974.1.27
by aaron.bentley at utoronto
Initial greedy fetch work |
1 |
# Copyright (C) 2005 by Canonical Ltd
|
2 |
||
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.
|
|
7 |
||
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.
|
|
12 |
||
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
|
|
1218
by Martin Pool
- fix up import |
16 |
|
1231
by Martin Pool
- more progress on fetch on top of weaves |
17 |
|
18 |
"""Copying of history from one branch to another.
|
|
19 |
||
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
|
|
23 |
branch.
|
|
24 |
||
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. Therefore, we first pull the XML and hold it in
|
|
30 |
memory until we've updated all of the files referenced.
|
|
31 |
"""
|
|
32 |
||
1534.1.31
by Robert Collins
Deprecated fetch.fetch and fetch.greedy_fetch for branch.fetch, and move the Repository.fetch internals to InterRepo and InterWeaveRepo. |
33 |
import bzrlib |
34 |
import bzrlib.errors as errors |
|
1563.2.10
by Robert Collins
Change weave store to be a versioned store, using WeaveFiles which maintain integrity without needing explicit 'put' operations. |
35 |
from bzrlib.errors import (InstallFailed, NoSuchRevision, |
1534.1.31
by Robert Collins
Deprecated fetch.fetch and fetch.greedy_fetch for branch.fetch, and move the Repository.fetch internals to InterRepo and InterWeaveRepo. |
36 |
MissingText) |
37 |
from bzrlib.trace import mutter |
|
38 |
from bzrlib.progress import ProgressBar |
|
39 |
from bzrlib.revision import NULL_REVISION |
|
40 |
from bzrlib.symbol_versioning import * |
|
41 |
||
42 |
||
1231
by Martin Pool
- more progress on fetch on top of weaves |
43 |
# TODO: Avoid repeatedly opening weaves so many times.
|
974.1.27
by aaron.bentley at utoronto
Initial greedy fetch work |
44 |
|
1238
by Martin Pool
- remove a lot of dead code from fetch |
45 |
# XXX: This doesn't handle ghost (not present in branch) revisions at
|
1240
by Martin Pool
- clean up fetch code and add progress bar |
46 |
# all yet. I'm not sure they really should be supported.
|
47 |
||
1262
by Martin Pool
- fetch should also copy ancestry records |
48 |
# NOTE: This doesn't copy revisions which may be present but not
|
49 |
# merged into the last revision. I'm not sure we want to do that.
|
|
1238
by Martin Pool
- remove a lot of dead code from fetch |
50 |
|
51 |
# - get a list of revisions that need to be pulled in
|
|
52 |
# - for each one, pull in that revision file
|
|
53 |
# and get the inventory, and store the inventory with right
|
|
54 |
# parents.
|
|
55 |
# - and get the ancestry, and store that with right parents too
|
|
56 |
# - and keep a note of all file ids and version seen
|
|
57 |
# - then go through all files; for each one get the weave,
|
|
58 |
# and add in all file versions
|
|
59 |
||
60 |
||
1534.1.31
by Robert Collins
Deprecated fetch.fetch and fetch.greedy_fetch for branch.fetch, and move the Repository.fetch internals to InterRepo and InterWeaveRepo. |
61 |
@deprecated_function(zero_eight) |
1260
by Martin Pool
- some updates for fetch/update function |
62 |
def greedy_fetch(to_branch, from_branch, revision=None, pb=None): |
1534.1.32
by Robert Collins
Deprecate direct use of fetch.Fetcher. |
63 |
"""Legacy API, please see branch.fetch(from_branch, last_revision, pb)."""
|
1219
by Martin Pool
- BROKEN: start refactoring fetch code to work well with weaves |
64 |
f = Fetcher(to_branch, from_branch, revision, pb) |
65 |
return f.count_copied, f.failed_revisions |
|
66 |
||
1534.1.32
by Robert Collins
Deprecate direct use of fetch.Fetcher. |
67 |
fetch = greedy_fetch |
68 |
||
1219
by Martin Pool
- BROKEN: start refactoring fetch code to work well with weaves |
69 |
|
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
70 |
class RepoFetcher(object): |
71 |
"""Pull revisions and texts from one repository to another.
