From: Iban Hatchond
Subject: cmucl 18c
Date: 
Message-ID: <3A9A905F.E124C29F@emi.u-bordeaux.fr>
Hi folk,

I trying to work with the cmucl 18c release (linux redhat kernel 2.4.0)
but I have a HUGE core dump.
It seams working well until I try to compile Free-CLIM or some other
stuff working with CLX, and creating a lot of garbage.
I success in all compile session, but at loading-time the interpret
crash systematically for an memory access violation.
What I'am sure, is that it should work with free-clim a least (it works
with 18b).
Any suggestion ?


PS : I worked "a lot" on a (very basic) web page for the window-manager.
so please be my guests
http://www.emi.u-bordeaux.fr/~hatchond/Eclipse
From: Pierre R. Mai
Subject: Re: cmucl 18c
Date: 
Message-ID: <87pug5ujd6.fsf@orion.bln.pmsf.de>
Iban Hatchond <········@emi.u-bordeaux.fr> writes:

> I trying to work with the cmucl 18c release (linux redhat kernel 2.4.0)
> but I have a HUGE core dump.
> It seams working well until I try to compile Free-CLIM or some other
> stuff working with CLX, and creating a lot of garbage.
> I success in all compile session, but at loading-time the interpret
> crash systematically for an memory access violation.
> What I'am sure, is that it should work with free-clim a least (it works
> with 18b).

Are you sure you have deleted all binary (i.e. *.x86f if you are on
x86, *.sparcf if you are on SPARC, etc.) files that were produced by
the 18b release?  Binaries compiled with 18b are not compatible with
18c, and could be the cause of your problems, since the FASL loader
doesn't do too many sanity checks.  There is work in progress to
remedy this situation...

> Any suggestion ?

Do a 

find . -name '*.x86f' | xargs rm

and try recompiling.  If your problems still persist, bug me...

> PS : I worked "a lot" on a (very basic) web page for the window-manager.
> so please be my guests
> http://www.emi.u-bordeaux.fr/~hatchond/Eclipse

Hmmm, just tried this with 18c, and I do get a SIGSEGV when the window
manager receives a configure request (because a new application has
started / a new window is being mapped).  Maybe there's a problem with
the definition of the configure request layout?

Other than that the WM seems to work correctly though, i.e. if you
start your clients before the WM, you get correct reparenting, window
dressing, move, resize, (un)iconify, (un)maximize, close, etc. all
seem to work fine.  Really looks great!

Regs, Pierre.

-- 
Pierre R. Mai <····@acm.org>                    http://www.pmsf.de/pmai/
 The most likely way for the world to be destroyed, most experts agree,
 is by accident. That's where we come in; we're computer professionals.
 We cause accidents.                           -- Nathaniel Borenstein