Last week tuesday we had a chance to bring up several APS Linac
IOCs under 3.13. Almost everything worked. We had just two hours
to try things. This was just before starting a scheduled 6 week
shutdown. The test gave Ned Arnold, the Application Developer
responsible for the Linac, confidence that he can convert to 3.13.
This is a major success since the Linac was the first and thus oldest
system at APS to start commissioning. Thus the special record
and device support was originally written several EPICS releases ago.
A few minor problems did arise that forced us to generated an internal
beta release. We are now starting work to convert the par (the next
APS accelerator after the linac) to 3.13. If things go well we hope
to convert some other APS/ASD systems during the current shutdown.
We should be generating a new beta release soon. At that time
updated release notes and updates to the Application Developer's
Guide will be available.
In addition to fixing minor problems two major problems are being
addressed for the next beta release.
The first is that GDCT does not work correctly when linked against
3.13. It partially works but often crashes badly. Several users
reported problems. When Jim Kowalkowski looked at the problem we
realized that there were still problems with the static database access
library. The main problem is that it did not provide for macro
definitions in menu or link fields. We are currently testing and are
close to having things work properly. In addition John Anderson has been
working on a tcl/tk replacement for the original DCT. Both a working
GDCT and the new tcl/tk DCT will be available for the next release.
The other major problem involves the UDF (undefined) field. The problem
was originally reported by Tim Mooney. Before 3.13 it was not possible
to distinguish between a link never given a value and a link given
a constant value. This forced some output records with DOL fields
to always set UDF false while initializing the record. Since we
can now distinguish between null and constant links the init code
was changed. The result is that some records could have UDF set true
forever. Janet Anderson has reviewed all record types and fixed
the problem. We also had to change dbPutString so that when it writes
to the VAL field it sets UDF FALSE.
So just when will we declare an official 3.13? My feeling now
is that we should wait until the end the current APS shutdown.
When we start up we will have several major APS/ASD subsystems
on 3.13. The EPICS developer's will do everyting they can to make
this a success including building new internal 3.13 releases
whenever problems arise. At the end of this process
we should have a stable 3.13 release.
I want to thank everyone who has started using the alpha and beta
releases. It is extremely important that we have testing at a variety
od sites. The EPICS developers take all reported bugs seriously!!
Marty Kraimer
- Navigate by Date:
- Prev:
EPICS Meeting Notes from CERN 9-96 Bob Dalesio
- Next:
FRC5CE task stack sizes Peregrine McGehee
- Index:
1994
1995
<1996>
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
- Navigate by Thread:
- Prev:
EPICS Meeting Notes from CERN 9-96 Bob Dalesio
- Next:
Re: R3.13 Update Marty Kraimer
- Index:
1994
1995
<1996>
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
|