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
28
from bzrlib.lazy_import import lazy_import
29
lazy_import(globals(), """
43
from bzrlib.revision import NULL_REVISION
44
from bzrlib.trace import mutter
47
class RepoFetcher(object):
48
"""Pull revisions and texts from one repository to another.
50
This should not be used directly, it's essential a object to encapsulate
51
the logic in InterRepository.fetch().
54
def __init__(self, to_repository, from_repository, last_revision=None,
55
pb=None, find_ghosts=True, fetch_spec=None):
56
"""Create a repo fetcher.
58
:param last_revision: If set, try to limit to the data this revision
60
:param find_ghosts: If True search the entire history for ghosts.
61
:param pb: ProgressBar object to use; deprecated and ignored.
62
This method will just create one on top of the stack.
65
symbol_versioning.warn(
66
symbol_versioning.deprecated_in((1, 14, 0))
67
% "pb parameter to RepoFetcher.__init__")
68
# and for simplicity it is in fact ignored
69
# repository.fetch has the responsibility for short-circuiting
70
# attempts to copy between a repository and itself.
71
self.to_repository = to_repository
72
self.from_repository = from_repository
73
self.sink = to_repository._get_sink()
74
# must not mutate self._last_revision as its potentially a shared instance
75
self._last_revision = last_revision
76
self._fetch_spec = fetch_spec
77
self.find_ghosts = find_ghosts
78
self.from_repository.lock_read()
79
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
80
self.from_repository, self.from_repository._format,
81
self.to_repository, self.to_repository._format)
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
pb = ui.ui_factory.nested_progress_bar()
102
pb.show_pct = pb.show_count = False
104
pb.update("Finding revisions", 0, 2)
105
search = self._revids_to_fetch()
108
pb.update("Fetching revisions", 1, 2)
109
self._fetch_everything_for_search(search)
113
def _fetch_everything_for_search(self, search):
114
"""Fetch all data for the given set of revisions."""
115
# The first phase is "file". We pass the progress bar for it directly
116
# into item_keys_introduced_by, which has more information about how
117
# that phase is progressing than we do. Progress updates for the other
118
# phases are taken care of in this function.
119
# XXX: there should be a clear owner of the progress reporting. Perhaps
120
# item_keys_introduced_by should have a richer API than it does at the
121
# moment, so that it can feed the progress information back to this
123
if (self.from_repository._format.rich_root_data and
124
not self.to_repository._format.rich_root_data):
125
raise errors.IncompatibleRepositories(
126
self.from_repository, self.to_repository,
127
"different rich-root support")
128
pb = ui.ui_factory.nested_progress_bar()
130
pb.update("Get stream source")
131
source = self.from_repository._get_source(
132
self.to_repository._format)
133
stream = source.get_stream(search)
134
from_format = self.from_repository._format
135
pb.update("Inserting stream")
136
resume_tokens, missing_keys = self.sink.insert_stream(
137
stream, from_format, [])
138
if self.to_repository._fallback_repositories:
140
self._parent_inventories(search.get_keys()))
142
pb.update("Missing keys")
143
stream = source.get_stream_for_missing_keys(missing_keys)
144
pb.update("Inserting missing keys")
145
resume_tokens, missing_keys = self.sink.insert_stream(
146
stream, from_format, resume_tokens)
148
raise AssertionError(
149
"second push failed to complete a fetch %r." % (
152
raise AssertionError(
153
"second push failed to commit the fetch %r." % (
155
pb.update("Finishing stream")
160
def _revids_to_fetch(self):
161
"""Determines the exact revisions needed from self.from_repository to
162
install self._last_revision in self.to_repository.
164
If no revisions need to be fetched, then this just returns None.
166
if self._fetch_spec is not None:
167
return self._fetch_spec
168
mutter('fetch up to rev {%s}', self._last_revision)
169
if self._last_revision is NULL_REVISION:
170
# explicit limit of no revisions needed
172
return self.to_repository.search_missing_revision_ids(
173
self.from_repository, self._last_revision,
174
find_ghosts=self.find_ghosts)
176
def _parent_inventories(self, revision_ids):
177
# Find all the parent revisions referenced by the stream, but
178
# not present in the stream, and make sure we send their
180
parent_maps = self.to_repository.get_parent_map(revision_ids)
182
map(parents.update, parent_maps.itervalues())
183
parents.discard(NULL_REVISION)
184
parents.difference_update(revision_ids)
185
missing_keys = set(('inventories', rev_id) for rev_id in parents)
189
class Inter1and2Helper(object):
190
"""Helper for operations that convert data from model 1 and 2
192
This is for use by fetchers and converters.
195
def __init__(self, source):
198
:param source: The repository data comes from
202
def iter_rev_trees(self, revs):
203
"""Iterate through RevisionTrees efficiently.
205
Additionally, the inventory's revision_id is set if unset.
