Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!uwm.edu!zaphod.mps.ohio-state.edu!sdd.hp.com!think!nrl-cmf!carrow From: carrow@nrl-cmf.UUCP (Steven Carrow) Newsgroups: comp.windows.x Subject: R4 server dying on Sparcs running 4.0 Summary: Honest, I did the fixincludes thing! Keywords: X11R4, Sun, Sparc, gcc Message-ID: <105@nrl-cmf.UUCP> Date: 26 May 90 03:56:10 GMT Organization: Society for the Creatively Deluded Lines: 46 Folks, My X server is curling up and dying on me. The particulars: * X11R4 on Sparcs (bw2 and cg3) running SunOS 4.0 * Compiled with cc where -pic was required, gcc elsewhere * Window manager: twm * Clients in .twmrc: xrdb, xsetroot, xhost, xterm Symptoms: * Server compiles successfully, but upon invoking xinit, the stippled screen comes up and stays up. If I have redirected stderr, my prompt eventually reappears, but "kbd_mode -a" is necessary to restore function; if stderr comes to the screen, a series of bizarre events occur: error messages come up, I am logged out, and the "login:" prompt comes up. The kicker here is that keystrokes then simply act as consecutive newlines: hit 'a', and the prompt scrolls up twice. The error messages are the usual: XIO: fatal IO error 32 (Broken pipe) on X server "unix:0.0" after 3 requests (2 known processed) with 0 events remaining. The connection was probably broken by a server shutdown or KillClient. except that subsequent messages replace "unix:0.0" with "(null)". The usual suspects (tentatively eliminated): * Fixincludes - I screwed that up earlier; fixed now. Not the same messages this time. * Shared libraries - No ld-type messages. * A botched client - I have tried this with various permutations of my .xinitrc; it never even gets to xsetroot. If anyone has any ideas, please email me; even a list of things to watch for on the Sparcs would be appreciated. Thanks loads. ***************************************************************************** * Steven A Carrow * ARPA: carrow@itd.nrl.navy.mil * * Code 5570 * UUCP: {backbone}!uunet! * * Naval Research Lab * nrl-css!carrow (I think) * * Washington, DC 20375 * Disclaimer: I can barely speak for * * (202) 767-9125 * myself, much less the government. * *****************************************************************************