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
"""Add a single text on top of the versioned file.
131
Must raise RevisionAlreadyPresent if the new version is
132
already present in file history.
134
Must raise RevisionNotPresent if any of the given parents are
135
not present in file history.
136
:param parent_texts: An optional dictionary containing the opaque
137
representations of some or all of the parents of
138
version_id to allow delta optimisations.
139
VERY IMPORTANT: the texts must be those returned
140
by add_lines or data corruption can be caused.
141
:return: An opaque representation of the inserted version which can be
142
provided back to future add_lines calls in the parent_texts
145
version_id = osutils.safe_revision_id(version_id)
146
parents = [osutils.safe_revision_id(v) for v in parents]
147
self._check_write_ok()
148
return self._add_lines(version_id, parents, lines, parent_texts)
150
def _add_lines(self, version_id, parents, lines, parent_texts):
151
"""Helper to do the class specific add_lines."""
152
raise NotImplementedError(self.add_lines)
154
def add_lines_with_ghosts(self, version_id, parents, lines,
156
"""Add lines to the versioned file, allowing ghosts to be present.
158
This takes the same parameters as add_lines.
160
version_id = osutils.safe_revision_id(version_id)
161
parents = [osutils.safe_revision_id(v) for v in parents]
162
self._check_write_ok()
163
return self._add_lines_with_ghosts(version_id, parents, lines,
166
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
167
"""Helper to do class specific add_lines_with_ghosts."""
168
raise NotImplementedError(self.add_lines_with_ghosts)
170
def check(self, progress_bar=None):
171
"""Check the versioned file for integrity."""
172
raise NotImplementedError(self.check)
174
def _check_lines_not_unicode(self, lines):
175
"""Check that lines being added to a versioned file are not unicode."""
177
if line.__class__ is not str:
178
raise errors.BzrBadParameterUnicode("lines")
180
def _check_lines_are_lines(self, lines):
181
"""Check that the lines really are full lines without inline EOL."""
183
if '\n' in line[:-1]:
184
raise errors.BzrBadParameterContainsNewline("lines")
186
def _check_write_ok(self):
187
"""Is the versioned file marked as 'finished' ? Raise if it is."""
189
raise errors.OutSideTransaction()
190
if self._access_mode != 'w':
191
raise errors.ReadOnlyObjectDirtiedError(self)
193
def enable_cache(self):
194
"""Tell this versioned file that it should cache any data it reads.
196
This is advisory, implementations do not have to support caching.
200
def clear_cache(self):
201
"""Remove any data cached in the versioned file object.
203
This only needs to be supported if caches are supported
207
def clone_text(self, new_version_id, old_version_id, parents):
208
"""Add an identical text to old_version_id as new_version_id.
210
Must raise RevisionNotPresent if the old version or any of the
211
parents are not present in file history.
213
Must raise RevisionAlreadyPresent if the new version is
214
already present in file history."""
215
new_version_id = osutils.safe_revision_id(new_version_id)
216
old_version_id = osutils.safe_revision_id(old_version_id)
217
self._check_write_ok()
218
return self._clone_text(new_version_id, old_version_id, parents)
220
def _clone_text(self, new_version_id, old_version_id, parents):
221
"""Helper function to do the _clone_text work."""
222
raise NotImplementedError(self.clone_text)
224
def create_empty(self, name, transport, mode=None):
225
"""Create a new versioned file of this exact type.
227
:param name: the file name
228
:param transport: the transport
229
:param mode: optional file mode.
231
raise NotImplementedError(self.create_empty)
233
def fix_parents(self, version_id, new_parents):
234
"""Fix the parents list for version.
236
This is done by appending a new version to the index
237
with identical data except for the parents list.
238
the parents list must be a superset of the current
241
version_id = osutils.safe_revision_id(version_id)
242
new_parents = [osutils.safe_revision_id(p) for p in new_parents]
243
self._check_write_ok()
244
return self._fix_parents(version_id, new_parents)
246
def _fix_parents(self, version_id, new_parents):
247
"""Helper for fix_parents."""
248
raise NotImplementedError(self.fix_parents)
250
def get_delta(self, version):
251
"""Get a delta for constructing version from some other version.
253
:return: (delta_parent, sha1, noeol, delta)
254
Where delta_parent is a version id or None to indicate no parent.
256
raise NotImplementedError(self.get_delta)
258
def get_deltas(self, version_ids):
259
"""Get multiple deltas at once for constructing versions.
