Issue 3440: adapter_activator (interceptors-rtf) Source: IONA (Mr. Matthew Newhook, ) Nature: Uncategorized Issue Severity: Summary: Should interceptors be called before POA adapter activators are called? Resolution: Based on the discussion, it appears that this was intentionally left undetermined. Revised Text: Actions taken: March 22, 2000: received issue January 9, 2001: closed issue Discussion: End of Annotations:===== Date: Wed, 22 Mar 2000 15:43:47 -0330 From: Matthew Newhook To: interceptors-ftf@omg.org Subject: adapter_activator Message-ID: <20000322154347.A26775@ooc.com> Mime-Version: 1.0 X-Mailer: Mutt 1.0pre3us Content-Type: text/plain; charset=us-ascii X-UIDL: &#K!!f;md9P=%"!'4Zd9 Hi, Should interceptors be called before POA adapter activators are called? Regards, Matthew -- Matthew Newhook E-Mail: mailto:matthew@ooc.com Software Designer WWW: http://www.ooc.com Object Oriented Concepts, Inc. Phone: (709) 738-3725 From: Paul Kyzivat To: interceptors-ftf@omg.org Subject: RE: adapter_activator Date: Wed, 22 Mar 2000 14:59:09 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2448.0) Content-Type: text/plain; charset="iso-8859-1" X-UIDL: dmb!!U!5e9BNfd9V1)!! I presume you are talking about the receive_request interception point? If so, the answer has to be NO, for the following reasons: - The receive_request point isn't supposed to be called until the arguments are available. - It isn't possible to know what type the target object is and unmarshal the arguments until the servant is available. - You can't find the servant until you have the POA. By similar logic, the interceptor can't be called until after the ServantManager has been called. Paul > -----Original Message----- > From: Matthew Newhook [mailto:matthew@ooc.com] > Sent: Wednesday, March 22, 2000 2:14 PM > To: interceptors-ftf@omg.org > Subject: adapter_activator > > > Hi, > > Should interceptors be called before POA adapter activators > are called? > > Regards, Matthew > -- > Matthew Newhook E-Mail: > mailto:matthew@ooc.com > Software Designer WWW: http://www.ooc.com > Object Oriented Concepts, Inc. Phone: (709) 738-3725 > X-Authentication-Warning: marcy.adiron.com: polar owned process doing > -bs Date: Wed, 22 Mar 2000 15:14:00 -0500 (EST) From: Polar Humenn To: Matthew Newhook cc: interceptors-ftf@omg.org Subject: Re: adapter_activator In-Reply-To: <20000322154347.A26775@ooc.com> Message-ID: > MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-UIDL: \Ded9?B-!!kY9!!Onad9 On Wed, 22 Mar 2000, Matthew Newhook wrote: > Hi, > > Should interceptors be called before POA adapter activators are > called? For Security, Access Control and Auditing the following are the requirements. 1. Getting the complete unique id (adapter and servant) of the servant within the server and the service contexts before the request is dispatched. On systems where the id is meaningfull, 1 might suffice. However, other systems need information from the Servant implemenation. Therefore, the servant must be located and activated and ready. An interceptor of this form might have intimate knowledge of the servant implemenation. Alternatively, a Policy could be placed on its POA that has intimate knowledge of the servants implementation to derive such a thing as a domain name. So, a policy placed on the POA that might answer query its servants: interface DomainNamePOAPolicy : CORBA::Policy { DomainName get_domain_name( in PortableServer::Servant serv ); }; Would be appropriate. So therefore I believe an interceptor should be called after POA activation and Servant activation? I guess we want to be right there just before the servant unmarshals the arguments. Does that mean another interceptor point? Cheers, -Polar > > Regards, Matthew > -- > Matthew Newhook E-Mail: > mailto:matthew@ooc.com > Software Designer WWW: http://www.ooc.com > Object Oriented Concepts, Inc. Phone: (709) 738-3725 > ------------------------------------------------------------------- Polar Humenn Adiron, LLC mailto:polar@adiron.com 2-212 CST Phone: 315-443-3171 Syracuse, NY 13244-4100 Fax: 315-443-4745 http://www.adiron.com Date: Wed, 22 Mar 2000 16:51:23 -0330 From: Matthew Newhook To: Paul Kyzivat Cc: interceptors-ftf@omg.org Subject: Re: adapter_activator Message-ID: <20000322165123.A27298@ooc.com> References: <9B164B713EE9D211B6DC0090273CEEA926BD2A@bos1.noblenet.com> Mime-Version: 1.0 X-Mailer: Mutt 1.0pre3us In-Reply-To: <9B164B713EE9D211B6DC0090273CEEA926BD2A@bos1.noblenet.com> Content-Type: text/plain; charset=us-ascii X-UIDL: 5,8e9(U[!!;>#!!["`!! Hi, On Wed, Mar 22, 2000 at 02:59:09PM -0500, Paul Kyzivat wrote: > I presume you are talking about the receive_request interception > point? No, I'm talking about the receive_request_service_contexts point. This point is called prior to servant managers since the service contexts have to be processed prior to that (for instance, transaction context has to be established). > If so, the answer has to be NO, for the following reasons: > - The receive_request point isn't supposed to be called until > the arguments are available. > - It isn't possible to know what type the target object is > and unmarshal the arguments until the servant is available. > - You can't find the servant until you have the POA. > > By similar logic, the interceptor can't be called until after the > ServantManager has been called. > > Paul Regards, Matthew -- Matthew Newhook E-Mail: mailto:matthew@ooc.com Software Designer WWW: http://www.ooc.com Object Oriented Concepts, Inc. Phone: (709) 738-3725 From: Paul Kyzivat To: interceptors-ftf@omg.org Subject: RE: adapter_activator Date: Wed, 22 Mar 2000 15:39:57 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2448.0) Content-Type: text/plain; charset="iso-8859-1" X-UIDL: ZKHe9K>Td9$ZGe9?l\!! > From: Matthew Newhook [mailto:matthew@ooc.com] > No, I'm talking about the receive_request_service_contexts point. This > point is called prior to servant managers since the service contexts > have to be processed prior to that (for instance, transaction context > has to be established). In that case, unless there is some compelling reason, I would prefer to leave the point of call undefined, to permit the most implementation latitude. Since AdapterActivators can be called for reasons other than an incoming request, calls to them should be considered independent of the interceptor path. Paul Date: Wed, 22 Mar 2000 12:42:00 -0800 From: Harold Carr X-Mailer: Mozilla 4.51 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 To: Matthew Newhook CC: interceptors-ftf@omg.org Subject: Re: adapter_activator References: <20000322154347.A26775@ooc.com> Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii X-UIDL: Z@>e9B`:e9*"/!![XDe9 If I recall we discussed this and decided not to specify receive_request viz-a-viz adapter activation. This came from IONA who wanted to ensure the ability to shortcircuit request demultiplexing... H Matthew Newhook wrote: > > Hi, > > Should interceptors be called before POA adapter activators are > called? > > Regards, Matthew > -- > Matthew Newhook E-Mail: > mailto:matthew@ooc.com > Software Designer WWW: http://www.ooc.com > Object Oriented Concepts, Inc. Phone: (709) 738-3725 Date: Wed, 22 Mar 2000 18:29:56 -0330 From: Matthew Newhook To: Paul Kyzivat Cc: interceptors-ftf@omg.org Subject: Re: adapter_activator Message-ID: <20000322182956.A27811@ooc.com> References: <9B164B713EE9D211B6DC0090273CEEA926BD2B@bos1.noblenet.com> Mime-Version: 1.0 X-Mailer: Mutt 1.0pre3us In-Reply-To: <9B164B713EE9D211B6DC0090273CEEA926BD2B@bos1.noblenet.com> Content-Type: text/plain; charset=us-ascii X-UIDL: 8Oi!!XF#"!UDdd9^m0!! Hi, On Wed, Mar 22, 2000 at 03:39:57PM -0500, Paul Kyzivat wrote: > > From: Matthew Newhook [mailto:matthew@ooc.com] > > > No, I'm talking about the receive_request_service_contexts > point. This > > point is called prior to servant managers since the service > contexts > > have to be processed prior to that (for instance, transaction > context > > has to be established). > > In that case, unless there is some compelling reason, I would prefer > to > leave the point of call undefined, to permit the most implementation > latitude. > > Since AdapterActivators can be called for reasons other than an > incoming > request, calls to them should be considered independent of the > interceptor > path. I guess what I'm primarily interested in is if the OTS (draft) specification has had anything to say on this. It was the OTS that was the the motivation to have servant managers be called *after* receive_request_service_contexts since the spec apparently has verbage that says that the servant managers will be called within a transaction context. > Paul Regards, Matthew -- Matthew Newhook E-Mail: mailto:matthew@ooc.com Software Designer WWW: http://www.ooc.com Object Oriented Concepts, Inc. Phone: (709) 738-3725 Date: Thu, 23 Mar 2000 16:41:51 +1000 (EST) From: Michi Henning To: Harold Carr cc: Matthew Newhook , interceptors-ftf@omg.org Subject: Re: adapter_activator In-Reply-To: <38D93018.22CC100C@sun.com> Message-ID: Organization: Object Oriented Concepts MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-UIDL: #A!!!75%e9$C;!!0"(!! On Wed, 22 Mar 2000, Harold Carr wrote: > If I recall we discussed this and decided not to specify > receive_request viz-a-viz adapter activation. This came > from IONA who wanted to ensure the ability to shortcircuit > request demultiplexing... I don't see how that is possible. For the arguments to be available in receive_request, the servant must have been located, which in turn implies that the adapter must have been activated, otherwise, there couldn't be a servant, but without a servant, the ORB can't unmarshal the arguments... Cheers, Michi. -- Michi Henning +61 7 3891 5744 Object Oriented Concepts +61 4 1118 2700 (mobile) Suite 4, 904 Stanley St +61 7 3891 5009 (fax) East Brisbane 4169 michi@ooc.com.au AUSTRALIA http://www.ooc.com.au/staff/michi-henning.html Date: Wed, 22 Mar 2000 23:17:30 -0800 From: Harold Carr X-Mailer: Mozilla 4.51 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 To: Michi Henning CC: Matthew Newhook , interceptors-ftf@omg.org Subject: Re: adapter_activator References: Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii X-UIDL: gY`!!Sa]d9Y#;e9`D&e9 Michi Henning wrote: > > On Wed, 22 Mar 2000, Harold Carr wrote: > > > If I recall we discussed this and decided not to specify > > receive_request viz-a-viz adapter activation. This came > > from IONA who wanted to ensure the ability to shortcircuit > > request demultiplexing... > > I don't see how that is possible. For the arguments to be available > in > receive_request, the servant must have been located, which in turn > implies > that the adapter must have been activated, otherwise, there couldn't > be > a servant, but without a servant, the ORB can't unmarshal the > arguments... Perhaps Matt Mihic or Bob Kukura could jump in here. I seem to remember them being reluctant to give details (i.e., trade secrets)... H