Re: [Bug: 21.5-b16] Core dump

Stephen> I don't recognize this one. It might be the "Lisp not Stephen> properly wrapped in redisplay" bug, but that usually actually Stephen> says so. No Lisp backtrace? You mean what is written to stdout when it dumps? I'll try to save that next time. Stephen> If you still have the core, can you run gdb, source .gdbinit Stephen> from the source directory, and type "lbt" (for "lisp back Stephen> trace")? I don't think it will work, you need a running Stephen> XEmacs. But it's worth a try. (gdb) lbt You can't do that without a process to debug. Expected I guess. Stephen> Also, can you do "frame 6", then "pobj form"? (gdb) frame 6 #6 0x080f76cd in Feval (form=1080187928) at eval.c:3498 3498 Vpending_warnings_tail = Qnil; /* perhaps not strictly necessary, (gdb) pobj form $1 = (struct Lisp_Symbol *) 0x40625c18 $2 = {lheader = {type = 4, mark = 0, c_readonly = 0, lisp_readonly = 0, unused = 0}, next = 0x0, name = 1081370452, value = 1080187928, function = 1079929420, plist = 1080189416} Symbol name: t Stephen> That should tell you some stuff about car_ and cdr_ in the Stephen> struct Lisp_Cons (the form should be a list), and then you Stephen> can do pobj $2.car_ to find out what (car form) is and so on. Well the car_-thing doesn't work so I guess it isn't a list. I don't know really what I'm doing anyway so I'll stop exploring this dump further now. Yours -- %% Mats