What’s New in Bazaar 2.3?

Bazaar 2.3 has been released on the 3rd of February 2011 and marks the start of another long-term-stable series. From here, we will only make bugfix releases on the 2.3 series (2.3.1, etc), while 2.4 will become our new development series. The 2.1 and 2.2 series will also continue to get bugfixes. (Currently 2.0 is planned to be EOLed circa September 2011.)

This document accumulates a high level summary of what’s changed. See the Bazaar Release Notes for a full list.

Users are encouraged to upgrade from the other stable series. This document outlines the improvements in Bazaar 2.3 vs Bazaar 2.2. As well as summarizing improvements made to the core product, it highlights enhancements within the broader Bazaar world of potential interest to those upgrading.

Bazaar 2.3.1 includes all the fixes in the un-released 2.0.7, 2.1.4 and 2.2.5 versions that weren’t included in 2.3.0 and fixes some bugs on its own.

Bazaar 2.3.2 is a bugfix release that was never released.

Bazaar 2.3.3 is a bugfix release including the fixes in 2.3.2 and fixing the test helpers deprecated by python-2.7.

Bazaar 2.3.4 is a bugfix release.

See the Bazaar Release Notes for details.

Bazaar 2.3 is fully compatible both locally and on the network with 2.0, 2.1, and 2.2. It can read and write repositories generated by all previous versions.

Changed Behaviour

  • Committing a new revision in a stacked branch is now supported, as long as you are using the current repository format (2a). It will preserve the stacking invariants, etc, so that fetching after commit is guaranteed to work. (John Arbash Meinel, #375013)
  • Support for some old development formats have been removed: development-rich-root, development6-rich-root, and development7-rich-root. These formats were always labelled experimental and not used unless the user specifically asked for them. If you have repositories using these old formats you should upgrade them to 2a using Bazaar 2.2. (Andrew Bennetts)
  • The default ignore file created by Bazaar will contain __pycache__, which is the name of the directory that will be used by Python to store bytecode files. (Andrea Corbellini, #626687)
  • The default sort order for the bzr tags command now uses a natural sort where numeric substrings are sorted numerically. The previous default was “asciibetical” where tags were sorted by the characters they contained. To get the old behavior, one can use bzr tags --sort=alpha. (Neil Martinsen-Burrell, #640760)
  • On platforms other than Windows and Mac OS X, Bazaar will use configuration files that live in $XDG_CONFIG_HOME/bazaar if that directory exists. This allows interested individuals to conform to the XDG Base Directory specification. The plugin location has not changed and is still ~/.bazaar/plugins. To use a different directory for plugins, use the environment variable BZR_PLUGIN_PATH. (Neil Martinsen-Burrell, #195397)
  • bzr upgrade now operates recursively when run on a shared repository, automatically upgrading the branches within it, and has grown additional options for showing what it will do and cleaning up after itself. (Ian Clatworthy, Matthew Fuller, #89830, #374734, #422450)

Launchpad integration

  • The lp: prefix will now use your known username (from bzr launchpad-login) to expand ~ to your username. For example: bzr launchpad-login user && bzr push lp:~/project/branch will now push to lp:~user/project/branch. (John Arbash Meinel)
  • Launchpad has announced that the edge.launchpad.net instance is deprecated and may be shut down in the future <http://blog.launchpad.net/general/edge-is-deprecated>. Bazaar has therefore been updated in this release to talk to the main (launchpad.net) servers, rather than the edge ones.

Performance improvements

  • bzr revert and bzr status are up to 15% faster on large trees with many changes by not repeatedly building a list of all file-ids. (Andrew Bennetts)
  • bzr send uses less memory. (John Arbash Meinel, #614576)
  • Fetches involving stacked branches and branches with tags now do slightly less I/O, and so does branching from an existing branch. This also improves the network performance of these operations. (Andrew Bennetts)
  • Inventory entries now consume less memory (on 32-bit Ubuntu file entries have dropped from 68 bytes to 40, and directory entries from 120 bytes to 48). This affects most operations, and depending on the size of the tree may substantially improve the speed of operations like bzr commit. (Andrew Bennetts)
  • Lower memory consumption when reading many chk index pages. Helpful for things like bzr co or bzr ls -R on large trees. (John Arbash Meinel)
  • When building new working trees, default to reading from the repository rather than the source tree unless explicitly requested. (via --files-from and --hardlink for bzr branch and bzr checkout. Generally, 2a format repositories extract content faster than seeking and reading content from another tree, especially in cold-cache situations. (John Arbash Meinel, #607298)

New revision specifiers

  • The mainline revision specifier has been added. It takes another revision spec as its input, and selects the revision which merged that revision into the mainline.

    For example, bzr log -vp -r mainline:1.2.3 will show the log of the revision that merged revision 1.2.3 into mainline, along with its status output and diff. (Aaron Bentley)

  • The annotate revision specifier has been added. It takes a path and a line as its input (in the form path:line), and selects the revision which introduced that line of that file.

    For example: bzr log -vp -r annotate:bzrlib/transform.py:500 will select the revision that introduced line 500 of transform.py, and display its log, status output and diff.

    It can be combined with mainline to select the revision that landed this line into trunk, like so: bzr log -vp -r mainline:annotate:bzrlib/transform.py:500 (Aaron Bentley)

Testing/Bug reporting

  • Shell-like scripts can now be run directly from the command line without writing a python test. This should help users adding reproducing recipes to bug reports. (Vincent Ladeuil)

Improved conflict handling

  • pull, merge or switch can lead to conflicts when deleting a versioned directory contains unversioned files. The cause of the conflict is that deleting the directory will orphan the unversioned files so the user needs to instruct bzr what do to do about these orpahns. This is controlled by setting the bzr.transform.orphan_policy configuration variable with a value of move. In this case the unversioned files are moved to a bzr-orphans directory at the root of the working tree. The default behaviour is specified (if needed) by setting the variable to conflict. (Vincent Ladeuil, #323111)
  • bzr resolve --take-this and bzr resolve --take-other can now be used for text conflicts. This will ignore the differences that were merged cleanly and replace the file with its content in the current branch (--take-this) or with its content in the merged branch (--take-other). (Vincent Ladeuil, #638451)
  • bzr resolve now provides more feedback about the conflicts just resolved and the remaining ones. (Vincent Ladeuil)


  • A beta version of the documentation is now available in GNU TexInfo format, used by emacs and the standalone info reader. (Vincent Ladeuil, #219334)


bzr can be configured via environment variables, command-line options and configurations files. We’ve started working on unifying this and give access to more options. The first step is a new bzr config command that can be used to display the active configuration options in the current working tree or branch as well as the ability to set or remove an option. Scripts can also use it to get only the value for a given option.

Further information

For more detailed information on the changes made, see the the Bazaar Release Notes for:

For a summary of changes made in earlier releases, see: