this week in ~canonical-bazaar

[Vincent Ladeuil]

Standup 2011-11-22

jelmer

* timezone-ignorant
* some reviews, including trying to finish spiv's
  faster-stacked-tree-build patch
* ported nautilus-bzr to GNOME3

* reading up on HPSS internals (it rocks!) and some more work
  eliminating HPSS VFS calls
 + main remaining HPSS calls: Repository.iter_inventories,
   Repository.get_deltas_for_revisions, commit builder
 + most other remaining VFS calls are related to bundles, check/reconcile
* some more minor launchpad-buildd fixes (now finally standalone, yay!)
* some more colo-related fixes (thanks mgz, for fixing several bugs in
  the URL handling)

NEXT

* review jam's branches
* wrap up bzr-builddeb multiple tarballs work

mgz

* one day on windows things (babune slave being red), several of the
  bugs involved resolved:
** tilde handling in URLs <http://pad.lv/842223>
** more test fallout from localised error messages <http://pad.lv/891582>
** splitting segment parameters from URLs <http://pad.lv/842233>
* investigated url handling in transports, which is still confusing
  (unclear internal usages)
* translation related bugs (some strings are still not exported),
  cleanup unused global options
* over weekend sent pageant fix upstream

NEXT:

* finally fixing localised error formatting? <http://pad.lv/273978>
* send mail to list about url and path handling in bzrlib

poolie

* more launchpad work: package builders gradually getting updated to new
  bzr and other related software,
* better web search results on launchpad
* fixed bug with details being lost when errors occur sending mail
 
NEXT:

* keep working on memory usage in bzr; looks pretty tractable 

vila

* 2.5b3 released
* .po merges in progress, need to setup some specific test framework as
  the workflow is a bit involved,
* config option expansion completed
* better understanding of http://pad.lv/874153, http://pad.lv/888629 and
  http://pad.lv/888631. In a nutshell, socket.close() should be used
  when we're done with a connection (including an idle one) and
  socket.shutdown() must be used when some error occurs. Being wrong
  between the two leads to hangs. The short term solution is to mark the
  offending tests as expected failures and wait for more evidence. 

NEXT:

* http://pad.lv/892992 fallout for locations.conf handling on windows
* http://pad.lv/884270 .po merges
* file bug for bzrlib.errors.ObjectNotLocked on the importer (fallout
  from running from bzr sources ?)
* 2.0.7 source freeze, last 2.0 release