Rashmi Rubdi wrote:
Thank you for your reply,
I will start reading those chapters and the presentation, I think it should be
fairly simple to understand since I'm already familiar with the concepts.
I think, now there's a new and better way to write JSPs using JSTL tags or
custom tags.
Perhaps, Cactus tests could be written with custom tags instead of directly
having scriptlets in JSPs.
sorry, I've been confusing
1. Cactus doesnt need JSP pages, but it does need to instantiate your
test case class in the client (the process running the junit ant task)
and the app server. Your results (including stack traces) get mixed with
other code on the client in the HTML junit reports.
2. the tests-in-JSP coding is stuff I do on the side. There you run
httpunit against the JSP pages -or point your browser at them- and see
the results in the browser. What you dont get is the stack traces
integrated in the junit reports.
This is what happy.jsp looks like
<%@ page import="d1.webapp.FindClass" %>
<%
//set the headers that disable caching.
//we do not want proxies or web browsers to cache the health pages,
//as they will then give a misleading impression.
response.setHeader("Cache-Control","no-cache");
response.setHeader("Pragma","no-cache");
response.setDateHeader ("Expires", 0);
%>
<html>
<head><title>Diary Happy Page</title></head>
<body>
<h2>Classes </h2>
<table>
<tr><td>
<%= FindClass.locate("d1.core.Event","no diary-core")%>
</td></tr>
<tr><td>
<%= FindClass.locate("com.sun.syndication.io.FeedException","No
Rome")%>
</td></tr>
<tr><td>
<%=
FindClass.locate("javax.persistence.TransactionRequiredException",
"Javax.persistence (the EJB3 persistence JAR) missing")%>
</td></tr>
<tr><td>
<%= FindClass.locate("org.jdom.JDOMException",
"JDom missing")%>
</td></tr>
<tr><td>
<%= FindClass.locate("org.apache.log4j.PatternLayout",
"Log4J missing")%>
</td></tr>
<tr><td>
<%= FindClass.locate("d1.webapp.servlets.LifecycleListener")%>
</td></tr>
<tr><td>
<%= FindClass.locate("d1.persist.session.CalendarFactory",
"diary-core-persist library missing")%>
</td></tr>
<tr><td>
<%= FindClass.locate("org.hsqldb.jdbcDriver","HSQL missing")%>
</td></tr>
<tr><td>
<%= FindClass.locate("d1.webapp.servlets.LifecycleListener")%>
</td></tr>
</table>
<h2>References </h2>
<table>
<tr><td>
<%= FindClass.ldapLookup("java:/DefaultDS")%>
</td></tr>
<tr><td>
<%= FindClass.ldapLookup("diary/CalendarSessionBean/local")%>
</td></tr>
</table>
<p>
The diary <i>enterprise</i> application is succesfully deployed
</p>
</body>
</html>
Finally yes, tags to run tests would be better. I have a <happy> tag
that lets you do all of that, but havent covered in the second edition
of the book as you have to look at how to create JSP tag libraries. Even
with java5 there's no annotation suite to do that, so you are stuck with
XDoclet, which is a project undergoing significant turmoil right now.
I also need to figure out how to remove the test part of the code inside a JSP
- I guess having it in an include file might help. But I will read about this
and gain a better understanding rather than guessing it.
look at cactus first; the test-in-JSP is another technique.
-steve
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
