home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.sun.misc
- Path: sparky!uunet!gatech!rpi!batcomputer!theory.TC.Cornell.EDU!lijewski
- From: lijewski@theory.TC.Cornell.EDU (Mike Lijewski)
- Subject: strange EINVAL error from write(2) instead of SIGPIPE
- Message-ID: <1993Jan11.234759.24945@tc.cornell.edu>
- Keywords: client/server EINVAL write SIGPIPE
- Sender: news@tc.cornell.edu
- Nntp-Posting-Host: theory.tc.cornell.edu
- Organization: Goddard Space Flight Center
- Date: Mon, 11 Jan 1993 23:47:59 GMT
- Lines: 23
-
-
- I'm getting a strange error in the server of a server/client pair
- running between two Sun IPXs under SunOS 4.1.1. The short of it is
- that write(2) occasionally returns an EINVAL error in the server when
- I kill the client, instead of it getting a SIGPIPE.
-
- The long version:
-
- The server forks with the child doing the processing. The child sets
- up a SIGPIPE catcher with sigaction(2) and then goes about its
- business writing a whole bunch of data. On the other machine when I
- kill the client with a ^C (I catch the signal and close the sockets
- nicely), I occasionally see write(2) fail with EINVAL instead of it
- exiting from the SIGPIPE catcher.
-
- Anyone have any clues as to what is going on? Thanks.
-
-
- --
- Mike Lijewski (H)301-982-5461
- Goddard Space Flight Center
- ARPA: lijewski@rosserv.gsfc.nasa.gov
- SMAIL: 446 Ridge Rd. Apt. 3, Greenbelt, MD 20770
-