Ok, I need some advice from the Sun-experts among you! I tried to track down the workshop/XPM/GC-related problems some more. So my plan was the following: I start xemacs under dbx and stop imediately in main. I use ldp to print the object, which gets corrupted later on (using the address from the coredump) -- cool, this is "auto-autoloads.el". Now I'd like to find out wheter there still is some pointer to this string, so I thought (after having browsed through the internals manual): let's do a GC and see if something changes: I call Fgarbage_collect() and... the machine locks up completly! I still could move the mouse but nothing else worked: no ^c, the desktop didn't react, telnet from the outside didn't work anymore. Oh, btw this is our main compute serve -- oops! frustrated users from all over the place converge right here and look at me accusingly... I decide to reboot the machine, type STOP-A, one of our part-time Sysadmin mentions sync and types it at the ok prompt -- sync starts doing its thing, times out -- kernel panics -- reboot doesn't work -- Ah shit! luckily another reboot brought the machine up cleanly. Now I have two questions: - what's evil about calling Fgarbage_collect right at the start of XEmacs -- I had anticipated that this might fail and kill xemacs, but - what the f**k is going on here, a dbx-session killing off the whole machine??? (fed up with debugging for today) Gunnar