207
Trees are retrieved in batches of 100, and then yielded in the order
210
:param revs: A list of revision ids
212
# In case that revs is not a list.
215
for tree in self.source.revision_trees(revs[:100]):
216
if tree.inventory.revision_id is None:
217
tree.inventory.revision_id = tree.get_revision_id()
221
def _find_root_ids(self, revs, parent_map, graph):
223
for tree in self.iter_rev_trees(revs):
224
revision_id = tree.inventory.root.revision
225
root_id = tree.get_root_id()
226
revision_root[revision_id] = root_id
227
# Find out which parents we don't already know root ids for
229
for revision_parents in parent_map.itervalues():
230
parents.update(revision_parents)
231
parents.difference_update(revision_root.keys() + [NULL_REVISION])
232
# Limit to revisions present in the versionedfile
233
parents = graph.get_parent_map(parents).keys()
234
for tree in self.iter_rev_trees(parents):
235
root_id = tree.get_root_id()
236
revision_root[tree.get_revision_id()] = root_id
239
def generate_root_texts(self, revs):
240
"""Generate VersionedFiles for all root ids.
242
:param revs: the revisions to include
244
graph = self.source.get_graph()
245
parent_map = graph.get_parent_map(revs)
246
rev_order = tsort.topo_sort(parent_map)
247
rev_id_to_root_id = self._find_root_ids(revs, parent_map, graph)
248
root_id_order = [(rev_id_to_root_id[rev_id], rev_id) for rev_id in
250
# Guaranteed stable, this groups all the file id operations together
251
# retaining topological order within the revisions of a file id.
252
# File id splits and joins would invalidate this, but they don't exist
253
# yet, and are unlikely to in non-rich-root environments anyway.
254
root_id_order.sort(key=operator.itemgetter(0))
255
# Create a record stream containing the roots to create.
257
graph = _get_rich_root_heads_graph(self.source_repo, revs)
258
new_roots_stream = _new_root_data_stream(
259
root_id_order, rev_id_to_root_id, parent_map, self.source, graph)
260
return [('texts', new_roots_stream)]
263
def _get_rich_root_heads_graph(source_repo, revision_ids):
264
"""Get a Graph object suitable for asking heads() for new rich roots."""
265
st = static_tuple.StaticTuple
266
revision_keys = [st(r_id).intern() for r_id in revision_ids]
267
known_graph = source_repo.revisions.get_known_graph_ancestry(
269
return _mod_graph.GraphThunkIdsToKeys(known_graph)
272
def _new_root_data_stream(
273
root_keys_to_create, rev_id_to_root_id_map, parent_map, repo, graph=None):
274
"""Generate a texts substream of synthesised root entries.
276
Used in fetches that do rich-root upgrades.
278
:param root_keys_to_create: iterable of (root_id, rev_id) pairs describing
279
the root entries to create.
280
:param rev_id_to_root_id_map: dict of known rev_id -> root_id mappings for
281
calculating the parents. If a parent rev_id is not found here then it
282
will be recalculated.
283
:param parent_map: a parent map for all the revisions in
285
:param graph: a graph to use instead of repo.get_graph().
287
for root_key in root_keys_to_create:
288
root_id, rev_id = root_key
289
parent_keys = _parent_keys_for_root_version(
290
root_id, rev_id, rev_id_to_root_id_map, parent_map, repo, graph)
291
yield versionedfile.FulltextContentFactory(
292
root_key, parent_keys, None, '')
295
def _parent_keys_for_root_version(
296
root_id, rev_id, rev_id_to_root_id_map, parent_map, repo, graph=None):
297
"""Get the parent keys for a given root id.
299
A helper function for _new_root_data_stream.
301
# Include direct parents of the revision, but only if they used the same
302
# root_id and are heads.
303
rev_parents = parent_map[rev_id]
305
for parent_id in rev_parents:
306
if parent_id == NULL_REVISION:
308
if parent_id not in rev_id_to_root_id_map:
309
# We probably didn't read this revision, go spend the extra effort
312
tree = repo.revision_tree(parent_id)
313
except errors.NoSuchRevision:
314
# Ghost, fill out rev_id_to_root_id in case we encounter this
316
# But set parent_root_id to None since we don't really know
317
parent_root_id = None
319
parent_root_id = tree.get_root_id()
320
rev_id_to_root_id_map[parent_id] = None
322
# rev_id_to_root_id_map[parent_id] = parent_root_id
323
# memory consumption maybe?
325
parent_root_id = rev_id_to_root_id_map[parent_id]
326
if root_id == parent_root_id:
327
# With stacking we _might_ want to refer to a non-local revision,
328
# but this code path only applies when we have the full content
329
# available, so ghosts really are ghosts, not just the edge of
331
parent_ids.append(parent_id)
333
# root_id may be in the parent anyway.
335
tree = repo.revision_tree(parent_id)
336
except errors.NoSuchRevision:
337
# ghost, can't refer to it.
341
parent_ids.append(tree.inventory[root_id].revision)
342
except errors.NoSuchId:
345
# Drop non-head parents
347
graph = repo.get_graph()
348
heads = graph.heads(parent_ids)
350
for parent_id in parent_ids:
351
if parent_id in heads and parent_id not in selected_ids:
352
selected_ids.append(parent_id)
353
parent_keys = [(root_id, parent_id) for parent_id in selected_ids]
added
removed
bzrlib/fetch.py
