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_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
81
"""Add a text to the versioned file via a pregenerated delta.
83
:param version_id: The version id being added.
84
:param parents: The parents of the version_id.
85
:param delta_parent: The parent this delta was created against.
86
:param sha1: The sha1 of the full text.
87
:param delta: The delta instructions. See get_delta for details.
89
version_id = osutils.safe_revision_id(version_id)
90
parents = [osutils.safe_revision_id(v) for v in parents]
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
left_matching_blocks=None, nostore_sha=None):
118
"""Add a single text on top of the versioned file.
120
Must raise RevisionAlreadyPresent if the new version is
121
already present in file history.
123
Must raise RevisionNotPresent if any of the given parents are
124
not present in file history.
125
:param parent_texts: An optional dictionary containing the opaque
126
representations of some or all of the parents of
127
version_id to allow delta optimisations.
128
VERY IMPORTANT: the texts must be those returned
129
by add_lines or data corruption can be caused.
130
:param left_matching_blocks: a hint about which areas are common
131
between the text and its left-hand-parent. The format is
132
the SequenceMatcher.get_matching_blocks format.
133
:param nostore_sha: Raise ExistingContent and do not add the lines to
134
the versioned file if the digest of the lines matches this.
135
:return: The text sha1, the number of bytes in the text, and an opaque
136
representation of the inserted version which can be provided
137
back to future add_lines calls in the parent_texts dictionary.
139
version_id = osutils.safe_revision_id(version_id)
140
parents = [osutils.safe_revision_id(v) for v in parents]
141
self._check_write_ok()
142
return self._add_lines(version_id, parents, lines, parent_texts,
143
left_matching_blocks, nostore_sha)
145
def _add_lines(self, version_id, parents, lines, parent_texts,
146
left_matching_blocks, nostore_sha):
147
"""Helper to do the class specific add_lines."""
148
raise NotImplementedError(self.add_lines)
150
def add_lines_with_ghosts(self, version_id, parents, lines,
151
parent_texts=None, nostore_sha=None):
152
"""Add lines to the versioned file, allowing ghosts to be present.
154
This takes the same parameters as add_lines and returns the same.
156
version_id = osutils.safe_revision_id(version_id)
157
parents = [osutils.safe_revision_id(v) for v in parents]
158
self._check_write_ok()
159
return self._add_lines_with_ghosts(version_id, parents, lines,
160
parent_texts, nostore_sha)
162
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_format_signature(self):
268
"""Get a text description of the data encoding in this file.
272
raise NotImplementedError(self.get_format_signature)
274
def make_mpdiffs(self, version_ids):
275
"""Create multiparent diffs for specified versions"""
276
knit_versions = set()
277
for version_id in version_ids:
278
knit_versions.add(version_id)
279
knit_versions.update(self.get_parents(version_id))
280
lines = dict(zip(knit_versions,
281
self._get_lf_split_line_list(knit_versions)))
283
for version_id in version_ids:
284
target = lines[version_id]
285
parents = [lines[p] for p in self.get_parents(version_id)]
287
left_parent_blocks = self._extract_blocks(version_id,
290
left_parent_blocks = None
291
diffs.append(multiparent.MultiParent.from_lines(target, parents,
295
def _extract_blocks(self, version_id, source, target):
298
def add_mpdiffs(self, records):
299
"""Add mpdiffs to this versionedfile
301
Records should be iterables of version, parents, expected_sha1,
302
mpdiff. mpdiff should be a MultiParent instance.
