1
# Copyright (C) 2005, 2006 Canonical Ltd
4
# Johan Rydberg <jrydberg@gnu.org>
6
# This program is free software; you can redistribute it and/or modify
7
# it under the terms of the GNU General Public License as published by
8
# the Free Software Foundation; either version 2 of the License, or
9
# (at your option) any later version.
11
# This program is distributed in the hope that it will be useful,
12
# but WITHOUT ANY WARRANTY; without even the implied warranty of
13
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14
# GNU General Public License for more details.
16
# You should have received a copy of the GNU General Public License
17
# along with this program; if not, write to the Free Software
18
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20
"""Versioned text file storage api."""
22
from bzrlib.lazy_import import lazy_import
23
lazy_import(globals(), """
33
from bzrlib.transport.memory import MemoryTransport
36
from cStringIO import StringIO
38
from bzrlib.inter import InterObject
39
from bzrlib.textmerge import TextMerge
42
class VersionedFile(object):
43
"""Versioned text file storage.
45
A versioned file manages versions of line-based text files,
46
keeping track of the originating version for each line.
48
To clients the "lines" of the file are represented as a list of
49
strings. These strings will typically have terminal newline
50
characters, but this is not required. In particular files commonly
51
do not have a newline at the end of the file.
53
Texts are identified by a version-id string.
56
def __init__(self, access_mode):
58
self._access_mode = access_mode
61
def check_not_reserved_id(version_id):
62
revision.check_not_reserved_id(version_id)
64
def copy_to(self, name, transport):
65
"""Copy this versioned file to name on transport."""
66
raise NotImplementedError(self.copy_to)
69
"""Return a unsorted list of versions."""
70
raise NotImplementedError(self.versions)
72
def has_ghost(self, version_id):
73
"""Returns whether version is present as a ghost."""
74
raise NotImplementedError(self.has_ghost)
76
def has_version(self, version_id):
77
"""Returns whether version is present."""
78
raise NotImplementedError(self.has_version)
80
def add_lines(self, version_id, parents, lines, parent_texts=None,
81
left_matching_blocks=None, nostore_sha=None, random_id=False,
83
"""Add a single text on top of the versioned file.
85
Must raise RevisionAlreadyPresent if the new version is
86
already present in file history.
88
Must raise RevisionNotPresent if any of the given parents are
89
not present in file history.
91
:param lines: A list of lines. Each line must be a bytestring. And all
92
of them except the last must be terminated with \n and contain no
93
other \n's. The last line may either contain no \n's or a single
94
terminated \n. If the lines list does meet this constraint the add
95
routine may error or may succeed - but you will be unable to read
96
the data back accurately. (Checking the lines have been split
97
correctly is expensive and extremely unlikely to catch bugs so it
98
is not done at runtime unless check_content is True.)
99
:param parent_texts: An optional dictionary containing the opaque
100
representations of some or all of the parents of version_id to
101
allow delta optimisations. VERY IMPORTANT: the texts must be those
102
returned by add_lines or data corruption can be caused.
103
:param left_matching_blocks: a hint about which areas are common
104
between the text and its left-hand-parent. The format is
105
the SequenceMatcher.get_matching_blocks format.
106
:param nostore_sha: Raise ExistingContent and do not add the lines to
107
the versioned file if the digest of the lines matches this.
108
:param random_id: If True a random id has been selected rather than
109
an id determined by some deterministic process such as a converter
110
from a foreign VCS. When True the backend may choose not to check
111
for uniqueness of the resulting key within the versioned file, so
112
this should only be done when the result is expected to be unique
114
:param check_content: If True, the lines supplied are verified to be
115
bytestrings that are correctly formed lines.
116
:return: The text sha1, the number of bytes in the text, and an opaque
117
representation of the inserted version which can be provided
118
back to future add_lines calls in the parent_texts dictionary.
