On Thu, Jan 29, 2009 at 8:07 PM, Emmanuel Lecharny <[email protected]> wrote: > Gregory Brownell wrote: >> >> Thanks Ashish. I'll try this out later today and let you know how it >> goes. >> >> Greg. >> >> Ashish wrote: >>> >>> On Thu, Jan 29, 2009 at 7:20 PM, Gregory Brownell >>> <[email protected]> wrote: >>> >>>> >>>> Ashish, >>>> >>>> Thanks again for looking into this. Note that this is a receive only >>>> application. The device(s) sending the packets have no capability to >>>> read >>>> anything. So I don't need to worry about writing at all. >>>> >>>> What's important here is the timeout. I suppose I could just write a >>>> timer >>>> that times out after N minutes and is interrupted and reset after each >>>> packet is received, and not worry about session closing/recycling. But >>>> when >>>> I think about how to handle this with multiple clients opening and >>>> closing >>>> all the time I get a headache. >>>> >>>> >>> >>> Ok, its interesting now :-) >>> >>> have 2 solutions for you >>> 1. Write Custom Session Recycler and set to Datagram Acceptor >>> >>> 2. See the ExpirationTimer class in MINA source and can refine the same. >>> >>> Effectively speaking, the timetolive variable is not set for >>> ExpiringObject. >>> >>> Emmanuel: Should we file a JIRA issue for this? > > Hmmm, unless this is a clear bug, or a suspected bug, or an improvement to > the lib, in other words, if it's more or less an usage discussion, it's > better to keep the convo on the ML.
The default implementation is not working :-( > However, if you think it worths some more documentation, then a JIRA is > fine. > > In any case, if you fell like it fits a JIRA, then create one. It's easy to > close, and it's more persistent than a mail ! > Let me keep it at back of my mind, and experiment more. Then shall create an Issue
