12
* ``bzr commit`` now doesn't print the revision number twice. (Matt
15
* New configuration option ``bugtracker_<tracker_abbrevation>_url`` to
16
define locations of bug trackers that are not directly supported by
17
bzr or a plugin. The URL will be treated as a template and ``{id}``
18
placeholders will be replaced by specific bug IDs. (Lukáš Lalinský)
20
* Support logging single merge revisions with short and line log formatters.
25
* New -Dhttp debug option reports http connections, requests and responses.
30
* Catch OSError 17 (file exists) in final phase of tree transform and show
32
(Alexander Belchenko, #111758)
34
* Catch ShortReadvErrors while using pycurl. Also make readv more robust by
35
allowing multiple GET requests to be issued if too many ranges are
37
(Vincent Ladeuil, #172701)
39
* Check for missing basis texts when fetching from packs to packs.
40
(John Arbash Meinel, #165290)
44
* Deprecate not passing a ``location`` argument to commit reporters'
45
``started`` methods. (Matt Nordhoff)
10
51
NOTES WHEN UPGRADING:
53
* The default repository format is now ``pack-0.92``. This
54
default is used when creating new repositories with ``init`` and
55
``init-repo``, and when branching over bzr+ssh or bzr+hpss.
56
(See https://bugs.launchpad.net/bugs/164626)
58
This format can be read and written by Bazaar 0.92 and later, and
59
data can be transferred to and from older formats.
61
To upgrade, please reconcile your repository (``bzr reconcile``), and then
62
upgrade (``bzr upgrade``).
64
``pack-0.92`` offers substantially better scaling and performance than the
65
previous knits format. Some operations are slower where the code already
66
had bad scaling characteristics under knits, the pack format makes such
67
operations more visible as part of being more scalable overall. We will
68
correct such operations over the coming releases and encourage the filing
69
of bugs on any operation which you observe to be slower in a packs
70
repository. One particular case that we do not intend to fix is pulling
71
data from a pack repository into a knit repository over a high latency
72
link; downgrading such data requires reinsertion of the file texts, and
73
this is a classic space/time tradeoff. The current implementation is
74
conservative on memory usage because we need to support converting data
75
from any tree without problems.
76
(Robert Collins, Martin Pool, #164476)
14
80
* Disable detection of plink.exe as possible ssh vendor. Plink vendor
31
97
tree roots that's recorded for all other directories.
32
98
(Aaron Bentley, #164639)
100
* ``pack-0.92`` repositories can now be reconciled.
101
(Robert Collins, #154173)
34
103
* ``switch`` command added for changing the branch a lightweight checkout
35
104
is associated with and updating the tree to reflect the latest content
36
105
accordingly. This command was previously part of the BzrTools plug-in.
37
106
(Ian Clatworthy, Aaron Bentley, David Allouche)
108
* ``reconfigure`` command can now convert branches, trees, or checkouts to
109
lightweight checkouts. (Aaron Bentley)
41
113
* Commit updates the state of the working tree via a delta rather than
43
115
this reduces the wall clock time for commit by roughly a 30%.
44
116
(Robert Collins, Martin Pool)
118
* Commit with many automatically found deleted paths no longer performs
119
linear scanning for the children of those paths during inventory
120
iteration. This should fix commit performance blowing out when many such
121
paths occur during commit. (Robert Collins, #156491)
46
123
* Fetch with pack repositories will no longer read the entire history graph.
47
124
(Robert Collins, #88319)
49
126
* Revert takes out an appropriate lock when reverting to a basis tree, and
50
127
does not read the basis inventory twice. (Robert Collins)
129
* Diff does not require an inventory to be generated on dirstate trees.
130
(Aaron Bentley, #149254)
132
* New annotate merge (--merge-type=weave) implementation is fast on
133
versionedfiles withough cached annotations, e.g. pack-0.92.
138
* ``bzr merge`` now warns when it encounters a criss-cross merge.
141
* ``bzr send`` now doesn't require the target e-mail address to be
142
specified on the command line if an interactive e-mail client is used.
