1
# Copyright (C) 2005 by 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
25
The copying is done in a slightly complicated order. We don't want to
26
add a revision to the store until everything it refers to is also
27
stored, so that if a revision is present we can totally recreate it.
28
However, we can't know what files are included in a revision until we
29
read its inventory. So we query the inventory store of the source for
30
the ids we need, and then pull those ids and finally actually join
35
import bzrlib.errors as errors
36
from bzrlib.errors import (InstallFailed, NoSuchRevision,
38
from bzrlib.trace import mutter
39
from bzrlib.progress import ProgressBar
40
from bzrlib.reconcile import RepoReconciler
41
from bzrlib.revision import NULL_REVISION
42
from bzrlib.symbol_versioning import *
45
# TODO: Avoid repeatedly opening weaves so many times.
47
# XXX: This doesn't handle ghost (not present in branch) revisions at
48
# all yet. I'm not sure they really should be supported.
50
# NOTE: This doesn't copy revisions which may be present but not
51
# merged into the last revision. I'm not sure we want to do that.
53
# - get a list of revisions that need to be pulled in
54
# - for each one, pull in that revision file
55
# and get the inventory, and store the inventory with right
57
# - and get the ancestry, and store that with right parents too
58
# - and keep a note of all file ids and version seen
59
# - then go through all files; for each one get the weave,
60
# and add in all file versions
63
@deprecated_function(zero_eight)
64
def greedy_fetch(to_branch, from_branch, revision=None, pb=None):
65
"""Legacy API, please see branch.fetch(from_branch, last_revision, pb)."""
66
f = Fetcher(to_branch, from_branch, revision, pb)
67
return f.count_copied, f.failed_revisions
72
class RepoFetcher(object):
73
"""Pull revisions and texts from one repository to another.
76
if set, try to limit to the data this revision references.
79
count_copied -- number of revisions copied
81
This should not be used directory, its essential a object to encapsulate
82
the logic in InterRepository.fetch().
84
def __init__(self, to_repository, from_repository, last_revision=None, pb=None):
86
self.failed_revisions = []
88
if to_repository.control_files._transport.base == from_repository.control_files._transport.base:
89
# check that last_revision is in 'from' and then return a no-operation.
90
if last_revision not in (None, NULL_REVISION):
91
from_repository.get_revision(last_revision)
93
self.to_repository = to_repository
94
self.from_repository = from_repository
95
# must not mutate self._last_revision as its potentially a shared instance
96
self._last_revision = last_revision
98
self.pb = bzrlib.ui.ui_factory.nested_progress_bar()
99
self.nested_pb = self.pb
102
self.nested_pb = None
103
self.from_repository.lock_read()
105
self.to_repository.lock_write()
109
if self.nested_pb is not None:
110
self.nested_pb.finished()
111
self.to_repository.unlock()
113
self.from_repository.unlock()
116
"""Primary worker function.
118
This initialises all the needed variables, and then fetches the
119
requested revisions, finally clearing the progress bar.
121
self.to_weaves = self.to_repository.weave_store
122
self.to_control = self.to_repository.control_weaves
123
self.from_weaves = self.from_repository.weave_store
124
self.from_control = self.from_repository.control_weaves
126
self.file_ids_names = {}
128
revs = self._revids_to_fetch()
131
self._fetch_weave_texts(revs)
132
self._fetch_inventory_weave(revs)
133
self._fetch_revision_texts(revs)
134
self.count_copied += len(revs)
138
def _revids_to_fetch(self):
139
self.pb.update('get destination history')
140
mutter('fetch up to rev {%s}', self._last_revision)
141
if self._last_revision is NULL_REVISION:
142
# explicit limit of no revisions needed
144
if (self._last_revision != None and
145
self.to_repository.has_revision(self._last_revision)):
149
return self.to_repository.missing_revision_ids(self.from_repository,
151
except errors.NoSuchRevision:
152
raise InstallFailed([self._last_revision])
154
def _fetch_weave_texts(self, revs):
155
file_ids = self.from_repository.fileid_involved_by_set(revs)
157
num_file_ids = len(file_ids)
158
for file_id in file_ids:
159
self.pb.update("merge weaves", count, num_file_ids)
161
to_weave = self.to_weaves.get_weave_or_empty(file_id,
162
self.to_repository.get_transaction())
164
if to_weave.num_versions() > 0:
165
# destination has contents, must merge
166
from_weave = self.from_weaves.get_weave(file_id,
167
self.from_repository.get_transaction())
168
# we fetch all the texts, because texts do
169
# not reference anything, and its cheap enough
170
to_weave.join(from_weave)
