply delta to lines."""
431
for start, end, count, delta_lines in delta:
432
lines[offset+start:offset+end] = delta_lines
433
offset = offset + (start - end) + count
436
def join(self, other, pb=None, msg=None, version_ids=None,
437
ignore_missing=False):
438
"""Integrate versions from other into this versioned file.
440
If version_ids is None all versions from other should be
441
incorporated into this versioned file.
443
Must raise RevisionNotPresent if any of the specified versions
444
are not present in the other files history unless ignore_missing
445
is supplied when they are silently skipped.
447
self._check_write_ok()
448
return InterVersionedFile.get(other, self).join(
454
def iter_lines_added_or_present_in_versions(self, version_ids=None,
456
"""Iterate over the lines in the versioned file from version_ids.
458
This may return lines from other versions, and does not return the
459
specific version marker at this point. The api may be changed
460
during development to include the version that the versioned file
461
thinks is relevant, but given that such hints are just guesses,
462
its better not to have it if we don't need it.
464
If a progress bar is supplied, it may be used to indicate progress.
465
The caller is responsible for cleaning up progress bars (because this
468
NOTES: Lines are normalised: they will all have \n terminators.
469
Lines are returned in arbitrary order.
471
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
473
def iter_parents(self, version_ids):
474
"""Iterate through the parents for many version ids.
476
:param version_ids: An iterable yielding version_ids.
477
:return: An iterator that yields (version_id, parents). Requested
478
version_ids not present in the versioned file are simply skipped.
479
The order is undefined, allowing for different optimisations in
480
the underlying implementation.
482
for version_id in version_ids:
484
yield version_id, tuple(self.get_parents(version_id))
485
except errors.RevisionNotPresent:
488
def transaction_finished(self):
489
"""The transaction that this file was opened in has finished.
491
This records self.finished = True and should cause all mutating
496
def plan_merge(self, ver_a, ver_b):
497
"""Return pseudo-annotation indicating how the two versions merge.
499
This is computed between versions a and b and their common
502
Weave lines present in none of them are skipped entirely.
505
killed-base Dead in base revision
506
killed-both Killed in each revision
509
unchanged Alive in both a and b (possibly created in both)
512
ghost-a Killed in a, unborn in b
513
ghost-b Killed in b, unborn in a
514
irrelevant Not in either revision
516
raise NotImplementedError(VersionedFile.plan_merge)
518
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
519
b_marker=TextMerge.B_MARKER):
520
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
523
class PlanWeaveMerge(TextMerge):
524
"""Weave merge that takes a plan as its input.
526
This exists so that VersionedFile.plan_merge is implementable.
527
Most callers will want to use WeaveMerge instead.
530
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
531
b_marker=TextMerge.B_MARKER):
532
TextMerge.__init__(self, a_marker, b_marker)
535
def _merge_struct(self):
540
def outstanding_struct():
541
if not lines_a and not lines_b:
543
elif ch_a and not ch_b:
546
elif ch_b and not ch_a:
548
elif lines_a == lines_b:
551
yield (lines_a, lines_b)
553
# We previously considered either 'unchanged' or 'killed-both' lines
554
# to be possible places to resynchronize. However, assuming agreement
555
# on killed-both lines may be too aggressive. -- mbp 20060324
556
for state, line in self.plan:
557
if state == 'unchanged':
558
# resync and flush queued conflicts changes if any
559
for struct in outstanding_struct():
565
if state == 'unchanged':
568
elif state == 'killed-a':
571
elif state == 'killed-b':
574
elif state == 'new-a':
577
elif state == 'new-b':
581
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
582
'killed-base', 'killed-both'), state
583
for struct in outstanding_struct():
587
class WeaveMerge(PlanWeaveMerge):
588
"""Weave merge that takes a VersionedFile and two versions as its input"""
590
def __init__(self, versionedfile, ver_a, ver_b,
591
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
592
plan = versionedfile.plan_merge(ver_a, ver_b)
593
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
596
class InterVersionedFile(InterObject):
597
"""This class represents operations taking place between two versionedfiles..
599
Its instances have methods like join, and contain
600
references to the source and target versionedfiles these operations can be
603
Often we will provide convenience methods on 'versionedfile' which carry out
604
operations with another versionedfile - they will always forward to
605
InterVersionedFile.get(other).method_name(parameters).
609
"""The available optimised InterVersionedFile types."""
611
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
612
"""Integrate versions from self.source into self.target.
614
If version_ids is None all versions from source should be
615
incorporated into this versioned file.
617
Must raise RevisionNotPresent if any of the specified versions
618
are not present in the other files history unless ignore_missing is
619
supplied when they are silently skipped.
622
# - if the target is empty, just add all the versions from
623
# source to target, otherwise:
624
# - make a temporary versioned file of type target
625
# - insert the source content into it one at a time
627
if not self.target.versions():
630
# Make a new target-format versioned file.
631
temp_source = self.target.create_empty("temp", MemoryTransport())
633
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
634
graph = self.source.get_graph(version_ids)
635
order = tsort.topo_sort(graph.items())
636
pb = ui.ui_factory.nested_progress_bar()
639
# TODO for incremental cross-format work:
640
# make a versioned file with the following content:
641
# all revisions we have been asked to join
642
# all their ancestors that are *not* in target already.
643
# the immediate parents of the above two sets, with
644
# empty parent lists - these versions are in target already
645
# and the incorrect version data will be ignored.
646
# TODO: for all ancestors that are present in target already,
647
# check them for consistent data, this requires moving sha1 from
649
# TODO: remove parent texts when they are not relevant any more for
650
# memory pressure reduction. RBC 20060313
651
# pb.update('Converting versioned data', 0, len(order))
652
for index, version in enumerate(order):
653
pb.update('Converting versioned data', index, len(order))
654
_, _, parent_text = target.add_lines(version,
655
self.source.get_parents(version),
656
self.source.get_lines(version),
657
parent_texts=parent_texts)
658
parent_texts[version] = parent_text
660
# this should hit the native code path for target
661
if target is not self.target:
662
return self.target.join(temp_source,
670
def _get_source_version_ids(self, version_ids, ignore_missing):
671
"""Determine the version ids to be used from self.source.
673
:param version_ids: The caller-supplied version ids to check. (None
674
for all). If None is in version_ids, it is stripped.
675
:param ignore_missing: if True, remove missing ids from the version
676
list. If False, raise RevisionNotPresent on
677
a missing version id.
678
:return: A set of version ids.
680
if version_ids is None:
681
# None cannot be in source.versions
682
return set(self.source.versions())
684
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
686
return set(self.source.versions()).intersection(set(version_ids))
688
new_version_ids = set()
689
for version in version_ids:
692
if not self.source.has_version(version):
693
raise errors.RevisionNotPresent(version, str(self.source))
695
new_version_ids.add(version)
696
return new_version_ids
added
removed
bzrlib/versionedfile.py
