[Kamailio-Devel] Patch: dialog module / is_known_dlg().

[Jerome Martin]

On Wed, 2008-11-12 at 18:22 -0500, Alex Balashov wrote:
> Jerome Martin wrote:
> > Alex,
> > 
> > On Wed, 2008-11-12 at 12:32 -0500, Alex Balashov wrote:
> >> Hello,
> >>
> >> I have attached a very simple patch to provide a new exported function 
> >> from the dialog module - is_known_dlg().  This can be used from script 
> >> to determine if the message being processed is part of an existing 
> >> dialog tracked by the dialog module, irrespective of the use of profiles.
> >>
> >> Because the dialog module is plugged into the RR / TMCB event loop it is 
> >> aware of all dialogs regardless of whether profiles are used.  Thus, 
> > 
> > I think this needs double checking, because the callback as registered
> > in the dialog module might not make dialog matching available before
> > t_relay or loose_route / record route is called ... I am saying this
> > from memory because I had to patch my dialog module in-house at some
> > point to make the dialog matching available upon entering the script
> > config (thus registering an other callback like this:
> > register_script_cb( dlg_onreqprescript, PRE_SCRIPT_CB|REQ_TYPE_CB,0)).
> > This is counter intuitive because the tmcb TMCB_REQUEST_IN, dlg_onreq
> > event, if I remember correctly, happens a bit later than what
> > intuitively expected, aka after t_relay I think. Again I haven't checked
> > this recently and might be wrong, just quoting from memory. 
> 
> Even if that's so, I don't know that this really injures the usefulness 
> of the patch in most conceivable use cases.  I don't have dialogs that 
> aren't RR'd and t_relay()'d out.  

Sure, but surely there are use cases where you need to know if the SIP
request belongs to a dialog BEFORE t_relaying it :-)



> But I suppose the clarification would 
> be desirable.
>