305
mpvf = multiparent.MultiMemoryVersionedFile()
307
for version, parent_ids, expected_sha1, mpdiff in records:
308
versions.append(version)
309
mpvf.add_diff(mpdiff, version, parent_ids)
310
needed_parents = set()
311
for version, parent_ids, expected_sha1, mpdiff in records:
312
needed_parents.update(p for p in parent_ids
313
if not mpvf.has_version(p))
314
for parent_id, lines in zip(needed_parents,
315
self._get_lf_split_line_list(needed_parents)):
316
mpvf.add_version(lines, parent_id, [])
317
for (version, parent_ids, expected_sha1, mpdiff), lines in\
318
zip(records, mpvf.get_line_list(versions)):
319
if len(parent_ids) == 1:
320
left_matching_blocks = list(mpdiff.get_matching_blocks(0,
321
mpvf.get_diff(parent_ids[0]).num_lines()))
323
left_matching_blocks = None
324
_, _, version_text = self.add_lines(version, parent_ids, lines,
325
vf_parents, left_matching_blocks=left_matching_blocks)
326
vf_parents[version] = version_text
327
for (version, parent_ids, expected_sha1, mpdiff), sha1 in\
328
zip(records, self.get_sha1s(versions)):
329
if expected_sha1 != sha1:
330
raise errors.VersionedFileInvalidChecksum(version)
332
def get_sha1(self, version_id):
333
"""Get the stored sha1 sum for the given revision.
335
:param name: The name of the version to lookup
337
raise NotImplementedError(self.get_sha1)
339
def get_sha1s(self, version_ids):
340
"""Get the stored sha1 sums for the given revisions.
342
:param version_ids: The names of the versions to lookup
343
:return: a list of sha1s in order according to the version_ids
345
raise NotImplementedError(self.get_sha1)
347
def get_suffixes(self):
348
"""Return the file suffixes associated with this versioned file."""
349
raise NotImplementedError(self.get_suffixes)
351
def get_text(self, version_id):
352
"""Return version contents as a text string.
354
Raises RevisionNotPresent if version is not present in
357
return ''.join(self.get_lines(version_id))
358
get_string = get_text
360
def get_texts(self, version_ids):
361
"""Return the texts of listed versions as a list of strings.
363
Raises RevisionNotPresent if version is not present in
366
return [''.join(self.get_lines(v)) for v in version_ids]
368
def get_lines(self, version_id):
369
"""Return version contents as a sequence of lines.
371
Raises RevisionNotPresent if version is not present in
374
raise NotImplementedError(self.get_lines)
376
def _get_lf_split_line_list(self, version_ids):
377
return [StringIO(t).readlines() for t in self.get_texts(version_ids)]
379
def get_ancestry(self, version_ids, topo_sorted=True):
380
"""Return a list of all ancestors of given version(s). This
381
will not include the null revision.
383
This list will not be topologically sorted if topo_sorted=False is
386
Must raise RevisionNotPresent if any of the given versions are
387
not present in file history."""
388
if isinstance(version_ids, basestring):
389
version_ids = [version_ids]
390
raise NotImplementedError(self.get_ancestry)
392
def get_ancestry_with_ghosts(self, version_ids):
393
"""Return a list of all ancestors of given version(s). This
394
will not include the null revision.
396
Must raise RevisionNotPresent if any of the given versions are
397
not present in file history.
399
Ghosts that are known about will be included in ancestry list,
400
but are not explicitly marked.
402
raise NotImplementedError(self.get_ancestry_with_ghosts)
404
def get_graph(self, version_ids=None):
405
"""Return a graph from the versioned file.
407
Ghosts are not listed or referenced in the graph.
408
:param version_ids: Versions to select.
409
None means retrieve all versions.
411
if version_ids is None:
412
return dict(self.iter_parents(self.versions()))
414
pending = set(osutils.safe_revision_id(v) for v in version_ids)
416
this_iteration = pending
418
for version, parents in self.iter_parents(this_iteration):
419
result[version] = parents
420
for parent in parents:
426
def get_graph_with_ghosts(self):
427
"""Return a graph for the entire versioned file.
429
Ghosts are referenced in parents list but are not
432
raise NotImplementedError(self.get_graph_with_ghosts)
434
def get_parents(self, version_id):
435
"""Return version names for parents of a version.
