Apply the collective.indexing profile when it is available. [maurits]
Removed the dependency on collective.indexing >= 1.1 as it was added to allow non-Archetypes content to be used, but it is giving problems with some normal Archetypes content. If you add collective.indexing to your buildout yourself, we still use it and register our own catalog queue processor with it. Please install it in your Plone Site yourself. [maurits]
Bug fix: when asked to return a maximum number of users, convert max_results to an integer. [maurits]
Remove deprecated workflow state category set status mapper. This should be implemented using a workflow based test in an IMembraneUserAuth implementation, not in the core Membrane code. [rossp, wichert]
Fix performance problems with the object_implements index using marker interfaces registered as utilities. [rossp]
The object_implements index used to use the ZCA to find out not only what interfaces an object provided, but what interfaces an object could be adapted to out to the second order (adapting two objects). Providing this degree of magical awareness proved to be a large performance problem.
Remove BBB method to migrate the list of membrane types from archetypes_tool to membrane_tool. [rossp]
Use ZCTextIndex for the SearchableText index. This fixes problems with unicode data. [wichert]
Modify the property plugin to handle property adapters returning None and pure dictionaries. Even though the PAS interface does not allow it this is common behaviour. [wichert]
Use collective.indexing to update the membrane_tool catalog data. This allows non-Archetypes content to be used. [wichert]
Refactor PAS plugins to only depend on the generic interfaces. This removes the dependency on Archetypes. [wichert]
Use plone.indexer to manager indexable attribuets. [wichert]
Update GenericSetup import handler to gracefully handle sites without a membrane_tool installed. [wichert]
Improve package description, RESTify the changelog. [wichert]
Move GenericSetup profile and step registration to zcml. [wichert]
Made Title in membrane_tool a ZCTextIndex too, so that enumerateGroups() of membrane_groups works without exact_match.
getUserName and getUserId indexes in membrane_tool are now of type ZCTextIndex, allowing us to do exact_match=False queries in MembraneUserManager.enumerateUsers.
PluggableAuthService.enumerateUsers recommends treating id and login as “contains” search tokens, but with our text index we can only do “starts with” searches. However, that’s much better than returning () for every call that has “exact_match=False”. [dpunktnpunkt]