1
# Copyright (C) 2005, 2006, 2008, 2009 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
18
"""Copying of history from one branch to another.
20
The basic plan is that every branch knows the history of everything
21
that has merged into it. As the first step of a merge, pull, or
22
branch operation we copy history from the source into the destination
33
from bzrlib.errors import InstallFailed
34
from bzrlib.progress import ProgressPhase
35
from bzrlib.revision import NULL_REVISION
36
from bzrlib.tsort import topo_sort
37
from bzrlib.trace import mutter
39
from bzrlib.versionedfile import FulltextContentFactory
42
class RepoFetcher(object):
43
"""Pull revisions and texts from one repository to another.
45
This should not be used directly, it's essential a object to encapsulate
46
the logic in InterRepository.fetch().
49
def __init__(self, to_repository, from_repository, last_revision=None,
50
pb=None, find_ghosts=True, fetch_spec=None):
51
"""Create a repo fetcher.
53
:param last_revision: If set, try to limit to the data this revision
55
:param find_ghosts: If True search the entire history for ghosts.
56
:param _write_group_acquired_callable: Don't use; this parameter only
57
exists to facilitate a hack done in InterPackRepo.fetch. We would
58
like to remove this parameter.
59
:param pb: ProgressBar object to use; deprecated and ignored.
60
This method will just create one on top of the stack.
63
symbol_versioning.warn(
64
symbol_versioning.deprecated_in((1, 14, 0))
65
% "pb parameter to RepoFetcher.__init__")
66
# and for simplicity it is in fact ignored
67
if to_repository.has_same_location(from_repository):
68
# repository.fetch should be taking care of this case.
69
raise errors.BzrError('RepoFetcher run '
70
'between two objects at the same location: '
71
'%r and %r' % (to_repository, from_repository))
72
self.to_repository = to_repository
73
self.from_repository = from_repository
74
self.sink = to_repository._get_sink()
75
# must not mutate self._last_revision as its potentially a shared instance
76
self._last_revision = last_revision
77
self._fetch_spec = fetch_spec
78
self.find_ghosts = find_ghosts
79
self.from_repository.lock_read()
80
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
81
self.from_repository, self.from_repository._format,
82
self.to_repository, self.to_repository._format)
86
self.from_repository.unlock()
89
"""Primary worker function.
91
This initialises all the needed variables, and then fetches the
92
requested revisions, finally clearing the progress bar.
94
# Roughly this is what we're aiming for fetch to become:
96
# missing = self.sink.insert_stream(self.source.get_stream(search))
98
# missing = self.sink.insert_stream(self.source.get_items(missing))
101
self.file_ids_names = {}
102
pb = bzrlib.ui.ui_factory.nested_progress_bar()
103
pb.show_pct = pb.show_count = False
105
pb.update("Finding revisions", 0, 2)
106
search = self._revids_to_fetch()
109
pb.update("Fetching revisions", 1, 2)
110
self._fetch_everything_for_search(search)
114
def _fetch_everything_for_search(self, search):
115
"""Fetch all data for the given set of revisions."""
116
# The first phase is "file". We pass the progress bar for it directly
117
# into item_keys_introduced_by, which has more information about how
118
# that phase is progressing than we do. Progress updates for the other
119
# phases are taken care of in this function.
120
# XXX: there should be a clear owner of the progress reporting. Perhaps
121
# item_keys_introduced_by should have a richer API than it does at the
122
# moment, so that it can feed the progress information back to this
124
if (self.from_repository._format.rich_root_data and
125
not self.to_repository._format.rich_root_data):
126
raise errors.IncompatibleRepositories(
127
self.from_repository, self.to_repository,
128
"different rich-root support")
129
pb = bzrlib.ui.ui_factory.nested_progress_bar()
131
pb.update("Get stream source")
132
source = self.from_repository._get_source(
133
self.to_repository._format)
134
stream = source.get_stream(search)
135
from_format = self.from_repository._format
136
pb.update("Inserting stream")
137
resume_tokens, missing_keys = self.sink.insert_stream(
138
stream, from_format, [])
139
if self.to_repository._fallback_repositories:
141
self._parent_inventories(search.get_keys()))
143
pb.update("Missing keys")
144
stream = source.get_stream_for_missing_keys(missing_keys)
145
pb.update("Inserting missing keys")
146
resume_tokens, missing_keys = self.sink.insert_stream(
147
stream, from_format, resume_tokens)
149
raise AssertionError(
150
"second push failed to complete a fetch %r." % (
153
raise AssertionError(
154
"second push failed to commit the fetch %r." % (
156
pb.update("Finishing stream")
161
def _revids_to_fetch(self):
162
"""Determines the exact revisions needed from self.from_repository to
163
install self._last_revision in self.to_repository.
