>This Rob thinks that they should as well, with the following question:
>How does a session actually expire?
>
>Will sessionids have a TTL as well? In which case a resultset with a TTL
>longer than the session's TTL is just incorrect.
>
>If they don't have a TTL, then are they just silently expired with no
>warning? In which case I would suggest that sessions shouldn't be expired
>while they have resultsets that still have time left to live as this is
>contrary to the only information presented to the client.
>
>Obviously if the implementation maps a session on to a Z Association and
>the connection to the Z server goes down, then the SRW server is in an
>error state regardless.
>
>(Is there a list of SR[WU] diagnostics somewhere, by the way?)
>
>Rob

I would say, that it is preferred that resultsets time out later than sessions but when it is the other way around, the client could save the original CQL to re-execute the query. 

Theo


--
      ,'/:.          Rob Sanderson ([log in to unmask])
    ,'-/::::.        http://www.o-r-g.org/~azaroth/ 
  ,'--/::(@)::.      Special Collections and Archives, extension 3142
,'---/::::::::::.    Twin Cathedrals:  telnet: liverpool.o-r-g.org 7777
____/:::::::::::::.              WWW:  http://liverpool.o-r-g.org:8000/ 
I L L U M I N A T I