1
# Copyright (C) 2005-2011 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
28
from bzrlib.lazy_import import lazy_import
29
lazy_import(globals(), """
40
from bzrlib.revision import NULL_REVISION
41
from bzrlib.trace import mutter
44
class RepoFetcher(object):
45
"""Pull revisions and texts from one repository to another.
47
This should not be used directly, it's essential a object to encapsulate
48
the logic in InterRepository.fetch().
51
def __init__(self, to_repository, from_repository, last_revision=None,
52
find_ghosts=True, fetch_spec=None):
53
"""Create a repo fetcher.
55
:param last_revision: If set, try to limit to the data this revision
57
:param find_ghosts: If True search the entire history for ghosts.
59
# repository.fetch has the responsibility for short-circuiting
60
# attempts to copy between a repository and itself.
61
self.to_repository = to_repository
62
self.from_repository = from_repository
63
self.sink = to_repository._get_sink()
64
# must not mutate self._last_revision as its potentially a shared instance
65
self._last_revision = last_revision
66
self._fetch_spec = fetch_spec
67
self.find_ghosts = find_ghosts
68
self.from_repository.lock_read()
69
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
70
self.from_repository, self.from_repository._format,
71
self.to_repository, self.to_repository._format)
75
self.from_repository.unlock()
78
"""Primary worker function.
80
This initialises all the needed variables, and then fetches the
81
requested revisions, finally clearing the progress bar.
83
# Roughly this is what we're aiming for fetch to become:
85
# missing = self.sink.insert_stream(self.source.get_stream(search))
87
# missing = self.sink.insert_stream(self.source.get_items(missing))
90
self.file_ids_names = {}
91
pb = ui.ui_factory.nested_progress_bar()
92
pb.show_pct = pb.show_count = False
94
pb.update("Finding revisions", 0, 2)
95
search = self._revids_to_fetch()
98
pb.update("Fetching revisions", 1, 2)
99
self._fetch_everything_for_search(search)
103
def _fetch_everything_for_search(self, search):
104
"""Fetch all data for the given set of revisions."""
105
# The first phase is "file". We pass the progress bar for it directly
106
# into item_keys_introduced_by, which has more information about how
107
# that phase is progressing than we do. Progress updates for the other
108
# phases are taken care of in this function.
109
# XXX: there should be a clear owner of the progress reporting. Perhaps
110
# item_keys_introduced_by should have a richer API than it does at the
111
# moment, so that it can feed the progress information back to this
113
if (self.from_repository._format.rich_root_data and
114
not self.to_repository._format.rich_root_data):
115
raise errors.IncompatibleRepositories(
116
self.from_repository, self.to_repository,
117
"different rich-root support")
118
pb = ui.ui_factory.nested_progress_bar()
120
pb.update("Get stream source")
121
source = self.from_repository._get_source(
122
self.to_repository._format)
123
stream = source.get_stream(search)
124
from_format = self.from_repository._format
125
pb.update("Inserting stream")
126
resume_tokens, missing_keys = self.sink.insert_stream(
127
stream, from_format, [])
129
pb.update("Missing keys")
130
stream = source.get_stream_for_missing_keys(missing_keys)
131
pb.update("Inserting missing keys")
132
resume_tokens, missing_keys = self.sink.insert_stream(
133
stream, from_format, resume_tokens)
135
raise AssertionError(
136
"second push failed to complete a fetch %r." % (
139
raise AssertionError(
140
"second push failed to commit the fetch %r." % (
142
pb.update("Finishing stream")
147
def _revids_to_fetch(self):
148
"""Determines the exact revisions needed from self.from_repository to
149
install self._last_revision in self.to_repository.
151
If no revisions need to be fetched, then this just returns None.
153
if self._fetch_spec is not None:
154
return self._fetch_spec
155
mutter('fetch up to rev {%s}', self._last_revision)
156
if self._last_revision is NULL_REVISION:
157
# explicit limit of no revisions needed
159
return self.to_repository.search_missing_revision_ids(
160
self.from_repository, self._last_revision,
161
find_ghosts=self.find_ghosts)
164
class Inter1and2Helper(object):
165
"""Helper for operations that convert data from model 1 and 2
167
This is for use by fetchers and converters.
170
# This is a class variable so that the test suite can override it.
171
known_graph_threshold = 100
173
def __init__(self, source):
176
:param source: The repository data comes from
180
def iter_rev_trees(self, revs):
181
"""Iterate through RevisionTrees efficiently.
183
Additionally, the inventory's revision_id is set if unset.