120
version_id = osutils.safe_revision_id(version_id)
121
parents = [osutils.safe_revision_id(v) for v in parents]
122
self._check_write_ok()
123
return self._add_lines(version_id, parents, lines, parent_texts,
124
left_matching_blocks, nostore_sha, random_id, check_content)
126
def _add_lines(self, version_id, parents, lines, parent_texts,
127
left_matching_blocks, nostore_sha, random_id, check_content):
128
"""Helper to do the class specific add_lines."""
129
raise NotImplementedError(self.add_lines)
131
def add_lines_with_ghosts(self, version_id, parents, lines,
132
parent_texts=None, nostore_sha=None, random_id=False,
134
"""Add lines to the versioned file, allowing ghosts to be present.
136
This takes the same parameters as add_lines and returns the same.
138
version_id = osutils.safe_revision_id(version_id)
139
parents = [osutils.safe_revision_id(v) for v in parents]
140
self._check_write_ok()
141
return self._add_lines_with_ghosts(version_id, parents, lines,
142
parent_texts, nostore_sha, random_id, check_content)
144
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts,
145
nostore_sha, random_id, check_content):
146
"""Helper to do class specific add_lines_with_ghosts."""
147
raise NotImplementedError(self.add_lines_with_ghosts)
149
def check(self, progress_bar=None):
150
"""Check the versioned file for integrity."""
151
raise NotImplementedError(self.check)
153
def _check_lines_not_unicode(self, lines):
154
"""Check that lines being added to a versioned file are not unicode."""
156
if line.__class__ is not str:
157
raise errors.BzrBadParameterUnicode("lines")
159
def _check_lines_are_lines(self, lines):
160
"""Check that the lines really are full lines without inline EOL."""
162
if '\n' in line[:-1]:
163
raise errors.BzrBadParameterContainsNewline("lines")
165
def _check_write_ok(self):
166
"""Is the versioned file marked as 'finished' ? Raise if it is."""
168
raise errors.OutSideTransaction()
169
if self._access_mode != 'w':
170
raise errors.ReadOnlyObjectDirtiedError(self)
172
def enable_cache(self):
173
"""Tell this versioned file that it should cache any data it reads.
175
This is advisory, implementations do not have to support caching.
179
def clear_cache(self):
180
"""Remove any data cached in the versioned file object.
182
This only needs to be supported if caches are supported
186
def clone_text(self, new_version_id, old_version_id, parents):
187
"""Add an identical text to old_version_id as new_version_id.
189
Must raise RevisionNotPresent if the old version or any of the
190
parents are not present in file history.
192
Must raise RevisionAlreadyPresent if the new version is
193
already present in file history."""
194
new_version_id = osutils.safe_revision_id(new_version_id)
195
old_version_id = osutils.safe_revision_id(old_version_id)
196
self._check_write_ok()
197
return self._clone_text(new_version_id, old_version_id, parents)
199
def _clone_text(self, new_version_id, old_version_id, parents):
200
"""Helper function to do the _clone_text work."""
201
raise NotImplementedError(self.clone_text)
203
def create_empty(self, name, transport, mode=None):
204
"""Create a new versioned file of this exact type.
206
:param name: the file name
207
:param transport: the transport
208
:param mode: optional file mode.
210
raise NotImplementedError(self.create_empty)
212
def get_format_signature(self):
213
"""Get a text description of the data encoding in this file.
217
raise NotImplementedError(self.get_format_signature)
219
def make_mpdiffs(self, version_ids):
220
"""Create multiparent diffs for specified versions"""
221
knit_versions = set()
222
for version_id in version_ids:
223
knit_versions.add(version_id)
224
knit_versions.update(self.get_parents(version_id))
225
lines = dict(zip(knit_versions,
226
self._get_lf_split_line_list(knit_versions)))
228
for version_id in version_ids:
229
target = lines[version_id]
230
parents = [lines[p] for p in self.get_parents(version_id)]
232
left_parent_blocks = self._extract_blocks(version_id,
235
left_parent_blocks = None
236
diffs.append(multiparent.MultiParent.from_lines(target, parents,
240
def _extract_blocks(self, version_id, source, target):
243
def add_mpdiffs(self, records):
244
"""Add mpdiffs to this versionedfile
246
Records should be iterables of version, parents, expected_sha1,
247
mpdiff. mpdiff should be a MultiParent instance.
