> -----Original Message-----
> From: Rocco Caputo [mailto:[EMAIL PROTECTED] 
> Sent: Monday, September 22, 2003 2:45 PM
> To: '[EMAIL PROTECTED]'
> Subject: Re: Session state registration
> 
> 
> On Mon, Sep 22, 2003 at 02:35:05PM -0500, Alan Hastings wrote:
> > I have long-lived sessions that (may) want to add or remove states after
the
> > session is created... the session object has a register_state method
that is
> > not documented, but looks functional if you peek under the covers.  I
> > thought I would check before wandering into this minefield... it looks
like
> > it would be ok to add or remove states, since it appears the SE_STATES
hash
> > is used in a "stateless" (har!) manner.  Any problems with removing
states
> > from the hash? 
> > 
> > I have other ways to do the job (it *is* perl, after all) but this might
be
> > cleaner.
> > 
> > Thanks for any insight --
> 
> Session's register_state() method is quasi-public.  It's got a public
> name (no leading underscore), but it's not documented.  Rather, it's
> published through Kernel's state() method, which acts on "the current
> session".  That's documented in POE::Kernel.
> 
> This is encapsulation through obfuscation. :)

It's almost there... The kernel->state() method has no provision for
removing a state, as far as I can tell.  In fact it takes pains to assure
that the handler is set before calling the session->register_state().  Is
that because there are issues with removing active states?  

Alan

> 
> -- 
> Rocco Caputo - [EMAIL PROTECTED] - http://poe.perl.org/
> 

Reply via email to