Ecaterina Moraru (Valica) wrote:
> On Tue, Dec 7, 2010 at 01:32, Ricardo Rodriguez [eBioTIC.] <
> [email protected]> wrote:
>
>   
>> Ecaterina Moraru (Valica) wrote:
>>     
>>> Hi,
>>>
>>> The behavior is correct because the checking order is: page > space >
>>>       
>> wiki
>>     
>>> (where a space-level setting can be superseded by a (higher ranking)
>>> page-level setting)
>>>
>>>       
>> That's OK but, as Wouter said, if the data model implements
>> real-inheritance, must not I expect that the rights explicitly granted
>> at space level are considered also as explicitly granted at document
>> level as a consequence of rights inheritance?
>>
>>     
>
> yes, but if you define a harder rule at page level, the space level one will
> be overlapsed by the new rule.
>
> Thanks,
> Caty
>
>   

I get your point. thanks. But I think it keep being counter-intuitive 
for me that once I grant some rights on a page to a group, the fact of 
explicitly adding some user (belonging or not to the previous group) 
rights to the same page prevents members of the original granted group, 
included the owner, to access the document.


>   
>> Thanks!
>>     
>>> The only exception I can think off that would help your usecase (but is
>>>       
>> not
>>     
>>> implemented) is to have additional special rights for the document
>>>       
>> Creator.
>>     
>>> Right now the creator gets DELETE right as an additional behavior. Maybe
>>>       
>> we
>>     
>>> should always grant VIEW and EDIT to the creator.
>>> This way, at least, he could fix the rights behavior (by giving rights
>>>       
>> also
>>     
>>> to GroupA).
>>>
>>> Another thing that is missing is a warning that by giving that right, the
>>> giver will lose it.
>>>
>>> If you want to read more about rights:
>>> - http://www.xwiki.org/xwiki/bin/view/FAQ/HowDoesRightsWork
>>> - http://dev.xwiki.org/xwiki/bin/view/Drafts/Access%20Rights
>>> - http://dev.xwiki.org/xwiki/bin/view/Drafts/XWikiRightServiceReversed
>>>
>>> Thanks,
>>> Caty
>>>
>>> On Fri, Nov 19, 2010 at 18:53, Wouter Boasson <[email protected]
>>> wrote:
>>>
>>>
>>>       
>>>> Hi,
>>>>
>>>> We ran into a rights problem, which might be the result of ignorance,
>>>>         
>> but
>>     
>>>> could also be caused by a perceptual omission in the rights model. The
>>>> following happened:
>>>>
>>>> 1. created space, with explicit rights on group 'GroupA' (this
>>>> automatically locks out users who are not a member of this group) => ok
>>>> 2. create/edit a page as user 'UserA', member of 'GroupA' => ok
>>>> 3. UserA (owner/creator of the document) grants view rights to user
>>>> 'UserB', NOT in GroupA => problems!
>>>>
>>>> Now the creator/owner of the document (UserA) can NOT view his own
>>>>         
>> document
>>     
>>>> anymore! Same for problem for every other user in 'GroupA'.
>>>>
>>>> I figured that this is correct from a certain point of view: an explicit
>>>> view for a specific user locks out all other users, but that includes
>>>>         
>> the
>>     
>>>> owner and all other users, including those in 'GroupA', with correct
>>>>         
>> rights
>>     
>>>> at the space level.
>>>> A possible solution is to grant GroupA explicitly at the same time you
>>>> grant a specific user access to a certain page, but people will forget
>>>>         
>> to do
>>     
>>>> so.
>>>>
>>>> My question is: did we do anything wrong, and is it possible to manage
>>>>         
>> the
>>     
>>>> rights in a way that prevents this counter-intuitive behaviour?
>>>>
>>>> I have the feeling that the rights model lacks real-inheritance: when
>>>> checking permissions for a user, it should return the permissions
>>>>         
>> including
>>     
>>>> that of the group as if it were his explicit permissions, also for pages
>>>> that inherit rights from the space. E.g.
>>>> hasView('UserA') should always return 'True' when the group he belongs
>>>>         
>> to
>>     
>>>> has view rights at the space level.
>>>> Now it apparently returns 'False' when there is an implicit override by
>>>> granting a user view rights. Or does inheritance from the space levels
>>>>         
>> stops
>>     
>>>> working as soon as there's any kind of override on a specific page?
>>>>
>>>> A possible but crude work-around could be using some intelligent trigger
>>>> functions in the database to explicitly add all rights from the space to
>>>>         
>> the
>>     
>>>> specific document as soon as an XWikiRights object is written, but
>>>>         
>> that's
>>     
>>>> kind of a last resort.
>>>>
>>>> Could you help me? I hope for a better solution!
>>>>
>>>> Thanks,
>>>> Wouter
>>>>
>>>>
>>>> Wouter Boasson (MSc)
>>>> Geo-IT Research and Coordination
>>>>
>>>> RIVM - National Institute for Public Health and the Environment
>>>> Expertise Centre for Methodology and Information Services
>>>>
>>>> Contact information
>>>> -----------------------
>>>> RIVM
>>>> VenZ/EMI, Pb 86
>>>> t.a.v. dhr. Drs. Wouter Boasson
>>>> Postbus 1
>>>> 3720 BA Bilthoven
>>>>
>>>> T +31(0)302748518
>>>> F +31(0)302744456
>>>> E [email protected]
>>>> mo - th
>>>>
>>>>
>>>> Disclaimer RIVM
>>>> _______________________________________________
>>>> users mailing list
>>>> [email protected]
>>>> http://lists.xwiki.org/mailman/listinfo/users
>>>>
>>>>
>>>>         
>>> _______________________________________________
>>> users mailing list
>>> [email protected]
>>> http://lists.xwiki.org/mailman/listinfo/users
>>>
>>>
>>>       
>> --
>> Ricardo Rodríguez
>> CTO
>> eBioTIC.
>> Life Sciences, Data Modeling and Information Management Systems
>>
>> _______________________________________________
>> users mailing list
>> [email protected]
>> http://lists.xwiki.org/mailman/listinfo/users
>>
>>     
> _______________________________________________
> users mailing list
> [email protected]
> http://lists.xwiki.org/mailman/listinfo/users
>
>   

-- 
Ricardo Rodríguez
CTO
eBioTIC.
Life Sciences, Data Modeling and Information Management Systems

_______________________________________________
users mailing list
[email protected]
http://lists.xwiki.org/mailman/listinfo/users

Reply via email to