Experimental Physics and
Industrial Control System

"Luchini, Kristi L. via Tech-talk" <tech-talk at aps.anl.gov> · Mon, 6 Mar 2023 16:31:56 +0000

Hi Michel,

I loaded you're test database below:
record(longout, "lo1") {
    field(DOL, "")
    field(VAL, "42")
}
record(longout, "lo2") {
    field(VAL, "42")
}
record(longout, "lo3") {
    field(DOL, "42")
}

Here's what I see:

- Kristi
-----Original Mess [luchini@lcls-dev3 ~]$ caget lo1
lo1                            0
[luchini@lcls-dev3 ~]$ caget lo2
lo2                            42
[luchini@lcls-dev3 ~]$ caget lo3
lo3                            42

-Kristi

From: Michael Davidsaver <mdavidsaver at gmail.com> 
Sent: Monday, March 6, 2023 8:14 AM
To: Luchini, Kristi L. <luchini at slac.stanford.edu>
Cc: tech-talk at aps.anl.gov
Subject: Re: base R7.0.3.1-1.0 db field DOL="" different behavior then in R3.14.12

On 3/3/23 23:38, Luchini, Kristi L. via Tech-talk wrote:
> Hello,
> 
> I’m upgrading an EPICS IOC Application from epics base R3.14.12 to R7.0.3.1-1.0, and I’ve noticed a different behavior on boot regarding the DOL field of an output record. For example for an aoRecord,  in R3.14.12 if you set DOL=”” on boot the VAL field will be restored.   If,  DOL is a link then the value of that PV will be restored in the ao.   However, in R7.0.3.1-1.0 , if DOL=”” a 0 is restored, independent of the VAL field.   Only if I don’t specify the DOL field in the db,  will the VAL field be restored on boot.

> Is there a reason this behavior has changed?  To me the R3.14.12 behavior seems to be correct.

If I have correctly understand the situation you describe, then this change was not intentional.

Can you see if this change results in the behavior you expect?

https://github.com/epics-base/epics-base/pull/335

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System