130
121
requested revisions, finally clearing the progress bar.
132
123
self.to_weaves = self.to_repository.weave_store
133
self.to_control = self.to_repository.control_weaves
134
124
self.from_weaves = self.from_repository.weave_store
135
self.from_control = self.from_repository.control_weaves
136
125
self.count_total = 0
137
126
self.file_ids_names = {}
138
127
pp = ProgressPhase('Fetch phase', 4, self.pb)
141
130
revs = self._revids_to_fetch()
145
self._fetch_weave_texts(revs)
147
self._fetch_inventory_weave(revs)
149
self._fetch_revision_texts(revs)
150
self.count_copied += len(revs)
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)
154
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.
155
188
mutter('fetch up to rev {%s}', self._last_revision)
156
189
if self._last_revision is NULL_REVISION:
157
190
# explicit limit of no revisions needed
166
199
except errors.NoSuchRevision:
167
200
raise InstallFailed([self._last_revision])
169
def _fetch_weave_texts(self, revs):
170
texts_pb = bzrlib.ui.ui_factory.nested_progress_bar()
172
# fileids_altered_by_revision_ids requires reading the inventory
173
# weave, we will need to read the inventory weave again when
174
# all this is done, so enable caching for that specific weave
175
inv_w = self.from_repository.get_inventory_weave()
177
file_ids = self.from_repository.fileids_altered_by_revision_ids(revs)
179
num_file_ids = len(file_ids)
180
for file_id, required_versions in file_ids.items():
181
texts_pb.update("fetch texts", count, num_file_ids)
183
to_weave = self.to_weaves.get_weave_or_empty(file_id,
184
self.to_repository.get_transaction())
185
from_weave = self.from_weaves.get_weave(file_id,
186
self.from_repository.get_transaction())
187
# we fetch all the texts, because texts do
188
# not reference anything, and its cheap enough
189
to_weave.join(from_weave, version_ids=required_versions)
190
# we don't need *all* of this data anymore, but we dont know
191
# what we do. This cache clearing will result in a new read
192
# of the knit data when we do the checkout, but probably we
193
# want to emit the needed data on the fly rather than at the
195
# the from weave should know not to cache data being joined,
196
# but its ok to ask it to clear.
197
from_weave.clear_cache()
198
to_weave.clear_cache()
202
def _fetch_inventory_weave(self, revs):
203
pb = bzrlib.ui.ui_factory.nested_progress_bar()
205
pb.update("fetch inventory", 0, 2)
206
to_weave = self.to_control.get_weave('inventory',
207
self.to_repository.get_transaction())
209
child_pb = bzrlib.ui.ui_factory.nested_progress_bar()
211
# just merge, this is optimisable and its means we don't
212
# copy unreferenced data such as not-needed inventories.
213
pb.update("fetch inventory", 1, 3)
214
from_weave = self.from_repository.get_inventory_weave()
215
pb.update("fetch inventory", 2, 3)
216
# we fetch only the referenced inventories because we do not
217
# know for unselected inventories whether all their required
218
# texts are present in the other repository - it could be
220
to_weave.join(from_weave, pb=child_pb, msg='merge inventory',
222
from_weave.clear_cache()
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.
229
250
class GenericRepoFetcher(RepoFetcher):
233
254
It triggers a reconciliation after fetching to ensure integrity.
236
def _fetch_revision_texts(self, revs):
257
def _fetch_revision_texts(self, revs, pb):
237
258
"""Fetch revision object texts"""
238
rev_pb = bzrlib.ui.ui_factory.nested_progress_bar()
240
to_txn = self.to_transaction = self.to_repository.get_transaction()
243
to_store = self.to_repository._revision_store
245
pb = bzrlib.ui.ui_factory.nested_progress_bar()
247
pb.update('copying revisions', count, total)
249
sig_text = self.from_repository.get_signature_text(rev)
250
to_store.add_revision_signature_text(rev, sig_text, to_txn)
251
except errors.NoSuchRevision:
254
to_store.add_revision(self.from_repository.get_revision(rev),
259
# fixup inventory if needed:
260
# this is expensive because we have no inverse index to current ghosts.
261
# but on local disk its a few seconds and sftp push is already insane.
263
# FIXME: repository should inform if this is needed.
264
self.to_repository.reconcile()
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()
269
282
class KnitRepoFetcher(RepoFetcher):
382
393
KnitRepoFetcher.__init__(self, to_repository, from_repository,
383
394
last_revision, pb)
385
def _fetch_weave_texts(self, revs):
386
KnitRepoFetcher._fetch_weave_texts(self, revs)
387
# Now generate a weave for the tree root
396
def _generate_root_texts(self, revs):
388
397
self.helper.generate_root_texts(revs)
390
def _fetch_inventory_weave(self, revs):
399
def _fetch_inventory_weave(self, revs, pb):
391
400
self.helper.regenerate_inventory(revs)
394
class Fetcher(object):
395
"""Backwards compatibility glue for branch.fetch()."""
397
@deprecated_method(zero_eight)
398
def __init__(self, to_branch, from_branch, last_revision=None, pb=None):
399
"""Please see branch.fetch()."""
400
to_branch.fetch(from_branch, last_revision, pb)