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
40
from bzrlib.symbol_versioning import (deprecated_function,
46
class VersionedFile(object):
47
"""Versioned text file storage.
49
A versioned file manages versions of line-based text files,
50
keeping track of the originating version for each line.
52
To clients the "lines" of the file are represented as a list of
53
strings. These strings will typically have terminal newline
54
characters, but this is not required. In particular files commonly
55
do not have a newline at the end of the file.
57
Texts are identified by a version-id string.
60
def __init__(self, access_mode):
62
self._access_mode = access_mode
65
def check_not_reserved_id(version_id):
66
revision.check_not_reserved_id(version_id)
68
def copy_to(self, name, transport):
69
"""Copy this versioned file to name on transport."""
70
raise NotImplementedError(self.copy_to)
72
@deprecated_method(zero_eight)
74
"""Return a list of all the versions in this versioned file.
76
Please use versionedfile.versions() now.
78
return self.versions()
81
"""Return a unsorted list of versions."""
82
raise NotImplementedError(self.versions)
84
def has_ghost(self, version_id):
85
"""Returns whether version is present as a ghost."""
86
raise NotImplementedError(self.has_ghost)
88
def has_version(self, version_id):
89
"""Returns whether version is present."""
90
raise NotImplementedError(self.has_version)
92
def add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
93
"""Add a text to the versioned file via a pregenerated delta.
95
:param version_id: The version id being added.
96
:param parents: The parents of the version_id.
97
:param delta_parent: The parent this delta was created against.
98
:param sha1: The sha1 of the full text.
99
:param delta: The delta instructions. See get_delta for details.
101
version_id = osutils.safe_revision_id(version_id)
102
parents = [osutils.safe_revision_id(v) for v in parents]
103
self._check_write_ok()
104
if self.has_version(version_id):
105
raise errors.RevisionAlreadyPresent(version_id, self)
106
return self._add_delta(version_id, parents, delta_parent, sha1, noeol, delta)
108
def _add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
109
"""Class specific routine to add a delta.
111
This generic version simply applies the delta to the delta_parent and
114
# strip annotation from delta
116
for start, stop, delta_len, delta_lines in delta:
117
new_delta.append((start, stop, delta_len, [text for origin, text in delta_lines]))
118
if delta_parent is not None:
119
parent_full = self.get_lines(delta_parent)
122
new_full = self._apply_delta(parent_full, new_delta)
123
# its impossible to have noeol on an empty file
124
if noeol and new_full[-1][-1] == '\n':
125
new_full[-1] = new_full[-1][:-1]
126
self.add_lines(version_id, parents, new_full)
128
def add_lines(self, version_id, parents, lines, parent_texts=None,
129
left_matching_blocks=None):
130
"""Add a single text on top of the versioned file.
132
Must raise RevisionAlreadyPresent if the new version is
133
already present in file history.
135
Must raise RevisionNotPresent if any of the given parents are
136
not present in file history.
137
:param parent_texts: An optional dictionary containing the opaque
138
representations of some or all of the parents of
139
version_id to allow delta optimisations.
140
VERY IMPORTANT: the texts must be those returned
141
by add_lines or data corruption can be caused.
142
:param left_matching_blocks: a hint about which areas are common
143
between the text and its left-hand-parent. The format is
144
the SequenceMatcher.get_matching_blocks format.
145
:return: An opaque representation of the inserted version which can be
146
provided back to future add_lines calls in the parent_texts
149
version_id = osutils.safe_revision_id(version_id)
150
parents = [osutils.safe_revision_id(v) for v in parents]
151
self._check_write_ok()
152
return self._add_lines(version_id, parents, lines, parent_texts,
153
left_matching_blocks)
155
def _add_lines(self, version_id, parents, lines, parent_texts,
156
left_matching_blocks):
157
"""Helper to do the class specific add_lines."""
158
raise NotImplementedError(self.add_lines)
160
def add_lines_with_ghosts(self, version_id, parents, lines,
162
"""Add lines to the versioned file, allowing ghosts to be present.
164
This takes the same parameters as add_lines.
166
version_id = osutils.safe_revision_id(version_id)
167
parents = [osutils.safe_revision_id(v) for v in parents]
168
self._check_write_ok()
169
return self._add_lines_with_ghosts(version_id, parents, lines,
172
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
173
"""Helper to do class specific add_lines_with_ghosts."""