437
Must raise RevisionNotPresent if version is not present in
440
raise NotImplementedError(self.get_parents)
442
def get_parents_with_ghosts(self, version_id):
443
"""Return version names for parents of version_id.
445
Will raise RevisionNotPresent if version_id is not present
448
Ghosts that are known about will be included in the parent list,
449
but are not explicitly marked.
451
raise NotImplementedError(self.get_parents_with_ghosts)
453
def annotate_iter(self, version_id):
454
"""Yield list of (version-id, line) pairs for the specified
457
Must raise RevisionNotPresent if any of the given versions are
458
not present in file history.
460
raise NotImplementedError(self.annotate_iter)
462
def annotate(self, version_id):
463
return list(self.annotate_iter(version_id))
465
def _apply_delta(self, lines, delta):
466
"""Apply delta to lines."""
469
for start, end, count, delta_lines in delta:
470
lines[offset+start:offset+end] = delta_lines
471
offset = offset + (start - end) + count
474
def join(self, other, pb=None, msg=None, version_ids=None,
475
ignore_missing=False):
476
"""Integrate versions from other into this versioned file.
478
If version_ids is None all versions from other should be
479
incorporated into this versioned file.
481
Must raise RevisionNotPresent if any of the specified versions
482
are not present in the other files history unless ignore_missing
483
is supplied when they are silently skipped.
485
self._check_write_ok()
486
return InterVersionedFile.get(other, self).join(
492
def iter_lines_added_or_present_in_versions(self, version_ids=None,
494
"""Iterate over the lines in the versioned file from version_ids.
496
This may return lines from other versions, and does not return the
497
specific version marker at this point. The api may be changed
498
during development to include the version that the versioned file
499
thinks is relevant, but given that such hints are just guesses,
500
its better not to have it if we don't need it.
502
If a progress bar is supplied, it may be used to indicate progress.
503
The caller is responsible for cleaning up progress bars (because this
506
NOTES: Lines are normalised: they will all have \n terminators.
507
Lines are returned in arbitrary order.
509
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
511
def iter_parents(self, version_ids):
512
"""Iterate through the parents for many version ids.
514
:param version_ids: An iterable yielding version_ids.
515
:return: An iterator that yields (version_id, parents). Requested
516
version_ids not present in the versioned file are simply skipped.
517
The order is undefined, allowing for different optimisations in
518
the underlying implementation.
520
for version_id in version_ids:
522
yield version_id, tuple(self.get_parents(version_id))
523
except errors.RevisionNotPresent:
526
def transaction_finished(self):
527
"""The transaction that this file was opened in has finished.
529
This records self.finished = True and should cause all mutating
534
def plan_merge(self, ver_a, ver_b):
535
"""Return pseudo-annotation indicating how the two versions merge.
537
This is computed between versions a and b and their common
540
Weave lines present in none of them are skipped entirely.
543
killed-base Dead in base revision
544
killed-both Killed in each revision
547
unchanged Alive in both a and b (possibly created in both)
550
ghost-a Killed in a, unborn in b
551
ghost-b Killed in b, unborn in a
552
irrelevant Not in either revision
554
raise NotImplementedError(VersionedFile.plan_merge)
556
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
557
b_marker=TextMerge.B_MARKER):
558
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
561
class PlanWeaveMerge(TextMerge):
562
"""Weave merge that takes a plan as its input.
564
This exists so that VersionedFile.plan_merge is implementable.
565
Most callers will want to use WeaveMerge instead.