185
Trees are retrieved in batches of 100, and then yielded in the order
188
:param revs: A list of revision ids
190
# In case that revs is not a list.
193
for tree in self.source.revision_trees(revs[:100]):
194
if tree.inventory.revision_id is None:
195
tree.inventory.revision_id = tree.get_revision_id()
199
def _find_root_ids(self, revs, parent_map, graph):
201
for tree in self.iter_rev_trees(revs):
202
revision_id = tree.inventory.root.revision
203
root_id = tree.get_root_id()
204
revision_root[revision_id] = root_id
205
# Find out which parents we don't already know root ids for
207
for revision_parents in parent_map.itervalues():
208
parents.update(revision_parents)
209
parents.difference_update(revision_root.keys() + [NULL_REVISION])
210
# Limit to revisions present in the versionedfile
211
parents = graph.get_parent_map(parents).keys()
212
for tree in self.iter_rev_trees(parents):
213
root_id = tree.get_root_id()
214
revision_root[tree.get_revision_id()] = root_id
217
def generate_root_texts(self, revs):
218
"""Generate VersionedFiles for all root ids.
220
:param revs: the revisions to include
222
graph = self.source.get_graph()
223
parent_map = graph.get_parent_map(revs)
224
rev_order = tsort.topo_sort(parent_map)
225
rev_id_to_root_id = self._find_root_ids(revs, parent_map, graph)
226
root_id_order = [(rev_id_to_root_id[rev_id], rev_id) for rev_id in
228
# Guaranteed stable, this groups all the file id operations together
229
# retaining topological order within the revisions of a file id.
230
# File id splits and joins would invalidate this, but they don't exist
231
# yet, and are unlikely to in non-rich-root environments anyway.
232
root_id_order.sort(key=operator.itemgetter(0))
233
# Create a record stream containing the roots to create.
234
if len(revs) > self.known_graph_threshold:
235
graph = self.source.get_known_graph_ancestry(revs)
236
new_roots_stream = _new_root_data_stream(
237
root_id_order, rev_id_to_root_id, parent_map, self.source, graph)
238
return [('texts', new_roots_stream)]
241
def _new_root_data_stream(
242
root_keys_to_create, rev_id_to_root_id_map, parent_map, repo, graph=None):
243
"""Generate a texts substream of synthesised root entries.
245
Used in fetches that do rich-root upgrades.
247
:param root_keys_to_create: iterable of (root_id, rev_id) pairs describing
248
the root entries to create.
249
:param rev_id_to_root_id_map: dict of known rev_id -> root_id mappings for
250
calculating the parents. If a parent rev_id is not found here then it
251
will be recalculated.
252
:param parent_map: a parent map for all the revisions in
254
:param graph: a graph to use instead of repo.get_graph().
256
for root_key in root_keys_to_create:
257
root_id, rev_id = root_key
258
parent_keys = _parent_keys_for_root_version(
259
root_id, rev_id, rev_id_to_root_id_map, parent_map, repo, graph)
260
yield versionedfile.FulltextContentFactory(
261
root_key, parent_keys, None, '')
264
def _parent_keys_for_root_version(
265
root_id, rev_id, rev_id_to_root_id_map, parent_map, repo, graph=None):
266
"""Get the parent keys for a given root id.
268
A helper function for _new_root_data_stream.
270
# Include direct parents of the revision, but only if they used the same
271
# root_id and are heads.
272
rev_parents = parent_map[rev_id]
274
for parent_id in rev_parents:
275
if parent_id == NULL_REVISION:
277
if parent_id not in rev_id_to_root_id_map:
278
# We probably didn't read this revision, go spend the extra effort
281
tree = repo.revision_tree(parent_id)
282
except errors.NoSuchRevision:
283
# Ghost, fill out rev_id_to_root_id in case we encounter this
285
# But set parent_root_id to None since we don't really know
286
parent_root_id = None
288
parent_root_id = tree.get_root_id()
289
rev_id_to_root_id_map[parent_id] = None
291
# rev_id_to_root_id_map[parent_id] = parent_root_id
292
# memory consumption maybe?
294
parent_root_id = rev_id_to_root_id_map[parent_id]
295
if root_id == parent_root_id:
296
# With stacking we _might_ want to refer to a non-local revision,
297
# but this code path only applies when we have the full content
298
# available, so ghosts really are ghosts, not just the edge of
300
parent_ids.append(parent_id)
302
# root_id may be in the parent anyway.
304
tree = repo.revision_tree(parent_id)
305
except errors.NoSuchRevision:
306
# ghost, can't refer to it.
310
parent_ids.append(tree.inventory[root_id].revision)
311
except errors.NoSuchId:
314
# Drop non-head parents
316
graph = repo.get_graph()
317
heads = graph.heads(parent_ids)
319
for parent_id in parent_ids:
320
if parent_id in heads and parent_id not in selected_ids:
321
selected_ids.append(parent_id)
322
parent_keys = [(root_id, parent_id) for parent_id in selected_ids]