Re: Adacore joins with Ferrous Systems to support Rust

[Andreas ZEURCHER <ZUERCHER_Andreas@outlook.com>]

On Friday, February 4, 2022 at 4:28:26 AM UTC-6, Luke A. Guest wrote:
> On 04/02/2022 03:20, Randy Brukardt wrote: 
> 
> >> Just like we didn't touch generic packages until much later in the year, 
> >> generic subprograms, yeah, because u_d was required early on. 
> >> 
> >>> objecting to this, I think he was right (philosophically; I've no idea 
> >>> about the practicality of alternate approaches). 
> >> 
> >> How else would you do controlled types? 
> > 
> > Ada 9x originally had a bunch of magic attributes (similar to streaming). It
> Now I want to know what these magic attributes were! Were they specific 
> to a version of OO? Or were they to enable finalization?

Randy, I agree with Luke:  were these intermediate design proposals lost entirely or have they (as still extant) have simply not been released publicly?  I suspect that at least some of these attributes have nowadays analogues in C++ smart pointer's & Objective-C/Swift's ARC {strong, weak, plain old data not needing finalization, pointed-to-object ownership, presence of finalization subroutine/function/procedure a.k.a. finalizer/destructor, whether this finalizer in a subtype displaces its parent's finalizer versus this finalizer in a subtype chains its finalizer to all of its ancestors' finalizers unwound from most-derived to underived-root, … and so forth}.  Or was Tucker's set of magic attributes going an entirely different direction?  That intermediate proposal under consideration back in the 1st half of the 1990s might be a quite interesting read (especially by a reader with an interest in envisioning an Ada-esque analogue of Rust's borrow-checker algorithm).