1
# Copyright (C) 2005-2010 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(), """
42
from bzrlib.revision import NULL_REVISION
43
from bzrlib.trace import mutter
46
class RepoFetcher(object):
47
"""Pull revisions and texts from one repository to another.
49
This should not be used directly, it's essential a object to encapsulate
50
the logic in InterRepository.fetch().
53
def __init__(self, to_repository, from_repository, last_revision=None,
54
find_ghosts=True, fetch_spec=None):
55
"""Create a repo fetcher.
57
:param last_revision: If set, try to limit to the data this revision
59
:param find_ghosts: If True search the entire history for ghosts.
61
# repository.fetch has the responsibility for short-circuiting
62
# attempts to copy between a repository and itself.
63
self.to_repository = to_repository
64
self.from_repository = from_repository
65
self.sink = to_repository._get_sink()
66
# must not mutate self._last_revision as its potentially a shared instance
67
self._last_revision = last_revision
68
self._fetch_spec = fetch_spec
69
self.find_ghosts = find_ghosts
70
self.from_repository.lock_read()
71
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
72
self.from_repository, self.from_repository._format,
73
self.to_repository, self.to_repository._format)
77
self.from_repository.unlock()
80
"""Primary worker function.
82
This initialises all the needed variables, and then fetches the
83
requested revisions, finally clearing the progress bar.
85
# Roughly this is what we're aiming for fetch to become:
87
# missing = self.sink.insert_stream(self.source.get_stream(search))
89
# missing = self.sink.insert_stream(self.source.get_items(missing))
92
self.file_ids_names = {}
93
pb = ui.ui_factory.nested_progress_bar()
94
pb.show_pct = pb.show_count = False
96
pb.update("Finding revisions", 0, 2)
97
search = self._revids_to_fetch()
100
pb.update("Fetching revisions", 1, 2)
101
self._fetch_everything_for_search(search)
105
def _fetch_everything_for_search(self, search):
106
"""Fetch all data for the given set of revisions."""
107
# The first phase is "file". We pass the progress bar for it directly
108
# into item_keys_introduced_by, which has more information about how
109
# that phase is progressing than we do. Progress updates for the other
110
# phases are taken care of in this function.
111
# XXX: there should be a clear owner of the progress reporting. Perhaps
112
# item_keys_introduced_by should have a richer API than it does at the
113
# moment, so that it can feed the progress information back to this
115
if (self.from_repository._format.rich_root_data and
116
not self.to_repository._format.rich_root_data):
117
raise errors.IncompatibleRepositories(
118
self.from_repository, self.to_repository,
119
"different rich-root support")
120
pb = ui.ui_factory.nested_progress_bar()
122
pb.update("Get stream source")
123
source = self.from_repository._get_source(
124
self.to_repository._format)
125
stream = source.get_stream(search)
126
from_format = self.from_repository._format
127
pb.update("Inserting stream")
128
resume_tokens, missing_keys = self.sink.insert_stream(
129
stream, from_format, [])
130
if self.to_repository._fallback_repositories:
132
self._parent_inventories(search.get_keys()))
134
pb.update("Missing keys")
135
stream = source.get_stream_for_missing_keys(missing_keys)
136
pb.update("Inserting missing keys")
137
resume_tokens, missing_keys = self.sink.insert_stream(
138
stream, from_format, resume_tokens)
140
raise AssertionError(
141
"second push failed to complete a fetch %r." % (
144
raise AssertionError(
145
"second push failed to commit the fetch %r." % (
147
pb.update("Finishing stream")
152
def _revids_to_fetch(self):
153
"""Determines the exact revisions needed from self.from_repository to
154
install self._last_revision in self.to_repository.
156
If no revisions need to be fetched, then this just returns None.
158
if self._fetch_spec is not None:
159
return self._fetch_spec
160
mutter('fetch up to rev {%s}', self._last_revision)
161
if self._last_revision is NULL_REVISION:
162
# explicit limit of no revisions needed
164
return self.to_repository.search_missing_revision_ids(
165
self.from_repository, self._last_revision,
166
find_ghosts=self.find_ghosts)
168
def _parent_inventories(self, revision_ids):
169
# Find all the parent revisions referenced by the stream, but
170
# not present in the stream, and make sure we send their
172
parent_maps = self.to_repository.get_parent_map(revision_ids)
174
map(parents.update, parent_maps.itervalues())
175
parents.discard(NULL_REVISION)
176
parents.difference_update(revision_ids)
177
missing_keys = set(('inventories', rev_id) for rev_id in parents)
181
class Inter1and2Helper(object):
182
"""Helper for operations that convert data from model 1 and 2
184
This is for use by fetchers and converters.
187
def __init__(self, source):
190
:param source: The repository data comes from
194
def iter_rev_trees(self, revs):
195
"""Iterate through RevisionTrees efficiently.
197
Additionally, the inventory's revision_id is set if unset.