|
|
72 |
||
73 |
last_revision
|
|
74 |
if set, try to limit to the data this revision references.
|
|
75 |
||
76 |
after running:
|
|
1260
by Martin Pool
- some updates for fetch/update function |
77 |
count_copied -- number of revisions copied
|
1534.1.33
by Robert Collins
Move copy_content_into into InterRepository and InterWeaveRepo, and disable the default codepath test as we have optimised paths for all current combinations. |
78 |
|
79 |
This should not be used directory, its essential a object to encapsulate
|
|
80 |
the logic in InterRepository.fetch().
|
|
1260
by Martin Pool
- some updates for fetch/update function |
81 |
"""
|
1534.4.50
by Robert Collins
Got the bzrdir api straightened out, plenty of refactoring to use it pending, but the api is up and running. |
82 |
def __init__(self, to_repository, from_repository, last_revision=None, pb=None): |
1534.1.31
by Robert Collins
Deprecated fetch.fetch and fetch.greedy_fetch for branch.fetch, and move the Repository.fetch internals to InterRepo and InterWeaveRepo. |
83 |
# result variables.
|
84 |
self.failed_revisions = [] |
|
85 |
self.count_copied = 0 |
|
1556.1.4
by Robert Collins
Add a new format for what will become knit, and the surrounding logic to upgrade repositories within metadirs, and tests for the same. |
86 |
if to_repository.control_files._transport.base == from_repository.control_files._transport.base: |
1534.6.13
by Robert Collins
Allow push/pull and branch between branches in the same shared repository. |
87 |
# check that last_revision is in 'from' and then return a no-operation.
|
88 |
if last_revision not in (None, NULL_REVISION): |
|
89 |
from_repository.get_revision(last_revision) |
|
90 |
return
|
|
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
91 |
self.to_repository = to_repository |
92 |
self.from_repository = from_repository |
|
93 |
# must not mutate self._last_revision as its potentially a shared instance
|
|
1185.65.27
by Robert Collins
Tweak storage towards mergability. |
94 |
self._last_revision = last_revision |
95 |
if pb is None: |
|
96 |
self.pb = bzrlib.ui.ui_factory.progress_bar() |
|
97 |
else: |
|
98 |
self.pb = pb |
|
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
99 |
self.from_repository.lock_read() |
1185.65.27
by Robert Collins
Tweak storage towards mergability. |
100 |
try: |
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
101 |
self.to_repository.lock_write() |
1185.65.27
by Robert Collins
Tweak storage towards mergability. |
102 |
try: |
103 |
self.__fetch() |
|
104 |
finally: |
|
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
105 |
self.to_repository.unlock() |
1185.65.27
by Robert Collins
Tweak storage towards mergability. |
106 |
finally: |
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
107 |
self.from_repository.unlock() |
1185.65.27
by Robert Collins
Tweak storage towards mergability. |
108 |
|
109 |
def __fetch(self): |
|
110 |
"""Primary worker function.
|
|
111 |
||
112 |
This initialises all the needed variables, and then fetches the
|
|
113 |
requested revisions, finally clearing the progress bar.
