Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Wed, 11 Jul 2007 10:22:07 -0500

Hi Ralph,

Ralph Lange wrote:

http://www.aps.anl.gov/epics/wiki/index.php/How_to_Set_Up_a_Soft_IOC_Framework_on_Linux#Make_the_Soft_IOC_Create_the_PID_File

shows the code I added to my soft IOCs.

Does anyone see any possible side-effects? Would that code be

Windows-proof? Should I integrate it into the example stuff in base to

make it standard?

Good idea, but why do this in main()? The nice thing about having a

startup script is that all ioc-specific configuration can all be put

into the st.cmd file. I don't know whether windows has a getpid()

command, but this is obviously not useful on RTEMS which also uses a

main() routine.

I would just register a command that generates the PID file at a

location given in an argument; since environment variables are available

as macros you can still use $(PIDFILE) if you want. The command should

also register an epicsAtExit() routine to delete the file at shutdown.

The only disadvantage of creating the PID file from st.cmd rather than

in main() is that it won't get created quite as early in the startup

sequence, but I doubt that the difference really matters.

I don't have any major objections to something like a 'mkpid' command

going into base somewhere, but it is more of a site-specific thing; in

our case at APS I would probably include it in our asdstd support

application which all our "standard" controls IOCs must include.

- Andrew
--
The right to be heard does not automatically include
the right to be taken seriously. -- Hubert H. Humphrey

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System