Solaris 8 network programming problem
- From: "sansp00" <Patrick.Sansoucy@xxxxxxxxx>
- Date: 31 Oct 2006 10:46:28 -0800
Hello,
First post here and I really hope someone will be able to give
pointers ... We have implemented a messaging library using the TCP
functionnality under Solaris 8.
We are stumped at a problem where we poll for events and when we read
the polled FD, the read operation returns 0, but the FD is NOT
deconnected on the other side. Both client and server application are
running on the same box and using loopback ips.
We noticed that ignoring the 0 read and waiting a bit, some data does
arrive on the FD ( ie read > 0 ). But always ignoring the 0 read
brings up the issue of knowing/trapping a client disconnection since
read 0 normally points to a closed socket.
The problem is not found in Solaris 6 ... We've been kicking at this
and trying workaround for the last year or so without any good
resolution methods
Can someone enlighten me ...
Thanks !
Patrick S.
.
- Follow-Ups:
- Re: Solaris 8 network programming problem
- From: Maxim Yegorushkin
- Re: Solaris 8 network programming problem
- Prev by Date: Re: Maximum limit of signal queue
- Next by Date: find a new process created in Linux
- Previous by thread: Re: Maximum limit of signal queue
- Next by thread: Re: Solaris 8 network programming problem
- Index(es):
Relevant Pages
|
