Description of Problem: After kudzu uninstalled my PS/2 mouse I noticed several gpm related problems : ( - the /etc/init.d/gpm file states in the comments that is uses the /etc/sysconfig/mouse as a configuration file. It forgets to mention /etc/sysconfig/gpm , which is also used ) - I discovered this after kudzu removed the mouse from the system : - the gpm executable is executed by the /etc/init.d/gpm script, even if no mouse is configured - the /etc/init.d/gpm script is executed at startup , even if no mouse is configured - after gpm is started , it fails with the following message : Starting console mouse services [OK] gpm: oops() invoked from gpm.c(978) /dev/mouse: No such file or directory Version-Release number of selected component (if applicable): standard redhat linux 7.2 installation. How Reproducible: Set up a system with one PS/2 mouse ( I gues other mouse types are good oo ). Deinstall the mouse. I set the "safe=yes" option in /etc/sysconfig/kudzu. After that on next boot kudzu will claim that the mouse was removed and it will offer to de-install the mouse configuration. I guess simply unplugging the mouse works too ( before the reboot )
* test tree Skipjack-re0320.0 (beta3) i586 * test hardware: ps/2 mouse, ps/2 keyboard, Compaq Proliant ML330 * booting at Runlevel 5 by default (gdm) I disconnected the ps/2 mouse on test machine. I rebooted. Kudzu came up and asked if I wanted to remove the mouse configuration. I removed it. No gpm nor X came up (the proper way to act). I changed the configuration file /etc/sysconfig/kudzu to SAFE=yes. I rebooted. X tried to start but it was unable to suceed: the X-reconfig tool show up, but it was unable to self-reconfigure (this is, obviously, OK: no mouse, no X) and, in the end, X and GDM had to be disabled, As I said, this is OK, expected behaviour. After pluging in the mouse back and rebooting, X comes alive right and well. But I get no Ooops nor other problem.
This was on a beta2 system ( 7.2.91 ) Installed beta2, with aPS/2 mouse. Then I disconnected the PS/2 mouse, rebooted. kudzu noticed the mouse was missing, asked about it, I selecte "Remove Configuration", then the system normaly booted. But : [root@localhost root]# service gpm status gpm dead but pid file exists [root@localhost root]# service gpm stop Shutting down console mouse services: [FAILED] [root@localhost root]# service gpm status gpm dead but subsys locked [root@localhost root]# Part of the problem is that /etc/init.d/gpm uses the following logic to determine if a mouse is present : MOUSECFG=/etc/sysconfig/mouse if [ -f "$MOUSECFG" ]; then . "$MOUSECFG" else echo $"(no mouse is configured)" exit 0 fi if [ "$MOUSETYPE" = "none" ]; then echo $"(no mouse is configured)" exit 0 fi But when unconfiguring the mouse, kudzu does not remove the /etc/sysconfig/mouse file, neither does it modify it to say MOUSETYPE=none. So the script tries to start the gpm subsystem. Another problem is that a lockfile and a pidfile for the gpm service are created , even if gpm fails to start ( which is due to /dev/mouse file missing, kudzu got at least this right ). Also tested on beta3 7.2.92, versions : kudzu-0.99.47-1 gpm-1.20.0-1 exactly the same behavior as with beta2
Same on beta4 ( 7.2.93 ) gpm-1.20.0-2 kudzu-0.99.48-1
Same in RHL 8.0 ( Psyche ) Versions : gpm-1.19.3-23 kudzu-0.99.69-1
Same in 8.0.93 ( Phoebe2 ) Versions : gpm-1.19.3-25 kudzu-0.99.97-1 Changing component to kudzu. Description of the problem , again : When the mouse is removed , kudzu does not delete /etc/sysconfig/mouse or alternatively, it does not change it to say MOUSETYPE="none" Can we mark this bug as public ?
Closing out bugs on older, no longer supported releases. Apologies for any lack of response. Please attempt to confirm with more recent releases.