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., 59 Temple Place, Suite 330, Boston, MA 02111-1307 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()
83
self.from_repository.unlock()
86
"""Primary worker function.
88
This initialises all the needed variables, and then fetches the
89
requested revisions, finally clearing the progress bar.
91
# Roughly this is what we're aiming for fetch to become:
93
# missing = self.sink.insert_stream(self.source.get_stream(search))
95
# missing = self.sink.insert_stream(self.source.get_items(missing))
98
self.file_ids_names = {}
99
pb = bzrlib.ui.ui_factory.nested_progress_bar()
100
pb.show_pct = pb.show_count = False
102
pb.update("Finding revisions", 0, 2)
103
search = self._revids_to_fetch()
106
pb.update("Fetching revisions", 1, 2)
107
self._fetch_everything_for_search(search)
111
def _fetch_everything_for_search(self, search):
112
"""Fetch all data for the given set of revisions."""
113
# The first phase is "file". We pass the progress bar for it directly
114
# into item_keys_introduced_by, which has more information about how
115
# that phase is progressing than we do. Progress updates for the other
116
# phases are taken care of in this function.
117
# XXX: there should be a clear owner of the progress reporting. Perhaps
118
# item_keys_introduced_by should have a richer API than it does at the
119
# moment, so that it can feed the progress information back to this
121
if (self.from_repository._format.rich_root_data and
122
not self.to_repository._format.rich_root_data):
123
raise errors.IncompatibleRepositories(
124
self.from_repository, self.to_repository,
125
"different rich-root support")
126
pb = bzrlib.ui.ui_factory.nested_progress_bar()
128
pb.update("Get stream source")
129
source = self.from_repository._get_source(
130
self.to_repository._format)
131
stream = source.get_stream(search)
132
from_format = self.from_repository._format
133
pb.update("Inserting stream")
134
resume_tokens, missing_keys = self.sink.insert_stream(
135
stream, from_format, [])
137
pb.update("Missing keys")
138
stream = source.get_stream_for_missing_keys(missing_keys)
139
pb.update("Inserting missing keys")
140
resume_tokens, missing_keys = self.sink.insert_stream(
141
stream, from_format, resume_tokens)
143
raise AssertionError(
144
"second push failed to complete a fetch %r." % (
147
raise AssertionError(
148
"second push failed to commit the fetch %r." % (
150
pb.update("Finishing stream")
155
def _revids_to_fetch(self):
156
"""Determines the exact revisions needed from self.from_repository to
157
install self._last_revision in self.to_repository.
159
If no revisions need to be fetched, then this just returns None.
161
if self._fetch_spec is not None:
162
return self._fetch_spec
163
if self._last_revision is NULL_REVISION:
164
# explicit limit of no revisions needed
166
if (self._last_revision is not None and
167
self.to_repository.has_revision(self._last_revision)):
170
return self.to_repository.search_missing_revision_ids(
171
self.from_repository, self._last_revision,
172
find_ghosts=self.find_ghosts)
173
except errors.NoSuchRevision, e:
174
raise InstallFailed([self._last_revision])
177
class Inter1and2Helper(object):
178
"""Helper for operations that convert data from model 1 and 2
180
This is for use by fetchers and converters.
183
def __init__(self, source):
186
:param source: The repository data comes from
190
def iter_rev_trees(self, revs):
191
"""Iterate through RevisionTrees efficiently.
193
Additionally, the inventory's revision_id is set if unset.
195
Trees are retrieved in batches of 100, and then yielded in the order
198
:param revs: A list of revision ids
200
# In case that revs is not a list.
203
for tree in self.source.revision_trees(revs[:100]):
204
if tree.inventory.revision_id is None:
205
tree.inventory.revision_id = tree.get_revision_id()
209
def _find_root_ids(self, revs, parent_map, graph):
211
planned_versions = {}
212
for tree in self.iter_rev_trees(revs):
213
revision_id = tree.inventory.root.revision
214
root_id = tree.get_root_id()
215
planned_versions.setdefault(root_id, []).append(revision_id)
216
revision_root[revision_id] = root_id
217
# Find out which parents we don't already know root ids for
219
for revision_parents in parent_map.itervalues():
220
parents.update(revision_parents)
221
parents.difference_update(revision_root.keys() + [NULL_REVISION])
222
# Limit to revisions present in the versionedfile
223
parents = graph.get_parent_map(parents).keys()
224
for tree in self.iter_rev_trees(parents):
225
root_id = tree.get_root_id()
226
revision_root[tree.get_revision_id()] = root_id
227
return revision_root, planned_versions
229
def generate_root_texts(self, revs):
230
"""Generate VersionedFiles for all root ids.
232
:param revs: the revisions to include
234
graph = self.source.get_graph()
235
parent_map = graph.get_parent_map(revs)
236
rev_order = topo_sort(parent_map)
237
rev_id_to_root_id, root_id_to_rev_ids = self._find_root_ids(
238
revs, parent_map, graph)
239
root_id_order = [(rev_id_to_root_id[rev_id], rev_id) for rev_id in
241
# Guaranteed stable, this groups all the file id operations together
242
# retaining topological order within the revisions of a file id.
243
# File id splits and joins would invalidate this, but they don't exist
244
# yet, and are unlikely to in non-rich-root environments anyway.
245
root_id_order.sort(key=operator.itemgetter(0))
246
# Create a record stream containing the roots to create.
248
for key in root_id_order:
249
root_id, rev_id = key
250
rev_parents = parent_map[rev_id]
251
# We drop revision parents with different file-ids, because
252
# that represents a rename of the root to a different location
253
# - its not actually a parent for us. (We could look for that
254
# file id in the revision tree at considerably more expense,
255
# but for now this is sufficient (and reconcile will catch and
256
# correct this anyway).
257
# When a parent revision is a ghost, we guess that its root id
258
# was unchanged (rather than trimming it from the parent list).
259
parent_keys = tuple((root_id, parent) for parent in rev_parents
260
if parent != NULL_REVISION and
261
rev_id_to_root_id.get(parent, root_id) == root_id)
262
yield FulltextContentFactory(key, parent_keys, None, '')
263
return [('texts', yield_roots())]