Friday, February 3, 2006, 3:52:36 AM, Nicolas Cannasse wrote:
> Doug Currie wrote:
>> Thursday, February 2, 2006, 12:00:09 PM, Nicolas Cannasse wrote:
>>
>>
>>>>>If still crashing after the update could try to follow the different
>>>>>calls being made and tell me in which function it's crashing ?
>>
>>
>> Another clue...
>>
>> neko.exe caused an Access Violation at location 10007cd7 in
>> module neko.dll Reading from location 0000000c.
> Are theses errors reproductible, or do random errors are hapenning
> depending of the runs ?
It seems to happen very early, every time, but since I am making small
changes to try to debug this, the info I get from gdb or DrMingw is
often a little different. The consistent part is: Segmentation fault
at a low memory address, and no stack.
> You might try to expand by-hand the macro DoCall so we can see with
> intruction is failing, although it seems a bit strange. It might be
> a GC-related issue (wrt your previous report).
I did see a call to *new once when I added this line
#define DoCall(this_arg,pc_args) \
...
} else if( val_tag(acc) == VAL_PRIMITIVE ) { \
fprintf(stderr, "c: %s\n", val_string(((vfunction*)acc)->module)); \
My latest attempt at gdb'ing lead to:
Program received signal SIGSEGV, Segmentation fault.
0x10007cd7 in neko_interp (vm=0x6f636124, m=0x7970, acc=0, pc=0x2b)
at vm/interp.c:649
649 DoCall(vm->vthis,pc[-1]);
(gdb) backtrace full
#0 0x10007cd7 in neko_interp (vm=0x6f636124, m=0x7970, acc=0, pc=0x2b)
at vm/interp.c:649
acc = 8
pc = (int_val *) 0xbc0018
sp = (int_val *) 0xa8eb04
csp = (int_val *) 0x0
vm = (neko_vm *) 0x65677324
m = (neko_module *) 0x2b
sp = (int_val *) 0x77c40e66
csp = (int_val *) 0x0
trap = (int_val *) 0xa8eb04
init_sp = 1651864435
old_ncalls = 35
old = {1818386724, 29801, 0, 35, 1651864417, 0, 0, 43, 1970495780, 98,
0, 43, 2053731169, 101, 0, 51}
#1 0x77c5fcc0 in gc_dll_iname ()
No symbol table info available.
(gdb)
e
--
Neko : One VM to run them all
(http://nekovm.org)
