There have been a few discussions over the years on IBM-MAIN as to
how to SLEEP for some period of time in REXX. I was asked today to look
at a dump of a "hung testcase" , and came across the following example
of how *not* to do it:
/* REXX */
/* WAIT 10 SECONDS FOR THE MESSAGE TO COMPLETE *
On Tue, 14 Jan 2014 16:12:31 -0500, Jim Mulder wrote:
> There have been a few discussions over the years on IBM-MAIN as to
>how to SLEEP for some period of time in REXX. I was asked today to look
>at a dump of a "hung testcase" , and came across the following example
>of how *not* to do it:
>
>
or if you really do have CPU cycles going spare, at least use TIME('E') !
On 14 January 2014 21:42, Paul Gilmartin wrote:
> On Tue, 14 Jan 2014 16:12:31 -0500, Jim Mulder wrote:
>
> > There have been a few discussions over the years on IBM-MAIN as to
> >how to SLEEP for some period of time in
>> address SYSCALL 'sleep 10'
You need a CALL SYSCALLS("ON")
first don't you?
Hope I spelled it correctly.
-
Ted MacNEIL
eamacn...@yahoo.ca
Twitter: @TedMacNEIL
--
For IBM-MAIN subscribe / signoff / archive access instructio
On 2014-01-14 15:54, Ted MacNEIL wrote:
>>> address SYSCALL 'sleep 10'
>
> You need a CALL SYSCALLS("ON")
> first don't you?
>
> Hope I spelled it correctly.
>
It's mostly automatic, but it does serve to define
about four hundred Rexx variables, which are unnecessary
for something as simple
I finally caved and stopped using the CPU hungry wait techniques.
The ADDRESS SYSCALL SLEEP works fine for whole seconds. We had a
requirement to wait quite a bit less, so I developed a called assembler
program that does a STIMER WAIT,DINTVL=DECTIME where DECTIME is CL8
'0500' for a 5 second w