!standard  D.15(24/2)                             10-04-24    AI05-0210-1/01
!class binding interpretation 10-04-24
!status work item 10-04-24
!status received 10-03-25
!priority Low
!difficulty Easy
!qualifier Error
!subject 
!summary 

The metric for Timing_Events only need consider the case where a single handler
attempts to execute at the time set.

!question

Should the metric D.15 (23/2) explicitly say that only one handler should execute
at the time set? (Yes.)

!recommendation

(See Summary.)

!wording

Modify D.15(24/2):

* An upper bound on the lateness of the execution of a handler. That is, the maximum
  time between when a handler is actually executed and the time specified when the event
  was set{, assuming no other handler executes at this time}.


!discussion



!corrigendum D.15(24/2)

@drepl
An upper bound on the lateness of the execution of a handler. That is, the
maximum time between when a handler is actually executed and the time specified
when the event was set.
@dby
An upper bound on the lateness of the execution of a handler. That is, the
maximum time between when a handler is actually executed and the time specified
when the event was set, assuming no other handler executes at this time.


!ACATS Test

Metrics are not testable.

!appendix

From: Randy Brukardt
Sent: Tuesday, February 23, 2010  5:22 PM

I'm going through AIs creating a spreadsheet of information so we can see the Big
Picture.

I happened to notice the following in the discussion of AI-20:

This latter case can't happen for fixed-point types, because use of an untagged
incomplete type as a non-access parameter is prohibited:
   function "*" (Left, Right : Incomp); -- Illegal!

Whomever wrote that never read AI-151 (probably because it hadn't been written yet).
Looks like we'll need another AI to fix the wording.

****************************************************************