174
raise NotImplementedError(self.add_lines_with_ghosts)
176
def check(self, progress_bar=None):
177
"""Check the versioned file for integrity."""
178
raise NotImplementedError(self.check)
180
def _check_lines_not_unicode(self, lines):
181
"""Check that lines being added to a versioned file are not unicode."""
183
if line.__class__ is not str:
184
raise errors.BzrBadParameterUnicode("lines")
186
def _check_lines_are_lines(self, lines):
187
"""Check that the lines really are full lines without inline EOL."""
189
if '\n' in line[:-1]:
190
raise errors.BzrBadParameterContainsNewline("lines")
192
def _check_write_ok(self):
193
"""Is the versioned file marked as 'finished' ? Raise if it is."""
195
raise errors.OutSideTransaction()
196
if self._access_mode != 'w':
197
raise errors.ReadOnlyObjectDirtiedError(self)
199
def enable_cache(self):
200
"""Tell this versioned file that it should cache any data it reads.
202
This is advisory, implementations do not have to support caching.
206
def clear_cache(self):
207
"""Remove any data cached in the versioned file object.
209
This only needs to be supported if caches are supported
213
def clone_text(self, new_version_id, old_version_id, parents):
214
"""Add an identical text to old_version_id as new_version_id.
216
Must raise RevisionNotPresent if the old version or any of the
217
parents are not present in file history.
219
Must raise RevisionAlreadyPresent if the new version is
220
already present in file history."""
221
new_version_id = osutils.safe_revision_id(new_version_id)
222
old_version_id = osutils.safe_revision_id(old_version_id)
223
self._check_write_ok()
224
return self._clone_text(new_version_id, old_version_id, parents)
226
def _clone_text(self, new_version_id, old_version_id, parents):
227
"""Helper function to do the _clone_text work."""
228
raise NotImplementedError(self.clone_text)
230
def create_empty(self, name, transport, mode=None):
231
"""Create a new versioned file of this exact type.
233
:param name: the file name
234
:param transport: the transport
235
:param mode: optional file mode.
237
raise NotImplementedError(self.create_empty)
239
def fix_parents(self, version_id, new_parents):
240
"""Fix the parents list for version.
242
This is done by appending a new version to the index
243
with identical data except for the parents list.
244
the parents list must be a superset of the current
247
version_id = osutils.safe_revision_id(version_id)
248
new_parents = [osutils.safe_revision_id(p) for p in new_parents]
249
self._check_write_ok()
250
return self._fix_parents(version_id, new_parents)
252
def _fix_parents(self, version_id, new_parents):
253
"""Helper for fix_parents."""
254
raise NotImplementedError(self.fix_parents)
256
def get_delta(self, version):
257
"""Get a delta for constructing version from some other version.
259
:return: (delta_parent, sha1, noeol, delta)
260
Where delta_parent is a version id or None to indicate no parent.
262
raise NotImplementedError(self.get_delta)
264
def get_deltas(self, version_ids):
265
"""Get multiple deltas at once for constructing versions.
267
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
268
Where delta_parent is a version id or None to indicate no parent, and
269
version_id is the version_id created by that delta.
272
for version_id in version_ids:
273
result[version_id] = self.get_delta(version_id)
276
def make_mpdiffs(self, version_ids):
277
"""Create multiparent diffs for specified versions"""
278
knit_versions = set()
279
for version_id in version_ids:
280
knit_versions.add(version_id)
281
knit_versions.update(self.get_parents(version_id))
282
lines = dict(zip(knit_versions,
283
self._get_lf_split_line_list(knit_versions)))
285
for version_id in version_ids:
286
target = lines[version_id]
287
parents = [lines[p] for p in self.get_parents(version_id)]
289
left_parent_blocks = self._extract_blocks(version_id,
292
left_parent_blocks = None
293
diffs.append(multiparent.MultiParent.from_lines(target, parents,
297
def _extract_blocks(self, version_id, source, target):
300
def add_mpdiffs(self, records):
301
"""Add mpdiffs to this versionedfile
303
Records should be iterables of version, parents, expected_sha1,
304
mpdiff. mpdiff should be a MultiParent instance.
