Re: sequencer programs in an IOC

[James F Ross <james.falconer.ross@gmail.com>]

Thanks guys,
Yes, the c code is not the problem.  There was a communication issue that I knew was causing problems.  That is fixed now, and I am getting a few of the desired html files written, but for some reason the main file is not being written.  It seems to be going running the program relating to the eemc (one of the early ones called by a seq command in the st.cmd file) but never moves on to the program after it.  There is no indication of error, but I'm looking at both the eemc source code and the bemc code (the next program called in the st.cmd file).

James

On Tue, Jan 22, 2013 at 3:19 PM, Mark Rivers <rivers@cars.uchicago.edu> wrote:

I would go a step further and say that the C code should be "transparent".  It is created in the O.$(ARCH) directory from the .stt code, and is then immediately compiler to an object file to be linked into your application.

The startup script calls a function that is defined in a dbd file via an entry like:
registrar(dxpMEDRegistrar)

That accesses the SNL program, dxpMED.st, whose definition is like the following:

program dxpMED("P=13GE2:med:,DXP=dxp,MCA=mca,N_DETECTORS=16,N_SCAS=16")

The O.linux-x86 directory for that module contains the following files:
corvette:dxp/dxpApp/src>ls -l O.linux-x86
total 19552
...
-rw-rw-r-- 1 epics epics  2043540 Dec 14 19:00 dxpMED.c
-rw-rw-r-- 1 epics epics     1004 Dec 14 19:00 dxpMED.d
-rw-rw-r-- 1 epics epics    97303 Dec 14 19:00 dxpMED.i
-rw-rw-r-- 1 epics epics  1805352 Dec 14 19:00 dxpMED.o

Mark

-----Original Message-----
From: tech-talk-bounces@aps.anl.gov [mailto:tech-talk-bounces@aps.anl.gov] On Behalf Of Jiro Fujita
Sent: Tuesday, January 22, 2013 2:10 PM
To: James F Ross
Cc: EPICS Tech Talk
Subject: Re: sequencer programs in an IOC

James,

In short, when you write a program in state notation language (that you use for EPICS sequencer) and make the IOC, it first gets covered to C code via SNC (SNL to C compiler).  Then, the resulting C code get compiled by GCC.  In general, you should not have to do much with the C code; certainly, anything we use at STAR soft IOC, we should not have to do much with it.


On Jan 22, 2013, at 9:56 AM, James F Ross wrote:

> Hello everyone,
> I'm trying to understand an IOC that generates a controls monitoring webpage at STAR.  For whatever reason, the IOC is not writing the html file that it should be (it has worked in the past).
>
> The source code is a number of .stt files, each of which handles variables for a different subsystem.  Ultimately all of these variables are brought together using a file called overall.stt, and the webpage is written by a file called overallhtml.stt.
>
> When the IOC is compiled, these .stt files produce .c files of the same name.  What I'm wondering is, how are these c files used?  Do they become binary files when the IOC is compiled?  What generally happens in an IOC between these c files and a sequencer program being called in the start up script for the IOC?
>
> The sequencer programs are loaded as follows in the IOC startup file:
> seq mot_gas
> seq trig
> seq eemc
> seq bemc
> seq tofm
> seq overall
> seq overallhtml
>
> Are these binary executable files that are being called, or does the startup script read the c code directly?  I can't find any binary files by these names on the computer, but I see no errors while compiling either.  If anyone has any insight I could use a hand.  I know you won't know specifics on my setup, but any help with sequencer programs in general would be nice.
>
> Thanks very much!
> James Ross
>