[Bug other/63504] [5 Regression] Issues found by --enable-checking=valgrind

["ccoutant at google dot com" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63504

--- Comment #14 from ccoutant at google dot com ---
> But then there is (question mainly on Cary) the .debug_types checksumming:
>
>     case dw_val_class_const_double:
>       CHECKSUM_ULEB128 (DW_FORM_block);
>       CHECKSUM_ULEB128 (sizeof (at->dw_attr_val.v.val_double));
>       CHECKSUM (at->dw_attr_val.v.val_double);
>       break;
>
>     case dw_val_class_wide_int:
>       CHECKSUM_ULEB128 (DW_FORM_block);
>       CHECKSUM_ULEB128 (sizeof (*at->dw_attr_val.v.val_wide));
>       CHECKSUM (*at->dw_attr_val.v.val_wide);
>       break;
>
> and that one ought to be governed by the DWARF standard, so I wonder how it can
> get away without converting for endianity etc.

There shouldn't be any location lists in .debug_types, so I don't
think the DWARF standard really matters for this case, and it would be
reasonable to use the same hashing you use for other purposes. Some
may slip through, unfortunately, so I covered this case just for
completeness.

-cary