1
# Copyright (C) 2005, 2006 Canonical Ltd
4
# Johan Rydberg <jrydberg@gnu.org>
6
# This program is free software; you can redistribute it and/or modify
7
# it under the terms of the GNU General Public License as published by
8
# the Free Software Foundation; either version 2 of the License, or
9
# (at your option) any later version.
11
# This program is distributed in the hope that it will be useful,
12
# but WITHOUT ANY WARRANTY; without even the implied warranty of
13
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14
# GNU General Public License for more details.
16
# You should have received a copy of the GNU General Public License
17
# along with this program; if not, write to the Free Software
18
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20
"""Versioned text file storage api."""
22
from bzrlib.lazy_import import lazy_import
23
lazy_import(globals(), """
33
from bzrlib.transport.memory import MemoryTransport
36
from cStringIO import StringIO
38
from bzrlib.inter import InterObject
39
from bzrlib.textmerge import TextMerge
42
class VersionedFile(object):
43
"""Versioned text file storage.
45
A versioned file manages versions of line-based text files,
46
keeping track of the originating version for each line.
48
To clients the "lines" of the file are represented as a list of
49
strings. These strings will typically have terminal newline
50
characters, but this is not required. In particular files commonly
51
do not have a newline at the end of the file.
53
Texts are identified by a version-id string.
56
def __init__(self, access_mode):
58
self._access_mode = access_mode
61
def check_not_reserved_id(version_id):
62
revision.check_not_reserved_id(version_id)
64
def copy_to(self, name, transport):
65
"""Copy this versioned file to name on transport."""
66
raise NotImplementedError(self.copy_to)
69
"""Return a unsorted list of versions."""
70
raise NotImplementedError(self.versions)
72
def has_ghost(self, version_id):
73
"""Returns whether version is present as a ghost."""
74
raise NotImplementedError(self.has_ghost)
76
def has_version(self, version_id):
77
"""Returns whether version is present."""
78
raise NotImplementedError(self.has_version)
80
def add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
81
"""Add a text to the versioned file via a pregenerated delta.
83
:param version_id: The version id being added.
84
:param parents: The parents of the version_id.
85
:param delta_parent: The parent this delta was created against.
86
:param sha1: The sha1 of the full text.
87
:param delta: The delta instructions. See get_delta for details.
89
version_id = osutils.safe_revision_id(version_id)
90
parents = [osutils.safe_revision_id(v) for v in parents]
91
self._check_write_ok()
92
if self.has_version(version_id):
93
raise errors.RevisionAlreadyPresent(version_id, self)
94
return self._add_delta(version_id, parents, delta_parent, sha1, noeol, delta)
96
def _add_delta(self, version_id, parents, delta_parent, sha1, noeol, delta):
97
"""Class specific routine to add a delta.
99
This generic version simply applies the delta to the delta_parent and
102
# strip annotation from delta
104
for start, stop, delta_len, delta_lines in delta:
105
new_delta.append((start, stop, delta_len, [text for origin, text in delta_lines]))
106
if delta_parent is not None:
107
parent_full = self.get_lines(delta_parent)
110
new_full = self._apply_delta(parent_full, new_delta)
111
# its impossible to have noeol on an empty file
112
if noeol and new_full[-1][-1] == '\n':
113
new_full[-1] = new_full[-1][:-1]
114
self.add_lines(version_id, parents, new_full)
116
def add_lines(self, version_id, parents, lines, parent_texts=None):
117
"""Add a single text on top of the versioned file.
119
Must raise RevisionAlreadyPresent if the new version is
120
already present in file history.
122
Must raise RevisionNotPresent if any of the given parents are
123
not present in file history.
124
:param parent_texts: An optional dictionary containing the opaque
125
representations of some or all of the parents of
126
version_id to allow delta optimisations.
127
VERY IMPORTANT: the texts must be those returned
128
by add_lines or data corruption can be caused.
129
:return: An opaque representation of the inserted version which can be
130
provided back to future add_lines calls in the parent_texts
133
version_id = osutils.safe_revision_id(version_id)
134
parents = [osutils.safe_revision_id(v) for v in parents]
135
self._check_write_ok()
136
return self._add_lines(version_id, parents, lines, parent_texts)
138
def _add_lines(self, version_id, parents, lines, parent_texts):
139
"""Helper to do the class specific add_lines."""