250
mpvf = multiparent.MultiMemoryVersionedFile()
252
for version, parent_ids, expected_sha1, mpdiff in records:
253
versions.append(version)
254
mpvf.add_diff(mpdiff, version, parent_ids)
255
needed_parents = set()
256
for version, parent_ids, expected_sha1, mpdiff in records:
257
needed_parents.update(p for p in parent_ids
258
if not mpvf.has_version(p))
259
for parent_id, lines in zip(needed_parents,
260
self._get_lf_split_line_list(needed_parents)):
261
mpvf.add_version(lines, parent_id, [])
262
for (version, parent_ids, expected_sha1, mpdiff), lines in\
263
zip(records, mpvf.get_line_list(versions)):
264
if len(parent_ids) == 1:
265
left_matching_blocks = list(mpdiff.get_matching_blocks(0,
266
mpvf.get_diff(parent_ids[0]).num_lines()))
268
left_matching_blocks = None
269
_, _, version_text = self.add_lines(version, parent_ids, lines,
270
vf_parents, left_matching_blocks=left_matching_blocks)
271
vf_parents[version] = version_text
272
for (version, parent_ids, expected_sha1, mpdiff), sha1 in\
273
zip(records, self.get_sha1s(versions)):
274
if expected_sha1 != sha1:
275
raise errors.VersionedFileInvalidChecksum(version)
277
def get_sha1(self, version_id):
278
"""Get the stored sha1 sum for the given revision.
280
:param name: The name of the version to lookup
282
raise NotImplementedError(self.get_sha1)
284
def get_sha1s(self, version_ids):
285
"""Get the stored sha1 sums for the given revisions.
287
:param version_ids: The names of the versions to lookup
288
:return: a list of sha1s in order according to the version_ids
290
raise NotImplementedError(self.get_sha1)
292
def get_suffixes(self):
293
"""Return the file suffixes associated with this versioned file."""
294
raise NotImplementedError(self.get_suffixes)
296
def get_text(self, version_id):
297
"""Return version contents as a text string.
299
Raises RevisionNotPresent if version is not present in
302
return ''.join(self.get_lines(version_id))
303
get_string = get_text
305
def get_texts(self, version_ids):
306
"""Return the texts of listed versions as a list of strings.
308
Raises RevisionNotPresent if version is not present in
311
return [''.join(self.get_lines(v)) for v in version_ids]
313
def get_lines(self, version_id):
314
"""Return version contents as a sequence of lines.
316
Raises RevisionNotPresent if version is not present in
319
raise NotImplementedError(self.get_lines)
321
def _get_lf_split_line_list(self, version_ids):
322
return [StringIO(t).readlines() for t in self.get_texts(version_ids)]
324
def get_ancestry(self, version_ids, topo_sorted=True):
325
"""Return a list of all ancestors of given version(s). This
326
will not include the null revision.
328
This list will not be topologically sorted if topo_sorted=False is
331
Must raise RevisionNotPresent if any of the given versions are
332
not present in file history."""
333
if isinstance(version_ids, basestring):
334
version_ids = [version_ids]
335
raise NotImplementedError(self.get_ancestry)
337
def get_ancestry_with_ghosts(self, version_ids):
338
"""Return a list of all ancestors of given version(s). This
339
will not include the null revision.
341
Must raise RevisionNotPresent if any of the given versions are
342
not present in file history.
344
Ghosts that are known about will be included in ancestry list,
345
but are not explicitly marked.
347
raise NotImplementedError(self.get_ancestry_with_ghosts)
349
def get_graph(self, version_ids=None):
350
"""Return a graph from the versioned file.
352
Ghosts are not listed or referenced in the graph.
353
:param version_ids: Versions to select.
354
None means retrieve all versions.
356
if version_ids is None:
357
return dict(self.iter_parents(self.versions()))
359
pending = set(osutils.safe_revision_id(v) for v in version_ids)
361
this_iteration = pending
363
for version, parents in self.iter_parents(this_iteration):
364
result[version] = parents
365
for parent in parents:
371
def get_graph_with_ghosts(self):
372
"""Return a graph for the entire versioned file.
