On Thu, Dec 1, 2011 at 12:26 PM, Pierre-Arnaud Marcelot <[email protected]> 
wrote:
> On 1 déc. 2011, at 11:08, Emmanuel Lecharny wrote:
>> Hi guys,
>>
>> following the last mails from Jason, it seems that having a bind() method in 
>> the Partition interface is a bt misleading. We most certainly don't need it, 
>> and currently, it does nothing. Moreover, the authenticationInterceptor does 
>> *not* call next.bind(), so there is no way the bind operation can percolate 
>> through the chain to the Partition.
>>
>> We discussed about this with Pierre Arnaud this morning, and I removed the 
>> method from the Partitio interface, t has no impact on the code so far (it's 
>> an experiment, I haven't committed anything yet).
>>
>> So do you think we can get rid of the bind() method in the Partition 
>> interface ?
>
> Analyzing the current situation and implementations, it looks like this 
> method isn't really used.
> Its only real implementation, in AbstractBTreePartition<ID>, throws an 
> LdapAuthenticationNotSupportedException and other implementations, in 
> DefaultPartitionNexus and SchemaPartition, are only following the call to the 
> wrapped partition.
> It really seems useless to me.
>
> +1 for the removal.

Yeah I agree. The only thing that used this was Penrose for some VD
functionality. However I don't think it logically makes sense to have
it here it was a hack for a hack.

+1

-- 
Best Regards,
-- Alex

Reply via email to