1
# Copyright (C) 2005, 2006 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,
38
from bzrlib.progress import ProgressPhase
39
from bzrlib.revision import is_null, NULL_REVISION
40
from bzrlib.symbol_versioning import (deprecated_function,
43
from bzrlib.trace import mutter
46
from bzrlib.lazy_import import lazy_import
48
# TODO: Avoid repeatedly opening weaves so many times.
50
# XXX: This doesn't handle ghost (not present in branch) revisions at
51
# all yet. I'm not sure they really should be supported.
53
# NOTE: This doesn't copy revisions which may be present but not
54
# merged into the last revision. I'm not sure we want to do that.
56
# - get a list of revisions that need to be pulled in
57
# - for each one, pull in that revision file
58
# and get the inventory, and store the inventory with right
60
# - and get the ancestry, and store that with right parents too
61
# - and keep a note of all file ids and version seen
62
# - then go through all files; for each one get the weave,
63
# and add in all file versions
66
class RepoFetcher(object):
67
"""Pull revisions and texts from one repository to another.
70
if set, try to limit to the data this revision references.
73
count_copied -- number of revisions copied
75
This should not be used directory, its essential a object to encapsulate
76
the logic in InterRepository.fetch().
78
def __init__(self, to_repository, from_repository, last_revision=None, pb=None):
80
self.failed_revisions = []
82
if to_repository.has_same_location(from_repository):
83
# check that last_revision is in 'from' and then return a
85
if last_revision is not None and not is_null(last_revision):
86
to_repository.get_revision(last_revision)
88
self.to_repository = to_repository
89
self.from_repository = from_repository
90
# must not mutate self._last_revision as its potentially a shared instance
91
self._last_revision = last_revision
93
self.pb = bzrlib.ui.ui_factory.nested_progress_bar()
94
self.nested_pb = self.pb
98
self.from_repository.lock_read()
100
self.to_repository.lock_write()
102
self.to_repository.start_write_group()
106
self.to_repository.abort_write_group()
109
self.to_repository.commit_write_group()
111
if self.nested_pb is not None:
112
self.nested_pb.finished()
113
self.to_repository.unlock()
115
self.from_repository.unlock()
118
"""Primary worker function.
120
This initialises all the needed variables, and then fetches the
121
requested revisions, finally clearing the progress bar.
123
self.to_weaves = self.to_repository.weave_store
124
self.from_weaves = self.from_repository.weave_store
126
self.file_ids_names = {}
127
pp = ProgressPhase('Fetch phase', 4, self.pb)
130
revs = self._revids_to_fetch()
131
self._fetch_everything_for_revisions(revs, pp)
135
def _fetch_everything_for_revisions(self, revs, pp):
136
"""Fetch all data for the given set of revisions."""
139
# The first phase is "file". We pass the progress bar for it directly
140
# into item_keys_introduced_by, which has more information about how
141
# that phase is progressing than we do. Progress updates for the other
142
# phases are taken care of in this function.
143
# XXX: there should be a clear owner of the progress reporting. Perhaps
144
# item_keys_introduced_by should have a richer API than it does at the
145
# moment, so that it can feed the progress information back to this
148
pb = bzrlib.ui.ui_factory.nested_progress_bar()
150
data_to_fetch = self.from_repository.item_keys_introduced_by(revs, pb)
151
for knit_kind, file_id, revisions in data_to_fetch:
152
if knit_kind != phase:
154
# Make a new progress bar for this phase
157
pb = bzrlib.ui.ui_factory.nested_progress_bar()
158
if knit_kind == "file":
159
self._fetch_weave_text(file_id, revisions)
160
elif knit_kind == "inventory":