374
Ghosts are referenced in parents list but are not
377
raise NotImplementedError(self.get_graph_with_ghosts)
379
def get_parents(self, version_id):
380
"""Return version names for parents of a version.
382
Must raise RevisionNotPresent if version is not present in
385
raise NotImplementedError(self.get_parents)
387
def get_parents_with_ghosts(self, version_id):
388
"""Return version names for parents of version_id.
390
Will raise RevisionNotPresent if version_id is not present
393
Ghosts that are known about will be included in the parent list,
394
but are not explicitly marked.
396
raise NotImplementedError(self.get_parents_with_ghosts)
398
def annotate_iter(self, version_id):
399
"""Yield list of (version-id, line) pairs for the specified
402
Must raise RevisionNotPresent if any of the given versions are
403
not present in file history.
405
raise NotImplementedError(self.annotate_iter)
407
def annotate(self, version_id):
408
return list(self.annotate_iter(version_id))
410
def _apply_delta(self, lines, delta):
411
"""Apply delta to lines."""
414
for start, end, count, delta_lines in delta:
415
lines[offset+start:offset+end] = delta_lines
416
offset = offset + (start - end) + count
419
def join(self, other, pb=None, msg=None, version_ids=None,
420
ignore_missing=False):
421
"""Integrate versions from other into this versioned file.
423
If version_ids is None all versions from other should be
424
incorporated into this versioned file.
426
Must raise RevisionNotPresent if any of the specified versions
427
are not present in the other files history unless ignore_missing
428
is supplied when they are silently skipped.
430
self._check_write_ok()
431
return InterVersionedFile.get(other, self).join(
437
def iter_lines_added_or_present_in_versions(self, version_ids=None,
439
"""Iterate over the lines in the versioned file from version_ids.
441
This may return lines from other versions, and does not return the
442
specific version marker at this point. The api may be changed
443
during development to include the version that the versioned file
444
thinks is relevant, but given that such hints are just guesses,
445
its better not to have it if we don't need it.
447
If a progress bar is supplied, it may be used to indicate progress.
448
The caller is responsible for cleaning up progress bars (because this
451
NOTES: Lines are normalised: they will all have \n terminators.
452
Lines are returned in arbitrary order.
454
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
456
def iter_parents(self, version_ids):
457
"""Iterate through the parents for many version ids.
459
:param version_ids: An iterable yielding version_ids.
460
:return: An iterator that yields (version_id, parents). Requested
461
version_ids not present in the versioned file are simply skipped.
462
The order is undefined, allowing for different optimisations in
463
the underlying implementation.
465
for version_id in version_ids:
467
yield version_id, tuple(self.get_parents(version_id))
468
except errors.RevisionNotPresent:
471
def transaction_finished(self):
472
"""The transaction that this file was opened in has finished.
474
This records self.finished = True and should cause all mutating
479
def plan_merge(self, ver_a, ver_b):
480
"""Return pseudo-annotation indicating how the two versions merge.
482
This is computed between versions a and b and their common
485
Weave lines present in none of them are skipped entirely.
488
killed-base Dead in base revision
489
killed-both Killed in each revision
492
unchanged Alive in both a and b (possibly created in both)
495
ghost-a Killed in a, unborn in b
496
ghost-b Killed in b, unborn in a
497
irrelevant Not in either revision
499
raise NotImplementedError(VersionedFile.plan_merge)
501
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
502
b_marker=TextMerge.B_MARKER):
503
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
506
class PlanWeaveMerge(TextMerge):
507
"""Weave merge that takes a plan as its input.
509
This exists so that VersionedFile.plan_merge is implementable.
510
Most callers will want to use WeaveMerge instead.
