6
5
.. contents:: List of Releases
9
2.1.0 series (not released yet)
10
###############################
18
* ``bzr+ssh`` and ``bzr`` paths can now be relative to home directories
19
specified in the URL. Paths starting with a path segment of ``~`` are
20
relative to the home directory of the user running the server, and paths
21
starting with ``~user`` are relative to the home directory of the named
22
user. For example, for a user "bob" with a home directory of
23
``/home/bob``, these URLs are all equivalent:
25
* ``bzr+ssh://bob@host/~/repo``
26
* ``bzr+ssh://bob@host/~bob/repo``
27
* ``bzr+ssh://bob@host/home/bob/repo``
29
If ``bzr serve`` was invoked with a ``--directory`` argument, then no
30
home directories outside that directory will be accessible via this
33
This is a feature of ``bzr serve``, so pre-2.1 clients will
34
automatically benefit from this feature when ``bzr`` on the server is
35
upgraded. (Andrew Bennetts, #109143)
37
* Give more control on BZR_PLUGIN_PATH by providing a way to refer to or
38
disable the user, site and core plugin directories.
39
(Vincent Ladeuil, #412930, #316192, #145612)
44
* Bazaar's native protocol code now correctly handles EINTR, which most
45
noticeably occurs if you break in to the debugger while connected to a
46
bzr+ssh server. You can now can continue from the debugger (by typing
47
'c') and the process continues. However, note that pressing C-\ in the
48
shell may still kill the SSH process, which is bug 162509, so you must
49
sent a signal to the bzr process specifically, for example by typing
50
``kill -QUIT PID`` in another shell. (Martin Pool, #341535)
52
* ``bzr check`` in pack-0.92, 1.6 and 1.9 format repositories will no
53
longer report incorrect errors about ``Missing inventory ('TREE_ROOT', ...)``
54
(Robert Collins, #416732)
56
* ``bzr info -v`` on a 2a format still claimed that it was a "Development
57
format" (John Arbash Meinel, #424392)
59
* ``bzr merge`` and ``bzr remove-tree`` now requires --force if pending
60
merges are present in the working tree.
61
(Vincent Ladeuil, #426344)
63
* Clearer message when Bazaar runs out of memory, instead of a ``MemoryError``
64
traceback. (Martin Pool, #109115)
66
* Conversion to 2a will create a single pack for all the new revisions (as
67
long as it ran without interruption). This improves both ``bzr upgrade``
68
and ``bzr pull`` or ``bzr merge`` from local branches in older formats.
69
The autopack logic that occurs every 100 revisions during local
70
conversions was not returning that pack's identifier, which resulted in
71
the partial packs created during the conversion not being consolidated
72
at the end of the conversion process. (Robert Collins, #423818)
74
* Don't give a warning on Windows when failing to import ``_readdir_pyx``
75
as it is never built. (John Arbash Meinel, #430645)
77
* Don't restrict the command name used to run the test suite.
78
(Vincent Ladeuil, #419950)
80
* Fetches from 2a to 2a are now again requested in 'groupcompress' order.
81
Groups that are seen as 'underutilized' will be repacked on-the-fly.
82
This means that when the source is fully packed, there is minimal
83
overhead during the fetch, but if the source is poorly packed the result
84
is a fairly well packed repository (not as good as 'bzr pack' but
85
good-enough.) (Robert Collins, John Arbash Meinel, #402652)
87
* Network streams now decode adjacent records of the same type into a
88
single stream, reducing layering churn. (Robert Collins)
90
* Make sure that we unlock the tree if we fail to create a TreeTransform
91
object when doing a merge, and there is limbo, or pending-deletions
92
directory. (Gary van der Merwe, #427773)
94
* Prevent some kinds of incomplete data from being committed to a 2a
95
repository, such as revisions without inventories or inventories without
96
chk_bytes root records.
97
(Andrew Bennetts, #423506)
99
* Registry objects should not use iteritems() when asked to use items().
100
(Vincent Ladeuil, #430510)
102
* Weave based repositories couldn't be cloned when committers were using
103
domains or user ids embedding '.sig'. Now they can.
104
(Matthew Fuller, Vincent Ladeuil, #430868)
109
* Bazaar gives a warning before exiting, and writes into ``.bzr.log``, if
110
compiled extensions can't be loaded. This typically indicates a
111
packaging or installation problem. In this case Bazaar will keep
112
running using pure-Python versions, but this may be substantially
113
slower. The warning can be disabled by setting
114
``ignore_missing_extensions = True`` in ``bazaar.conf``.
115
See also <https://answers.launchpad.net/bzr/+faq/703>.
116
(Martin Pool, #406113, #430529)
121
* Describe the new shell-like test feature. (Vincent Ladeuil)
123
* Help on hooks no longer says 'Not deprecated' for hooks that are
124
currently supported. (Ian Clatworthy, #422415)
129
* ``ProgressTask.note`` is deprecated.
