this week in ~canonical-bazaar

[Vincent Ladeuil]

Standup 2011-12-06

jelmer
------

* Did more work on HPSS (more verbs, less roundtrips)
* Lots more on multiple tarball support, now really fixed, will put up
  for review when last few test failures are resolved (the resulting
  revision graph is now sane),
* recipe builders mostly green (again) except for bzr-hg,

NEXT:
-----
* more launchpad stuff,
* finishing HPSS stuff,
* update whatsnew for colocated branches for 2.5b4 release

mgz
---

* built windows installers for 2.5b3 (including shipping bzr-git !),
  fixed various issues in the process
* Attempt at resolving issue with bundling of runtime libraries in
  windows installer <http://pad.lv/881203>
* Addressed a problem sftp test with a fix in selftest and the test
  itself
* Better handling of bad inputs in trace logging <http://pad.lv/714449>
* Progress on running under other locales
** Confirmed test suite passes now
** Background coding for decoding system error messages
** More to come
* Access to unicode environment on windows <http://pad.lv/262874>
* Fix regression in reporting unversioned files <http://pad.lv/898408>

NEXT:
-----
* More work on other locales

vila
----

* package importer:
** bzrlib.errors.ObjectNotLocked http://pad.lv/897690 fixed (9
   failures), newer bzr-builddeb (revno 650) deployed on jubany
** ValueError: WorkingTree.set_root_id with fileid=None (4 failures),
   already fixed in bzr.dev by jelmer, newer bzr.dev (revno 6326)
   deployed on jubany
** AssertionError:<module>:run_script:main:find_unimported_versions:check,
   'requeue --full' worked for some packages, fail with the same error
   for others after encountering the wrong freshness check.
** the wrong freshness check is now disabled on jubany via
   launchpad.verbosity = off in bazaar.conf, we probably want a better
   way to achieve that,
** key OSError
   <module>:main:import_package:import_package:_do_import_package:import_upstream:import_dir:import_archive:_import_archive:finalize:delete_any:_delete_file_or_dir
   turned into various different failures, some related to turning a
   symlink into a directory (or vice-versa),
** nexuiz-data needs more than 24h to complete which conflicts with the
   time quota, http://pad.lv/899308 filed,
** nexuiz-data is requeued after being killed, there is probably a wrong
   signature recorded which makes kills appear as launchpad downtimes,
** killing a process was spamming log http://pad.lv/899231
* imprecise deprecation warnings http://pad.lv/897718 fixed, opens the
  door to better error reporting for test scripts,
* patch pilot this week

NEXT:
-------

* bzr-2.5b4 freeze on Thursday
* http:/pad.lv/874153 add a specific -E flag that skips the failing test
 by default and keep it active on babune so we separate the fallouts for
 all downstreams from keeping track of the issue internally
* patch pilot
* http://pad.lv/899308 so nexuiz-data can complete