On Fri, 3 Jul 2015 14:42:31 -0500, George Kozakos wrote:

>> In my time zone, the leap second occurred at 17:59:60.  So, I wonder
>> about the z/OS STIMER macro:
>>
>> If, at 17:59:59 I had issued STIMER WAIT,LT=[18:000:01]
>>     would the wait have expired in 3 seconds?
>
>Yes, if you scheduled the leap second via STP or ETR.
>It would expire in 2 seconds if you aren't using leap seconds
> 
OK.  I imagine that there's a queue of Clock Comparator Register
values, and that Leap Second processing scans that queue and
adds one second to each value.

>> If, at 17:59:59 I had issued STIMER WAIT,BINTVL=[3 seconds]
>>     would the wait have expired at 18:00:01?
>
>No, it would expire at 18:00:02 regardless of whether you are using
>leap seconds
> 
And since the Leap Second code is ignorant of whether an entry
corresponds to a clock time or a duration, it assumes the former
and adds the second regardless.  Particularly silly if the wait
spans midnight since the syntax of STIMER macro provides no
way to specify a time past midnight, so it must have been a
duration.

So it actually waits for 4 seconds rather than the 3 requested.

>> Unless the answer to both questions is "Yes," there's a bug.
>> 
There's a bug.

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to