Session Management inside TBO, SBO and Documentum Aspects

Raghu_A

I want to create a custom aspect which overrides DfDocument's doSave() method. Withing this override I want to perform the

following operations:-
a. To be able to store a custom file within the Docbase which has initiated the call to this aspect.
b To be able to store a custom file within any Docbase.
c. To be able to store a custom file within any Docbases in a federated environment.

However to perform the store operation I would need access to the IdfSession object. To fetch this object I referred to "EMC® Documentum® Foundation Classes Version 6". I found the following 3 ways to get the session object however, I am not sure what is the suggested solution in all the three cases a, b, and c. .
- 1. Use this.getSession(): In this case we cannot release the session as mentioned in the above documentation under "Getting sessions inside TBOs". Won't this create probs?
- 2. Use IDfTypedObject.getSessionManager(): this way we can get a new session and release it. Right? However, this sessionManager would be associated to a particular Docbase only. Right? Can we change this? Or Even better then this would be the third option as far as we feel.
- 3. Use the following:-
IDfClient localClient = DfClient.getLocalClient();
sessionManager = localClient.newSessionManager();
IDfLoginInfo loginInfo = new DfLoginInfo();
loginInfo.setUser(userName);
loginInfo.setPassword(password);
sessionManager.setIdentity(docbaseName, loginInfo);
docbaseSession = sessionManager.getSession(docbaseName);

Is it correct to say that for my requirement a.) method 2 is best supported and for requirements b.) and c.) method 3 is the best supported?

shailshah

I think method 2 is a variation of method 1. Method 1 provides a convenient way of accessing session without having to worry about releasing it. Method 3 is to connect to a different docbase. In most cases, you will use method 1 (same docbase) or method 3 (for a different docbase); method 2 serves a little purpose for TBOs and could be less efficient.

To answer your questions, I would use method 1 for req a  and method 3 for req b and c. For req b and c, you can optimize your code and default to method 1 if the docbase happens to be the "current" docbase; else use method 3 (if docbase is different than current). Agree?