https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68010
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Priority|P3 |P2
Last reconfirmed|2021-08-13 00:00:00 |2024-10-14
--- Comment #13 from Richard Biener <rguenth at gcc dot gnu.org> ---
The BINFO is allocated from
#0 0x00000000012abfdf in copy_binfo (binfo=<tree 0x0>,
type=<record_type 0x7ffff69b82a0 C>,
t=<record_type 0x7ffff69b8f18 StructDecoder>, igo_prev=0x7fffffffd270,
virt=0) at ../../src/gcc/gcc/cp/tree.cc:2090
#1 0x0000000000fc5650 in xref_basetypes (
ref=<record_type 0x7ffff69b8f18 StructDecoder>,
base_list=<tree_list 0x7ffff69d1028>) at ../../src/gcc/gcc/cp/decl.cc:17048
#2 0x000000000114b7de in cp_parser_class_head (parser=0x7ffff69b8b28,
nested_name_specifier_p=0x7fffffffd5a7)
at ../../src/gcc/gcc/cp/parser.cc:28045
#3 0x0000000001148db3 in cp_parser_class_specifier (parser=0x7ffff69b8b28)
at ../../src/gcc/gcc/cp/parser.cc:27094
and ->binfo.offset is never initialized and thus left NULL_TREE.
But dwarf2out.cc definitely expects something there.
It's easy to plug the "hole" here, for example by doing nothing for this
BINFO in add_data_member_location_attribute. But I have no idea what's
expected here in the end.
The inhertitance DIE produced sofar is
DIE 0: DW_TAG_inheritance (0x7ffff68334b0)
abbrev id: 0 offset: 0 mark: 0
DW_AT_type: die -> 0 (0x7ffff6833500)
Jason?