261
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
262
Where delta_parent is a version id or None to indicate no parent, and
263
version_id is the version_id created by that delta.
266
for version_id in version_ids:
267
result[version_id] = self.get_delta(version_id)
270
def make_mpdiffs(self, version_ids):
271
"""Create multiparent diffs for specified versions"""
272
knit_versions = set()
273
for version_id in version_ids:
274
knit_versions.add(version_id)
275
knit_versions.update(self.get_parents(version_id))
276
lines = dict(zip(knit_versions,
277
self._get_lf_split_line_list(knit_versions)))
279
for version_id in version_ids:
280
target = lines[version_id]
281
parents = [lines[p] for p in self.get_parents(version_id)]
283
left_parent_blocks = self._extract_blocks(version_id,
286
left_parent_blocks = None
287
diffs.append(multiparent.MultiParent.from_lines(target, parents,
291
def _extract_blocks(self, version_id, source, target):
294
def add_mpdiffs(self, records):
295
"""Add mpdiffs to this versionedfile
297
Records should be iterables of version, parents, expected_sha1,
298
mpdiff. mpdiff should be a MultiParent instance.
301
for version, parents, expected_sha1, mpdiff in records:
302
mpvf = multiparent.MultiMemoryVersionedFile()
303
needed_parents = [p for p in parents if not mpvf.has_version(p)]
304
parent_lines = self._get_lf_split_line_list(needed_parents)
305
for parent_id, lines in zip(needed_parents, parent_lines):
306
mpvf.add_version(lines, parent_id, [])
307
mpvf.add_diff(mpdiff, version, parents)
308
lines = mpvf.get_line_list([version])[0]
309
version_text = self.add_lines(version, parents, lines, vf_parents)
310
vf_parents[version] = version_text
311
if expected_sha1 != self.get_sha1(version):
312
raise errors.VersionedFileInvalidChecksum(version)
314
def get_sha1(self, version_id):
315
"""Get the stored sha1 sum for the given revision.
317
:param name: The name of the version to lookup
319
raise NotImplementedError(self.get_sha1)
321
def get_sha1s(self, version_ids):
322
"""Get the stored sha1 sums for the given revisions.
324
:param version_ids: The names of the versions to lookup
325
:return: a list of sha1s in order according to the version_ids
327
raise NotImplementedError(self.get_sha1)
329
def get_suffixes(self):
330
"""Return the file suffixes associated with this versioned file."""
331
raise NotImplementedError(self.get_suffixes)
333
def get_text(self, version_id):
334
"""Return version contents as a text string.
336
Raises RevisionNotPresent if version is not present in
339
return ''.join(self.get_lines(version_id))
340
get_string = get_text
342
def get_texts(self, version_ids):
343
"""Return the texts of listed versions as a list of strings.
345
Raises RevisionNotPresent if version is not present in
348
return [''.join(self.get_lines(v)) for v in version_ids]
350
def get_lines(self, version_id):
351
"""Return version contents as a sequence of lines.
353
Raises RevisionNotPresent if version is not present in
356
raise NotImplementedError(self.get_lines)
358
def _get_lf_split_line_list(self, version_ids):
359
return [StringIO(t).readlines() for t in self.get_texts(version_ids)]
361
def get_ancestry(self, version_ids, topo_sorted=True):
362
"""Return a list of all ancestors of given version(s). This
363
will not include the null revision.
365
This list will not be topologically sorted if topo_sorted=False is
368
Must raise RevisionNotPresent if any of the given versions are
369
not present in file history."""
370
if isinstance(version_ids, basestring):
371
version_ids = [version_ids]
372
raise NotImplementedError(self.get_ancestry)
374
def get_ancestry_with_ghosts(self, version_ids):
375
"""Return a list of all ancestors of given version(s). This
376
will not include the null revision.
378
Must raise RevisionNotPresent if any of the given versions are
379
not present in file history.
381
Ghosts that are known about will be included in ancestry list,
382
but are not explicitly marked.
384
raise NotImplementedError(self.get_ancestry_with_ghosts)
386
def get_graph(self, version_ids=None):
387
"""Return a graph from the versioned file.
389
Ghosts are not listed or referenced in the graph.
390
:param version_ids: Versions to select.
391
None means retrieve all versions.