140
raise NotImplementedError(self.add_lines)
142
def add_lines_with_ghosts(self, version_id, parents, lines,
144
"""Add lines to the versioned file, allowing ghosts to be present.
146
This takes the same parameters as add_lines.
148
version_id = osutils.safe_revision_id(version_id)
149
parents = [osutils.safe_revision_id(v) for v in parents]
150
self._check_write_ok()
151
return self._add_lines_with_ghosts(version_id, parents, lines,
154
def _add_lines_with_ghosts(self, version_id, parents, lines, parent_texts):
155
"""Helper to do class specific add_lines_with_ghosts."""
156
raise NotImplementedError(self.add_lines_with_ghosts)
158
def check(self, progress_bar=None):
159
"""Check the versioned file for integrity."""
160
raise NotImplementedError(self.check)
162
def _check_lines_not_unicode(self, lines):
163
"""Check that lines being added to a versioned file are not unicode."""
165
if line.__class__ is not str:
166
raise errors.BzrBadParameterUnicode("lines")
168
def _check_lines_are_lines(self, lines):
169
"""Check that the lines really are full lines without inline EOL."""
171
if '\n' in line[:-1]:
172
raise errors.BzrBadParameterContainsNewline("lines")
174
def _check_write_ok(self):
175
"""Is the versioned file marked as 'finished' ? Raise if it is."""
177
raise errors.OutSideTransaction()
178
if self._access_mode != 'w':
179
raise errors.ReadOnlyObjectDirtiedError(self)
181
def enable_cache(self):
182
"""Tell this versioned file that it should cache any data it reads.
184
This is advisory, implementations do not have to support caching.
188
def clear_cache(self):
189
"""Remove any data cached in the versioned file object.
191
This only needs to be supported if caches are supported
195
def clone_text(self, new_version_id, old_version_id, parents):
196
"""Add an identical text to old_version_id as new_version_id.
198
Must raise RevisionNotPresent if the old version or any of the
199
parents are not present in file history.
201
Must raise RevisionAlreadyPresent if the new version is
202
already present in file history."""
203
new_version_id = osutils.safe_revision_id(new_version_id)
204
old_version_id = osutils.safe_revision_id(old_version_id)
205
self._check_write_ok()
206
return self._clone_text(new_version_id, old_version_id, parents)
208
def _clone_text(self, new_version_id, old_version_id, parents):
209
"""Helper function to do the _clone_text work."""
210
raise NotImplementedError(self.clone_text)
212
def create_empty(self, name, transport, mode=None):
213
"""Create a new versioned file of this exact type.
215
:param name: the file name
216
:param transport: the transport
217
:param mode: optional file mode.
219
raise NotImplementedError(self.create_empty)
221
def fix_parents(self, version_id, new_parents):
222
"""Fix the parents list for version.
224
This is done by appending a new version to the index
225
with identical data except for the parents list.
226
the parents list must be a superset of the current
229
version_id = osutils.safe_revision_id(version_id)
230
new_parents = [osutils.safe_revision_id(p) for p in new_parents]
231
self._check_write_ok()
232
return self._fix_parents(version_id, new_parents)
234
def _fix_parents(self, version_id, new_parents):
235
"""Helper for fix_parents."""
236
raise NotImplementedError(self.fix_parents)
238
def get_delta(self, version):
239
"""Get a delta for constructing version from some other version.
241
:return: (delta_parent, sha1, noeol, delta)
242
Where delta_parent is a version id or None to indicate no parent.
244
raise NotImplementedError(self.get_delta)
246
def get_deltas(self, version_ids):
247
"""Get multiple deltas at once for constructing versions.
249
:return: dict(version_id:(delta_parent, sha1, noeol, delta))
250
Where delta_parent is a version id or None to indicate no parent, and
251
version_id is the version_id created by that delta.