162
# Once we've processed all the files, then we generate the root
163
# texts (if necessary), then we process the inventory. It's a
164
# bit distasteful to have knit_kind == "inventory" mean this,
165
# perhaps it should happen on the first non-"file" knit, in case
166
# it's not always inventory?
167
self._generate_root_texts(revs)
168
self._fetch_inventory_weave(revs, pb)
169
elif knit_kind == "signatures":
170
# Nothing to do here; this will be taken care of when
171
# _fetch_revision_texts happens.
173
elif knit_kind == "revisions":
174
self._fetch_revision_texts(revs, pb)
176
raise AssertionError("Unknown knit kind %r" % knit_kind)
180
self.count_copied += len(revs)
182
def _revids_to_fetch(self):
183
"""Determines the exact revisions needed from self.from_repository to
184
install self._last_revision in self.to_repository.
186
If no revisions need to be fetched, then this just returns None.
188
mutter('fetch up to rev {%s}', self._last_revision)
189
if self._last_revision is NULL_REVISION:
190
# explicit limit of no revisions needed
192
if (self._last_revision is not None and
193
self.to_repository.has_revision(self._last_revision)):
197
return self.to_repository.missing_revision_ids(self.from_repository,
199
except errors.NoSuchRevision:
200
raise InstallFailed([self._last_revision])
202
def _fetch_weave_text(self, file_id, required_versions):
203
to_weave = self.to_weaves.get_weave_or_empty(file_id,
204
self.to_repository.get_transaction())
205
from_weave = self.from_weaves.get_weave(file_id,
206
self.from_repository.get_transaction())
207
# we fetch all the texts, because texts do
208
# not reference anything, and its cheap enough
209
to_weave.join(from_weave, version_ids=required_versions)
210
# we don't need *all* of this data anymore, but we dont know
211
# what we do. This cache clearing will result in a new read
212
# of the knit data when we do the checkout, but probably we
213
# want to emit the needed data on the fly rather than at the
215
# the from weave should know not to cache data being joined,
216
# but its ok to ask it to clear.
217
from_weave.clear_cache()
218
to_weave.clear_cache()
220
def _fetch_inventory_weave(self, revs, pb):
221
pb.update("fetch inventory", 0, 2)
222
to_weave = self.to_repository.get_inventory_weave()
223
child_pb = bzrlib.ui.ui_factory.nested_progress_bar()
225
# just merge, this is optimisable and its means we don't
226
# copy unreferenced data such as not-needed inventories.
227
pb.update("fetch inventory", 1, 3)
228
from_weave = self.from_repository.get_inventory_weave()
229
pb.update("fetch inventory", 2, 3)
230
# we fetch only the referenced inventories because we do not
231
# know for unselected inventories whether all their required
232
# texts are present in the other repository - it could be
234
to_weave.join(from_weave, pb=child_pb, msg='merge inventory',
236
from_weave.clear_cache()
240
def _generate_root_texts(self, revs):
241
"""This will be called by __fetch between fetching weave texts and
242
fetching the inventory weave.
244
Subclasses should override this if they need to generate root texts
245
after fetching weave texts.
250
class GenericRepoFetcher(RepoFetcher):
251
"""This is a generic repo to repo fetcher.
253
This makes minimal assumptions about repo layout and contents.
254
It triggers a reconciliation after fetching to ensure integrity.
257
def _fetch_revision_texts(self, revs, pb):
258
"""Fetch revision object texts"""
259
to_txn = self.to_transaction = self.to_repository.get_transaction()
262
to_store = self.to_repository._revision_store
264
pb.update('copying revisions', count, total)
266
sig_text = self.from_repository.get_signature_text(rev)
267
to_store.add_revision_signature_text(rev, sig_text, to_txn)
268
except errors.NoSuchRevision:
271
to_store.add_revision(self.from_repository.get_revision(rev),
274
# fixup inventory if needed:
275
# this is expensive because we have no inverse index to current ghosts.
276
# but on local disk its a few seconds and sftp push is already insane.
278
# FIXME: repository should inform if this is needed.
279
self.to_repository.reconcile()
282
class KnitRepoFetcher(RepoFetcher):
283
"""This is a knit format repository specific fetcher.