|
|
114 |
"""
|
|
1185.67.2
by Aaron Bentley
Renamed Branch.storage to Branch.repository |
115 |
self.to_weaves = self.to_repository.weave_store |
116 |
self.to_control = self.to_repository.control_weaves |
|
117 |
self.from_weaves = self.from_repository.weave_store |
|
118 |
self.from_control = self.from_repository.control_weaves |
|
1240
by Martin Pool
- clean up fetch code and add progress bar |
119 |
self.count_total = 0 |
1185.33.55
by Martin Pool
[patch] weave fetch optimizations (Goffredo Baroncelli) |
120 |
self.file_ids_names = {} |
1392
by Robert Collins
reinstate testfetch test case |
121 |
try: |
1185.65.30
by Robert Collins
Merge integration. |
122 |
revs = self._revids_to_fetch() |
1185.64.3
by Goffredo Baroncelli
This patch changes the fetch code. Before, the original code expanded every inventory and |
123 |
# nothing to do
|
124 |
if revs: |
|
1534.1.11
by Robert Collins
Format tweaks for latest merges. |
125 |
self._fetch_weave_texts(revs) |
126 |
self._fetch_inventory_weave(revs) |
|
127 |
self._fetch_revision_texts(revs) |
|
1185.64.3
by Goffredo Baroncelli
This patch changes the fetch code. Before, the original code expanded every inventory and |
128 |
self.count_copied += len(revs) |
1185.12.15
by Aaron Bentley
Cleared progress bar properly after fetch |
129 |
finally: |
130 |
self.pb.clear() |
|
1231
by Martin Pool
- more progress on fetch on top of weaves |
131 |
|
1185.65.30
by Robert Collins
Merge integration. |
132 |
def _revids_to_fetch(self): |
1534.1.31
by Robert Collins
Deprecated fetch.fetch and fetch.greedy_fetch for branch.fetch, and move the Repository.fetch internals to InterRepo and InterWeaveRepo. |
133 |
self.pb.update('get destination history') |
1185.65.27
by Robert Collins
Tweak storage towards mergability. |
134 |
mutter('fetch up to rev {%s}', self._last_revision) |
1534.4.50
by Robert Collins
Got the bzrdir api straightened out, plenty of refactoring to use it pending, but the api is up and running. |
135 |
if self._last_revision is NULL_REVISION: |
136 |
# explicit limit of no revisions needed
|
|
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
137 |
return None |
1534.4.50
by Robert Collins
Got the bzrdir api straightened out, plenty of refactoring to use it pending, but the api is up and running. |
138 |
if (self._last_revision != None and |
1185.65.27
by Robert Collins
Tweak storage towards mergability. |
139 |
self.to_repository.has_revision(self._last_revision)): |
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
140 |
return None |
141 |
||
1417.1.13
by Robert Collins
do not download remote ancestry.weave if the target revision we are stopping at is in our local store |
142 |
try: |
1534.4.50
by Robert Collins
Got the bzrdir api straightened out, plenty of refactoring to use it pending, but the api is up and running. |
143 |
return self.to_repository.missing_revision_ids(self.from_repository, |
144 |
self._last_revision) |
|
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
145 |
except errors.NoSuchRevision: |
1185.65.27
by Robert Collins
Tweak storage towards mergability. |
146 |
raise InstallFailed([self._last_revision]) |
1185.64.3
by Goffredo Baroncelli
This patch changes the fetch code. Before, the original code expanded every inventory and |
147 |
|
1534.1.11
by Robert Collins
Format tweaks for latest merges. |
148 |
def _fetch_revision_texts(self, revs): |
1185.65.30
by Robert Collins
Merge integration. |
149 |
self.to_repository.revision_store.copy_multi( |
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
150 |
self.from_repository.revision_store, |
151 |
revs, |
|
152 |
pb=self.pb) |
|
1185.64.3
by Goffredo Baroncelli
This patch changes the fetch code. Before, the original code expanded every inventory and |
153 |
|
1534.1.11
by Robert Collins
Format tweaks for latest merges. |
154 |
def _fetch_weave_texts(self, revs): |
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
155 |
file_ids = self.from_repository.fileid_involved_by_set(revs) |
1185.64.3
by Goffredo Baroncelli
This patch changes the fetch code. Before, the original code expanded every inventory and |
156 |
count = 0 |
157 |
num_file_ids = len(file_ids) |
|
158 |
for file_id in file_ids: |
|
1185.77.1
by John Arbash Meinel
Simplifying what bzr displays when merging |
159 |
