13
End user testing of the 2a format revealed two serious bugs. The first,
14
#365615, caused bzr to raise AbsentContentFactory errors when autopacking.
15
This meant that commits or pushes to 2a-format repositories failed
18
The second bug, #390563, caused the smart server to raise AbsentContentFactory
19
when streaming 2a stacked 2a-format branches. This particularly affected
20
branches stored on Launchpad in the 2a format.
22
Both of these bugs cause command failures only, neither of them cause data
23
corruption or data loss. And, of course, both of these bugs are now fixed.
28
* We now properly request a more minimal set of file texts when fetching
29
multiple revisions. (Robert Collins, John Arbash Meinel, #390563)
31
* Repositories using CHK pages (which includes the new 2a format) will no
32
longer error during commit or push operations when an autopack operation
33
is triggered. (Robert Collins, #365615)
35
* ``chk_map.iter_interesting_nodes`` now properly uses the *intersection*
36
of referenced nodes rather than the *union* to determine what
37
uninteresting pages we still need to look at. Prior to this,
38
incrementally pushing to stacked branch would push the minimal data, but
39
fetching everything would request extra texts. There are some unhandled
40
cases wrt trees of different depths, but this fixes the common cases.
41
(Robert Collins, John Arbash Meinel, #390563)
43
* ``GroupCompress`` repositories now take advantage of the pack hints
44
parameter to permit cross-format fetching to incrementally pack the
45
converted data. (Robert Collins)
47
* ``Repository.commit_write_group`` now returns opaque data about what
48
was committed, for passing to the ``Repository.pack``. Repositories
49
without atomic commits will still return None. (Robert Collins)
51
* ``Repository.pack`` now takes an optional ``hint`` parameter
52
which will support doing partial packs for repositories that can do
53
that. (Robert Collins)
55
* RepositoryFormat has a new attribute 'pack_compresses' which is True
56
when doing a pack operation changes the compression of content in the
57
repository. (Robert Collins)
59
* ``StreamSink`` and ``InterDifferingSerialiser`` will call
60
``Repository.pack`` with the hint returned by
61
``Repository.commit_write_group`` if the formats were different and the
62
repository can increase compression by doing a pack operation.
63
(Robert Collins, #376748)
68
:Codename: yesterday-in-california
72
This version of Bazaar contains the beta release of the new ``2a`` repository
73
format, suitable for testing by fearless, advanced users. This format or an
74
updated version of it will become the default format in Bazaar 2.0. Please
75
read the NEWS entry before even thinking about upgrading to the new format.
77
Also included are speedups for many operations on huge projects, a bug fix for
78
pushing stacked new stacked branches to smart servers and the usual bevy of
79
bug fixes and improvements.
82
Changes from 1.16rc1 to 1.16final
83
*********************************
85
* Fix the nested tree flag check so that upgrade from development formats to
86
2a can work correctly.
87
(Jelmer Vernooij, #388727)
89
* Automatic format upgrades triggered by default stacking policies on a
90
1.16rc1 (or later) smart server work again.
91
(Andrew Bennetts, #388675)
97
* Display prompt on stderr (instead of stdout) when querying users so
98
that the output of commands can be safely redirected.
99
(Vincent Ladeuil, #376582)
105
* A new repository format ``2a`` has been added. This is a beta release
106
of the brisbane-core (aka group-compress) project. This format now
107
suitable for wider testing by advanced users willing to deal with some
108
bugs. We would appreciate test reports, either positive or negative.
109
Format 2a is substantially smaller and faster for many operations on
110
many trees. This format or an updated version will become the default
113
This is a rich-root format, so this repository format can be used with
114
bzr-svn. Bazaar branches in previous non-rich-root formats can be
115
converted (including by merge, push and pull) to format 2a, but not vice
116
versa. We recommend upgrading previous development formats to 2a.
118
Upgrading to this format can take considerable time because it expands
119
and more concisely repacks the full history.
121
If you use stacked branches, you must upgrade the stacked branches
122
before the stacked-on branches. (See <https://bugs.launchpad.net/bugs/374735>)
124
* ``--development7-rich-root`` is a new dev format, similar to ``--dev6``
125
but using a Revision serializer using bencode rather than XML.
126
(Jelmer Vernooij, John Arbash Meinel)
128
* mail_client=claws now supports --body (and message body hooks). Also uses
129
configured from address. (Barry Warsaw)
135
* ``--development6-rich-root`` can now stack. (Modulo some smart-server
136
bugs with stacking and non default formats.)
