Re: JAAS Interface to sourceid
[email protected] Wed, 26 Mar 2003 17:57:28 -0500
| Newsgroups | gmane.comp.sourceid.sso.devel |
|---|---|
| Message-ID | <OF6B5E876C.5FCB190A-ON85256CF5.007C6840-85256CF5.007E23DD@imclone.com> |
Just to clarify my point on the Socket/HTTP thing...then I'll let it go. To me a browser represents a TCP/IP enabled process/application that can push bytes to and from a server. A browser in the definition of the Liberty protocol (which I'm not looking up because I'm lazy) is probably referring to this process/application. The HTTP side of it is a well known text protocol that the browser can interpret and execute with some degree of reliability. My point is that emulating the behavior of a browser with a socket library and some patchwork logic is completely feasible. A full blown browser with email support, html WUSIWYG editor, instance messenger, is not required to fulfill the requirement of browser. I believe this dead horse is well beat. I read your thread regarding JAAS, Tomcat, the Authenticators, and the general lack of a unified security mechanism in Java. Without JSR 196 around, moving credentials around from web container to ejb container to stand alone app is very primitive and/or proprietary. I agree with your assessment and I'm not interested (here anyway) to solve that problem. So, with that aside, this is a developers mailing list and I'm a developer who is willing to contribute. LECP is scheduled to be delivered shortly. Is it possible to get my hands on it, perhaps a CVS server somewhere? Also, what about refactoring out any reference to a servlet in org.sourceid.sso.util? I've done it the last beta drop with some success. This leaves the reading of the configuration and other functions free to be implemented in a non servlet-esq environment. Let me know how I should contribute these and other changes. Also, do you have any test cases for testing this? Thanks, Noah