From: ts <decoux@...>
Date: 2004-08-29T00:36:02+09:00
Subject: Re: [PATCH] Subtle bug in bignum.c

>>>>> "t" == ts  <decoux@moulon.inra.fr> writes:

t> 1663        z = bignew(l2, RBIGNUM(x)->sign || RBIGNUM(y)->sign);
t> (gdb) info register ebx
t> ebx            0x8117700        135362304
t> (gdb) 

t>  it will call bignew() which will call rb_gc(), but in ebx it has
t>  RBIGNUM(x)->digits. This mean that when ruby will mark the registers it
    ^^^^^^^^^^^^^^^^^^
t>  will not mark the Bignum created by rb_big_clone() and the GC remove the
t>  Bignum created. 

 In reality this is RBIGNUM(y)->digits which is is in ebx (the result is
 the same)

(gdb) p *(struct RBignum *)y
$13 = {basic = {flags = 13, klass = 1074419244}, sign = 1 '\001', len = 1, 
  digits = 0x8117700}
(gdb) 


Guy Decoux