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
19
from cStringIO import StringIO
22
import bzrlib.errors as errors
23
from bzrlib.errors import (InstallFailed, NoSuchRevision, WeaveError,
25
from bzrlib.trace import mutter, note, warning
26
from bzrlib.branch import Branch
27
from bzrlib.progress import ProgressBar
28
from bzrlib.xml5 import serializer_v5
29
from bzrlib.osutils import sha_string, split_lines
31
"""Copying of history from one branch to another.
33
The basic plan is that every branch knows the history of everything
34
that has merged into it. As the first step of a merge, pull, or
35
branch operation we copy history from the source into the destination
38
The copying is done in a slightly complicated order. We don't want to
39
add a revision to the store until everything it refers to is also
40
stored, so that if a revision is present we can totally recreate it.
41
However, we can't know what files are included in a revision until we
42
read its inventory. Therefore, we first pull the XML and hold it in
43
memory until we've updated all of the files referenced.
46
# TODO: Avoid repeatedly opening weaves so many times.
48
# XXX: This doesn't handle ghost (not present in branch) revisions at
49
# all yet. I'm not sure they really should be supported.
51
# NOTE: This doesn't copy revisions which may be present but not
52
# merged into the last revision. I'm not sure we want to do that.
54
# - get a list of revisions that need to be pulled in
55
# - for each one, pull in that revision file
56
# and get the inventory, and store the inventory with right
58
# - and get the ancestry, and store that with right parents too
59
# - and keep a note of all file ids and version seen
60
# - then go through all files; for each one get the weave,
61
# and add in all file versions
65
def greedy_fetch(to_branch, from_branch, revision=None, pb=None):
66
f = Fetcher(to_branch, from_branch, revision, pb)
67
return f.count_copied, f.failed_revisions
71
class Fetcher(object):
72
"""Pull revisions and texts from one branch to another.
74
This doesn't update the destination's history; that can be done
75
separately if desired.
78
If set, pull only up to this revision_id.
82
last_revision -- if last_revision
83
is given it will be that, otherwise the last revision of
86
count_copied -- number of revisions copied
88
count_weaves -- number of file weaves copied
90
def __init__(self, to_branch, from_branch, last_revision=None, pb=None):
91
if to_branch == from_branch:
92
raise Exception("can't fetch from a branch to itself")
93
self.to_branch = to_branch
94
self.to_repository = to_branch.repository
95
self.to_weaves = self.to_repository.weave_store
96
self.to_control = self.to_repository.control_weaves
97
self.from_branch = from_branch
98
self.from_repository = from_branch.repository
99
self.from_weaves = self.from_repository.weave_store
100
self.from_control = self.from_repository.control_weaves
101
self.failed_revisions = []
102
self.count_copied = 0
104
self.count_weaves = 0
105
self.copied_file_ids = set()
106
self.file_ids_names = {}
108
self.pb = bzrlib.ui.ui_factory.progress_bar()
111
self.from_branch.lock_read()
113
self._fetch_revisions(last_revision)
115
self.from_branch.unlock()
118
def _fetch_revisions(self, last_revision):
119
self.last_revision = self._find_last_revision(last_revision)
120
mutter('fetch up to rev {%s}', self.last_revision)
121
if (self.last_revision is not None and
122
self.to_repository.has_revision(self.last_revision)):
125
revs_to_fetch = self._compare_ancestries()
127
raise InstallFailed([self.last_revision])
128
self._copy_revisions(revs_to_fetch)
129
self.new_ancestry = revs_to_fetch
131
def _find_last_revision(self, last_revision):
132
"""Find the limiting source revision.
134
Every ancestor of that revision will be merged across.
136
Returns the revision_id, or returns None if there's no history
137
in the source branch."""
140
self.pb.update('get source history')
141
from_history = self.from_branch.revision_history()
142
self.pb.update('get destination history')
144
return from_history[-1]
146
return None # no history in the source branch
149
def _compare_ancestries(self):
150
"""Get a list of revisions that must be copied.
152
That is, every revision that's in the ancestry of the source
153
branch and not in the destination branch."""
