Experimental Physics and
Industrial Control System

Eric Norum <[email protected]> · Thu, 28 Feb 2013 09:44:42 -0800

I've never been brave enough to hack code into the waveform record support.  Instead I FLNK the waveform record to a longin record which gets its value from the waveform record NORD field and posts monitors as required.

On Feb 28, 2013, at 9:34 AM, Andrew Johnson <[email protected]> wrote:

> Hi Gary,
> 
> On 2013-02-28 Carr, Gary wrote:
>> I have an application that needs to monitor changes in the NORD field of a
>> waveform record. Since we are running Channel Access 3.14.11, I cannot get
>> a monitor on the field, so I use a simple "get" to poll the field in Java.
>> Basically:
>> 
>> Channel ic = context.createChannel("pvname.NORD");
>> context.pendIO(timeout);
>> ...
>> int val = ((INT) ic.get(DBRType.INT,1)).getValue();
>> context.pendIO(timeout);
>> 
>> Using jca context this works fine either reading thechannel directly, or
>> through a CA gateway. If I use a caj context in the same code, it works
>> fine reading the channel directly, but if I read through a CA gateway I
>> never see the NORD field change when the waveform processes.
> 
> That's because the gateway sets up a monitor on the channel and uses that 
> value when you do a ca_get from it.  It does that to try and reduce the load 
> on the IOCs if lots of scripts do ca_get operations on random channels.
> 
> There may be a flag you can give the gateway to prevent that behavior (but it 
> will increase the load on your IOCs), or you could patch the waveform record 
> in Base to make it post monitors on NORD — we've done that in the 3.15 branch, 
> and it's just a one-line change (line-numbers may vary):
> 
> === modified file 'src/std/rec/waveformRecord.c'
> --- src/std/rec/waveformRecord.c        2010-10-05 19:27:37 +0000
> +++ src/std/rec/waveformRecord.c        2012-02-17 23:32:25 +0000
> @@ -191,6 +191,7 @@
>     if (prec->nord > prec->nelm)
>         prec->nord = prec->nelm;
> 
> +    db_post_events(prec, &prec->nord, DBE_VALUE | DBE_LOG);
>     return 0;
> }
> 
> 
> - Andrew
> -- 
> There is no such thing as a free lunch.  When invited for lunch,
> it is best to check if you are there to eat, or to be eaten.
> -- Clive Robinson
> 

-- 
Eric Norum
[email protected]

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System