Experimental Physics and
Industrial Control System

Michael Davidsaver <[email protected]> · Tue, 18 Oct 2011 11:09:06 -0400

I've stumbled onto an interesting behavior of aSub. I want to write a

function which takes a CHAR array as input (a long string). Sometimes

the string can legitimately be empty (length 0). What I find is that

when the the input should go from non-empty to empty it does not, and

instead keeps the previous value.

In aSubRecord.c I see the following

for (i = 0; i < NUM_ARGS; i++) {
        long nRequest = (&prec->noa)[i];

 status = dbGetLink(&(&prec->inpa)[i], (&prec->fta)[i],

(&prec->a)[i], 0,

&nRequest);
        if (nRequest > 0)
            (&prec->nea)[i] = nRequest;
        if (status)
            return status;
    }

http://bazaar.launchpad.net/~epics-core/epics-base/3.15/view/head:/src/std/rec/aSubRecord.c#L321

I think what I am seeing is due to the "if (nRequest > 0)" test.

Is it ever the case that status==0 and nRequest==0 is used to signal an

error? If not then can this be changed? Perhaps to:

for (i = 0; i < NUM_ARGS; i++) {
        long nRequest = (&prec->noa)[i];

 status = dbGetLink(&(&prec->inpa)[i], (&prec->fta)[i],

(&prec->a)[i], 0,

&nRequest);
        if (status)
            return status;
        else
            (&prec->nea)[i] = nRequest;
    }

It looks like subArray does the right thing. However, soft device

support for waveform, and aai don't. They actually ignore status

completely!

http://bazaar.launchpad.net/~epics-core/epics-base/3.15/view/head:/src/std/dev/devWfSoft.c#L72

Regards,
Michael

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System