154
self.pb.update('get source ancestry')
155
from_repository = self.from_branch.repository
156
self.from_ancestry = from_repository.get_ancestry(self.last_revision)
158
dest_last_rev = self.to_branch.last_revision()
159
self.pb.update('get destination ancestry')
161
to_repository = self.to_branch.repository
162
dest_ancestry = to_repository.get_ancestry(dest_last_rev)
165
ss = set(dest_ancestry)
167
for rev_id in self.from_ancestry:
169
to_fetch.append(rev_id)
170
mutter('need to get revision {%s}', rev_id)
171
mutter('need to get %d revisions in total', len(to_fetch))
172
self.count_total = len(to_fetch)
175
def _copy_revisions(self, revs_to_fetch):
177
for rev_id in revs_to_fetch:
181
if self.to_repository.has_revision(rev_id):
183
self.pb.update('copy revision', i, self.count_total)
184
self._copy_one_revision(rev_id)
185
self.count_copied += 1
188
def _copy_one_revision(self, rev_id):
189
"""Copy revision and everything referenced by it."""
190
mutter('copying revision {%s}', rev_id)
191
rev_xml = self.from_repository.get_revision_xml(rev_id)
192
inv_xml = self.from_repository.get_inventory_xml(rev_id)
193
rev = serializer_v5.read_revision_from_string(rev_xml)
194
inv = serializer_v5.read_inventory_from_string(inv_xml)
195
assert rev.revision_id == rev_id
196
assert rev.inventory_sha1 == sha_string(inv_xml)
197
mutter(' commiter %s, %d parents',
200
self._copy_new_texts(rev_id, inv)
201
parents = rev.parent_ids
202
new_parents = copy(parents)
203
for parent in parents:
204
if not self.to_repository.has_revision(parent):
205
new_parents.pop(new_parents.index(parent))
206
self._copy_inventory(rev_id, inv_xml, new_parents)
207
self.to_repository.revision_store.add(StringIO(rev_xml), rev_id)
208
mutter('copied revision %s', rev_id)
210
def _copy_inventory(self, rev_id, inv_xml, parent_ids):
211
self.to_control.add_text('inventory', rev_id,
212
split_lines(inv_xml), parent_ids,
213
self.to_repository.get_transaction())
215
def _copy_new_texts(self, rev_id, inv):
216
"""Copy any new texts occuring in this revision."""
217
# TODO: Rather than writing out weaves every time, hold them
218
# in memory until everything's done? But this way is nicer
219
# if it's interrupted.
220
for path, ie in inv.iter_entries():
221
self._copy_one_weave(rev_id, ie.file_id, ie.revision)
223
def _copy_one_weave(self, rev_id, file_id, text_revision):
224
"""Copy one file weave, esuring the result contains text_revision."""
225
# check if the revision is already there
226
if file_id in self.file_ids_names.keys( ) and \
227
text_revision in self.file_ids_names[file_id]:
229
to_weave = self.to_weaves.get_weave_or_empty(file_id,
230
self.to_repository.get_transaction())
231
if not file_id in self.file_ids_names.keys( ):
232
self.file_ids_names[file_id] = to_weave.names( )
233
if text_revision in to_weave:
235
from_weave = self.from_weaves.get_weave(file_id,
236
self.from_branch.repository.get_transaction())
237
if text_revision not in from_weave:
238
raise MissingText(self.from_branch, text_revision, file_id)
239
mutter('copy file {%s} modified in {%s}', file_id, rev_id)
241
if to_weave.numversions() > 0:
242
# destination has contents, must merge
244
to_weave.join(from_weave)
245
except errors.WeaveParentMismatch:
246
to_weave.reweave(from_weave)
248
# destination is empty, just replace it
249
to_weave = from_weave.copy( )
250
self.to_weaves.put_weave(file_id, to_weave,
251
self.to_repository.get_transaction())
252
self.count_weaves += 1
253
self.copied_file_ids.add(file_id)
254
self.file_ids_names[file_id] = to_weave.names()
255
mutter('copied file {%s}', file_id)
added
removed
bzrlib/fetch.py
