Re: [higgins-dev] IContext.close() and underlying connections

["Tom Doman" <TDoman@xxxxxxxxxx>]

I like the idea as well.  We should make sure the prescription of
"return to new" is clear in it's meaning for implementors or they might
think they're compelled to be newer than is required.

Tom

>>> "Jim Sermersheim" <jimse@xxxxxxxxxx> 8/14/2006 1:12 PM >>>
>>> "Tom Doman" <TDoman@xxxxxxxxxx> 8/14/06 1:04 PM >>>
<snip>

>One question, I don't see how the close operation can return the
>Context to the "new" state if, as Jim suggested, we might maintain an
>open connection under the CP covers.
A context's "new" state may be such that the "connection" is open.
It's
really just implementation details at that point I think. The CP may
even consider a context to be in a new state regardless of whether or
not there's already an established (yet not authenticated) connection.

 
I like the idea of prescribing authentication behavior, yet leaving
connection behavior to the CP's.
 
JIm