307
mpvf = multiparent.MultiMemoryVersionedFile()
309
for version, parent_ids, expected_sha1, mpdiff in records:
310
versions.append(version)
311
mpvf.add_diff(mpdiff, version, parent_ids)
312
needed_parents = set()
313
for version, parent_ids, expected_sha1, mpdiff in records:
314
needed_parents.update(p for p in parent_ids
315
if not mpvf.has_version(p))
316
for parent_id, lines in zip(needed_parents,
317
self._get_lf_split_line_list(needed_parents)):
318
mpvf.add_version(lines, parent_id, [])
319
for (version, parent_ids, expected_sha1, mpdiff), lines in\
320
zip(records, mpvf.get_line_list(versions)):
321
if len(parent_ids) == 1:
322
left_matching_blocks = list(mpdiff.get_matching_blocks(0,
323
mpvf.get_diff(parent_ids[0]).num_lines()))
325
left_matching_blocks = None
326
version_text = self.add_lines(version, parent_ids, lines,
327
vf_parents, left_matching_blocks=left_matching_blocks)
328
vf_parents[version] = version_text
329
for (version, parent_ids, expected_sha1, mpdiff), sha1 in\
330
zip(records, self.get_sha1s(versions)):
331
if expected_sha1 != sha1:
332
raise errors.VersionedFileInvalidChecksum(version)
334
def get_sha1(self, version_id):
335
"""Get the stored sha1 sum for the given revision.
337
:param name: The name of the version to lookup
339
raise NotImplementedError(self.get_sha1)
341
def get_sha1s(self, version_ids):
342
"""Get the stored sha1 sums for the given revisions.
344
:param version_ids: The names of the versions to lookup
345
:return: a list of sha1s in order according to the version_ids
347
raise NotImplementedError(self.get_sha1)
349
def get_suffixes(self):
350
"""Return the file suffixes associated with this versioned file."""
351
raise NotImplementedError(self.get_suffixes)
353
def get_text(self, version_id):
354
"""Return version contents as a text string.
356
Raises RevisionNotPresent if version is not present in
359
return ''.join(self.get_lines(version_id))
360
get_string = get_text
362
def get_texts(self, version_ids):
363
"""Return the texts of listed versions as a list of strings.
365
Raises RevisionNotPresent if version is not present in
368
return [''.join(self.get_lines(v)) for v in version_ids]
370
def get_lines(self, version_id):
371
"""Return version contents as a sequence of lines.
373
Raises RevisionNotPresent if version is not present in
376
raise NotImplementedError(self.get_lines)
378
def _get_lf_split_line_list(self, version_ids):
379
return [StringIO(t).readlines() for t in self.get_texts(version_ids)]
381
def get_ancestry(self, version_ids, topo_sorted=True):
382
"""Return a list of all ancestors of given version(s). This
383
will not include the null revision.
385
This list will not be topologically sorted if topo_sorted=False is
388
Must raise RevisionNotPresent if any of the given versions are
389
not present in file history."""
390
if isinstance(version_ids, basestring):
391
version_ids = [version_ids]
392
raise NotImplementedError(self.get_ancestry)
394
def get_ancestry_with_ghosts(self, version_ids):
395
"""Return a list of all ancestors of given version(s). This
396
will not include the null revision.
398
Must raise RevisionNotPresent if any of the given versions are
399
not present in file history.
401
Ghosts that are known about will be included in ancestry list,
402
but are not explicitly marked.
404
raise NotImplementedError(self.get_ancestry_with_ghosts)
406
def get_graph(self, version_ids=None):
407
"""Return a graph from the versioned file.
409
Ghosts are not listed or referenced in the graph.
410
:param version_ids: Versions to select.
411
None means retrieve all versions.
413
if version_ids is None:
414
return dict(self.iter_parents(self.versions()))
416
pending = set(osutils.safe_revision_id(v) for v in version_ids)
418
this_iteration = pending
420
for version, parents in self.iter_parents(this_iteration):
421
result[version] = parents
422
for parent in parents:
428
def get_graph_with_ghosts(self):
429
"""Return a graph for the entire versioned file.
431
Ghosts are referenced in parents list but are not
434
raise NotImplementedError(self.get_graph_with_ghosts)
436
@deprecated_method(zero_eight)
437
def parent_names(self, version):
438
"""Return version names for parents of a version.
440
See get_parents for the current api.
442
return self.get_parents(version)
444
def get_parents(self, version_id):
445
"""Return version names for parents of a version.
