Hi Marc,

Funny you should mention performance, because I'm starting to have an issue
with it.
My WebTest suite is growing, and I'm finding that, as it approaches the end,
it slows
to a crawl and practically doesn't finish at all.

I've noticed the Out of Memory recommendations in the TroubleShooting
WebTests
section of the WebTest site.  I do have my ANT_OPTS set as recommended.
This
has not resolved the issue.

Typically, my tests look like this ...

<?xml version="1.0"?>
<!DOCTYPE project SYSTEM "../dtd/Project.dtd">
<project name="demo.beboWebTest" default="test">
    <property name="BirthDay" value="17"/>
    <property name="BirthMonth" value="10"/>
    <property name="BirthYear" value="1980"/>
    <property name="FirstName1" value="Alan"/>
    .
    .
    .
    <target name="test">
        <webtest name="Register First User">
            <config
                protocol="http"
                host="testbranch.bebo.com">
                <header name="User-Agent" value="Mozilla/5.0 (Macintosh; U;
Intel Mac OS X; en-US; rv:1.8.1.14) Gecko/20080404 Firefox/2.0.0.14"/>
                <header name="Accept-Language"
value="en;q=0.7,en-us;q=0.3"/>
                <header name="ThrowExceptionOnScriptError" value="false"/>
                <header name="X-BeboAutomatedTestAgent" value="WebTest"/>
            </config>
            <steps>
                <invoke url="InviteJoin.jsp?Member=N&amp;AutomatedTest=Y"
description="Register First User"/>
                <setInputField name="FirstName" value="${FirstName1}"/>
                <setInputField name="LastName" value="${LastName1}"/>
                <setRadioButton name="GenderCd" value="${GenderCd1}"/>
                <setSelectField name="BirthDay" value="${BirthDay}"/>
                <setSelectField name="BirthMonth" value="${BirthMonth}"/>
                <setSelectField name="BirthYear" value="${BirthYear}"/>
                <setInputField name="Email" value="${User1}"/>
                <setInputField name="Password" value="${Password}"/>
                <clickButton name="xxx"/>
                <verifyText text="${Check your email address book below}"/>
            </steps>
        </webtest>
        <webtest name="Sign Out First User">
            <config
                protocol="http"
                host="testbranch.bebo.com">
                <header name="User-Agent" value="Mozilla/5.0 (Macintosh; U;
Intel Mac OS X; en-US; rv:1.8.1.14) Gecko/20080404 Firefox/2.0.0.14"/>
                <header name="Accept-Language"
value="en;q=0.7,en-us;q=0.3"/>
                <header name="ThrowExceptionOnScriptError" value="false"/>
                <header name="X-BeboAutomatedTestAgent" value="WebTest"/>
            </config>
            <steps>
                <invoke url="SignOut.jsp" description="Sign Out First
User"/>
                <verifyText text="${Signed Out}"/>
            </steps>
        </webtest>
        .
        .
        .
    </target>
</project>

The properties section above is truncated and has gotten quite long in my
actual
test suite.  Then, I keep adding more and more webtests as shown above.  Am
I structuring this whole improperly?  In the WebTest site, I see the
recommendation
about not calling taskdef more than once .. but I don't know what this means
(in general I'm no ANT expert).  Also, I do not know how to use more than
one file
for collecting reports and still have a unified regression test.  Any
suggestions
would be greatly appreciated.

Thank You,
Kenneth Hopf





On 5/15/08, Marc Guillemot <[EMAIL PROTECTED]> wrote:
>
> Hi,
>
> 3 developers of FINN.no (http://www.finn.no/) have created a new project
> named Celerity (http://celerity.rubyforge.org/), that wraps HTMLUnit in
> a Watir compatible API to use in JRuby. The motivation of the project
> was the performance problems they had running their Watir test suite.
>
> They have now published some benchmarks:
> http://celerity.rubyforge.org/benchmarks.html
>
> the time reduction is really impressive: it varies between 69% and 99,8%!
>
> I imagine that the overhead of having WebTest around HtmlUnit is not
> more significant than the one of JRuby and therefore that similar speed
> differences could be messed between WebTest and Watir.
>
> Now it would be interesting to see how fast other tools are, like for
> instance Selenium. My first guess is that it is even slower than Watir
> when running in IE.
>
> Cheers,
> Marc.
>
> --
> Blog: http://mguillem.wordpress.com
>
> _______________________________________________
> WebTest mailing list
> [email protected]
> http://lists.canoo.com/mailman/listinfo/webtest
>

Reply via email to