137
(John Arbash Meinel, #373455)
139
* ``--development6-rich-root`` delays generating a delta index for the
140
first object inserted into a group. This has a beneficial impact on
141
``bzr commit`` since each committed texts goes to its own group. For
142
committing a 90MB file, it drops peak memory by about 200MB, and speeds
143
up commit from 7s => 4s. (John Arbash Meinel)
145
* Numerous operations are now faster for huge projects, i.e. those
146
with a large number of files and/or a large number of revisions,
147
particularly when the latest development format is used. These
148
operations (and improvements on OpenOffice.org) include:
150
* branch in a shared repository (2X faster)
151
* branch --no-tree (100X faster)
157
* Pyrex version of ``bencode`` support. This provides optimized support
158
for both encoding and decoding, and is now found at ``bzrlib.bencode``.
159
``bzrlib.utils.bencode`` is now deprecated.
160
(Alexander Belchenko, Jelmer Vernooij, John Arbash Meinel)
166
* Bazaar can now pass attachment files to the mutt email client.
167
(Edwin Grubbs, #384158)
169
* Better message in ``bzr add`` output suggesting using ``bzr ignored`` to
170
see which files can also be added. (Jason Spashett, #76616)
172
* ``bzr pull -r 123`` from a stacked branch on a smart server no longer fails.
173
Also, the ``Branch.revision_history()`` API now works in the same
174
situation. (Andrew Bennetts, #380314)
176
* ``bzr serve`` on Windows no longer displays a traceback simply because a
177
TCP client disconnected. (Andrew Bennetts)
179
* Clarify the rules for locking and fallback repositories. Fix bugs in how
180
``RemoteRepository`` was handling fallbacks along with the
181
``_real_repository``. (Andrew Bennetts, John Arbash Meinel, #375496)
183
* Fix a small bug with fetching revisions w/ ghosts into a new stacked
184
branch. Not often triggered, because it required ghosts to be part of
185
the fetched revisions, not in the stacked-on ancestry.
188
* Fix status and commit to work with content filtered trees, addressing
189
numerous bad bugs with line-ending support. (Ian Clatworthy, #362030)
191
* Fix problem of "directory not empty" when contending for a lock over
192
sftp. (Martin Pool, #340352)
194
* Fix rule handling so that eol is optional, not mandatory.
195
(Ian Clatworthy, #379370)
197
* Pushing a new stacked branch to a 1.15 smart server was broken due to a
198
bug in the ``BzrDirFormat.initialize_ex`` smart verb. This is fixed in
199
1.16, but required changes to the network protocol, so the
200
``BzrDirFormat.initialize_ex`` verb has been removed and replaced with a
201
corrected ``BzrDirFormat.initialize_ex_1.16`` verb. 1.15 clients will
202
still work with a 1.16 server as they will fallback to slower (and
204
(Jonathan Lange, Robert Collins, Andrew Bennetts, #385132)
206
* Reconcile can now deal with text revisions that originated in revisions
207
that are ghosts. (Jelmer Vernooij, #336749)
209
* Support cloning of branches with ghosts in the left hand side history.
210
(Jelmer Vernooij, #248540)
212
* The ''bzr diff'' now catches OSError from osutils.rmtree and logs a
213
helpful message to the trace file, unless the temp directory really was
214
removed (which would be very strange). Since the diff operation has
215
succeeded from the user's perspective, no output is written to stderr
216
or stdout. (Maritza Mendez, #363837)
218
* Translate errors received from a smart server in response to a
219
``BzrDirFormat.initialize`` or ``BzrDirFormat.initialize_ex`` request.
220
This was causing tracebacks even for mundane errors like
221
``PermissionDenied``. (Andrew Bennetts, #381329)
226
* Added directory structure and started translation of docs in Russian.
227
(Alexey Shtokalo, Alexander Iljin, Alexander Belchenko, Dmitry Vasiliev,
233
* Added osutils.parent_directories(). (Ian Clatworthy)
235
* ``bzrlib.progress.ProgressBar``, ``ChildProgress``, ``DotsProgressBar``,
236
``TTYProgressBar`` and ``child_progress`` are now deprecated; use
237
``ui_factory.nested_progress_bar`` instead. (Martin Pool)
239
* ``graph.StackedParentsProvider`` is now a public API, replacing
240
``graph._StackedParentsProvider``. The api is now considered stable and ready
241
for external users. (Gary van der Merwe)
243
* ``bzrlib.user_encoding`` is deprecated in favor of
244
``get_user_encoding``. (Alexander Belchenko)
246
* TreeTransformBase no longer assumes that limbo is provided via disk.
247
DiskTreeTransform now provides disk functionality. (Aaron Bentley)
252
* Remove ``weave.py`` script for accessing internals of old weave-format
253
repositories. (Martin Pool)
258
* ``make check`` no longer repeats the test run in ``LANG=C``.
259
(Martin Pool, #386180)
261
* The number of cores is now correctly detected on OSX. (John Szakmeister)
263
* The number of cores is also detected on Solaris and win32. (Vincent Ladeuil)
265
* The number of cores is also detected on FreeBSD. (Matthew Fuller)
269
vim: tw=74 ft=rst ff=unix
added
removed
doc/en/release-notes/bzr-1.16.txt
