jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefan Guggisberg <stefan.guggisb...@gmail.com>
Subject Re: Question about same name siblings and authorization
Date Tue, 03 Jul 2012 09:39:35 GMT
On Tue, Jul 3, 2012 at 11:19 AM, Bart van der Schans
<b.vanderschans@onehippo.com> wrote:
> On Tue, Jul 3, 2012 at 11:06 AM, Jukka Zitting <jukka.zitting@gmail.com> wrote:
>> Hi,
>>
>> On Tue, Jul 3, 2012 at 10:55 AM, Bart van der Schans
>> <b.vanderschans@onehippo.com> wrote:
>>> We running into a (potential) issue with same name siblings and
>>> authorization and I would like to have some feedback before we try to
>>> fix this (if needed at all).
>>
>> Your problem boils down to having "same name siblings" and
>> "authorization" in the same sentence, or perhaps even having same name
>> siblings in the first place. You're probably better off if you try to
>> avoid the SNS feature entirely.
>
> We are very aware of this, but unfortunately we are dealing with a bit
> of legacy here :-(
>
>>> Which is a bit odd, because how can the session know there is more
>>> than one item sibling? Does anybody knows what the spec has to say
>>> about this?
>>
>> The spec leaves this open by explicitly allowing access control to
>> leave gaps in the SNS sequence. It's not an elegant solution, but
>> AFAIUI nobody really considered such interactions when the SNS feature
>> was specified (IIUC it came up just as a logical consequence of the
>> XML import feature).
>
> Ah, thanks Jukka and Stefan for the quick replies.
>
> So the "compacts indices" mode for SNS is also allowed (as stated in
> paragraph $22.5), but I suspect that may take quite some effort to
> implement this in the current code base. And even then it probably
> will come at the cost of some performance.

i would guess that implementing session-local index compacting would
result in pretty nasty code. apart from that i don't see any benefit in
doing so as the current behavior is clearly spec-compliant.

cheers
stefan

>
> Regards,
> Bart

Mime
View raw message