[Bug debug/19406] [4.0/4.1 Regression] ICE: in force_decl_die, at dwarf2out.c:12442

[jakub at gcc dot gnu dot org]

------- Additional Comments From jakub at gcc dot gnu dot org  2005-03-24 21:40 -------
It seems rth's review got lost in the archives :(, so I'll cut'n'paste
from my mailbox:

> 2005-01-18  Jakub Jelinek  <jakub@redhat.com>
>
>       PR c++/19406
>       * dwarf2out.c (gen_type_die_for_member): Handle FIELD_DECL.
>       (dwarf2out_imported_module_or_decl): Use gen_type_die_for_member
>       for FIELD_DECLs.
>
>       * g++.dg/debug/using1.C: New test.

I don't think this is the right approach.

Presumably we're suppressing debug for this class because it is "unused"?
If so, I can't think except that's a bug.  In this test case, A is used
twice -- once through the using and once through the inheritence.

Basically, I think it's wrong to try to emit just one field of a class.
If we need a field for some reason, then we need the entire class to put
it in proper context.

--

Jason replied to this that the compiler is already doing something similar,
see http://gcc.gnu.org/ml/gcc-patches/2005-02/msg00170.html
Daniel was wondering if gdb can handle it and I have tested that gdb
indeed handles this: http://gcc.gnu.org/ml/gcc-patches/2005-02/msg01230.html

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jason at redhat dot com, rth
                   |                            |at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19406