Re: Implementing Rust's borrow checked pointers

[Optikos <optikos@verizon.net>]

On Tuesday, September 24, 2019 at 2:13:33 PM UTC-5, Keith Thompson wrote:
> Lucretia <laguest9000@googlemail.com> writes:
> > On Tuesday, 24 September 2019 12:23:30 UTC+1, Optikos  wrote:
> >
> >> > begin
> >> >    A.all ... ; -- raises exception.
> >> 
> >> No, to be as useful as Rust's borrow checker, instead of raising
> >> exception, it needs to be a compile-time error.  The compiler needs
> >> to maintain a whole-program directed graph at compile-time, not defer
> >> a detection-based localized analysis to run-time.
> >
> > Yes, the compiler would raise that exception at compile time. This
> > idea that all exceptions are raised at runtime is false and you should
> > check the AARM.
> 
> Can you provide a specific citation?
> 
> Certainly a compiler can diagnose an error at compile time, but I've
> never heard that referred to as an "exception".  And a compiler can
> generate code that unconditionally raises an exception, but that code is
> executed at run time.
> 
> (I've seen compile-time exceptions, but they were compiler bugs, not
> diagnostics for the program being compiled.)
> 
> -- 
> Keith Thompson (The_Other_Keith) kst-u@mib.org  <http://www.ghoti.net/~kst>
> Will write code for food.
> void Void(void) { Void(); } /* The recursive call of the void */

Keith, I already did in a prior posting above, but you apparently blocked me, so your loss.  Here is the answer to your question again for all other readers:  §1.1.5 Classification of •Errors “If such an •error• situation is certain to arise in every execution of a construct, then an implementation is •••allowed (although not required)••• to report this fact at •compilation• time.”