Re: Are delay statements abortable?

[Tucker Taft <stt@averstar.com>]

davidb42@my-deja.com wrote:
> 
> Are delay statements interruptable/abortable?

Yes.

> 
> I was expecting delay statements in tasks (or in the abortable part of
> an asynchronous select construct) to be brought to a premature end on
> aborting the task (or on the triggering event).

They should be.
> 
> However, in the Ada implementation I am using, the abortion only seems
> to take effect at the expiry of the delay.

This would not qualify as a "real-time-systems-annex" supporting
implementation in Ada 95.  "Real-time" implementations are required
to abort preemptively in Ada 95 (see RM95 D.6(2)).
> 
> For example, in the program below, the main task completes after 25
> seconds, but the dependent task (and thus the program as a whole)
> doesn't terminate until 40 seconds (at expiry of the 2nd 20 second
> delay).
> 
> Is this right?

In a non-real-time implementation, this is "correct" but not particularly
helpful.  In a "real-time" Ada 95 implementation, this would be incorrect.

>         David Bowerman
>         CSC Australia.
-- 
-Tucker Taft   stt@averstar.com   http://www.averstar.com/~stt/
Technical Director, Distributed IT Solutions  (www.averstar.com/tools)
AverStar (formerly Intermetrics, Inc.)   Burlington, MA  USA