https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62194
--- Comment #3 from Josh Triplett <josh at joshtriplett dot org> --- (In reply to Josh Triplett from comment #2) > One alternative implementation: if GCC supported a "property" attribute, > specifying (optional) functions to get or set the property (with > compile-time errors if attempting to get/set a property for which the > corresponding function does not exist), that would completely address this > issue as well: just define a no-op setter and either a constant or > non-existent getter. One issue with this approach: it would require the property mechanism to handle initializers somehow. So, on second thought, I think "deadfield" is still the right way to go.
