[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [wsrm] message headers for Ack and Fault ??
Sunil Kunisetty wrote: > > > Tom Rutt wrote: > >> Sunil Kunisetty wrote: >> >> >Tom Rutt wrote: >> > >> > >> > >> >>Sunil Kunisetty wrote: >> >> >> >> >> >> >> >>>> >> >>>> >> >>>I don't think that restriction is warranted. What is important is >> the Poll >> >>>request is initiated by the Sender in a different >> transaction/connection >> >>>than the original message. How the response (Ack or Fault) is sent >> >>>is an implementation issue and doesn't effect our protocol one way >> >>>or other. >> >>> >> >>> >> >>> >> >>I do not understand. The justification for poll is when the sender >> >>cannot accept http requests. >> >> >> >> >> > >> > While that is the primary justification, that's not the only one. >> For your >> > recollection, other use cases are avoid resends & for thin clients. >> > >> > Also, that would have been the prominent one if Poll was only used for >> > that scenario, and not with other patterns. However, now that we allow >> > Poll to be used for Callbacks and Response patterns also, then the >> use >> > case I suggested is much more valid. >> > >> > >> I agree that the poll request can be used to obtain status infomation >> for reliable messages sent using any reply pattern. >> >> The point I am making is that the poll request is always (by definition >> and design) mapped to an underlying request/response protocol. >> >> Thus there is no problem distinguishing a delayed callback response >> (which is always on an underlying request) from a poll response >> (which is always on a underlying response). >> > _As I mentioned earlier_, poll response need to be on the underlying > response. > The requirement we had was that poll request be sent on a separate > transaction/connection than that of the original reliable message. > Requiring > it to be on the same response is an un-necessary implementation > restriction. > > Also, when the ACK or Fault comes to the RMP/SOAP processing later, > it will be (most likely) agnostic to that fact that it was carried > thru' a underlying > request or underlying response. > Currently the Reply To element is stated as only applying to the callback reply pattern. If you really want to open up a callback response for the Poll request, then the protocol would have to be change to explicitly allow this. If this is the case, changing the poll response to have its own header (separate from the response header) with the addition of fault information (as I sent in a prevous mail) would enable such a protocol change. Tom Rutt > > > >> >> >> ---------------------------------------------------- >> Tom Rutt email: tom@coastin.com; trutt@fsw.fujitsu.com >> Tel: +1 732 801 5744 Fax: +1 732 774 5133 >> >> To unsubscribe from this mailing list (and be removed from the roster >> of the OASIS TC), go to >> http://www.oasis-open.org/apps/org/workgroup/wsrm/members/leave_workgroup.php. >> -- ---------------------------------------------------- Tom Rutt email: tom@coastin.com; trutt@fsw.fujitsu.com Tel: +1 732 801 5744 Fax: +1 732 774 5133
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]