132
* ``bzrlib.user_encoding`` has been removed; use
133
``bzrlib.osutils.get_user_encoding`` instead. (Martin Pool)
135
* ``bzrlib.tests`` now uses ``stopTestRun`` for its ``TestResult``
136
subclasses - the same as python's unittest module. (Robert Collins)
138
* ``bzrlib.trace.log_error``, ``error`` and ``info`` have been deprecated.
144
* ``BTreeLeafParser.extract_key`` has been tweaked slightly to reduce
145
mallocs while parsing the index (approx 3=>1 mallocs per key read).
146
This results in a 10% speedup while reading an index.
149
* The ``bzrlib.lsprof`` module has a new class ``BzrProfiler`` which makes
150
profiling in some situations like callbacks and generators easier.
156
* Passing ``--lsprof-tests -v`` to bzr selftest will cause lsprof output to
157
be output for every test. Note that this is very verbose! (Robert Collins)
159
* Setting ``BZR_TEST_PDB=1`` when running selftest will cause a pdb
160
post_mortem to be triggered when a test failure occurs. (Robert Collins)
162
* Shell-like tests can now be written. Code in ``bzrlib/tests/script.py`` ,
163
documentation in ``developers/testing.txt`` for details.
166
* Test parameterisation now does a shallow copy, not a deep copy of the test
167
to be parameterised. This is not expected to break external use of test
168
parameterisation, and is substantially faster. (Robert Collins)
170
* Tests that try to open a bzr dir on an arbitrary transport will now
171
fail unless they have explicitly permitted the transport via
172
``self.permit_url``. The standard test factories such as ``self.get_url``
173
will permit the urls they provide automatically, so only exceptional
174
tests should need to do this. (Robert Collins)
176
* The break-in test no longer cares about clean shutdown of the child,
177
instead it is happy if the debugger starts up. (Robert Collins)
179
* The full test suite is expected to pass when the C extensions are not
180
present. (Vincent Ladeuil, #430749)
183
bzr 2.0.1 (Not Released Yet)
184
############################
189
* Improve the time for ``bzr log DIR`` for 2a format repositories.
190
We had been using the same code path as for <2a formats, which required
191
iterating over all objects in all revisions.
192
(John Arbash Meinel, #374730)
194
* Make sure that we unlock the tree if we fail to create a TreeTransform
195
object when doing a merge, and there is limbo, or pending-deletions
196
directory. (Gary van der Merwe, #427773)
203
:Codename: Instant Karma
205
This release of Bazaar makes the 2a (previously 'brisbane-core') format
206
the default when new branches or repositories are created. This format is
207
substantially smaller and faster for many operations. Most of the work in
208
this release focuses on bug fixes and stabilization, covering both 2a and
209
previous formats. (See the Upgrade Guide for information on migrating
212
This release also improves the documentation content and presentation,
213
including adding Windows HtmlHelp manuals.
215
The Bazaar team decided that 2.0 will be a long-term supported release,
216
with bugfix-only 2.0.x releases based on it, continuing for at least six
217
months or until the following stable release.
219
Changes from 2.0.0rc2 to final
220
******************************
222
* Officially branded as 2.0.0 rather than 2.0 to clarify between things
223
that "want to happen on the 2.0.x stable series" versus things that want
224
to "land in 2.0.0". (Changes how bzrlib._format_version_tuple() handles
225
micro = 0.) (John Arbash Meinel)
231
:2.0.0rc2: 2009-09-10
236
* Added post_commit hook for mutable trees. This allows the keywords
237
plugin to expand keywords on files changed by the commit.
238
(Ian Clatworthy, #408841)
243
* Bazaar's native protocol code now correctly handles EINTR, which most
244
noticeably occurs if you break in to the debugger while connected to a
245
bzr+ssh server. You can now can continue from the debugger (by typing
246
'c') and the process continues. However, note that pressing C-\ in the
247
shell may still kill the SSH process, which is bug 162509, so you must
248
sent a signal to the bzr process specifically, for example by typing
249
``kill -QUIT PID`` in another shell. (Martin Pool, #341535)
251
* ``bzr check`` in pack-0.92, 1.6 and 1.9 format repositories will no
252
longer report incorrect errors about ``Missing inventory ('TREE_ROOT', ...)``
253
(Robert Collins, #416732)
255
* ``bzr info -v`` on a 2a format still claimed that it was a "Development
256
format" (John Arbash Meinel, #424392)
258
* ``bzr log stacked-branch`` shows the full log including
259
revisions that are in the fallback repository. (Regressed in 2.0rc1).
260
(John Arbash Meinel, #419241)
262
* Clearer message when Bazaar runs out of memory, instead of a ``MemoryError``
263
traceback. (Martin Pool, #109115)
265
* Conversion to 2a will create a single pack for all the new revisions (as
266
long as it ran without interruption). This improves both ``bzr upgrade``
267
and ``bzr pull`` or ``bzr merge`` from local branches in older formats.