254
for version_id in version_ids:
255
result[version_id] = self.get_delta(version_id)
258
def make_mpdiffs(self, version_ids):
259
"""Create multiparent diffs for specified versions"""
260
knit_versions = set()
261
for version_id in version_ids:
262
knit_versions.add(version_id)
263
knit_versions.update(self.get_parents(version_id))
264
lines = dict(zip(knit_versions,
265
self._get_lf_split_line_list(knit_versions)))
267
for version_id in version_ids:
268
target = lines[version_id]
269
parents = [lines[p] for p in self.get_parents(version_id)]
271
left_parent_blocks = self._extract_blocks(version_id,
274
left_parent_blocks = None
275
diffs.append(multiparent.MultiParent.from_lines(target, parents,
279
def _extract_blocks(self, version_id, source, target):
282
def add_mpdiffs(self, records):
283
"""Add mpdiffs to this versionedfile
285
Records should be iterables of version, parents, expected_sha1,
286
mpdiff. mpdiff should be a MultiParent instance.
289
for version, parents, expected_sha1, mpdiff in records:
290
mpvf = multiparent.MultiMemoryVersionedFile()
291
needed_parents = [p for p in parents if not mpvf.has_version(p)]
292
parent_lines = self._get_lf_split_line_list(needed_parents)
293
for parent_id, lines in zip(needed_parents, parent_lines):
294
mpvf.add_version(lines, parent_id, [])
295
mpvf.add_diff(mpdiff, version, parents)
296
lines = mpvf.get_line_list([version])[0]
297
version_text = self.add_lines(version, parents, lines, vf_parents)
298
vf_parents[version] = version_text
299
if expected_sha1 != self.get_sha1(version):
300
raise errors.VersionedFileInvalidChecksum(version)
302
def get_sha1(self, version_id):
303
"""Get the stored sha1 sum for the given revision.
305
:param name: The name of the version to lookup
307
raise NotImplementedError(self.get_sha1)
309
def get_sha1s(self, version_ids):
310
"""Get the stored sha1 sums for the given revisions.
312
:param version_ids: The names of the versions to lookup
313
:return: a list of sha1s in order according to the version_ids
315
raise NotImplementedError(self.get_sha1)
317
def get_suffixes(self):
318
"""Return the file suffixes associated with this versioned file."""
319
raise NotImplementedError(self.get_suffixes)
321
def get_text(self, version_id):
322
"""Return version contents as a text string.
324
Raises RevisionNotPresent if version is not present in
327
return ''.join(self.get_lines(version_id))
328
get_string = get_text
330
def get_texts(self, version_ids):
331
"""Return the texts of listed versions as a list of strings.
333
Raises RevisionNotPresent if version is not present in
336
return [''.join(self.get_lines(v)) for v in version_ids]
338
def get_lines(self, version_id):
339
"""Return version contents as a sequence of lines.
341
Raises RevisionNotPresent if version is not present in
344
raise NotImplementedError(self.get_lines)
346
def _get_lf_split_line_list(self, version_ids):
347
return [StringIO(t).readlines() for t in self.get_texts(version_ids)]
349
def get_ancestry(self, version_ids, topo_sorted=True):
350
"""Return a list of all ancestors of given version(s). This
351
will not include the null revision.
353
This list will not be topologically sorted if topo_sorted=False is
356
Must raise RevisionNotPresent if any of the given versions are
357
not present in file history."""
358
if isinstance(version_ids, basestring):
359
version_ids = [version_ids]
360
raise NotImplementedError(self.get_ancestry)
362
def get_ancestry_with_ghosts(self, version_ids):
363
"""Return a list of all ancestors of given version(s). This
364
will not include the null revision.
366
Must raise RevisionNotPresent if any of the given versions are
367
not present in file history.
369
Ghosts that are known about will be included in ancestry list,
370
but are not explicitly marked.
372
raise NotImplementedError(self.get_ancestry_with_ghosts)
374
def get_graph(self, version_ids=None):
375
"""Return a graph from the versioned file.
377
Ghosts are not listed or referenced in the graph.
378
:param version_ids: Versions to select.
379
None means retrieve all versions.
381
if version_ids is None:
382
return dict(self.iter_parents(self.versions()))
384
pending = set(osutils.safe_revision_id(v) for v in version_ids)
386
this_iteration = pending
388
for version, parents in self.iter_parents(this_iteration):
389
result[version] = parents
390
for parent in parents:
396
def get_graph_with_ghosts(self):
397
"""Return a graph for the entire versioned file.
399
Ghosts are referenced in parents list but are not
402
raise NotImplementedError(self.get_graph_with_ghosts)
404
def get_parents(self, version_id):
405
"""Return version names for parents of a version.
