I thought I had sent this email, but I had some mailer troubles, so I
think I didn't go through... :(
Hello, and thanks for the answer!
I was looking through the code this afternoon a little. I was thinking
that maybe I could implement this myself into repoze.what, however, I do
have some questions you might be able to answer. Is it necessary that
all items in a section be of the same type: can a group have items which
are users and others which are groups?
-> yes: no problem, but how do you distinguish the item types? and would
you have a favorite way of modelling this in sqlalchemy? what would a
good generic way be to look for nested groups?
-> no: does this mean I would have to revamp the base class for sources?
or would you rather an approach favoring a secondary groups source,
containing the groups nesting?
Again, thanks for your insight and quick answer !
Gustavo Narea a écrit :
> Bonjour, Yannick.
> The predicate checkers are fine, the problem is that the so-called source
> adapters  don't support nested groups yet.
> I myself need this too, so it should be implemented at some point. Although
> unfortunately it won't happen in the next few weeks, because I won't have
> enough time. :/
> - Gustavo.
>  http://what.repoze.org/docs/1.x/Manual/GettingStarted.html
> brehon said:
>> I would like to use repoze.what (in a turbogears 2.0 context)in conjunction
>> with nested groups. For instance, I would like to be able to have a group
>> belonging to another group, and so on. Users could be a part of any group
>> also of course. This is not trivial using the current system: if user is
>> part of group G1 and G1 is part of group G0, removing G1 from G0 should
>> remove U from G0,, if and only if U was not also a "direct" member of G0.
>> The various repoze.what functions do not support nested groups as far as I
>> can see, but I may be wrong. What would your insight be regarding this
Repoze-dev mailing list