Thanks, it helps me a lot :)

On 27/02/2008 18:39, Rob Miller wrote:
> Yoan Le Clanche wrote:
> > Hi everyone,
> > 
> > Is it possible to create a new content type to manage groups with
> > remember?
> 
> yup.
> 
> > I know membrane can do it, but remember seems to overwrite this
> > functionnality. So what can i do to use my own content type?
> 
> membrane provides the toolkit to use content objects as groups, but as you
> say 
the default remember implementation doesn't take advantage of it. 
> here's a 
rough outline of what you'd need to do:
> 
> - decide what defines the relationship btn a member and a group (e.g. they 
> could be linked via AT or plone.relations references, or possibly the member 
> object could simply store a list of group ids, and the group store a list of 
> member ids).  whatever you decide has to be bidirectional... i.e. both the 
> member and the group objects must be aware of the relationship
> 
> - provide custom member type, subclassing remember's base type with the 
> following customizations:
> 
>    o groups field: this may become a ReferenceField if you use AT references 
> or plone.relations to relate the groups to the members.
> 
>    o setGroup / getGroup methods: the groups field may need custom accessor 
> and/or mutator methods based on how you manage the group <-> member
> relationships.

>    o valid_groups method: this method provides the set of all available 
> groups, it may need to be customized to correctly generate this list.
> 
> your group class needs to provide membrane's IGroup interface.  if you want
> to 
use one of membrane's default implementations for managing group <->
> member 
relationships (using AT references) then you may need to provide
> another 
interface, e.g. IGroupsProvider or ISelectedGroupsProvider.
> 
> also, i can't say for certain how well the regular Plone group management 
> interface will work.  if you want to retain that interface but have it 
> interact w/ your content-based groups you may need to customize some of that 
> code as well.
> 
> hope this helps,
> 
> -r