407
Must raise RevisionNotPresent if version is not present in
410
raise NotImplementedError(self.get_parents)
412
def get_parents_with_ghosts(self, version_id):
413
"""Return version names for parents of version_id.
415
Will raise RevisionNotPresent if version_id is not present
418
Ghosts that are known about will be included in the parent list,
419
but are not explicitly marked.
421
raise NotImplementedError(self.get_parents_with_ghosts)
423
def annotate_iter(self, version_id):
424
"""Yield list of (version-id, line) pairs for the specified
427
Must raise RevisionNotPresent if any of the given versions are
428
not present in file history.
430
raise NotImplementedError(self.annotate_iter)
432
def annotate(self, version_id):
433
return list(self.annotate_iter(version_id))
435
def _apply_delta(self, lines, delta):
436
"""Apply delta to lines."""
439
for start, end, count, delta_lines in delta:
440
lines[offset+start:offset+end] = delta_lines
441
offset = offset + (start - end) + count
444
def join(self, other, pb=None, msg=None, version_ids=None,
445
ignore_missing=False):
446
"""Integrate versions from other into this versioned file.
448
If version_ids is None all versions from other should be
449
incorporated into this versioned file.
451
Must raise RevisionNotPresent if any of the specified versions
452
are not present in the other files history unless ignore_missing
453
is supplied when they are silently skipped.
455
self._check_write_ok()
456
return InterVersionedFile.get(other, self).join(
462
def iter_lines_added_or_present_in_versions(self, version_ids=None,
464
"""Iterate over the lines in the versioned file from version_ids.
466
This may return lines from other versions, and does not return the
467
specific version marker at this point. The api may be changed
468
during development to include the version that the versioned file
469
thinks is relevant, but given that such hints are just guesses,
470
its better not to have it if we don't need it.
472
If a progress bar is supplied, it may be used to indicate progress.
473
The caller is responsible for cleaning up progress bars (because this
476
NOTES: Lines are normalised: they will all have \n terminators.
477
Lines are returned in arbitrary order.
479
raise NotImplementedError(self.iter_lines_added_or_present_in_versions)
481
def iter_parents(self, version_ids):
482
"""Iterate through the parents for many version ids.
484
:param version_ids: An iterable yielding version_ids.
485
:return: An iterator that yields (version_id, parents). Requested
486
version_ids not present in the versioned file are simply skipped.
487
The order is undefined, allowing for different optimisations in
488
the underlying implementation.
490
for version_id in version_ids:
492
yield version_id, tuple(self.get_parents(version_id))
493
except errors.RevisionNotPresent:
496
def transaction_finished(self):
497
"""The transaction that this file was opened in has finished.
499
This records self.finished = True and should cause all mutating
504
def plan_merge(self, ver_a, ver_b):
505
"""Return pseudo-annotation indicating how the two versions merge.
507
This is computed between versions a and b and their common
510
Weave lines present in none of them are skipped entirely.
513
killed-base Dead in base revision
514
killed-both Killed in each revision
517
unchanged Alive in both a and b (possibly created in both)
520
ghost-a Killed in a, unborn in b
521
ghost-b Killed in b, unborn in a
522
irrelevant Not in either revision
524
raise NotImplementedError(VersionedFile.plan_merge)
526
def weave_merge(self, plan, a_marker=TextMerge.A_MARKER,
527
b_marker=TextMerge.B_MARKER):
528
return PlanWeaveMerge(plan, a_marker, b_marker).merge_lines()[0]
531
class PlanWeaveMerge(TextMerge):
532
"""Weave merge that takes a plan as its input.
534
This exists so that VersionedFile.plan_merge is implementable.
535
Most callers will want to use WeaveMerge instead.