394
if version_ids is None:
395
for version in self.versions():
396
result[version] = self.get_parents(version)
398
pending = set(osutils.safe_revision_id(v) for v in version_ids)
400
version = pending.pop()
401
if version in result:
403
parents = self.get_parents(version)
404
for parent in parents:
408
result[version] = parents
411
def get_graph_with_ghosts(self):
412
"""Return a graph for the entire versioned file.
414
Ghosts are referenced in parents list but are not
417
raise NotImplementedError(self.get_graph_with_ghosts)
419
@deprecated_method(zero_eight)
420
def parent_names(self, version):
421
"""Return version names for parents of a version.
423
See get_parents for the current api.
425
return self.get_parents(version)
427
def get_parents(self, version_id):
428
"""Return version names for parents of a version.
430
Must raise RevisionNotPresent if version is not present in
433
raise NotImplementedError(self.get_parents)
435
def get_parents_with_ghosts(self, version_id):
436
"""Return version names for parents of version_id.
438
Will raise RevisionNotPresent if version_id is not present
441
Ghosts that are known about will be included in the parent list,
442
but are not explicitly marked.
444
raise NotImplementedError(self.get_parents_with_ghosts)
446
def annotate_iter(self, version_id):
447
"""Yield list of (version-id, line) pairs for the specified
450
Must raise RevisionNotPresent if any of the given versions are
451
not present in file history.
453
raise NotImplementedError(self.annotate_iter)
455
def annotate(self, version_id):
456
return list(self.annotate_iter(version_id))
458
def _apply_delta(self, lines, delta):
459
"""Apply delta to lines."""
462
for start, end, count, delta_lines in delta:
463
lines[offset+start:offset+end] = delta_lines
464
offset = offset + (start - end) + count
467
def join(self, other, pb=None, msg=None, version_ids=None,
468
ignore_missing=False):
469
"""Integrate versions from other into this versioned file.
471
If version_ids is None all versions from other should be
472
incorporated into this versioned file.
474
Must raise RevisionNotPresent if any of the specified versions
475
are not present in the other files history unless ignore_missing
476
is supplied when they are silently skipped.
478
self._check_write_ok()
479
return InterVersionedFile.get(other, self).join(
485
def iter_lines_added_or_present_in_versions(self, version_ids=None,
487
"""Iterate over the lines in the versioned file from version_ids.
489
This may return lines from other versions, and does not return the
490
specific version marker at this point. The api may be changed
491
during development to include the version that the versioned file
492
thinks is relevant, but given that such hints are just guesses,
493
its better not to have it if we don't need it.
495
If a progress bar is supplied, it may be used to indicate progress.
496
The caller is responsible for cleaning up progress bars (because this
499
NOTES: Lines are normalised: they will all have \n terminators.
500
Lines are returned in arbitrary order.
502
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
504
def transaction_finished(self):
505
"""The transaction that this file was opened in has finished.
507
This records self.finished = True and should cause all mutating
512
@deprecated_method(zero_eight)
513
def walk(self, version_ids=None):
514
"""Walk the versioned file as a weave-like structure, for
515
versions relative to version_ids. Yields sequence of (lineno,
516
insert, deletes, text) for each relevant line.
518
Must raise RevisionNotPresent if any of the specified versions
519
are not present in the file history.
521
:param version_ids: the version_ids to walk with respect to. If not
522
supplied the entire weave-like structure is walked.
524
walk is deprecated in favour of iter_lines_added_or_present_in_versions
526
raise NotImplementedError(self.walk)
528
@deprecated_method(zero_eight)
529
def iter_names(self):
530
"""Walk the names list."""
531
return iter(self.versions())
533
def plan_merge(self, ver_a, ver_b):
534
"""Return pseudo-annotation indicating how the two versions merge.
536
This is computed between versions a and b and their common
539
Weave lines present in none of them are skipped entirely.
542
killed-base Dead in base revision
543
killed-both Killed in each revision
546
unchanged Alive in both a and b (possibly created in both)
549
ghost-a Killed in a, unborn in b
550
ghost-b Killed in b, unborn in a
551
irrelevant Not in either revision
553
raise NotImplementedError(VersionedFile.plan_merge)
555
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
556
b_marker=TextMerge.B_MARKER):
557
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
560
class PlanWeaveMerge(TextMerge):
561
"""Weave merge that takes a plan as its input.
563
This exists so that VersionedFile.plan_merge is implementable.
564
Most callers will want to use WeaveMerge instead.
