home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.protocols.iso
- Path: sparky!uunet!news.univie.ac.at!scsing.switch.ch!univ-lyon1.fr!ghost.dsi.unimi.it!rpi!uwm.edu!linac!att!cbnewse!cbnewsd!att-out!walter!porthos!base.bellcore.com!rhp
- From: rhp@base.bellcore.com (Randy Parker)
- Subject: Re: CMIS get-response question - please help !
- Organization: Bellcore, Livingston, NJ
- Date: Wed, 27 Jan 93 15:14:53 GMT
- Message-ID: <1993Jan27.151453.4446@porthos.cc.bellcore.com>
- References: <86010007@nsmdserv.CND.HP.COM>
- Sender: netnews@porthos.cc.bellcore.com (USENET System Software)
- Lines: 15
-
- |> The issue comes up when you have errors related to some attributes
- |> of an object. If the answer to the question above is 'yes', one could send
- |> back a 'GET-LIST-ERROR' as a CMIS-LINKED-REPLY. If the answer is 'no', one
- |> has to send back a CMIS-SERVICE-ERROR if the reply involves one objext, but
- |> a CMIS-LINKED-REPLY if the reply involves more than one object.
-
- This issue came up during interoperability testing. I do not believe there is
- anything in the standard preventing you from sending the linked error rsp.
-
- |> I like to have just one type of reply in this case and let that be
- |> a CMIS-LINKED-REPLY whether the reply involves one object or more.
-
- This is how the agent did it. They used the process fail error instead of
- the get list error. I do not know if this was officially clarified (from ISO).
- I can just say that my experience is that your interpretation is ok.
-