285
This differs from the GenericRepoFetcher by not doing a
286
reconciliation after copying, and using knit joining to
290
def _fetch_revision_texts(self, revs, pb):
291
# may need to be a InterRevisionStore call here.
292
from_transaction = self.from_repository.get_transaction()
293
to_transaction = self.to_repository.get_transaction()
294
to_sf = self.to_repository._revision_store.get_signature_file(
296
from_sf = self.from_repository._revision_store.get_signature_file(
298
to_sf.join(from_sf, version_ids=revs, ignore_missing=True)
299
to_rf = self.to_repository._revision_store.get_revision_file(
301
from_rf = self.from_repository._revision_store.get_revision_file(
303
to_rf.join(from_rf, version_ids=revs)
306
class Inter1and2Helper(object):
307
"""Helper for operations that convert data from model 1 and 2
309
This is for use by fetchers and converters.
312
def __init__(self, source, target):
315
:param source: The repository data comes from
316
:param target: The repository data goes to
321
def iter_rev_trees(self, revs):
322
"""Iterate through RevisionTrees efficiently.
324
Additionally, the inventory's revision_id is set if unset.
326
Trees are retrieved in batches of 100, and then yielded in the order
329
:param revs: A list of revision ids
332
for tree in self.source.revision_trees(revs[:100]):
333
if tree.inventory.revision_id is None:
334
tree.inventory.revision_id = tree.get_revision_id()
338
def generate_root_texts(self, revs):
339
"""Generate VersionedFiles for all root ids.
341
:param revs: the revisions to include
343
inventory_weave = self.source.get_inventory_weave()
346
to_store = self.target.weave_store
347
for tree in self.iter_rev_trees(revs):
348
revision_id = tree.inventory.root.revision
349
root_id = tree.inventory.root.file_id
350
parents = inventory_weave.get_parents(revision_id)
351
if root_id not in versionedfile:
352
versionedfile[root_id] = to_store.get_weave_or_empty(root_id,
353
self.target.get_transaction())
354
parent_texts[root_id] = versionedfile[root_id].add_lines(
355
revision_id, parents, [], parent_texts)
357
def regenerate_inventory(self, revs):
358
"""Generate a new inventory versionedfile in target, convertin data.
360
The inventory is retrieved from the source, (deserializing it), and
361
stored in the target (reserializing it in a different format).
362
:param revs: The revisions to include
364
inventory_weave = self.source.get_inventory_weave()
365
for tree in self.iter_rev_trees(revs):
366
parents = inventory_weave.get_parents(tree.get_revision_id())
367
self.target.add_inventory(tree.get_revision_id(), tree.inventory,
371
class Model1toKnit2Fetcher(GenericRepoFetcher):
372
"""Fetch from a Model1 repository into a Knit2 repository
374
def __init__(self, to_repository, from_repository, last_revision=None,
376
self.helper = Inter1and2Helper(from_repository, to_repository)
377
GenericRepoFetcher.__init__(self, to_repository, from_repository,
380
def _generate_root_texts(self, revs):
381
self.helper.generate_root_texts(revs)
383
def _fetch_inventory_weave(self, revs, pb):
384
self.helper.regenerate_inventory(revs)
387
class Knit1to2Fetcher(KnitRepoFetcher):
388
"""Fetch from a Knit1 repository into a Knit2 repository"""
390
def __init__(self, to_repository, from_repository, last_revision=None,
392
self.helper = Inter1and2Helper(from_repository, to_repository)
393
KnitRepoFetcher.__init__(self, to_repository, from_repository,
396
def _generate_root_texts(self, revs):
397
self.helper.generate_root_texts(revs)
399
def _fetch_inventory_weave(self, revs, pb):
400
self.helper.regenerate_inventory(revs)
added
removed
bzrlib/fetch.py
