[Bug ada/72740] gnat.dg/specs/access[12].ads ICE when compiling with -g

["rguenther at suse dot de" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=72740

--- Comment #2 from rguenther at suse dot de <rguenther at suse dot de> ---
On Wed, 10 Aug 2016, ebotcazou at gcc dot gnu.org wrote:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=72740
> 
> Eric Botcazou <ebotcazou at gcc dot gnu.org> changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>              Status|UNCONFIRMED                 |NEW
>    Last reconfirmed|                            |2016-08-10
>                  CC|                            |ebotcazou at gcc dot gnu.org
>      Ever confirmed|0                           |1
> 
> --- Comment #1 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
> > the issue is that we have recursive pointer types and blow the stack when
> > dwarf2out calls verify_type which calls variably_modified_type_p:
> > 
> > (gdb) p type->typed.type->typed.type
> > $3 = <pointer_type 0x7ffff6525a80 access2__inc>
> > (gdb) p type->typed.type->typed.type->typed.type
> > $4 = <pointer_type 0x7ffff6525888 access2__priv>
> > (gdb) p type->typed.type->typed.type->typed.type->typed.type
> > $5 = <pointer_type 0x7ffff6525a80 access2__inc>
> > 
> > not sure how to best represent such a structure.  Maybe
> > variably_modified_type_p needs adjustment?
> 
> Do you mean something similar to what Jan added in get_alias_set?

Yeah, or maybe mark the "forward" reference (the "backedge") tree type
in some way in the FE?  POINTER_TYPE_CYCLE_CLOSING_P?