"J-P. Rosen" <rosen@adalog.fr> wrote in message 
news:ou8lod$3iq$3@dont-email.me...
> Le 12/11/2017 à 05:33, Robert Eachus a écrit :
>> On Monday, November 6, 2017 at 1:31:07 PM UTC-5, Simon Wright wrote:
>>> Robert Eachus <rieachus@comcast.net> writes:
>>>
>>>>    entry Write (D: in Data) is
>>>>    begin
>>>>       The_Data := D;
>>>>       Some_Data := True;
>>>>       Print(D);
>>>>       delay 0.1;
>>>>    end Write;
>>>
>> Having said that, the delay 0.1 is very naughty, but on most hardware it 
>> will be implemented as a counting loop.  (Make it delay 0.0001 if that 
>> makes you happier. ;-)   It is not illegal, and is only erroneous if it 
>> results in a scheduling action.
>>
> Delay is potentially blocking operation, so this is definitely a bounded
> error. But it is allowed for a bounded error situation to give the
> proper result...

Unless pragma Detect_Blocking is given (which should have been the default 
in Ada, IMHO -- predictable is always better than "might work -- or not"). 
In that case, raising Program_Error is required.

If the new aspect Nonblocking is used, the delay is statically illegal. 
(That *really* should have been the default, but it took us 20 years to 
figure out the right set of rules -- too late to mandate.)

                              Randy.