Experimental Physics and
Industrial Control System

Eric Norum <[email protected]> · Mon, 29 Mar 2010 11:15:57 -0700

On Mar 29, 2010, at 10:51 AM, Mark Rivers wrote:

> Hi Eric,
> 
> Thanks for the information.  It seems like renaming "thread" to
> epicsThreadShow would be more consistent.

I agree.
Andrew, can you please make the one-line change to iocsh/libComRegister.c:

--- ../../iocsh/libComRegister.c	2009-02-24 12:10:57.000000000 -0800
+++ /tmp/blah	2010-03-29 11:09:00.000000000 -0700
@@ -204,7 +204,7 @@
 /* thread (thread information) */
 static const iocshArg threadArg0 = { "[-level] [thread ...]", iocshArgArgv};
 static const iocshArg * const threadArgs[1] = { &threadArg0 };
-static const iocshFuncDef threadFuncDef = {"thread",1,threadArgs};
+static const iocshFuncDef threadFuncDef = {"epicsThreadShow",1,threadArgs};
 static void threadCallFunc(const iocshArgBuf *args)
 {
     int i = 1;

and the corresponding change to the app dev guide 'show' command?

> 
> More importantly, the "-level" argument does not seem to provide any
> additional information, at least on Linux or vxWorks.  It would be very
> helpful if higher levels could print additional information like stack
> usage, etc.

Yep, none of the operating-system specific implementations of epicsThreadShow appear to do anything with the 'level' argument.

> 
> On vxWorks the following command:
> 
> thread -5 taskwd
> 
> does not restrict its output to the taskwd thread.  In fact it just
> prints out the entire thread list twice.

Was that to the iocsh or to the vxWorks shell?  The latter would need quotes around the task name.  I suspect that you were just lucky to get the task list twice and not to get an access fault!


Should the epicsThreadShowAll command just go away?  It's function is identical to that of epicsThreadShow with no thread-name (or thread-id) arguments.
-- 
Eric Norum
[email protected]







Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System