447
Must raise RevisionNotPresent if version is not present in
450
raise NotImplementedError(self.get_parents)
452
def get_parents_with_ghosts(self, version_id):
453
"""Return version names for parents of version_id.
455
Will raise RevisionNotPresent if version_id is not present
458
Ghosts that are known about will be included in the parent list,
459
but are not explicitly marked.
461
raise NotImplementedError(self.get_parents_with_ghosts)
463
def annotate_iter(self, version_id):
464
"""Yield list of (version-id, line) pairs for the specified
467
Must raise RevisionNotPresent if any of the given versions are
468
not present in file history.
470
raise NotImplementedError(self.annotate_iter)
472
def annotate(self, version_id):
473
return list(self.annotate_iter(version_id))
475
def _apply_delta(self, lines, delta):
476
"""Apply delta to lines."""
479
for start, end, count, delta_lines in delta:
480
lines[offset+start:offset+end] = delta_lines
481
offset = offset + (start - end) + count
484
def join(self, other, pb=None, msg=None, version_ids=None,
485
ignore_missing=False):
486
"""Integrate versions from other into this versioned file.
488
If version_ids is None all versions from other should be
489
incorporated into this versioned file.
491
Must raise RevisionNotPresent if any of the specified versions
492
are not present in the other files history unless ignore_missing
493
is supplied when they are silently skipped.
495
self._check_write_ok()
496
return InterVersionedFile.get(other, self).join(
502
def iter_lines_added_or_present_in_versions(self, version_ids=None,
504
"""Iterate over the lines in the versioned file from version_ids.
506
This may return lines from other versions, and does not return the
507
specific version marker at this point. The api may be changed
508
during development to include the version that the versioned file
509
thinks is relevant, but given that such hints are just guesses,
510
its better not to have it if we don't need it.
512
If a progress bar is supplied, it may be used to indicate progress.
513
The caller is responsible for cleaning up progress bars (because this
516
NOTES: Lines are normalised: they will all have \n terminators.
517
Lines are returned in arbitrary order.
519
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
521
def iter_parents(self, version_ids):
522
"""Iterate through the parents for many version ids.
524
:param version_ids: An iterable yielding version_ids.
525
:return: An iterator that yields (version_id, parents). Requested
526
version_ids not present in the versioned file are simply skipped.
527
The order is undefined, allowing for different optimisations in
528
the underlying implementation.
530
for version_id in version_ids:
532
yield version_id, tuple(self.get_parents(version_id))
533
except errors.RevisionNotPresent:
536
def transaction_finished(self):
537
"""The transaction that this file was opened in has finished.
539
This records self.finished = True and should cause all mutating
544
@deprecated_method(zero_eight)
545
def walk(self, version_ids=None):
546
"""Walk the versioned file as a weave-like structure, for
547
versions relative to version_ids. Yields sequence of (lineno,
548
insert, deletes, text) for each relevant line.
550
Must raise RevisionNotPresent if any of the specified versions
551
are not present in the file history.
553
:param version_ids: the version_ids to walk with respect to. If not
554
supplied the entire weave-like structure is walked.
556
walk is deprecated in favour of iter_lines_added_or_present_in_versions
558
raise NotImplementedError(self.walk)
560
@deprecated_method(zero_eight)
561
def iter_names(self):
562
"""Walk the names list."""
563
return iter(self.versions())
565
def plan_merge(self, ver_a, ver_b):
566
"""Return pseudo-annotation indicating how the two versions merge.
568
This is computed between versions a and b and their common
571
Weave lines present in none of them are skipped entirely.
574
killed-base Dead in base revision
575
killed-both Killed in each revision
578
unchanged Alive in both a and b (possibly created in both)
581
ghost-a Killed in a, unborn in b
582
ghost-b Killed in b, unborn in a
583
irrelevant Not in either revision
585
raise NotImplementedError(VersionedFile.plan_merge)
587
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
588
b_marker=TextMerge.B_MARKER):
589
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
592
class PlanWeaveMerge(TextMerge):
593
"""Weave merge that takes a plan as its input.
595
This exists so that VersionedFile.plan_merge is implementable.
596
Most callers will want to use WeaveMerge instead.
