Experimental Physics and
Industrial Control System

Mark Rivers <[email protected]> · Tue, 13 Mar 2012 22:13:57 +0000

Thanks to Eric for pointing this out, I was not aware of it.

I have just made 2 changes to the asyn module to enable I/O redirection using both the "dbior" and "asynReport" commands.

I tested it with an areaDetector driver and it seems to work fine:

epics> dbior > dbior.out
epics> asynReport > asynReport.out

This is now committed to the Subversion repository, where it can be accessed from the Web, and it will be in the R4-19 release.

The asynTrace command already has a more powerful I/O redirection capability:

asynSetTraceFile portName addr filename

This allows one to redirect the trace (debugging) output at the level of a specific device, i.e. a single motor axis, etc.

Mark

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Eric Norum
Sent: Tuesday, March 13, 2012 4:23 PM
To: Allison, Stephanie
Cc: EPICS Tech Talk
Subject: Re: iocshCmd, redirection, and function pointers

For I/O redirection to work a source file must include epicsStdioRedirect.h.  
It's likely that lots of support modules and drivers have not yet done so.

On Mar 13, 2012, at 2:11 PM, Allison, Stephanie wrote:

> Hello,
> 
> When using redirection of output to a file in iocshCmd where the command is one that calls report functions via function pointer like dbior:
> 
> iocshCmd("dbior > /dat/dbiorOutput")
> 
> the output file contains only the output (printf) from the dbior routine itself but not output (printf) from each driver (called using a function pointer in drvet).
> 
> Is this a limitation of the iocsh, mixing C and C++, or the non-thread-safeness (as advertised in the app dev guide) of iocshCmd?  Or am I doing something else wrong?  It  happens on linux-x86_64, RTEMS 4.9.4, and vxWorks 5.4.
> 
> Thank you,
> Stephanie Allison
> 

-- 
Eric Norum
[email protected]

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System