https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118837
--- Comment #6 from Tom Tromey <tromey at gcc dot gnu.org> ---
I also looked through the LLVM DWARF writer.
In part it is a bit nicer because it does:
void DwarfUnit::addConstantValue(DIE &Die, bool Unsigned, uint64_t Val) {
// FIXME: This is a bit conservative/simple - it emits negative values always
// sign extended to 64 bits rather than minimizing the number of bytes.
addUInt(Die, dwarf::DW_AT_const_value,
Unsigned ? dwarf::DW_FORM_udata : dwarf::DW_FORM_sdata, Val);
}
That is, for this attribute at least, it doesn't try to compute
the narrowest possible form.
Looking at the other "confused" attributes:
if (Offset < 0)
addSInt(MemberDie, dwarf::DW_AT_bit_offset, dwarf::DW_FORM_sdata,
Offset);
else
addUInt(MemberDie, dwarf::DW_AT_bit_offset, std::nullopt,
(uint64_t)Offset);
.. i.e., LLVM does the same as GCC here.
And finally:
if (DD->getDwarfVersion() == 3)
addUInt(MemberDie, dwarf::DW_AT_data_member_location,
dwarf::DW_FORM_udata, OffsetInBytes);
else
addUInt(MemberDie, dwarf::DW_AT_data_member_location, std::nullopt,
OffsetInBytes);
... which is fine since it at least agrees with GCC that narrow
forms should not sign-extend.
