Experimental Physics and
Industrial Control System

"Zelazny, Michael Stanley" <[email protected]> · Wed, 11 Sep 2013 08:23:04 -0700

FWIW I have a machine physicist who wants to define the TSEL on the fly, I would like to use autosave/restore, but due to this behavior, I can't.  the TSEL doesn't restore properly.

________________________________________
From: [email protected] [[email protected]] On Behalf Of Ralph Lange [[email protected]]
Sent: Wednesday, September 11, 2013 7:30 AM
To: EPICS Tech-Talk
Subject: Re: TSEL question

On 11.09.2013 16:11, Ralph Lange wrote:
> On 11.09.2013 16:08, Ralph Lange wrote:
>> On 11.09.2013 15:55, Zelazny, Michael Stanley wrote:
>>> Q: Why does the output of caget list ANOTHER:PV.VAL when I set it to
>>> ANOTHER:PV.TIME?
>>
>> Because the .TIME field is not accessible by Channel Access. The
>> timestamp can only be obtained by using a DBR type that includes the
>> timestamp on any valid CA-accessible field of a record.
>
> And this is what the record does. It sets up a link to the VAL field,
> but uses the time stamp it gets.

Using your example:

 > softIoc -d abc.db
Starting iocInit
############################################################################
## EPICS R3.14.12.3-6 $Date: Mon 2012-12-17 14:11:47 -0600$
## EPICS Base built May 22 2013
############################################################################
iocRun: All initialization complete
epics> dbgf MY:PV.TSEL
DBR_STRING:          "ANOTHER:PV.VAL NPP NMS"

I.e., the link in the database is set up pointing to the VAL field, the
db library hard-codes the time field in recGblGetTimeStamp().

Cheers,
~Ralph

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System