Mailing-List: contact watchdog-dev-help@jakarta.apache.org; run by ezmlm
From: "Scott Stirling" <sstirling@mediaone.net>
To: <watchdog-dev@jakarta.apache.org>
Subject: RE: Am I the first?? 
Date: Wed, 15 Nov 2000 20:44:56 -0500
Message-ID: <NEBBJKPJKLBGBGIOBGAGGEFKCBAA.sstirling@mediaone.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Importance: Normal
In-Reply-To: <NEBBJKPJKLBGBGIOBGAGOEFJCBAA.sstirling@mediaone.net>

Speaking of good links -- there seems to be some of the sort of thing I was
hoping for here:

http://c2.com/cgi/wiki?ServletTesting

and

http://www.linuxcare.com.au/mirrors/c2wiki/RecentChanges/EnhancingJunit.html

I've read the "TestInfect your EJBs" article
(http://www.javaworld.com/javaworld/jw-05-2000/jw-0526-testinfect_p.html)
before, but it didn't hit me right, I guess.  I'm looking for some drop dead
simple cases using servlets first, so maybe I'll work my way up to the EJBs
from there.  Anyway, check these out, and I'll check them out and maybe come
up with something.

Scott Stirling
West Newton, MA

-----Original Message-----
From: Scott Stirling [mailto:sstirling@mediaone.net]
Sent: Wednesday, November 15, 2000 8:36 PM
To: watchdog-dev@jakarta.apache.org
Subject: RE: Am I the first??


Hi Dan,

I too am getting into JUnit.  I still don't quite "get" how to write tests
for it.  I'm a QA engineer for JRun, and I'd like to use JUnit for a lot of
testing, but none of the examples I've found on the Web so far show me
exactly what I'm looking for (e.g., a full blown servlet testing example
would be easiest to start with, I think).  I need to test servlets, EJBs and
regular classes.  If you've found any great resources on it, please let me
know.

Watchdog is a spec-compliance test suite.  It's a boat load of servlets and
JSPs that test those two specs (Servlet 2.2 and JSP 1.1 right now),
paragraph by paragraph, basically.  There's a simple HTTP client that drives
the tests, which are packaged as web-apps.  The client kicks off requests to
a server, such as Tomcat or JRun, that's got the Watchdog apps deployed.
Then a log is generated with hundreds of success and/or failure messages.

Ant is a very cool tool, and I've pretty much got that under my belt.  Now
if I could just wrap my head around JUnit and start making use of it.

Scott Stirling
West Newton, MA


-----Original Message-----
From: Dan Connelly [mailto:dsconnelly@adelphia.net]
Sent: Wednesday, November 15, 2000 10:15 AM
To: watchdog-dev@jakarta.apache.org
Subject: Am I the first??


I'm new.  This list seems to be empty, so maybe nobody is listening.  Or is
ezmlm just broken??

OK.  Let me first of all make a confession.   I haven't looked at the
existing distribution except the README and the doc/design, so maybe I am
missing the point.   But just what is the point of watchdog??   The
sub-project web page is far from clear on this.

Basically, I just looking into jarkarta-watchdog because I am interested in
frameworks where I could use JUnit testing during servlet development, XP
style.   See http://junit.org/

This goes beyond merely testing for conformance with JSP and Servlet API
specifications, of course.

In addition, I would run the unit test suites repeatedly to gather
statistical success/failure/error data which I would post to a MySQL
database.   Continuous testing seems necessary to me in multi-threaded code
because the interactions are more statistical than deterministic.   One
examines the statistics using ad hoc queries, status displays or "Hey You"
alerters, all using JSP/XML technology.  Developers and managers would
select their own anxiety thresholds for panic, when the production system
would be shutdown.

The scope of this work must include some provision for "black box" testing
of "legacy" applications EJBed to the servlets.  This exploits the Java/EJB
and XP advantage, seems to me, because the testing suites are closer to the
programmers.

Well, this scheme comes closer to my notion of a Watchdog then just API
conformance.   Any comments?

Dan Connelly, Burlington VT
(the other one, not Dan Connolly, W3C)