Bus error on sparc64 running OpenBSD

[Mario Domenech Goulart]

$ ~/chickens/4.7.2/bin/csi                        
[panic] `##sys#error-hook' is not defined - the `library' unit was probably not linked with this executable - execution terminated

$ gdb ~/chickens/4.7.2/bin/csi
GNU gdb 6.3
Copyright 2004 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "sparc64-unknown-openbsd4.6"...
(gdb) run
Starting program: /home/bunny351/chickens/4.7.2/bin/csi 

Program received signal SIGBUS, Bus error.
0x0000000050939e28 in f_7560 (c=4, t0=0, t1=1162484112) at expand.c:8844
8844    t5=C_a_i_list(&a,4,((C_word*)((C_word*)t0)[8])[1],((C_word*)((C_word*)t0)[7])[1],((C_word*)t0)[6],t4);
(gdb)

This is on gcc13.fsffrance.org

ssh -i .ssh/gcc-compile-farm -p 9204 bunny351@gcc13.fsffrance.org

The sources are in ~/chicken-4.7.2. I used

gmake PLATFORM=bsd PREFIX=/home/bunny351/chickens/4.7.2 DEBUGBUILD=1 HACKED_APPLY=0 install

to build chicken

[felix winkelmann]

I was able to get the sparc64 build back to working on Linux, but currently I'm clueless as to why this fails on OpenBSD. I tried various things (higher nursery settings, disabling the (broken) apply hack, etc. but to no avail).

[felix winkelmann]

gcc64 has been updated on the compile-farm - the test should be repeated.

[felix winkelmann]

Executables built on this machine behave totally erratic, it is nearly impossible to reproduce a failure on a second run. Errors occurring are stack overflows, errors during initialization of the library ("'##sys#error-hook' not defined") and segmentation faults.

I have no idea what's going on here.

[felix winkelmann]

This machine seems to be seriously broken. See:

​https://mail.gna.org/public/gcc-cfarm-users/2012-03/msg00002.html