>Hmm, so probably exception is thrown only if password is in hashed form.
>I'll try to check this.
I'm testing it now, too.

>I'm not sure if putting ThreadLocal inside callback object is a good idea
:/
>Maybe its better to create external object which will contain
>ThreadLocal object and have same usefull data accessing methods. You
>have to also remeber to clear this variable at the end of request.
Fair enough.

>For password in plain form it should work, but for hashed form you can
>encounter a few problems. (sending plain password should be avoided
>IMHO, so at least ssl should be used)
>I usually  keep user data during msg processing ( in ThreadLocal or
>appContexts , depends on api ) so every service method has easy access
>to it . 
Yes, I'm planning on an SSL connection.

I'll let you know how the test pans out.  I'll submit a test case via jira.
--
View this message in context: 
http://www.nabble.com/WS-Security-and-UserTokens-t1543793.html#a4249407
Sent from the XFire - User forum at Nabble.com.

Reply via email to