1
# Copyright (C) 2005, 2006 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, ProgressPhase
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 = {}
127
pp = ProgressPhase('Fetch phase', 4, self.pb)
130
revs = self._revids_to_fetch()
134
self._fetch_weave_texts(revs)
136
self._fetch_inventory_weave(revs)
138
self._fetch_revision_texts(revs)
139
self.count_copied += len(revs)
143
def _revids_to_fetch(self):
144
mutter('fetch up to rev {%s}', self._last_revision)
145
if self._last_revision is NULL_REVISION:
146
# explicit limit of no revisions needed
148
if (self._last_revision != None and
149
self.to_repository.has_revision(self._last_revision)):
153
return self.to_repository.missing_revision_ids(self.from_repository,
155
except errors.NoSuchRevision:
156
raise InstallFailed([self._last_revision])
158
def _fetch_weave_texts(self, revs):
159
texts_pb = bzrlib.ui.ui_factory.nested_progress_bar()
161
file_ids = self.from_repository.fileids_altered_by_revision_ids(revs)
163
num_file_ids = len(file_ids)
164
for file_id, required_versions in file_ids.items():
165
texts_pb.update("fetch texts", count, num_file_ids)
167
to_weave = self.to_weaves.get_weave_or_empty(file_id,
168
self.to_repository.get_transaction())
169
from_weave = self.from_weaves.get_weave(file_id,
170
self.from_repository.get_transaction())
171
# we fetch all the texts, because texts do
172
# not reference anything, and its cheap enough
173
to_weave.join(from_weave, version_ids=required_versions)
174
# we don't need *all* of this data anymore, but we dont know
175
# what we do. This cache clearing will result in a new read
176
# of the knit data when we do the checkout, but probably we
177
# want to emit the needed data on the fly rather than at the
179
# the from weave should know not to cache data being joined,
180
# but its ok to ask it to clear.
181
from_weave.clear_cache()
182
to_weave.clear_cache()
186
def _fetch_inventory_weave(self, revs):
187
pb = bzrlib.ui.ui_factory.nested_progress_bar()
189
pb.update("fetch inventory", 0, 2)
190
to_weave = self.to_control.get_weave('inventory',
191
self.to_repository.get_transaction())
193
child_pb = bzrlib.ui.ui_factory.nested_progress_bar()
195
# just merge, this is optimisable and its means we don't
196
# copy unreferenced data such as not-needed inventories.
197
pb.update("fetch inventory", 1, 3)
198
from_weave = self.from_repository.get_inventory_weave()
199
pb.update("fetch inventory", 2, 3)
200
# we fetch only the referenced inventories because we do not
201
# know for unselected inventories whether all their required
202
# texts are present in the other repository - it could be
204
to_weave.join(from_weave, pb=child_pb, msg='merge inventory',
212
class GenericRepoFetcher(RepoFetcher):
213
"""This is a generic repo to repo fetcher.
215
This makes minimal assumptions about repo layout and contents.
216
It triggers a reconciliation after fetching to ensure integrity.
219
def _fetch_revision_texts(self, revs):
220
"""Fetch revision object texts"""
221
rev_pb = bzrlib.ui.ui_factory.nested_progress_bar()
223
to_txn = self.to_transaction = self.to_repository.get_transaction()
226
to_store = self.to_repository._revision_store
228
pb = bzrlib.ui.ui_factory.nested_progress_bar()
230
pb.update('copying revisions', count, total)
232
sig_text = self.from_repository.get_signature_text(rev)
233
to_store.add_revision_signature_text(rev, sig_text, to_txn)
234
except errors.NoSuchRevision:
237
to_store.add_revision(self.from_repository.get_revision(rev),
242
# fixup inventory if needed:
243
# this is expensive because we have no inverse index to current ghosts.
244
# but on local disk its a few seconds and sftp push is already insane.
246
# FIXME: repository should inform if this is needed.
247
self.to_repository.reconcile()
252
class KnitRepoFetcher(RepoFetcher):
253
"""This is a knit format repository specific fetcher.
255
This differs from the GenericRepoFetcher by not doing a
256
reconciliation after copying, and using knit joining to
260
def _fetch_revision_texts(self, revs):
261
# may need to be a InterRevisionStore call here.
262
from_transaction = self.from_repository.get_transaction()
263
to_transaction = self.to_repository.get_transaction()
264
to_sf = self.to_repository._revision_store.get_signature_file(
266
from_sf = self.from_repository._revision_store.get_signature_file(
268
to_sf.join(from_sf, version_ids=revs, ignore_missing=True)
269
to_rf = self.to_repository._revision_store.get_revision_file(
271
from_rf = self.from_repository._revision_store.get_revision_file(
273
to_rf.join(from_rf, version_ids=revs)
276
class Fetcher(object):
277
"""Backwards compatibility glue for branch.fetch()."""
279
@deprecated_method(zero_eight)
280
def __init__(self, to_branch, from_branch, last_revision=None, pb=None):
281
"""Please see branch.fetch()."""
282
to_branch.fetch(from_branch, last_revision, pb)
added
removed
bzrlib/fetch.py
