Hi, I created SLING-2276 [1] and attached a potential but yet untested patch....
Regards Felix [1] https://issues.apache.org/jira/browse/SLING-2276 Am 10.11.2011 um 00:20 schrieb Justin Edelson: > On Wed, Nov 9, 2011 at 3:03 PM, Felix Meschberger <[email protected]> wrote: >> Hi, >> >> Am 09.11.2011 um 22:28 schrieb Justin Edelson: >> >>> On Wed, Nov 9, 2011 at 1:02 PM, Felix Meschberger <[email protected]> >>> wrote: >>>> Hi >>>> >>>> Moving to dev@... (bcc only to users@) >>>> >>>> Am 09.11.2011 um 18:24 schrieb Felix Meschberger: >>>> >>>>>> >>>>>> In addition, I suspect that while the requirements of the DavEx >>>>>> servlet with respect to Sling Authentication are not entirely unique >>>>>> and that *if* there are changes required in Sling Authentication to >>>>>> support the DavEx servlet will benefit other use cases. >>>>> >>>>> Well, there is a unique "feature" which is called "missing-auth-mapping". >>>>> I am not entirely sure, that I want to add support for such >>>>> functionality. Because in essence, this would just replace one >>>>> "anonymous" user with another one (with maybe the added functionality of >>>>> being able to disable that user...). >>>> >>>> On second thought .... >>>> >>>> The location to implement this would probably be the >>>> JcrResourceResolverFactory implementing the getResourceResolver method. >>>> This could then do analogous to the missing-auth-mapping. >>>> >>>> Still not sure, whether we would want to go there ... >>> >>> I was thinking it belonged in SlingAuthenticator's >>> getAnonymousResolver() method. >>> >>> Could be something as simple as: >>> if (request.getAttribute("anonymous.userid") { >>> ResourceResolver resolver = >>> resourceResolverFactory.getResourceResolver(Collections.singletonMap(ResourceResolverFactory.USER_IMPERSONATION, >>> request.getAttribute("anonymous.userid")); >>> } else { >>> ResourceResolver resolver = >>> resourceResolverFactory.getResourceResolver(null); >>> } >> >> The drawback of this solution would be to not be able to easily leverage JCR >> 2 GuestCredentials.... Do we need that ? > > IIUC, in the case Markus is describing (which is where you would set > the anonymous.userid attribute), GuestCredentials wouldn't work. > > In essence, if wanted to put a non-blank, non-null value in > missing-auth-mapping, that's the same value which would go into this > anonymous.userid (which maybe would be better named > missing.auth.userid) attribute. > > Justin > >> >> Regards >> Felix >> >> >>> >>> Justin >>> >>>> >>>> Regards >>>> Felix >> >>