538
def __init__(self, plan, a_marker=TextMerge.A_MARKER,
539
b_marker=TextMerge.B_MARKER):
540
TextMerge.__init__(self, a_marker, b_marker)
543
def _merge_struct(self):
548
def outstanding_struct():
549
if not lines_a and not lines_b:
551
elif ch_a and not ch_b:
554
elif ch_b and not ch_a:
556
elif lines_a == lines_b:
559
yield (lines_a, lines_b)
561
# We previously considered either 'unchanged' or 'killed-both' lines
562
# to be possible places to resynchronize. However, assuming agreement
563
# on killed-both lines may be too aggressive. -- mbp 20060324
564
for state, line in self.plan:
565
if state == 'unchanged':
566
# resync and flush queued conflicts changes if any
567
for struct in outstanding_struct():
573
if state == 'unchanged':
576
elif state == 'killed-a':
579
elif state == 'killed-b':
582
elif state == 'new-a':
585
elif state == 'new-b':
589
assert state in ('irrelevant', 'ghost-a', 'ghost-b',
590
'killed-base', 'killed-both'), state
591
for struct in outstanding_struct():
595
class WeaveMerge(PlanWeaveMerge):
596
"""Weave merge that takes a VersionedFile and two versions as its input"""
598
def __init__(self, versionedfile, ver_a, ver_b,
599
a_marker=PlanWeaveMerge.A_MARKER, b_marker=PlanWeaveMerge.B_MARKER):
600
plan = versionedfile.plan_merge(ver_a, ver_b)
601
PlanWeaveMerge.__init__(self, plan, a_marker, b_marker)
604
class InterVersionedFile(InterObject):
605
"""This class represents operations taking place between two versionedfiles..
607
Its instances have methods like join, and contain
608
references to the source and target versionedfiles these operations can be
611
Often we will provide convenience methods on 'versionedfile' which carry out
612
operations with another versionedfile - they will always forward to
613
InterVersionedFile.get(other).method_name(parameters).
617
"""The available optimised InterVersionedFile types."""
619
def join(self, pb=None, msg=None, version_ids=None, ignore_missing=False):
620
"""Integrate versions from self.source into self.target.
622
If version_ids is None all versions from source should be
623
incorporated into this versioned file.
625
Must raise RevisionNotPresent if any of the specified versions
626
are not present in the other files history unless ignore_missing is
627
supplied when they are silently skipped.
630
# - if the target is empty, just add all the versions from
631
# source to target, otherwise:
632
# - make a temporary versioned file of type target
633
# - insert the source content into it one at a time
635
if not self.target.versions():
638
# Make a new target-format versioned file.
639
temp_source = self.target.create_empty("temp", MemoryTransport())
641
version_ids = self._get_source_version_ids(version_ids, ignore_missing)
642
graph = self.source.get_graph(version_ids)
643
order = tsort.topo_sort(graph.items())
644
pb = ui.ui_factory.nested_progress_bar()
647
# TODO for incremental cross-format work:
648
# make a versioned file with the following content:
649
# all revisions we have been asked to join
650
# all their ancestors that are *not* in target already.
651
# the immediate parents of the above two sets, with
652
# empty parent lists - these versions are in target already
653
# and the incorrect version data will be ignored.
654
# TODO: for all ancestors that are present in target already,
655
# check them for consistent data, this requires moving sha1 from
657
# TODO: remove parent texts when they are not relevant any more for
658
# memory pressure reduction. RBC 20060313
659
# pb.update('Converting versioned data', 0, len(order))
660
# deltas = self.source.get_deltas(order)
661
for index, version in enumerate(order):
662
pb.update('Converting versioned data', index, len(order))
663
parent_text = target.add_lines(version,
664
self.source.get_parents(version),
665
self.source.get_lines(version),
666
parent_texts=parent_texts)
667
parent_texts[version] = parent_text
668
#delta_parent, sha1, noeol, delta = deltas[version]
669
#target.add_delta(version,
670
# self.source.get_parents(version),
675
#target.get_lines(version)
677
# this should hit the native code path for target
678
if target is not self.target:
679
return self.target.join(temp_source,
687
def _get_source_version_ids(self, version_ids, ignore_missing):
688
"""Determine the version ids to be used from self.source.
690
:param version_ids: The caller-supplied version ids to check. (None
691
for all). If None is in version_ids, it is stripped.
692
:param ignore_missing: if True, remove missing ids from the version
693
list. If False, raise RevisionNotPresent on
694
a missing version id.
695
:return: A set of version ids.
697
if version_ids is None:
698
# None cannot be in source.versions
699
return set(self.source.versions())
701
version_ids = [osutils.safe_revision_id(v) for v in version_ids]
703
return set(self.source.versions()).intersection(set(version_ids))
705
new_version_ids = set()
706
for version in version_ids:
709
if not self.source.has_version(version):
710
raise errors.RevisionNotPresent(version, str(self.source))
712
new_version_ids.add(version)
713
return new_version_ids
added
removed
bzrlib/versionedfile.py
