[Bug ada/124226] [13/14/15/16 regression] missing implicit dereference for access-to-protected used as prefix

[ebotcazou at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124226

--- Comment #18 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
> I agree that WONTFIX isn't a good resolution (and this one has been fixed
> as I can see), but there are for example a lot of ice-on-invalid-code,
> error-recovery "fuzzed" bugs that are NEW (aka confirmed) but will likely
> never be fixed due to resource issues.  They do clutter the bug database as
> well, so it's also not good to keep such bugs open for an infinite amount of
> time (bugs need attention from time to time, if only to re-confirm they
> still exist).

Yes but, more importantly, almost any change made to the compiler may
potentially introduce new issues so, by fuzzing the compiler, you can create an
avalanche effect that will soon be out of control.