599
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
600
b_marker=TextMerge.B_MARKER):
601
TextMerge.__init__(self, a_marker, b_marker)
604
def _merge_struct(self):
609
def outstanding_struct():
610
if not lines_a and not lines_b:
612
elif ch_a and not ch_b:
615
elif ch_b and not ch_a:
617
elif lines_a == lines_b:
620
yield (lines_a, lines_b)
622
# We previously considered either 'unchanged' or 'killed-both' lines
623
# to be possible places to resynchronize. However, assuming agreement
624
# on killed-both lines may be too aggressive. -- mbp 20060324
625
for state, line in self.plan:
626
if state == 'unchanged':
627
# resync and flush queued conflicts changes if any
628
for struct in outstanding_struct():
634
if state == 'unchanged':
637
elif state == 'killed-a':
640
elif state == 'killed-b':
643
elif state == 'new-a':
646
elif state == 'new-b':
650
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
651
'killed-base', 'killed-both'), state
652
for struct in outstanding_struct():
656
class WeaveMerge(PlanWeaveMerge):
657
"""Weave merge that takes a VersionedFile and two versions as its input"""
659
def __init__(self, versionedfile, ver_a, ver_b,
660
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
661
plan = versionedfile.plan_merge(ver_a, ver_b)
662
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
665
class InterVersionedFile(InterObject):
666
"""This class represents operations taking place between two versionedfiles..
668
Its instances have methods like join, and contain
669
references to the source and target versionedfiles these operations can be
672
Often we will provide convenience methods on 'versionedfile' which carry out
673
operations with another versionedfile - they will always forward to
674
InterVersionedFile.get(other).method_name(parameters).
678
"""The available optimised InterVersionedFile types."""
680
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
681
"""Integrate versions from self.source into self.target.
683
If version_ids is None all versions from source should be
684
incorporated into this versioned file.
686
Must raise RevisionNotPresent if any of the specified versions
687
are not present in the other files history unless ignore_missing is
688
supplied when they are silently skipped.
691
# - if the target is empty, just add all the versions from
692
# source to target, otherwise:
693
# - make a temporary versioned file of type target
694
# - insert the source content into it one at a time
696
if not self.target.versions():
699
# Make a new target-format versioned file.
700
temp_source = self.target.create_empty("temp", MemoryTransport())
702
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
703
graph = self.source.get_graph(version_ids)
704
order = tsort.topo_sort(graph.items())
705
pb = ui.ui_factory.nested_progress_bar()
708
# TODO for incremental cross-format work:
709
# make a versioned file with the following content:
710
# all revisions we have been asked to join
711
# all their ancestors that are *not* in target already.
712
# the immediate parents of the above two sets, with
713
# empty parent lists - these versions are in target already
714
# and the incorrect version data will be ignored.
715
# TODO: for all ancestors that are present in target already,
716
# check them for consistent data, this requires moving sha1 from
718
# TODO: remove parent texts when they are not relevant any more for
719
# memory pressure reduction. RBC 20060313
720
# pb.update('Converting versioned data', 0, len(order))
721
# deltas = self.source.get_deltas(order)
722
for index, version in enumerate(order):
723
pb.update('Converting versioned data', index, len(order))
724
parent_text = target.add_lines(version,
725
self.source.get_parents(version),
726
self.source.get_lines(version),
727
parent_texts=parent_texts)
728
parent_texts[version] = parent_text
729
#delta_parent, sha1, noeol, delta = deltas[version]
730
#target.add_delta(version,
731
# self.source.get_parents(version),
736
#target.get_lines(version)
738
# this should hit the native code path for target
739
if target is not self.target:
740
return self.target.join(temp_source,
748
def _get_source_version_ids(self, version_ids, ignore_missing):
749
"""Determine the version ids to be used from self.source.
751
:param version_ids: The caller-supplied version ids to check. (None
752
for all). If None is in version_ids, it is stripped.
753
:param ignore_missing: if True, remove missing ids from the version
754
list. If False, raise RevisionNotPresent on
755
a missing version id.
756
:return: A set of version ids.
758
if version_ids is None:
759
# None cannot be in source.versions
760
return set(self.source.versions())
762
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
764
return set(self.source.versions()).intersection(set(version_ids))
766
new_version_ids = set()
767
for version in version_ids:
770
if not self.source.has_version(version):
771
raise errors.RevisionNotPresent(version, str(self.source))
773
new_version_ids.add(version)
774
return new_version_ids
added
removed
bzrlib/versionedfile.py