268
The autopack logic that occurs every 100 revisions during local
269
conversions was not returning that pack's identifier, which resulted in
270
the partial packs created during the conversion not being consolidated
271
at the end of the conversion process. (Robert Collins, #423818)
273
* Fetches from 2a to 2a are now again requested in 'groupcompress' order.
274
Groups that are seen as 'underutilized' will be repacked on-the-fly.
275
This means that when the source is fully packed, there is minimal
276
overhead during the fetch, but if the source is poorly packed the result
277
is a fairly well packed repository (not as good as 'bzr pack' but
278
good-enough.) (Robert Collins, John Arbash Meinel, #402652)
280
* Fix a potential segmentation fault when doing 'log' of a branch that had
281
ghosts in its mainline. (Evaluating None as a tuple is bad.)
282
(John Arbash Meinel, #419241)
284
* ``groupcompress`` sort order is now more stable, rather than relying on
285
``topo_sort`` ordering. The implementation is now
286
``KnownGraph.gc_sort``. (John Arbash Meinel)
288
* Local data conversion will generate correct deltas. This is a critical
289
bugfix vs 2.0rc1, and all 2.0rc1 users should upgrade to 2.0rc2 before
290
converting repositories. (Robert Collins, #422849)
292
* Network streams now decode adjacent records of the same type into a
293
single stream, reducing layering churn. (Robert Collins)
295
* Prevent some kinds of incomplete data from being committed to a 2a
296
repository, such as revisions without inventories, a missing chk_bytes
297
record for an inventory, or a missing text referenced by an inventory.
298
(Andrew Bennetts, #423506, #406687)
303
* Fix assertion error about "_remember_remote_is_before" when pushing to
305
(Andrew Bennetts, #418931)
307
* Help on hooks no longer says 'Not deprecated' for hooks that are
308
currently supported. (Ian Clatworthy, #422415)
310
* PDF and CHM (Windows HtmlHelp) formats are now supported for the
311
user documentation. The HTML documentation is better broken up into
312
topics. (Ian Clatworthy)
314
* The developer and foreign language documents are now separated
315
out so that searching in the HTML and CHM files produces more
316
useful results. (Ian Clatworthy)
318
* The main table of contents now provides links to the new Migration Docs
319
and Plugins Guide. (Ian Clatworthy)
325
:Codename: no worries
326
:2.0.0rc1: 2009-08-26
12
328
Compatibility Breaks
13
329
********************
51
393
* ``bzr push`` locally on windows will no longer give a locking error with
52
394
dirstate based formats. (Robert Collins)
396
* ``bzr shelve`` and ``bzr unshelve`` now work on windows.
397
(Robert Collins, #305006)
399
* Commit of specific files no longer prevents using the the iter_changes
400
codepath. On 2a repositories, commit of specific files should now be as
401
fast, or slightly faster, than a full commit. (Robert Collins)
403
* The internal core code that handles specific file operations like
404
``bzr st FILENAME`` or ``bzr commit FILENAME`` has been changed to
405
include the parent directories if they have altered, and when a
406
directory stops being a directory its children are always included. This
407
fixes a number of causes for ``InconsistentDelta`` errors, and permits
408
faster commit of specific paths. (Robert Collins, #347649)
413
* New developer documentation for content filtering.
419
* ``bzrlib.shelf_ui`` has had the ``from_args`` convenience methods of its
420
classes changed to manage lock lifetime of the trees they open in a way
421
consistent with reader-exclusive locks. (Robert Collins, #305006)
429
:Codename: nein nein nein!
432
This release fixes two small but worthwhile bugs relevant to users on
433
Microsoft Windows: some commands that failed on with locking errors will
434
now work, and a bug that caused poor performance after committing a file
435
with line-ending conversion has now been fixed. It also fixes a bug in
436
pushing to older servers.
441
* Fixed a problem where using content filtering and especially end-of-line
442
conversion will commit too many copies a file.
443
(Martin Pool, #415508)
445
* Fix assertion error about ``_remember_remote_is_before`` in
446
``set_tags_bytes`` when pushing to older smart servers.
447
(Andrew Bennetts, Alexander Belchenko, #418931)
452
* ``bzr push`` locally on Windows will no longer give a locking error with
453
dirstate based formats. (Robert Collins)
455
* ``bzr shelve`` and ``bzr unshelve`` now work on Windows.
456
(Robert Collins, #305006)
461
* ``bzrlib.shelf_ui`` has had the ``from_args`` convenience methods of its
462
classes changed to manage lock lifetime of the trees they open in a way
463
consistent with reader-exclusive locks. (Robert Collins, #305006)
465
* ``Tree.path_content_summary`` may return a size of None, when called on
466
a tree with content filtering where the size of the canonical form
467
cannot be cheaply determined. (Martin Pool)
469
* When manually creating transport servers in test cases, a new helper
470
``TestCase.start_server`` that registers a cleanup and starts the server
471
should be used. (Robert Collins)