bug#18769: 24.4; assertion failed in emacs-24.4 on cygwin x86_64

[Eli Zaretskii]

> From: Jon TURNEY <jon.turney@dronecode.org.uk>
> Date: Sun, 19 Oct 2014 16:37:44 +0100
> 
> AS discussed at https://cygwin.com/ml/cygwin/2014-10/msg00233.html, I've
> been getting assertion failures with emacs on cygwin x86_64.
> 
> I've seen asserts in a few different places. These always seem to occur
> while I am doing something in emacs, never when it is idle.
> 
> Output of 'bt full' and 'xbacktrace' attached.
> 
> [2:text/plain Hide]
> 
> #0  die (msg=0x100a88cd4 <DEFAULT_REHASH_SIZE+124> "OBJFWDP (a)", 
> file=0x100a88c60 <DEFAULT_REHASH_SIZE+8> 
> "/home/kbrown/src/cygemacs/emacs-24.4-0.x86_64/src/emacs-24.4/src/data.c", 
> line=136) at /usr/src/debug/emacs-24.4-0/src/alloc.c:6838
> No locals.
> #1  0x00000001005f61c9 in XOBJFWD (a=0x100b24ed0 <o_fwd>) at 
> /usr/src/debug/emacs-24.4-0/src/data.c:136
> No locals.
> #2  0x00000001005f7d62 in store_symval_forwarding (valcontents=0x100b24ed0 
> <o_fwd>, newval=25783361702, buf=0x0) at 
> /usr/src/debug/emacs-24.4-0/src/data.c:998
> 
> [...]
> 
> I tried to investigate a bit more, but I must be misunderstanding
> something, because I don't understand why this assertion fired:
> 
> (gdb) up
> #1  0x00000001005f61c9 in XOBJFWD (a=0x100b24ed0 <o_fwd>) at
> /usr/src/debug/emacs-24.4-0/src/data.c:136
> 136       eassert (OBJFWDP (a));
> (gdb) p a
> $10 = (union Lisp_Fwd *) 0x100b24ed0 <o_fwd>
> (gdb) p *a
> $11 = {
>   u_intfwd = {
>       type = Lisp_Fwd_Obj,
>           intvar = 0x100b2f5d8 <globals+184>
>             },
>   u_boolfwd = {
>        type = Lisp_Fwd_Obj,
>           boolvar = 0x100b2f5d8 <globals+184>
>               },
>   u_objfwd = {
>        type = Lisp_Fwd_Obj,
>           objvar = 0x100b2f5d8 <globals+184>
>              },
>   u_buffer_objfwd = {
>        type = Lisp_Fwd_Obj,
>           offset = 0,
>           predicate = 4306695640
>               },
>   u_kboard_objfwd = {
>             type = Lisp_Fwd_Obj,
>             offset = 0
>               }
>       }
> (gdb) p a->u_intfwd.type
> $12 = Lisp_Fwd_Obj

Could be one more of those unexplained assertion violations, see bug
#18438.  And if that's so, then the problem isn't limited to the
Cygwin-w32 builds.