About ClonedServerEntty

[Emmanuel Lecharny <el...@apache.org>]

Hi,

we are using this data structure to feed it with the existing entry when 
a request enters the chain. It's really useful for many operations like 
modification, as it avoids many lookup all along the chains.

The problem I have with this structure is that the content does not 
reflect the reality. We store two elements :
- the original entry (the one which is in the server when the operation 
enters the chain)
- and what we call the ClonedEntry

This second element is most certainly useless while going down the 
chain, as if it's cloned, it's cloned only once. Then any modification 
done on this data will impact the rest of the chain. My personal guess 
is that we should clone the original entry anytime we want to do 
temporarily modifications, like what we do in the schema interceptor.

We also need to store the resulting modification once it has been 
applied on the server : for a rename operation, we must update some AT 
on the backend (ModifiersName, ModifyTimeStamp), and we must do that on 
the modified entry. That means at some point we store this modified 
entry into this structure.

I would suggest we rename the clonedEntry to modifiedEntry, and that 
this element is fed in the backend, and used when returning.

thoughts ?

-- 
--
cordialement, regards,
Emmanuel Lécharny
www.iktek.com
directory.apache.org