So, the sourcecode directory is a bit of a wart for us. It manages a few % of our dependencies, and this requires special handling - it increases the learning curve, and its quite unlike the stock python tools of (buildout|virtualenv|eggs). It also promotes a form of laziness for us - we have lots of inventory (fix committed bugs) in loggerhead, which we maintain, but we haven't done a release on it for ~ a year - and this occurs because we don't need to do a release (or anything release-like) to update our copy.
We have (broadly) three sets of things in sourcecode: - things that we have no other answer for: mailman and moustache.js - things that we need a small bit of work to be able to do differently - the bzr plugins - https://bugs.launchpad.net/bzr/+bug/927920 - and things that have no cause to be in sourcecode at all (testtools, testresources). I propose that we declare sourcecode/ deprecated effective immediately (this means no new stuff in there at all). I would then file bugs for the things that need to move out of sourcecode, and we can get to them as rage permits. I'm not sure what to do for mailman and moustache.js; for mailman I would like us to decouple it more thoroughly (so we can use the system mailman, for instance); for moustache.js I'd love input from our js afficionados about the standard practices 'out there'. -Rob _______________________________________________ Mailing list: https://launchpad.net/~launchpad-dev Post to : launchpad-dev@lists.launchpad.net Unsubscribe : https://launchpad.net/~launchpad-dev More help : https://help.launchpad.net/ListHelp
