Re: longout record issue

[Dirk Zimoch <dirk.zimoch@psi.ch>]

Hi Emmanuel,

That looks like the problem I have reported in 2007 (and which is still "open" 
in Mantis):

Summary  	 gcc 3.2.3 bug: long integer values are corruped in R3.14.9

Description 	When writing integer values with a CA client (e.g. medm) to a long 
integer channel (e.g. longout record), the upper bytes get corrupted. The lower 
2 bytes are OK.
I can't see the error when the client is caput from base.
Client and IOC run on a linux-x86 machine.
The problem is new in 3.14.9.

See http://www.aps.anl.gov/epics/mantis/view_bug_page.php?f_id=0000291

It happens when using g++ 3.2.3 with option -O3.

Dirk


Emmanuel Mayssat wrote:
> Hello all,
> I have noticed oddities with my longout records.
> Particularly when I use the soft Channel record.
>
> record(longout, "myLO") {
> 	field(VAL ,"100")
> 	field(SCAN, "Passive")
> 	field(DTYP, "Soft Channel")
> }
>
> Now if I create a medm text entry attached to this PV, and attempt to
> modify the PV, it updates but not as expected. For example, if I enter
> the value 10, the PV is updated to -11100023 (or something like that)
>
> I am using medm 3.1.1 compiled against EPICS 3.14.9.
>
> I have other longout records which are not "Soft Channel".
> They correctly update with the same medm widget.
>
> Any idea?
> --
> Emmanuel

--
Dr. Dirk Zimoch
Paul Scherrer Institut, WBGB/006
5232 Villigen PSI, Switzerland
Phone +41 56 310 5182