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. If absent or None,
60
this method will just create one on top of the stack.
62
if to_repository.has_same_location(from_repository):
63
# repository.fetch should be taking care of this case.
64
raise errors.BzrError('RepoFetcher run '
65
'between two objects at the same location: '
66
'%r and %r' % (to_repository, from_repository))
67
self.to_repository = to_repository
68
self.from_repository = from_repository
69
self.sink = to_repository._get_sink()
70
# must not mutate self._last_revision as its potentially a shared instance
71
self._last_revision = last_revision
72
self._fetch_spec = fetch_spec
73
self.find_ghosts = find_ghosts
75
self.pb = bzrlib.ui.ui_factory.nested_progress_bar()
77
symbol_versioning.warn(
78
symbol_versioning.deprecated_in((1, 14, 0))
79
% "pb parameter to RepoFetcher.__init__")
81
self.from_repository.lock_read()
85
self.from_repository.unlock()
88
"""Primary worker function.
90
This initialises all the needed variables, and then fetches the
91
requested revisions, finally clearing the progress bar.
93
# Roughly this is what we're aiming for fetch to become:
95
# missing = self.sink.insert_stream(self.source.get_stream(search))
97
# missing = self.sink.insert_stream(self.source.get_items(missing))
100
self.file_ids_names = {}
101
pp = ProgressPhase('Transferring', 4, self.pb)
104
search = self._revids_to_fetch()
107
self._fetch_everything_for_search(search, pp)
111
def _fetch_everything_for_search(self, search, pp):
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
self.pb = bzrlib.ui.ui_factory.nested_progress_bar()
128
source = self.from_repository._get_source(
129
self.to_repository._format)
130
stream = source.get_stream(search)
131
from_format = self.from_repository._format
132
resume_tokens, missing_keys = self.sink.insert_stream(
133
stream, from_format, [])
135
stream = source.get_stream_for_missing_keys(missing_keys)
136
resume_tokens, missing_keys = self.sink.insert_stream(
137
stream, from_format, resume_tokens)
139
raise AssertionError(
140
"second push failed to complete a fetch %r." % (
143
raise AssertionError(
144
"second push failed to commit the fetch %r." % (
148
if self.pb is not None:
151
def _revids_to_fetch(self):
152
"""Determines the exact revisions needed from self.from_repository to
153
install self._last_revision in self.to_repository.
155
If no revisions need to be fetched, then this just returns None.
157
if self._fetch_spec is not None:
158
return self._fetch_spec
159
mutter('fetch up to rev {%s}', self._last_revision)
160
if self._last_revision is NULL_REVISION:
161
# explicit limit of no revisions needed
163
if (self._last_revision is not None and
164
self.to_repository.has_revision(self._last_revision)):
167
return self.to_repository.search_missing_revision_ids(
168
self.from_repository, self._last_revision,
169
find_ghosts=self.find_ghosts)
170
except errors.NoSuchRevision, e:
171
raise InstallFailed([self._last_revision])
174
class Inter1and2Helper(object):
175
"""Helper for operations that convert data from model 1 and 2
177
This is for use by fetchers and converters.
180
def __init__(self, source):
183
:param source: The repository data comes from
187
def iter_rev_trees(self, revs):
188
"""Iterate through RevisionTrees efficiently.
190
Additionally, the inventory's revision_id is set if unset.
192
Trees are retrieved in batches of 100, and then yielded in the order
195
:param revs: A list of revision ids
197
# In case that revs is not a list.
200
for tree in self.source.revision_trees(revs[:100]):
201
if tree.inventory.revision_id is None:
202
tree.inventory.revision_id = tree.get_revision_id()
206
def _find_root_ids(self, revs, parent_map, graph):
208
planned_versions = {}
209
for tree in self.iter_rev_trees(revs):
210
revision_id = tree.inventory.root.revision
211
root_id = tree.get_root_id()
212
planned_versions.setdefault(root_id, []).append(revision_id)
213
revision_root[revision_id] = root_id
214
# Find out which parents we don't already know root ids for
216
for revision_parents in parent_map.itervalues():
217
parents.update(revision_parents)
218
parents.difference_update(revision_root.keys() + [NULL_REVISION])
219
# Limit to revisions present in the versionedfile
220
parents = graph.get_parent_map(parents).keys()
221
for tree in self.iter_rev_trees(parents):
222
root_id = tree.get_root_id()
223
revision_root[tree.get_revision_id()] = root_id
224
return revision_root, planned_versions
226
def generate_root_texts(self, revs):
227
"""Generate VersionedFiles for all root ids.
229
:param revs: the revisions to include
231
graph = self.source.get_graph()
232
parent_map = graph.get_parent_map(revs)
233
rev_order = topo_sort(parent_map)
234
rev_id_to_root_id, root_id_to_rev_ids = self._find_root_ids(
235
revs, parent_map, graph)
236
root_id_order = [(rev_id_to_root_id[rev_id], rev_id) for rev_id in
238
# Guaranteed stable, this groups all the file id operations together
239
# retaining topological order within the revisions of a file id.
240
# File id splits and joins would invalidate this, but they don't exist
241
# yet, and are unlikely to in non-rich-root environments anyway.
242
root_id_order.sort(key=operator.itemgetter(0))
243
# Create a record stream containing the roots to create.
245
for key in root_id_order:
246
root_id, rev_id = key
247
rev_parents = parent_map[rev_id]
248
# We drop revision parents with different file-ids, because
249
# that represents a rename of the root to a different location
250
# - its not actually a parent for us. (We could look for that
251
# file id in the revision tree at considerably more expense,
252
# but for now this is sufficient (and reconcile will catch and
253
# correct this anyway).
254
# When a parent revision is a ghost, we guess that its root id
255
# was unchanged (rather than trimming it from the parent list).
256
parent_keys = tuple((root_id, parent) for parent in rev_parents
257
if parent != NULL_REVISION and
258
rev_id_to_root_id.get(parent, root_id) == root_id)
259
yield FulltextContentFactory(key, parent_keys, None, '')
260
return [('texts', yield_roots())]
added
removed
bzrlib/fetch.py
