Re: null records and box

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

"Georg Bauhaus" <bauhaus@futureapps.de> wrote in message
news:44942492$0$11080$9b4e6d93@newsread4.arcor-online.net...
> John Barnes writes about a null record type, "The aggregate
> for the (null) value ... can be written as (null record)
> or even as (others => <>)." (Barnes 2006, p.146)
>
> Is the latter part is now rubbed out by LRM 4.3.1(15),
> so that (null record) is the one way to express a null
> record aggregate?
>
> (GNAT would be right, then.)

That looks like a bug in Ada 2005. It certainly was my intent that you could
write "(others => <>)" for a null record (I rewrote this text for a number
of reasons in the later stages of the Amendment work; I don't ever recall
thinking that 4.3.1(15) prevented that). I think the "null record" syntax is
a mistake, in that it doesn't look like the syntax you would expect for an
aggregate.

Anyway, I'll check with the ARG (just because it was *my* intent -- and one
tht John also expected -- doesn't mean that it was really intended).

                                 Randy.