54
145
* ``bzr tags`` now prints the revision number for each tag, instead of
55
146
the revision id, unless --show-ids is passed. In addition, tags can be
56
147
sorted chronologically instead of lexicographically with --sort=time.
57
148
(Adeodato Simó, #120231)
59
* New configuration option ``bugtracker_<tracker_abbrevation>_url`` to
60
define locations of bug trackers that are not directly supported by
61
bzr or a plugin. The URL will be treated as a template and ``{id}``
62
placeholders will be replaced by specific bug IDs. (Lukáš Lalinský)
150
* Windows standalone version of bzr is able to load system-wide plugins from
151
"plugins" subdirectory in installation directory. In addition standalone
152
installer write to the registry (HKLM\SOFTWARE\Bazaar) useful info
153
about paths and bzr version. (Alexander Belchenko, #129298)
68
* Fix possible error in insert_data_stream when copying between
69
pack repositories over bzr+ssh or bzr+http.
70
KnitVersionedFile.get_data_stream now makes sure that requested
71
compression parents are sent before any delta hunks that depend
73
(Martin Pool, #164637)
75
159
* A progress bar has been added for knitpack -> knitpack fetching.
76
(Robert Collins, #157789)
160
(Robert Collins, #157789, #159147)
162
* Branching from a branch via smart server now preserves the repository
163
format. (Andrew Bennetts, #164626)
78
165
* ``commit`` is now able to invoke an external editor in a non-ascii
79
166
directory. (Daniel Watkins, #84043)
81
168
* Catch connection errors for ftp.
82
169
(Vincent Ladeuil, #164567)
171
* ``check`` no longer reports spurious unreferenced text versions.
172
(Robert Collins, John A Meinel, #162931, #165071)
84
174
* Conflicts are now resolved recursively by ``revert``.
85
175
(Aaron Bentley, #102739)
94
184
retrying an http request or some programming errors may be masked.
95
185
(Vincent Ladeuil, #160012)
187
* Fix ``bzr rm`` to not delete modified and ignored files.
188
(Lukáš Lalinský, #172598)
97
190
* Fix exception when revisionspec contains merge revisons but log
98
191
formatter doesn't support merge revisions. (Kent Gibson, #148908)
103
196
* Fix multiple connections during checkout --lightweight.
104
197
(Vincent Ladeuil, #159150)
199
* Fix possible error in insert_data_stream when copying between
200
pack repositories over bzr+ssh or bzr+http.
201
KnitVersionedFile.get_data_stream now makes sure that requested
202
compression parents are sent before any delta hunks that depend
204
(Martin Pool, #164637)
106
206
* Fix typo in limiting offsets coalescing for http, leading to
107
207
whole files being downloaded instead of parts.
108
208
(Vincent Ladeuil, #165061)
129
233
* Obsolete packs are now cleaned up by pack and autopack operations.
130
234
(Robert Collins, #153789)
236
* Operations pulling data from a smart server where the underlying
237
repositories are not both annotated/both unannotated will now work.
238
(Robert Collins, #165304).
132
240
* Reconcile now shows progress bars. (Robert Collins, #159351)
134
242
* ``RemoteBranch`` was not initializing ``self._revision_id_to_revno_map``
182
290
(line, version_id) tuples. This change has been made to aid reconcile and
183
291
fetch operations. (Robert Collins)
293
* ``bzrlib.repository.get_versioned_file_checker`` is now private.
296
* The Repository format registry default has been removed; it was previously
297
obsoleted by the bzrdir format default, which implies a default repository
187
303
* Added ``ContainerSerialiser`` and ``ContainerPushParser`` to
254
380
* Commit no longer checks for new text keys during insertion when the
255
381
revision id was deterministically unique. (Robert Collins)
257
* Commit with many automatically found deleted paths no longer performs
258
linear scanning for the children of those paths during inventory
259
iteration. This should fix commit performance blowing out when many such
260
paths occur during commit. (Robert Collins, #156491)
262
383
* Committing a change which is not a merge and does not change the number of
263
384
files in the tree is faster by utilising the data about whether files are
264
385
changed to determine if the tree is unchanged rather than recalculating
added
removed
NEWS
