Re: Intervention needed?

["Randy Brukardt" <randy@rrsoftware.com>]

"Florian Weimer" <fw@deneb.enyo.de> wrote in message 
news:87zhpdgvhl.fsf@mid.deneb.enyo.de...
>* Randy Brukardt:
>
>>>"Florian Weimer" <fw@deneb.enyo.de> wrote in message
>>>news:87sgv5k1j5.fsf@mid.deneb.enyo.de...
>>>* Randy Brukardt:
>>>
>>>> But a better question is whether the Rust borrow checker allows
>>>> building proper ADTs for most data structures.
>>>
>>> The Rust standard library uses unsafe extensively in the
>>> implementation of containers.  I'm not sure if this is a problem.
>>> Most languages have this property (that the standard library cannot be
>>> implemented in the language itself).
>>>
>>> For container libraries, the only widely-used counter-example (of
>>> which I'm aware) is pre-generic Java.
>>
>> You're saying Ada isn't widely used? And I thought the STL was 
>> implemented
>> in C++.
>
> The GNAT implementation doesn't seem particularly memory-safe to me,
> looking at Ada.Containers.Bounded_Ordered_Maps (which picked more or
> less at random).

You said "Most languages have this property (that the standard library 
cannot be implemented in the language itself)." You didn't say anything 
about memory safety -- and it wouldn't make sense anyway, since most 
languages don't have "memory safety" (whatever that is) in the first place.

For Ada, the entire idea of the containers was to allow implementations to 
have as much (or as little) memory safety as they want. There's none that an 
implementation can provide with access types.

                               Randy.