567
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
568
b_marker=TextMerge.B_MARKER):
569
TextMerge.__init__(self, a_marker, b_marker)
572
def _merge_struct(self):
577
def outstanding_struct():
578
if not lines_a and not lines_b:
580
elif ch_a and not ch_b:
583
elif ch_b and not ch_a:
585
elif lines_a == lines_b:
588
yield (lines_a, lines_b)
590
# We previously considered either 'unchanged' or 'killed-both' lines
591
# to be possible places to resynchronize. However, assuming agreement
592
# on killed-both lines may be too aggressive. -- mbp 20060324
593
for state, line in self.plan:
594
if state == 'unchanged':
595
# resync and flush queued conflicts changes if any
596
for struct in outstanding_struct():
602
if state == 'unchanged':
605
elif state == 'killed-a':
608
elif state == 'killed-b':
611
elif state == 'new-a':
614
elif state == 'new-b':
618
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
619
'killed-base', 'killed-both'), state
620
for struct in outstanding_struct():
624
class WeaveMerge(PlanWeaveMerge):
625
"""Weave merge that takes a VersionedFile and two versions as its input"""
627
def __init__(self, versionedfile, ver_a, ver_b,
628
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
629
plan = versionedfile.plan_merge(ver_a, ver_b)
630
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
633
class InterVersionedFile(InterObject):
634
"""This class represents operations taking place between two versionedfiles..
636
Its instances have methods like join, and contain
637
references to the source and target versionedfiles these operations can be
640
Often we will provide convenience methods on 'versionedfile' which carry out
641
operations with another versionedfile - they will always forward to
642
InterVersionedFile.get(other).method_name(parameters).
646
"""The available optimised InterVersionedFile types."""
648
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
649
"""Integrate versions from self.source into self.target.
651
If version_ids is None all versions from source should be
652
incorporated into this versioned file.
654
Must raise RevisionNotPresent if any of the specified versions
655
are not present in the other files history unless ignore_missing is
656
supplied when they are silently skipped.
659
# - if the target is empty, just add all the versions from
660
# source to target, otherwise:
661
# - make a temporary versioned file of type target
662
# - insert the source content into it one at a time
664
if not self.target.versions():
667
# Make a new target-format versioned file.
668
temp_source = self.target.create_empty("temp", MemoryTransport())
670
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
671
graph = self.source.get_graph(version_ids)
672
order = tsort.topo_sort(graph.items())
673
pb = ui.ui_factory.nested_progress_bar()
676
# TODO for incremental cross-format work:
677
# make a versioned file with the following content:
678
# all revisions we have been asked to join
679
# all their ancestors that are *not* in target already.
680
# the immediate parents of the above two sets, with
681
# empty parent lists - these versions are in target already
682
# and the incorrect version data will be ignored.
683
# TODO: for all ancestors that are present in target already,
684
# check them for consistent data, this requires moving sha1 from
686
# TODO: remove parent texts when they are not relevant any more for
687
# memory pressure reduction. RBC 20060313
688
# pb.update('Converting versioned data', 0, len(order))
689
# deltas = self.source.get_deltas(order)
690
for index, version in enumerate(order):
691
pb.update('Converting versioned data', index, len(order))
692
parent_text = target.add_lines(version,
693
self.source.get_parents(version),
694
self.source.get_lines(version),
695
parent_texts=parent_texts)
696
parent_texts[version] = parent_text
697
#delta_parent, sha1, noeol, delta = deltas[version]
698
#target.add_delta(version,
699
# self.source.get_parents(version),
704
#target.get_lines(version)
706
# this should hit the native code path for target
707
if target is not self.target:
708
return self.target.join(temp_source,
716
def _get_source_version_ids(self, version_ids, ignore_missing):
717
"""Determine the version ids to be used from self.source.
719
:param version_ids: The caller-supplied version ids to check. (None
720
for all). If None is in version_ids, it is stripped.
721
:param ignore_missing: if True, remove missing ids from the version
722
list. If False, raise RevisionNotPresent on
723
a missing version id.
724
:return: A set of version ids.
726
if version_ids is None:
727
# None cannot be in source.versions
728
return set(self.source.versions())
730
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
732
return set(self.source.versions()).intersection(set(version_ids))
734
new_version_ids = set()
735
for version in version_ids:
738
if not self.source.has_version(version):
739
raise errors.RevisionNotPresent(version, str(self.source))
741
new_version_ids.add(version)
742
return new_version_ids
added
removed
bzrlib/versionedfile.py
