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
:return: An opaque representation of the inserted version which can be
143
provided back to future add_lines calls in the parent_texts
146
version_id = osutils.safe_revision_id(version_id)
147
parents = [osutils.safe_revision_id(v) for v in parents]
148
self._check_write_ok()
149
return self._add_lines(version_id, parents, lines, parent_texts,
150
left_matching_blocks)
152
def _add_lines(self, version_id, parents, lines, parent_texts,
153
left_matching_blocks):
154
"""Helper to do the class specific add_lines."""
155
raise NotImplementedError(self.add_lines)
157
def add_lines_with_ghosts(self, version_id, parents, lines,
159
"""Add lines to the versioned file, allowing ghosts to be present.
161
This takes the same parameters as add_lines.
163
version_id = osutils.safe_revision_id(version_id)
164
parents = [osutils.safe_revision_id(v) for v in parents]
165
self._check_write_ok()
166
return self._add_lines_with_ghosts(version_id, parents, lines,
169
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
170
"""Helper to do class specific add_lines_with_ghosts."""
171
raise NotImplementedError(self.add_lines_with_ghosts)
173
def check(self, progress_bar=None):
174
"""Check the versioned file for integrity."""
175
raise NotImplementedError(self.check)
177
def _check_lines_not_unicode(self, lines):
178
"""Check that lines being added to a versioned file are not unicode."""
180
if line.__class__ is not str:
181
raise errors.BzrBadParameterUnicode("lines")
183
def _check_lines_are_lines(self, lines):
184
"""Check that the lines really are full lines without inline EOL."""
186
if '\n' in line[:-1]:
187
raise errors.BzrBadParameterContainsNewline("lines")
189
def _check_write_ok(self):
190
"""Is the versioned file marked as 'finished' ? Raise if it is."""
192
raise errors.OutSideTransaction()
193
if self._access_mode != 'w':
194
raise errors.ReadOnlyObjectDirtiedError(self)
196
def enable_cache(self):
197
"""Tell this versioned file that it should cache any data it reads.
199
This is advisory, implementations do not have to support caching.
203
def clear_cache(self):
204
"""Remove any data cached in the versioned file object.
206
This only needs to be supported if caches are supported
210
def clone_text(self, new_version_id, old_version_id, parents):
211
"""Add an identical text to old_version_id as new_version_id.
213
Must raise RevisionNotPresent if the old version or any of the
214
parents are not present in file history.
216
Must raise RevisionAlreadyPresent if the new version is
217
already present in file history."""
218
new_version_id = osutils.safe_revision_id(new_version_id)
219
old_version_id = osutils.safe_revision_id(old_version_id)
220
self._check_write_ok()
221
return self._clone_text(new_version_id, old_version_id, parents)
223
def _clone_text(self, new_version_id, old_version_id, parents):
224
"""Helper function to do the _clone_text work."""
225
raise NotImplementedError(self.clone_text)
227
def create_empty(self, name, transport, mode=None):
228
"""Create a new versioned file of this exact type.
230
:param name: the file name
231
:param transport: the transport
232
:param mode: optional file mode.
234
raise NotImplementedError(self.create_empty)
236
def fix_parents(self, version_id, new_parents):
237
"""Fix the parents list for version.
239
This is done by appending a new version to the index
240
with identical data except for the parents list.
241
the parents list must be a superset of the current
244
version_id = osutils.safe_revision_id(version_id)
245
new_parents = [osutils.safe_revision_id(p) for p in new_parents]
246
self._check_write_ok()
247
return self._fix_parents(version_id, new_parents)
249
def _fix_parents(self, version_id, new_parents):
250
"""Helper for fix_parents."""
251
raise NotImplementedError(self.fix_parents)
253
def get_delta(self, version):
254
"""Get a delta for constructing version from some other version.
256
:return: (delta_parent, sha1, noeol, delta)
257
Where delta_parent is a version id or None to indicate no parent.
259
raise NotImplementedError(self.get_delta)
261
def get_deltas(self, version_ids):
262
"""Get multiple deltas at once for constructing versions.
264
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
265
Where delta_parent is a version id or None to indicate no parent, and
266
version_id is the version_id created by that delta.
269
for version_id in version_ids:
270
result[version_id] = self.get_delta(version_id)
273
def make_mpdiffs(self, version_ids):
274
"""Create multiparent diffs for specified versions"""
275
knit_versions = set()
276
for version_id in version_ids:
277
knit_versions.add(version_id)
278
knit_versions.update(self.get_parents(version_id))
279
lines = dict(zip(knit_versions,
280
self._get_lf_split_line_list(knit_versions)))
282
for version_id in version_ids:
283
target = lines[version_id]
284
parents = [lines[p] for p in self.get_parents(version_id)]
286
left_parent_blocks = self._extract_blocks(version_id,
289
left_parent_blocks = None
290
diffs.append(multiparent.MultiParent.from_lines(target, parents,
294
def _extract_blocks(self, version_id, source, target):
297
def add_mpdiffs(self, records):
298
"""Add mpdiffs to this versionedfile
300
Records should be iterables of version, parents, expected_sha1,
301
mpdiff. mpdiff should be a MultiParent instance.