self.pb.update("merge weaves", count, num_file_ids) |
1185.64.3
by Goffredo Baroncelli
This patch changes the fetch code. Before, the original code expanded every inventory and |
160 |
count +=1 |
161 |
to_weave = self.to_weaves.get_weave_or_empty(file_id, |
|
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
162 |
self.to_repository.get_transaction()) |
1185.64.3
by Goffredo Baroncelli
This patch changes the fetch code. Before, the original code expanded every inventory and |
163 |
|
164 |
if to_weave.numversions() > 0: |
|
165 |
# destination has contents, must merge
|
|
1563.2.10
by Robert Collins
Change weave store to be a versioned store, using WeaveFiles which maintain integrity without needing explicit 'put' operations. |
166 |
from_weave = self.from_weaves.get_weave(file_id, |
167 |
self.from_repository.get_transaction()) |
|
1563.2.11
by Robert Collins
Consolidate reweave and join as we have no separate usage, make reweave tests apply to all versionedfile implementations and deprecate the old reweave apis. |
168 |
to_weave.join(from_weave) |
1185.64.3
by Goffredo Baroncelli
This patch changes the fetch code. Before, the original code expanded every inventory and |
169 |
else: |
170 |
# destination is empty, just replace it
|
|
1563.2.10
by Robert Collins
Change weave store to be a versioned store, using WeaveFiles which maintain integrity without needing explicit 'put' operations. |
171 |
self.to_weaves.copy_multi(self.from_weaves, [file_id], self.pb) |
1534.1.11
by Robert Collins
Format tweaks for latest merges. |
172 |
self.pb.clear() |
1185.64.3
by Goffredo Baroncelli
This patch changes the fetch code. Before, the original code expanded every inventory and |
173 |
|
1534.1.11
by Robert Collins
Format tweaks for latest merges. |
174 |
def _fetch_inventory_weave(self, revs): |
1551.3.11
by Aaron Bentley
Merge from Robert |
175 |
self.pb.update("inventory fetch", 0, 2) |
176 |
to_weave = self.to_control.get_weave('inventory', |
|
1534.4.43
by Robert Collins
Merge from integration |
177 |
self.to_repository.get_transaction()) |
1185.64.4
by Goffredo Baroncelli
Apply some style changes on the basis of Robert suggest |
178 |
|
1185.64.3
by Goffredo Baroncelli
This patch changes the fetch code. Before, the original code expanded every inventory and |
179 |
if to_weave.numversions() > 0: |
180 |
# destination has contents, must merge
|
|
1563.2.10
by Robert Collins
Change weave store to be a versioned store, using WeaveFiles which maintain integrity without needing explicit 'put' operations. |
181 |
self.pb.update("inventory fetch", 1, 2) |
182 |
from_weave = self.from_repository.get_inventory_weave() |
|
183 |
self.pb.update("inventory fetch", 2, 2) |
|
1563.2.11
by Robert Collins
Consolidate reweave and join as we have no separate usage, make reweave tests apply to all versionedfile implementations and deprecate the old reweave apis. |
184 |
to_weave.join(from_weave, pb=self.pb, msg='merge inventory') |
1185.64.3
by Goffredo Baroncelli
This patch changes the fetch code. Before, the original code expanded every inventory and |
185 |
else: |
186 |
# destination is empty, just replace it
|
|
1563.2.10
by Robert Collins
Change weave store to be a versioned store, using WeaveFiles which maintain integrity without needing explicit 'put' operations. |
187 |
self.to_control.copy_multi(self.from_control, |
188 |
['inventory'], |
|
189 |
self.pb, |
|
190 |
self.from_repository.get_transaction()) |
|
1185.64.4
by Goffredo Baroncelli
Apply some style changes on the basis of Robert suggest |
191 |
|
1534.1.11
by Robert Collins
Format tweaks for latest merges. |
192 |
self.pb.clear() |
1240
by Martin Pool
- clean up fetch code and add progress bar |
193 |
|
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
194 |
|
195 |
class Fetcher(object): |
|
1534.1.32
by Robert Collins
Deprecate direct use of fetch.Fetcher. |
196 |
"""Backwards compatability glue for branch.fetch()."""
|
197 |
||
198 |
@deprecated_method(zero_eight) |
|
1534.4.41
by Robert Collins
Branch now uses BzrDir reasonably sanely. |
199 |
def __init__(self, to_branch, from_branch, last_revision=None, pb=None): |
1534.1.32
by Robert Collins
Deprecate direct use of fetch.Fetcher. |
200 |
"""Please see branch.fetch()."""
|
201 |
to_branch.fetch(from_branch, last_revision, pb) |