165
If no revisions need to be fetched, then this just returns None.
167
if self._fetch_spec is not None:
168
return self._fetch_spec
169
mutter('fetch up to rev {%s}', self._last_revision)
170
if self._last_revision is NULL_REVISION:
171
# explicit limit of no revisions needed
173
if (self._last_revision is not None and
174
self.to_repository.has_revision(self._last_revision)):
177
return self.to_repository.search_missing_revision_ids(
178
self.from_repository, self._last_revision,
179
find_ghosts=self.find_ghosts)
180
except errors.NoSuchRevision, e:
181
raise InstallFailed([self._last_revision])
183
def _parent_inventories(self, revision_ids):
184
# Find all the parent revisions referenced by the stream, but
185
# not present in the stream, and make sure we send their
187
parent_maps = self.to_repository.get_parent_map(revision_ids)
189
map(parents.update, parent_maps.itervalues())
190
parents.discard(NULL_REVISION)
191
parents.difference_update(revision_ids)
192
missing_keys = set(('inventories', rev_id) for rev_id in parents)
196
class Inter1and2Helper(object):
197
"""Helper for operations that convert data from model 1 and 2
199
This is for use by fetchers and converters.
202
def __init__(self, source):
205
:param source: The repository data comes from
209
def iter_rev_trees(self, revs):
210
"""Iterate through RevisionTrees efficiently.
212
Additionally, the inventory's revision_id is set if unset.
214
Trees are retrieved in batches of 100, and then yielded in the order
217
:param revs: A list of revision ids
219
# In case that revs is not a list.
222
for tree in self.source.revision_trees(revs[:100]):
223
if tree.inventory.revision_id is None:
224
tree.inventory.revision_id = tree.get_revision_id()
228
def _find_root_ids(self, revs, parent_map, graph):
230
planned_versions = {}
231
for tree in self.iter_rev_trees(revs):
232
revision_id = tree.inventory.root.revision
233
root_id = tree.get_root_id()
234
planned_versions.setdefault(root_id, []).append(revision_id)
235
revision_root[revision_id] = root_id
236
# Find out which parents we don't already know root ids for
238
for revision_parents in parent_map.itervalues():
239
parents.update(revision_parents)
240
parents.difference_update(revision_root.keys() + [NULL_REVISION])
241
# Limit to revisions present in the versionedfile
242
parents = graph.get_parent_map(parents).keys()
243
for tree in self.iter_rev_trees(parents):
244
root_id = tree.get_root_id()
245
revision_root[tree.get_revision_id()] = root_id
246
return revision_root, planned_versions
248
def generate_root_texts(self, revs):
249
"""Generate VersionedFiles for all root ids.
251
:param revs: the revisions to include
253
graph = self.source.get_graph()
254
parent_map = graph.get_parent_map(revs)
255
rev_order = topo_sort(parent_map)
256
rev_id_to_root_id, root_id_to_rev_ids = self._find_root_ids(
257
revs, parent_map, graph)
258
root_id_order = [(rev_id_to_root_id[rev_id], rev_id) for rev_id in
260
# Guaranteed stable, this groups all the file id operations together
261
# retaining topological order within the revisions of a file id.
262
# File id splits and joins would invalidate this, but they don't exist
263
# yet, and are unlikely to in non-rich-root environments anyway.
264
root_id_order.sort(key=operator.itemgetter(0))
265
# Create a record stream containing the roots to create.
267
for key in root_id_order:
268
root_id, rev_id = key
269
rev_parents = parent_map[rev_id]
270
# We drop revision parents with different file-ids, because
271
# that represents a rename of the root to a different location
272
# - its not actually a parent for us. (We could look for that
273
# file id in the revision tree at considerably more expense,
274
# but for now this is sufficient (and reconcile will catch and
275
# correct this anyway).
276
# When a parent revision is a ghost, we guess that its root id
277
# was unchanged (rather than trimming it from the parent list).
278
parent_keys = tuple((root_id, parent) for parent in rev_parents
279
if parent != NULL_REVISION and
280
rev_id_to_root_id.get(parent, root_id) == root_id)
281
yield FulltextContentFactory(key, parent_keys, None, '')
282
return [('texts', yield_roots())]