Re: [STUMP] crashing on startup

[Carlos Konstanski]

Am 25.02.2010 08:47, schrieb Paddy Mullen:
> I am relatively new to lisp and stumpwm but I have liked it so far.
> 
> After getting my setup honed, I hadn't restarted in about a week,
> yesterday I had to restart stumpwm for another reason, I now cannot
> restart it.
> 
> I removed everything from my stumpwm.start
> everything except debugging info from my .stumpwmrc
> 
> 
> (in-package :stumpwm)
> (setf *debug-level* 20)
> (redirect-all-output (data-dir-file "debug-output" "txt"))
> 
> 
> I get the following error in debug-output.txt
> line 8510:39:55 Outputting a message:
> ������� ^2*Welcome to The ^BStump^b ^BW^bindow ^BM^banager!
> ������� Press ^5*C-t ?^2* for help.
> 10:39:55 Window #<XLIB:WINDOW :0 400009> is dock-type. Placing in mode-line.
> 
> debugger invoked on a TYPE-ERROR in thread #<THREAD "initial thread"
> RUNNING {100371F9E1}>:
> � The value 525/2 is not of type (UNSIGNED-BYTE 29).
> 
> I am running ubuntu 910, sbcl 1.0.29.11, on a lenovo t500 in 64 bit mode
> 
> what other information would be helpful to debug this problem?
> 
> 
> 
> _______________________________________________
> Stumpwm-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/stumpwm-devel

If it were me, and I were trying to debug stump, I would run it from an SBCL
prompt, as opposed to the normal way with the built-and-saved image. Then I
would have the benefit of the debugger.

If you login to your machine with a failsafe session (choose this from the
Sessions menu option at the GDM login screen), you get nothing but a blank
screen with a 80x25 borderless xterm. From there you can type "sbcl", and
then, from the REPL, start stump with:

(asdf:operate 'asdf:load-op 'stumpwm)
(stumpwm:stumpwm)

If you want to get fancier, you could login to your machine with your normal
desktop manager (gnome most likely), and then start an Xnest session with a
failsafe session, and run stump in that under SLIME. (I wonder if VNC would
also work for this?) But the first method is simpler, and will provide a
backtrace.

The goal is to figure out where the program is when the error occurs.

Carlos