568
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
569
b_marker=TextMerge.B_MARKER):
570
TextMerge.__init__(self, a_marker, b_marker)
573
def _merge_struct(self):
578
def outstanding_struct():
579
if not lines_a and not lines_b:
581
elif ch_a and not ch_b:
584
elif ch_b and not ch_a:
586
elif lines_a == lines_b:
589
yield (lines_a, lines_b)
591
# We previously considered either 'unchanged' or 'killed-both' lines
592
# to be possible places to resynchronize. However, assuming agreement
593
# on killed-both lines may be too aggressive. -- mbp 20060324
594
for state, line in self.plan:
595
if state == 'unchanged':
596
# resync and flush queued conflicts changes if any
597
for struct in outstanding_struct():
603
if state == 'unchanged':
606
elif state == 'killed-a':
609
elif state == 'killed-b':
612
elif state == 'new-a':
615
elif state == 'new-b':
619
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
620
'killed-base', 'killed-both'), state
621
for struct in outstanding_struct():
625
class WeaveMerge(PlanWeaveMerge):
626
"""Weave merge that takes a VersionedFile and two versions as its input"""
628
def __init__(self, versionedfile, ver_a, ver_b,
629
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
630
plan = versionedfile.plan_merge(ver_a, ver_b)
631
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
634
class InterVersionedFile(InterObject):
635
"""This class represents operations taking place between two versionedfiles..
637
Its instances have methods like join, and contain
638
references to the source and target versionedfiles these operations can be
641
Often we will provide convenience methods on 'versionedfile' which carry out
642
operations with another versionedfile - they will always forward to
643
InterVersionedFile.get(other).method_name(parameters).
647
"""The available optimised InterVersionedFile types."""
649
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
650
"""Integrate versions from self.source into self.target.
652
If version_ids is None all versions from source should be
653
incorporated into this versioned file.
655
Must raise RevisionNotPresent if any of the specified versions
656
are not present in the other files history unless ignore_missing is
657
supplied when they are silently skipped.
660
# - if the target is empty, just add all the versions from
661
# source to target, otherwise:
662
# - make a temporary versioned file of type target
663
# - insert the source content into it one at a time
665
if not self.target.versions():
668
# Make a new target-format versioned file.
669
temp_source = self.target.create_empty("temp", MemoryTransport())
671
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
672
graph = self.source.get_graph(version_ids)
673
order = tsort.topo_sort(graph.items())
674
pb = ui.ui_factory.nested_progress_bar()
677
# TODO for incremental cross-format work:
678
# make a versioned file with the following content:
679
# all revisions we have been asked to join
680
# all their ancestors that are *not* in target already.
681
# the immediate parents of the above two sets, with
682
# empty parent lists - these versions are in target already
683
# and the incorrect version data will be ignored.
684
# TODO: for all ancestors that are present in target already,
685
# check them for consistent data, this requires moving sha1 from
687
# TODO: remove parent texts when they are not relevant any more for
688
# memory pressure reduction. RBC 20060313
689
# pb.update('Converting versioned data', 0, len(order))
690
# deltas = self.source.get_deltas(order)
691
for index, version in enumerate(order):
692
pb.update('Converting versioned data', index, len(order))
693
_, _, parent_text = target.add_lines(version,
694
self.source.get_parents(version),
695
self.source.get_lines(version),
696
parent_texts=parent_texts)
697
parent_texts[version] = parent_text
698
#delta_parent, sha1, noeol, delta = deltas[version]
699
#target.add_delta(version,
700
# self.source.get_parents(version),
705
#target.get_lines(version)
707
# this should hit the native code path for target
708
if target is not self.target:
709
return self.target.join(temp_source,
717
def _get_source_version_ids(self, version_ids, ignore_missing):
718
"""Determine the version ids to be used from self.source.
720
:param version_ids: The caller-supplied version ids to check. (None
721
for all). If None is in version_ids, it is stripped.
722
:param ignore_missing: if True, remove missing ids from the version
723
list. If False, raise RevisionNotPresent on
724
a missing version id.
725
:return: A set of version ids.
727
if version_ids is None:
728
# None cannot be in source.versions
729
return set(self.source.versions())
731
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
733
return set(self.source.versions()).intersection(set(version_ids))
735
new_version_ids = set()
736
for version in version_ids:
739
if not self.source.has_version(version):
740
raise errors.RevisionNotPresent(version, str(self.source))
742
new_version_ids.add(version)
743
return new_version_ids
added
removed
bzrlib/versionedfile.py
