+* Fri Dec 16 2011 Thierry Parmentelat <thierry.parmentelat@sophia.inria.fr> - sfa-2.0-6
+- bugfix: sfi was not sending call_id with ListResources to v2 servers
+- SFA_API_DEBUG replaced with SFA_API_LOGLEVEL
+- PlDriver / PlShell : PLCAPI methods now explicitly go to the shell
+
+* Wed Dec 14 2011 Thierry Parmentelat <thierry.parmentelat@sophia.inria.fr> - sfa-2.0-5
+- client: sfi -a / -p deprecated (use -s instead)
+- client: sfi cleaned up
+- client: sfi has backward support for APIv1 aggregates again
+- server: only APIv2 is supported and should be rather strict
+- server: settings for turning on/off caching in sm or am
+- server: plc-dependant code has moved from aggregate to pldriver
+- server: driver interface extended accordingly
+
+* Fri Dec 09 2011 Thierry Parmentelat <thierry.parmentelat@sophia.inria.fr> - sfa-2.0-4
+- screwed up previous tag
+
+* Fri Dec 09 2011 Thierry Parmentelat <thierry.parmentelat@sophia.inria.fr> - sfa-2.0-3
+- client side revisited with a bootstrap library
+- client side has a new source layout
+- various (nasty) bug fixes wrt options and call_id
+
+* Tue Dec 06 2011 Thierry Parmentelat <thierry.parmentelat@sophia.inria.fr> - sfa-2.0-2
+- various fixes in rspecs for sfav1&slice tags
+- uses 'geni_rspec_version' and not just 'rspec_version'
+- example flavour for the max testbed
+- embryo for an sfa client library
+- topology.py moved into plc
+- sql: table is named records; record_types are enforced
+- sql: table creation cleaned up
+
+* Wed Nov 30 2011 Thierry Parmentelat <thierry.parmentelat@sophia.inria.fr> - sfa-2.0-1
+- cleaned up all references to SFA_*_TYPE in config
+- enable cache at the aggregate by default
+- registry now uses the driver in a sensible way (see managers/driver.py)
+- slice manager supports sfav1/pgv2 neighbours
+- get_key renamed into get_key_from_incoming_ip
+- new sfa.storage module for record/table and all db-related stuff
+- db schema in sfa.storage.sfa.sql
+- init.d and cron.d move one step up
+- cleaned up rspec/ directory
+- add deps to pyopenssl and myplc-config
+- start support for new API (return value)
+- plc.remoteshell removed, use plshell instead
+- plshell uses a 'capability' auth method whenever possible
+- various tweaks in rspec elements
+- made dependency on sfatables softer
+
+* Thu Nov 24 2011 Thierry Parmentelat <thierry.parmentelat@sophia.inria.fr> - sfa-1.1-5
+- sfa should now be started *before* the initial import
+- sfa to use its own database (default sfa) - can run without myplc
+- server calls support optional 'options'
+- client sends options in argument when needed
+- fix infinite getattr recursion in elements/element.py
+- error codes in line with geni
+
+* Fri Nov 18 2011 Thierry Parmentelat <thierry.parmentelat@sophia.inria.fr> - sfa-1.1-4
+- fixed links and attributes in rspecs
+- minor cleanup in the API methods, and more consistent names in manager methods
+
+* Thu Nov 17 2011 Thierry Parmentelat <thierry.parmentelat@sophia.inria.fr> - sfa-1.1-3
+- ongoing refoactoring towards more genericity
+- passes tests again although known issues remain with attributes/tags
+
+* Mon Nov 07 2011 Thierry Parmentelat <thierry.parmentelat@sophia.inria.fr> - sfa-1.1-2
+- checkpoint tag: use SFA_GENERIC_FLAVOUR instead of SFA_*_TYPE
+- improvements in the pgv2 rspecs
+- driver separated from api
+- code starts moving around where it belongs
+- sfascan caches getversion across invokations
+- vini topology extracted as a config file
+
+* Fri Oct 28 2011 Thierry Parmentelat <thierry.parmentelat@sophia.inria.fr> - sfa-1.1-1
+- first support for protogeni rspecs is working
+- vini no longer needs a specific manager
+- refactoring underway towards more flexible/generic architecture
+