Question

I’m trying to implement the following:

I have an Importer class which registers so called Processors, and then executes them by iterating over them and calling execute (Processor is an interface and contains a method void execute()).

What those processors do is connect to an LDAP, and do certain tasks, e.g. retrieves a list of persons and stores them into a database, or reads certain privileges and maps them to groups.

This all works out quite well… except I just don’t know how to handle the LDAP connection.

Here are some ideas:

1. initialize the connection in the constructor (and assign it to a field) and provide a dispose()` method, which closes the connection or
2. do not initialize the connection field in the constructor, but initialize and close it in the execute() method or
3. create the connection in the execute method and pass it to all methods which are called
4. even uglier (create the connection in the Importer and pass it to all processors, then, after execution, close it)

Answer 1

Handle the connection to the LDAP server with a Strategy method that is a required parameter. This de-couples and isolates the code. You should prefer the UnboundID LDAP SDK for this work – and the SDK supports the ability to create an in-memory directory server so there is no need to use “mock” or “fake” connections. Also, you can examine the blog post “LDAP: Programming Practices”.