|
From: Frank K. <fbk...@ve...> - 2007-07-12 09:04:22
|
Evenbit wrote:
> On Jul 11, 8:55 pm, Frank Kotler <fbk...@ve...> wrote:
>
>>strace ./luxasm.dbg 2> lux.tra gives me:
>>
>>execve("./luxasm.dbg", ["luxasm.dbg"], [/* 36 vars */]) = 0
>>brk(0) = 0x806186c
>>brk(0x8361880) = 0x8361880
>>write(1, "luxasm.dbg", 10) = 10
>>write(1, " [options]\n\t-i <name> Fi"..., 191) = 191
>>_exit(1) = ?
>>
>>Can you try that and let me know where it stops?
>
>
> execve("./luxasm.dbg", ["./luxasm.dbg"], [/* 36 vars */]) = 0
> --- SIGSEGV (Segmentation fault) @ 0 (0) ---
> +++ killed by SIGSEGV +++
Oh, my! Very early...
procedure start, n, v, e
locals counter, name
mov eax, [ local v ] ; Initialise subsystems
mov eax, [ eax ]
cld
mov [ local name ], eax
mov [ luxasm_command ], eax
call Memory.initialise
The "brk(0)" occurs in Memory.initialise, so we're not even getting that
far. I suspect "mov eax, [eax]", which suggests we're not finding
"argv[0]" - the program name. A wild guess, without being able to
duplicate it. If "they" have changed the startup stack from one version
to another, we're totally gefukt. Oh, my!
In spite of the fact that LuxAsm is generally considered to be "dead",
*I* consider this valuable feedback (if not entirely welcome news).
Thanks, Nathan!
Best,
Frank
|