Re: Rust's temporal safety for Ada/SPARK

[Dennis Lee Bieber <wlfraed@ix.netcom.com>]

On Sun, 14 May 2017 22:59:55 +0300, Niklas Holsti
<niklas.holsti@tidorum.invalid> declaimed the following:

>Even in applications where heap allocation is forbidden, there are 
>usually some dynamically allocated resources -- elements of "resource 
>pools" such as message buffers -- with the corresponding 
>application-defined "reference" data types, and the same problems of 
>managing allocations over time. I don't know if Rust's memory-management 
>scheme extends to such non-heap "references, however.

	My last employment system environment came down to allocation only
during the initialization of the application, no allocation permitted once
the application transitioned to "running", an NO DEALLOCATION ALLOWED.

	That left only two potential erroneous situations: duplicating a
reference such that two parts of the code were independently accessing the
same memory, and the converse, "dropping" a reference such that the memory
could no longer be accessed. The latter being the less likely to get out
into the wild, as running out of a buffers from a pool would tend to be
found in testing, given the limited number of buffers available.
-- 
	Wulfraed                 Dennis Lee Bieber         AF6VN
    wlfraed@ix.netcom.com    HTTP://wlfraed.home.netcom.com/