304
mpvf = multiparent.MultiMemoryVersionedFile()
306
for version, parent_ids, expected_sha1, mpdiff in records:
307
versions.append(version)
308
mpvf.add_diff(mpdiff, version, parent_ids)
309
needed_parents = set()
310
for version, parent_ids, expected_sha1, mpdiff in records:
311
needed_parents.update(p for p in parent_ids
312
if not mpvf.has_version(p))
313
for parent_id, lines in zip(needed_parents,
314
self._get_lf_split_line_list(needed_parents)):
315
mpvf.add_version(lines, parent_id, [])
316
for (version, parent_ids, expected_sha1, mpdiff), lines in\
317
zip(records, mpvf.get_line_list(versions)):
318
if len(parent_ids) == 1:
319
left_matching_blocks = list(mpdiff.get_matching_blocks(0,
320
mpvf.get_diff(parent_ids[0]).num_lines()))
322
left_matching_blocks = None
323
version_text = self.add_lines(version, parent_ids, lines,
324
vf_parents, left_matching_blocks = left_matching_blocks)
325
vf_parents[version] = version_text
326
for (version, parent_ids, expected_sha1, mpdiff), sha1 in\
327
zip(records, self.get_sha1s(versions)):
328
if expected_sha1 != sha1:
329
raise errors.VersionedFileInvalidChecksum(version)
331
def get_sha1(self, version_id):
332
"""Get the stored sha1 sum for the given revision.
334
:param name: The name of the version to lookup
336
raise NotImplementedError(self.get_sha1)
338
def get_sha1s(self, version_ids):
339
"""Get the stored sha1 sums for the given revisions.
341
:param version_ids: The names of the versions to lookup
342
:return: a list of sha1s in order according to the version_ids
344
raise NotImplementedError(self.get_sha1)
346
def get_suffixes(self):
347
"""Return the file suffixes associated with this versioned file."""
348
raise NotImplementedError(self.get_suffixes)
350
def get_text(self, version_id):
351
"""Return version contents as a text string.
353
Raises RevisionNotPresent if version is not present in
356
return ''.join(self.get_lines(version_id))
357
get_string = get_text
359
def get_texts(self, version_ids):
360
"""Return the texts of listed versions as a list of strings.
362
Raises RevisionNotPresent if version is not present in
365
return [''.join(self.get_lines(v)) for v in version_ids]
367
def get_lines(self, version_id):
368
"""Return version contents as a sequence of lines.
370
Raises RevisionNotPresent if version is not present in
373
raise NotImplementedError(self.get_lines)
375
def _get_lf_split_line_list(self, version_ids):
376
return [StringIO(t).readlines() for t in self.get_texts(version_ids)]
378
def get_ancestry(self, version_ids, topo_sorted=True):
379
"""Return a list of all ancestors of given version(s). This
380
will not include the null revision.
382
This list will not be topologically sorted if topo_sorted=False is
385
Must raise RevisionNotPresent if any of the given versions are
386
not present in file history."""
387
if isinstance(version_ids, basestring):
388
version_ids = [version_ids]
389
raise NotImplementedError(self.get_ancestry)
391
def get_ancestry_with_ghosts(self, version_ids):
392
"""Return a list of all ancestors of given version(s). This
393
will not include the null revision.
395
Must raise RevisionNotPresent if any of the given versions are
396
not present in file history.
398
Ghosts that are known about will be included in ancestry list,
399
but are not explicitly marked.
401
raise NotImplementedError(self.get_ancestry_with_ghosts)
403
def get_graph(self, version_ids=None):
404
"""Return a graph from the versioned file.
406
Ghosts are not listed or referenced in the graph.
407
:param version_ids: Versions to select.
408
None means retrieve all versions.
411
if version_ids is None:
412
for version in self.versions():
413
result[version] = self.get_parents(version)
415
pending = set(osutils.safe_revision_id(v) for v in version_ids)
417
version = pending.pop()
418
if version in result:
420
parents = self.get_parents(version)
421
for parent in parents:
425
result[version] = 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 transaction_finished(self):
522
"""The transaction that this file was opened in has finished.
524
This records self.finished = True and should cause all mutating
529
@deprecated_method(zero_eight)
530
def walk(self, version_ids=None):
531
"""Walk the versioned file as a weave-like structure, for
532
versions relative to version_ids. Yields sequence of (lineno,
533
insert, deletes, text) for each relevant line.
535
Must raise RevisionNotPresent if any of the specified versions
536
are not present in the file history.
538
:param version_ids: the version_ids to walk with respect to. If not
539
supplied the entire weave-like structure is walked.
541
walk is deprecated in favour of iter_lines_added_or_present_in_versions
543
raise NotImplementedError(self.walk)
545
@deprecated_method(zero_eight)
546
def iter_names(self):
547
"""Walk the names list."""
548
return iter(self.versions())
550
def plan_merge(self, ver_a, ver_b):
551
"""Return pseudo-annotation indicating how the two versions merge.
553
This is computed between versions a and b and their common
556
Weave lines present in none of them are skipped entirely.
