Experimental Physics and
Industrial Control System

"Jeff Hill" <[email protected]> · Thu, 9 Aug 2007 18:06:00 -0600

I have reproduced the bug, and prepared a fix, but not finished testing it.
The bug was caused by the fix for Mantis 275 (introduced in R3.14.9). The
cause was the access rights response arriving after, and not before as is
required, the channel create response message.

> I guess this issue tells us that our pre-release tests have 
> to be improved.

The CA regression tests would have caught this bug, but unfortunately the
bug does not occur unless AS is enabled (an access security file must be
installed to reproduce the problem).

The CA regression tests go through a broad brush of the input request
permutations that can be issued from the CA interface, but what is difficult
to automate w/o improved testing infrastructure is the other IOC side input
variables that can be varied.

Jeff

-----Original Message-----
From: [email protected] [mailto:[email protected]]
On Behalf Of Andrew Johnson
Sent: Wednesday, August 08, 2007 4:48 PM
To: Ralph Lange
Cc: EPICS Core Talk
Subject: Re: Pre-Release Tests of EPICS Base

Ralph Lange wrote:
> I guess this issue tells us that our pre-release tests have to be
improved.

Agreed.  I've not had time to try and replicate your bug report yet, but 
I may be able to do so tomorrow.

> Marty's test suite has tests for access security that should have shown 
> this problem. As long as I was doing things for HP-UX, I was running all 
> of Marty's tests against Linux, HP and 68k-vxWorks.
> Testing things only by sample and on Linux (what I did for 3.14.9) seems 
> definitely not enough.

I'm pretty sure I did run the mrkSoftTest access security stuff before 
the 3.14.9 release, but like a number of of those tests it's not fully 
automated and I'm not completely sure I understood what I was supposed 
to do and/or see.  We need a fully automated test suite, but converting 
mrkSoftTest into a series of automated tests will unfortunately take 
grunt work.

- Andrew
-- 
When a distinguished but elderly scientist states that something is
possible, he is almost certainly right.  When he states that something
is impossible, he is very probably wrong.  -- Arthur C. Clarke

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System