Problem has steadily worsened. Started with the "C library" reports that I've seen mentioned here, then system started reporting the "can't open widget" message. Now when I try to start linuxconf from the terminal I get "broken pipe". When I try to start it again I get Error message from readmin :Xlib: connection to ":0.0" refused by server Error message from readmin :Xlib: Invalid MIT-MAGIC-COOKIE-1 key Error message from readmin : Error message from readmin :Gtk-WARNING **: cannot open diplay: :0.0 readmin (GUI frontend) exiting abnormaly # echo $DISPLAY produces: :0.0 # xauth list $DISPLAY returns mylogin-name.my.isp.com/unix:0 MIT-MAGIC-COOKIE-1 eb3999f524929996aee16e970bddeo67c # linuxconf --text works fine (so far) Another symptom I'm fairly sure is related: I added a user account (gwalsh). Then I changed it, then completely deleted it --wiping all account info from the disk. Around this time linuxconf started getting weirder. Now when I create a new acount using the same user name (gwalsh), I can login, gnome starts, but there is *no* panel. This happens only with this one username, all other accounts are fine. It occurs even if I completely delete the account and reboot. It occurs even if I create the account using "linuxconf --text" Obviously a file is corrupted and that username and gnome don't get along any longer. I can switch the desktop in that account to KDE and the account is fine. Unfortunately, I'm a complete unix/linux novice and 6.1 has been my first install, so I can't really tell you more than that. I'd love to know how to fix this particular account. <gripe on> I also can't begin to tell you what acomplete and total pain in the ass this experience has been. How did you guys ever ship a release with that munged up installer? Didn't anyone test it? I bet I installed at least *20* times over a period of three days before I finally decided to wipe out my NTFS partition and got the install to work. Then it screwed up my Win98 MBR. I've had serious troubles getting Disk Druid to recognize linux partitions, and now this. For a total newbie, this has been a real drag. Trust me on this. For an outfit that just did a major IPO it seems like your quality control is kinda loose. </gripe off> Thanks for listening. Let me know if there is anything I can do besides wait for a fix and/or reinstall
When you removed the account, did you select the option to remove the account's data at the same time? It sounds as if you have created the new account and reused the same home directory, somehow having given the user a different UID than before. With a different UID or GID owning your files, strange things would happen. Please check that the owner of your files is your user by running: ls -ln $HOME and comparing the uid and gid listed before the file sizes with the output of id -a
I'm seeing similar problems using ssh with X forwarding. I log in with ssh machine -l root which lets me in OK and X forwarding works for all other programs than linuxconf. I tried xterm, xbiff, gcombust. When I start linuxconf I get # linuxconf X11 connection rejected because of wrong authentication at Thu Jul 27 13:52:09 2000. a Rejected connection at Thu Jul 27 13:52:09 2000: X11 connection from machine.domain.dk port 1293 Error message from remadmin :Gdk-ERROR **: X connection to machine.domain.dk:10.0 broken (explicit kill or server shutdown). The Xauth details are: # xauth Using authority file /tmp/Xauthv12345 xauth> list machine.domain.dk:10 MIT-MAGIC-COOKIE-1 21e98dd1234567893dd8f2cd0f2a4fef xauth> quit So perhaps linuxconf is getting confused because it looks like a local server, but you have to use TCP, you can't use Unix domain sockets (AFAIK) to talk to the ssh forwarding port. I did an strace of linuxconf, and here is what looks most likely to be vaguely relevant: lstat("/bin/remadmin", {st_mode=S_IFREG|0755, st_size=835, ...}) = 0 socketpair(PF_UNIX, SOCK_STREAM, 1, [14, 15]) = 0 pipe([16, 17]) = 0 pipe([18, 19]) = 0 rt_sigaction(SIGCHLD, {0x80b61e4, [], SA_RESTART|0x4000000}, {SIG_DFL}, 8) = 0 fork() = 12373 close(17) = 0 close(19) = 0 close(15) = 0 fcntl(14, F_GETFL) = 0x2 (flags O_RDWR) fstat(14, {st_mode=S_IFSOCK|0777, st_size=0, ...}) = 0 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x402a3000 _llseek(14, 0, 0xbfffda78, SEEK_CUR) = -1 ESPIPE (Illegal seek) write(14, "Version 3 \"Invalid GUI protocol "..., 41) = 41 select(19, [14 16 18], NULL, NULL, {1000000, 0}) = 1 (in [14], left {999999, 980000}) read(14, "prefer treemenu\n", 9999) = 16 write(14, "23 main-0-1 \"Command line usage\""..., 1024) = 1024 write(14, "inuxconf_dialout site ppp-device"..., 1024) = 1024 write(14, " domain [--startuid uid ] [--dal"..., 856) = 856 select(19, [14 16 18], NULL, NULL, {1000000, 0}) = 1 (in [16], left {999999, 980000}) read(16, "Gdk-ERROR **: X connection to br"..., 9999) = 95 write(2, "Error message from remadmin :Gdk"..., 124Error message from remadmin :Gdk-ERROR **: X connection to brage.mjolner.dk:10.0 broken (explicit kill or server shutdown). ) = 124 select(19, [14 16 18], NULL, NULL, {1000000, 0}) = 3 (in [14 16 18], left {1000000, 0}) --- SIGCHLD (Child exited) --- In contrast, xbiff, which works does a access("/tmp/Xauthv12345", R_OK) = 0 open("/tmp/Xauthv12345", O_RDONLY) = 4 read(4, "\0\0", 2) = 2 read(4, "\0\4", 2) = 2 read(4, "\n\0\0\7", 4) = 4 read(4, "\0\2", 2) = 2 read(4, "10", 2) = 2 read(4, "\0\22", 2) = 2 read(4, "MIT-MAGIC-COOKIE-1", 18) = 18 I can't find any references to the Xauth file when I strace linuxconf. The environment file XAUTHORITY=/tmp/Xauthv12345 is set. Versions on the machine I am ssh'ing to: openssh-server-1.2pre11-1 linuxconf-1.17r2-6 mkxauth-1.7-12 pam-0.72-6 gcombust-0.1.28-1 gtk+-1.2.6-7 redhat-release-6.2-1 and on the machine I am ssh'ing from: XFree86-3.3.5-1.6.0 ssh-1.2.27-7i redhat-release-6.0-1
Closing because we don't ship linuxconf anymore
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.