199
Trees are retrieved in batches of 100, and then yielded in the order
202
:param revs: A list of revision ids
204
# In case that revs is not a list.
207
for tree in self.source.revision_trees(revs[:100]):
208
if tree.inventory.revision_id is None:
209
tree.inventory.revision_id = tree.get_revision_id()
213
def _find_root_ids(self, revs, parent_map, graph):
215
for tree in self.iter_rev_trees(revs):
216
revision_id = tree.inventory.root.revision
217
root_id = tree.get_root_id()
218
revision_root[revision_id] = root_id
219
# Find out which parents we don't already know root ids for
221
for revision_parents in parent_map.itervalues():
222
parents.update(revision_parents)
223
parents.difference_update(revision_root.keys() + [NULL_REVISION])
224
# Limit to revisions present in the versionedfile
225
parents = graph.get_parent_map(parents).keys()
226
for tree in self.iter_rev_trees(parents):
227
root_id = tree.get_root_id()
228
revision_root[tree.get_revision_id()] = root_id
231
def generate_root_texts(self, revs):
232
"""Generate VersionedFiles for all root ids.
234
:param revs: the revisions to include
236
graph = self.source.get_graph()
237
parent_map = graph.get_parent_map(revs)
238
rev_order = tsort.topo_sort(parent_map)
239
rev_id_to_root_id = self._find_root_ids(revs, parent_map, graph)
240
root_id_order = [(rev_id_to_root_id[rev_id], rev_id) for rev_id in
242
# Guaranteed stable, this groups all the file id operations together
243
# retaining topological order within the revisions of a file id.
244
# File id splits and joins would invalidate this, but they don't exist
245
# yet, and are unlikely to in non-rich-root environments anyway.
246
root_id_order.sort(key=operator.itemgetter(0))
247
# Create a record stream containing the roots to create.
249
# XXX: not covered by tests, should have a flag to always run
250
# this. -- mbp 20100129
251
graph = _get_rich_root_heads_graph(self.source, revs)
252
new_roots_stream = _new_root_data_stream(
253
root_id_order, rev_id_to_root_id, parent_map, self.source, graph)
254
return [('texts', new_roots_stream)]
257
def _get_rich_root_heads_graph(source_repo, revision_ids):
258
"""Get a Graph object suitable for asking heads() for new rich roots."""
259
st = static_tuple.StaticTuple
260
revision_keys = [st(r_id).intern() for r_id in revision_ids]
261
known_graph = source_repo.revisions.get_known_graph_ancestry(
263
return _mod_graph.GraphThunkIdsToKeys(known_graph)
266
def _new_root_data_stream(
267
root_keys_to_create, rev_id_to_root_id_map, parent_map, repo, graph=None):
268
"""Generate a texts substream of synthesised root entries.
270
Used in fetches that do rich-root upgrades.
272
:param root_keys_to_create: iterable of (root_id, rev_id) pairs describing
273
the root entries to create.
274
:param rev_id_to_root_id_map: dict of known rev_id -> root_id mappings for
275
calculating the parents. If a parent rev_id is not found here then it
276
will be recalculated.
277
:param parent_map: a parent map for all the revisions in
279
:param graph: a graph to use instead of repo.get_graph().
281
for root_key in root_keys_to_create:
282
root_id, rev_id = root_key
283
parent_keys = _parent_keys_for_root_version(
284
root_id, rev_id, rev_id_to_root_id_map, parent_map, repo, graph)
285
yield versionedfile.FulltextContentFactory(
286
root_key, parent_keys, None, '')
289
def _parent_keys_for_root_version(
290
root_id, rev_id, rev_id_to_root_id_map, parent_map, repo, graph=None):
291
"""Get the parent keys for a given root id.
293
A helper function for _new_root_data_stream.
295
# Include direct parents of the revision, but only if they used the same
296
# root_id and are heads.
297
rev_parents = parent_map[rev_id]
299
for parent_id in rev_parents:
300
if parent_id == NULL_REVISION:
302
if parent_id not in rev_id_to_root_id_map:
303
# We probably didn't read this revision, go spend the extra effort
306
tree = repo.revision_tree(parent_id)
307
except errors.NoSuchRevision:
308
# Ghost, fill out rev_id_to_root_id in case we encounter this
310
# But set parent_root_id to None since we don't really know
311
parent_root_id = None
313
parent_root_id = tree.get_root_id()
314
rev_id_to_root_id_map[parent_id] = None
316
# rev_id_to_root_id_map[parent_id] = parent_root_id
317
# memory consumption maybe?
319
parent_root_id = rev_id_to_root_id_map[parent_id]
320
if root_id == parent_root_id:
321
# With stacking we _might_ want to refer to a non-local revision,
322
# but this code path only applies when we have the full content
323
# available, so ghosts really are ghosts, not just the edge of
325
parent_ids.append(parent_id)
327
# root_id may be in the parent anyway.
329
tree = repo.revision_tree(parent_id)
330
except errors.NoSuchRevision:
331
# ghost, can't refer to it.
335
parent_ids.append(tree.inventory[root_id].revision)
336
except errors.NoSuchId:
339
# Drop non-head parents
341
graph = repo.get_graph()
342
heads = graph.heads(parent_ids)
344
for parent_id in parent_ids:
345
if parent_id in heads and parent_id not in selected_ids:
346
selected_ids.append(parent_id)
347
parent_keys = [(root_id, parent_id) for parent_id in selected_ids]