172
# destination is empty, just copy it.
173
# this copies all the texts, which is useful and
174
# on per-file basis quite cheap.
175
self.to_weaves.copy_multi(self.from_weaves, [file_id], self.pb,
176
self.from_repository.get_transaction(),
177
self.to_repository.get_transaction())
180
def _fetch_inventory_weave(self, revs):
181
self.pb.update("inventory fetch", 0, 2)
182
to_weave = self.to_control.get_weave('inventory',
183
self.to_repository.get_transaction())
185
# just merge, this is optimisable and its means we dont
186
# copy unreferenced data such as not-needed inventories.
187
self.pb.update("inventory fetch", 1, 2)
188
from_weave = self.from_repository.get_inventory_weave()
189
self.pb.update("inventory fetch", 2, 2)
190
# we fetch only the referenced inventories because we do not
191
# know for unselected inventories whether all their required
192
# texts are present in the other repository - it could be
194
to_weave.join(from_weave, pb=self.pb, msg='merge inventory',
199
class GenericRepoFetcher(RepoFetcher):
200
"""This is a generic repo to repo fetcher.
202
This makes minimal assumptions about repo layout and contents.
203
It triggers a reconciliation after fetching to ensure integrity.
206
def _fetch_revision_texts(self, revs):
207
self.to_transaction = self.to_repository.get_transaction()
211
self.pb.update('copying revisions', count, total)
213
sig_text = self.from_repository.get_signature_text(rev)
214
self.to_repository._revision_store.add_revision_signature_text(
215
rev, sig_text, self.to_transaction)
216
except errors.NoSuchRevision:
219
self.to_repository._revision_store.add_revision(
220
self.from_repository.get_revision(rev),
223
self.pb.update('copying revisions', count, total)
224
# fixup inventory if needed:
225
# this is expensive because we have no inverse index to current ghosts.
226
# but on local disk its a few seconds and sftp push is already insane.
228
# FIXME: repository should inform if this is needed.
229
self.to_repository.reconcile()
232
class KnitRepoFetcher(RepoFetcher):
233
"""This is a knit format repository specific fetcher.
235
This differs from the GenericRepoFetcher by not doing a
236
reconciliation after copying, and using knit joining to
240
def _fetch_revision_texts(self, revs):
241
# may need to be a InterRevisionStore call here.
242
from_transaction = self.from_repository.get_transaction()
243
to_transaction = self.to_repository.get_transaction()
244
to_sf = self.to_repository._revision_store.get_signature_file(
246
from_sf = self.from_repository._revision_store.get_signature_file(
248
to_sf.join(from_sf, version_ids=revs, pb=self.pb, ignore_missing=True)
249
to_rf = self.to_repository._revision_store.get_revision_file(
251
from_rf = self.from_repository._revision_store.get_revision_file(
253
to_rf.join(from_rf, version_ids=revs, pb=self.pb)
256
class Fetcher(object):
257
"""Backwards compatability glue for branch.fetch()."""
259
@deprecated_method(zero_eight)
260
def __init__(self, to_branch, from_branch, last_revision=None, pb=None):
261
"""Please see branch.fetch()."""
262
to_branch.fetch(from_branch, last_revision, pb)
added
removed
bzrlib/fetch.py
