segfault during "make stage1"

[Miklat]

first, thanks for the effort put into rscheme. From the docs, it looks
like a fine system.

Now for the problem: I've tried this using both gcc3.4 and gcc4, either
way this is what I get at some point:

rshell/rs -image tmp/system.bas -c.repl system.img
make[1]: *** [system.img] Segmentation fault
make[1]: Leaving directory `....../rs-0.7.3.3/stage0'

Any suggestions?

Thanks

[Miklat]

My apologies, I didn't notice the warning on the front page until now.

[Donovan Kolbly]

Someday we'll figure out what the right solution to the gcc 3.3 problem
is... Or everybody will move on to gcc 4.0, which has its own problems
:-)

On Sat, 31 Dec 2005, Miklat wrote:

>
> My apologies, I didn't notice the warning on the front page until now.
>
>

-- Donovan Kolbly ( d.ko...@rscheme.org
( http://www.rscheme.org/~donovan/

[yn.rou...@gmail.com]

I experienced the same problem.
so I am unable to compile and use rscheme 0.7.3.4 b1.

[Donovan Kolbly]

Did you try the --enable-debug trick?

--
-- Donovan

[yn.rou...@gmail.com]

Donovan Kolbly wrote:
> Did you try the --enable-debug trick?
>

Yep, that didn't work.
Here is the gdb backtrace:

Program received signal SIGSEGV, Segmentation fault.
IRC_foundPointer (ircInfo=0x808b368, ptr=0x806992c) at traverse.c:237
237 if (sc->gen->genNum == 7) {
(gdb) bt
#0 IRC_foundPointer (ircInfo=0x808b368, ptr=0x806992c) at
traverse.c:237
#1 0x0805ce89 in do_traversal (gen=0x808b368, time=9776) at
gcclient2.ci:31
#2 0x0805d3ce in run_gen (gen=0x808b368, time=30000) at
traverse.c:1003
#3 0x0805d63b in IRC_safePoint (heap=0x808b368) at traverse.c:1058
#4 0x08054ad0 in quasi_interp_spin (f=0x80651fe <bci_0>) at
gcserver1.ci:22
#5 0x08051b6b in run (cont=<value optimized out>) at entry.c:142
#6 0x08052189 in call_scheme (closure=134799411, num_args=3) at
entry.c:99
#7 0x080652db in rscheme_std_main (argc=3, argv=0xbf9c7304,
modules=0x808b368, default_image=0x808b368 "") at main.c:127
#8 0x0804ea12 in main (argc=134787944, argv=0x808b368) at shell.c:47

I can help to debug if you can't reproduce this bug.

[Donovan Kolbly]

Hmm, I will see if I can reproduce. This is on gcc4 32-bit x86?

--
-- Donovan

[Donovan Kolbly]

I was able to build and run, albeit with lots of warnings, with gcc4
on x86 (this is on a Fedora Core 4 fresh install) using
--enable-debug.

What is your config?

--
-- Donovan

[yn.rou...@gmail.com]

x86
gcc 4.0.3 20060115
ubuntu dapper drake

In fact I am building the source package version 0.7.3.4.b1-2

the debian/ubuntu package has some specific patches, I will try to
disable them and rebuild

[Donovan Kolbly]

I have an ubuntu system, but it only knows about gcc 4.0-0pre6ubuntu7

I can try to build there if you can advise me on how to upgrade it to that
set of packages.

-- Donovan Kolbly ( d.ko...@rscheme.org
( http://www.rscheme.org/~donovan/

[yn.rou...@gmail.com]

You're probably using ubuntu breezy, the currently stable version.
I am working on Dapper Drake the devel version.

You just need to change breezy with dapper in your
/etc/apt/sources.list file

however you might not want to use a unstable system, in this case you
can set up a dapper chroot with deboostrap, or I can give you access to
my box.

[yn.rou...@gmail.com]

To create a dapper chroot:

sudo debootstrap dapper /path/to/chroot
http://archive.ubuntu.com/ubuntu

if debootstrap doesn't know dapper you need the
/usr/lib/debootstrap/scripts/dapper file
you can get it here http://www.pleiades.fr.eu.org/dapper

to have a really working chroot do:
mount -o bind /proc //proc
mount -o bind /tmp //tmp
mount -o bind /dev //dev
mount -o bind /proc/bus/usb //proc/bus/usb
mount -o bind /home //home
then enter the chroot
chroot /parh/to/chroot /bin/bash

[yn.rou...@gmail.com]

err little error,
the mount line should rather be:
mount -o bind /proc /path/to/chroot/proc