513
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
514
b_marker=TextMerge.B_MARKER):
515
TextMerge.__init__(self, a_marker, b_marker)
518
def _merge_struct(self):
523
def outstanding_struct():
524
if not lines_a and not lines_b:
526
elif ch_a and not ch_b:
529
elif ch_b and not ch_a:
531
elif lines_a == lines_b:
534
yield (lines_a, lines_b)
536
# We previously considered either 'unchanged' or 'killed-both' lines
537
# to be possible places to resynchronize. However, assuming agreement
538
# on killed-both lines may be too aggressive. -- mbp 20060324
539
for state, line in self.plan:
540
if state == 'unchanged':
541
# resync and flush queued conflicts changes if any
542
for struct in outstanding_struct():
548
if state == 'unchanged':
551
elif state == 'killed-a':
554
elif state == 'killed-b':
557
elif state == 'new-a':
560
elif state == 'new-b':
564
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
565
'killed-base', 'killed-both'), state
566
for struct in outstanding_struct():
570
class WeaveMerge(PlanWeaveMerge):
571
"""Weave merge that takes a VersionedFile and two versions as its input"""
573
def __init__(self, versionedfile, ver_a, ver_b,
574
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
575
plan = versionedfile.plan_merge(ver_a, ver_b)
576
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
579
class InterVersionedFile(InterObject):
580
"""This class represents operations taking place between two versionedfiles..
582
Its instances have methods like join, and contain
583
references to the source and target versionedfiles these operations can be
586
Often we will provide convenience methods on 'versionedfile' which carry out
587
operations with another versionedfile - they will always forward to
588
InterVersionedFile.get(other).method_name(parameters).
592
"""The available optimised InterVersionedFile types."""
594
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
595
"""Integrate versions from self.source into self.target.
597
If version_ids is None all versions from source should be
598
incorporated into this versioned file.
600
Must raise RevisionNotPresent if any of the specified versions
601
are not present in the other files history unless ignore_missing is
602
supplied when they are silently skipped.
605
# - if the target is empty, just add all the versions from
606
# source to target, otherwise:
607
# - make a temporary versioned file of type target
608
# - insert the source content into it one at a time
610
if not self.target.versions():
613
# Make a new target-format versioned file.
614
temp_source = self.target.create_empty("temp", MemoryTransport())
616
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
617
graph = self.source.get_graph(version_ids)
618
order = tsort.topo_sort(graph.items())
619
pb = ui.ui_factory.nested_progress_bar()
622
# TODO for incremental cross-format work:
623
# make a versioned file with the following content:
624
# all revisions we have been asked to join
625
# all their ancestors that are *not* in target already.
626
# the immediate parents of the above two sets, with
627
# empty parent lists - these versions are in target already
628
# and the incorrect version data will be ignored.
629
# TODO: for all ancestors that are present in target already,
630
# check them for consistent data, this requires moving sha1 from
632
# TODO: remove parent texts when they are not relevant any more for
633
# memory pressure reduction. RBC 20060313
634
# pb.update('Converting versioned data', 0, len(order))
635
for index, version in enumerate(order):
636
pb.update('Converting versioned data', index, len(order))
637
_, _, parent_text = target.add_lines(version,
638
self.source.get_parents(version),
639
self.source.get_lines(version),
640
parent_texts=parent_texts)
641
parent_texts[version] = parent_text
643
# this should hit the native code path for target
644
if target is not self.target:
645
return self.target.join(temp_source,
653
def _get_source_version_ids(self, version_ids, ignore_missing):
654
"""Determine the version ids to be used from self.source.
656
:param version_ids: The caller-supplied version ids to check. (None
657
for all). If None is in version_ids, it is stripped.
658
:param ignore_missing: if True, remove missing ids from the version
659
list. If False, raise RevisionNotPresent on
660
a missing version id.
661
:return: A set of version ids.
663
if version_ids is None:
664
# None cannot be in source.versions
665
return set(self.source.versions())
667
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
669
return set(self.source.versions()).intersection(set(version_ids))
671
new_version_ids = set()
672
for version in version_ids:
675
if not self.source.has_version(version):
676
raise errors.RevisionNotPresent(version, str(self.source))
678
new_version_ids.add(version)
679
return new_version_ids
added
removed
bzrlib/versionedfile.py
