Hi David, I'm not thinking about interceptor specific data, but generic informations. To me, we have a lot of structure which are associated during the interceptor initialization which are not associated with the interceptor itself, but as the operation needs them, we have added them in the init phase. Not sure this was the best idea, because that means you have to create a lot of static variables all over the interceptors.
I'm not sure that the cost of passing a new element in the opContext will be measurable... Anyways, we have to discuss further what I have in mind, because it's a little bit fuzzy for me too atm ;) On 10/3/07, David Jencks <[EMAIL PROTECTED]> wrote: > I tend to think this is not a very good idea or else I don't > understand it. I think the interceptors will be clearest if they are > injected (by spring or external setup code) with whatever information > they need to do their job that is not request specific and the > opContext objects are setup with only the request-specific > information. Likewise the opContext object will be quickest to setup > the less information they have in them. > > thanks > david jencks > > On Oct 3, 2007, at 5:45 AM, Emmanuel Lecharny (JIRA) wrote: > > > Injecting more information in the opContext structure > > ----------------------------------------------------- > > > > Key: DIRSERVER-1081 > > URL: https://issues.apache.org/jira/browse/ > > DIRSERVER-1081 > > Project: Directory ApacheDS > > Issue Type: Improvement > > Affects Versions: 1.5.1 > > Reporter: Emmanuel Lecharny > > Fix For: 2.0.0 > > > > > > As each server operations now use this opContext for every > > operations, it would be useful to inject more information into this > > context, like access to the registries, or other things which are > > currently stored locally in each interceptors during the > > initialization phaze. > > > > It will be more elegant, will allow global modification of the > > server without having to reinit all the interceptors, and could > > help to define specific operation's context for a request (for > > instance, if we think about an operation associated to a specific > > partition, we could perfectly inject the partition associated > > registries, allowing us to have per-partition regristries, instead > > of one single registries set for all the partition - just an idea) > > > > -- > > This message is automatically generated by JIRA. > > - > > You can reply to this email to add a comment to the issue online. > > > > -- Regards, Cordialement, Emmanuel Lécharny www.iktek.com