559
killed-base Dead in base revision
560
killed-both Killed in each revision
563
unchanged Alive in both a and b (possibly created in both)
566
ghost-a Killed in a, unborn in b
567
ghost-b Killed in b, unborn in a
568
irrelevant Not in either revision
570
raise NotImplementedError(VersionedFile.plan_merge)
572
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
573
b_marker=TextMerge.B_MARKER):
574
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
577
class PlanWeaveMerge(TextMerge):
578
"""Weave merge that takes a plan as its input.
580
This exists so that VersionedFile.plan_merge is implementable.
581
Most callers will want to use WeaveMerge instead.
584
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
585
b_marker=TextMerge.B_MARKER):
586
TextMerge.__init__(self, a_marker, b_marker)
589
def _merge_struct(self):
594
def outstanding_struct():
595
if not lines_a and not lines_b:
597
elif ch_a and not ch_b:
600
elif ch_b and not ch_a:
602
elif lines_a == lines_b:
605
yield (lines_a, lines_b)
607
# We previously considered either 'unchanged' or 'killed-both' lines
608
# to be possible places to resynchronize. However, assuming agreement
609
# on killed-both lines may be too aggressive. -- mbp 20060324
610
for state, line in self.plan:
611
if state == 'unchanged':
612
# resync and flush queued conflicts changes if any
613
for struct in outstanding_struct():
619
if state == 'unchanged':
622
elif state == 'killed-a':
625
elif state == 'killed-b':
628
elif state == 'new-a':
631
elif state == 'new-b':
635
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
636
'killed-base', 'killed-both'), state
637
for struct in outstanding_struct():
641
class WeaveMerge(PlanWeaveMerge):
642
"""Weave merge that takes a VersionedFile and two versions as its input"""
644
def __init__(self, versionedfile, ver_a, ver_b,
645
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
646
plan = versionedfile.plan_merge(ver_a, ver_b)
647
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
650
class InterVersionedFile(InterObject):
651
"""This class represents operations taking place between two versionedfiles..
653
Its instances have methods like join, and contain
654
references to the source and target versionedfiles these operations can be
657
Often we will provide convenience methods on 'versionedfile' which carry out
658
operations with another versionedfile - they will always forward to
659
InterVersionedFile.get(other).method_name(parameters).
663
"""The available optimised InterVersionedFile types."""
665
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
666
"""Integrate versions from self.source into self.target.
668
If version_ids is None all versions from source should be
669
incorporated into this versioned file.
671
Must raise RevisionNotPresent if any of the specified versions
672
are not present in the other files history unless ignore_missing is
673
supplied when they are silently skipped.
676
# - if the target is empty, just add all the versions from
677
# source to target, otherwise:
678
# - make a temporary versioned file of type target
679
# - insert the source content into it one at a time
681
if not self.target.versions():
684
# Make a new target-format versioned file.
685
temp_source = self.target.create_empty("temp", MemoryTransport())
687
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
688
graph = self.source.get_graph(version_ids)
689
order = tsort.topo_sort(graph.items())
690
pb = ui.ui_factory.nested_progress_bar()
693
# TODO for incremental cross-format work:
694
# make a versioned file with the following content:
695
# all revisions we have been asked to join
696
# all their ancestors that are *not* in target already.
697
# the immediate parents of the above two sets, with
698
# empty parent lists - these versions are in target already
699
# and the incorrect version data will be ignored.
700
# TODO: for all ancestors that are present in target already,
701
# check them for consistent data, this requires moving sha1 from
703
# TODO: remove parent texts when they are not relevant any more for
704
# memory pressure reduction. RBC 20060313
705
# pb.update('Converting versioned data', 0, len(order))
706
# deltas = self.source.get_deltas(order)
707
for index, version in enumerate(order):
708
pb.update('Converting versioned data', index, len(order))
709
parent_text = target.add_lines(version,
710
self.source.get_parents(version),
711
self.source.get_lines(version),
712
parent_texts=parent_texts)
713
parent_texts[version] = parent_text
714
#delta_parent, sha1, noeol, delta = deltas[version]
715
#target.add_delta(version,
716
# self.source.get_parents(version),
721
#target.get_lines(version)
723
# this should hit the native code path for target
724
if target is not self.target:
725
return self.target.join(temp_source,
733
def _get_source_version_ids(self, version_ids, ignore_missing):
734
"""Determine the version ids to be used from self.source.
736
:param version_ids: The caller-supplied version ids to check. (None
737
for all). If None is in version_ids, it is stripped.
738
:param ignore_missing: if True, remove missing ids from the version
739
list. If False, raise RevisionNotPresent on
740
a missing version id.
741
:return: A set of version ids.
743
if version_ids is None:
744
# None cannot be in source.versions
745
return set(self.source.versions())
747
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
749
return set(self.source.versions()).intersection(set(version_ids))
751
new_version_ids = set()
752
for version in version_ids:
755
if not self.source.has_version(version):
756
raise errors.RevisionNotPresent(version, str(self.source))
758
new_version_ids.add(version)
759
return new_version_ids
added
removed
bzrlib/versionedfile.py
