incubator-jspwiki-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Glen Mazza <glen.ma...@gmail.com>
Subject Re: JSPWiki test failing
Date Sun, 28 Oct 2012 00:29:18 GMT
Never mind on the 2nd half of what I was saying below, switching away to 
brand new pages not being used in the other tests works fine (coding 
error on my part.)  The first part of the problem still holds, I'll 
submit a patch...

Glen

On 10/27/2012 12:32 PM, Glen Mazza wrote:
> Hi, trying to build JSPWiki trunk (I'm using Oracle JDK 7 on Ubuntu 
> Linux) is bringing up JUnit test failures with PageViewPluginTest's 
> testShowCountsBasic().  To test this repeatedly, I first modify the 
> Ant build.xml to run this test alone:
>
> Index: build.xml
> ===================================================================
> --- build.xml    (revision 1402816)
> +++ build.xml    (working copy)
> @@ -844,7 +844,7 @@
>          <formatter type="xml" usefile="yes" />
>          <batchtest todir="${tests.reports}">
>             <fileset dir="${tests.src}">
> -                <include name="**/*Test.java" />
> +                <include name="**/PageViewPluginTest.java" />
>                  <exclude name="**/AllTest*java" />
>                  <include name="**/StressTestSpeed.java" 
> if="tests.stress.enabled"/>
>                  <exclude name="org/apache/wiki/web/*.*" />
>
>
> It fails with these results:
>
> <ul><li>Test Page 01 (8 views)
> </li></ul><ul><li>Test Page 02 (12 views)
> </li></ul><ul><li>Test Page 03 (2 views)
> </li></ul><ul><li>Test Page 04 (2 views)
> </li></ul><ul><li>Test Page Excluded (1 views)
> </li></ul>
>
> Namely, test page #1 should have only 2 views and test page #2 should 
> have three.  I *thought* the problem was somehow the pages weren't 
> being deleted and the increased viewcounts were because the numbers 
> were being increased from other tests within this class. But that may 
> not be the only problem, because  when I switched away from the common 
> Page 1 and Page 2 and went with brand-new pages that weren't being 
> used in the other test cases, the same problem occurred 50% of the 
> time (pass-fail-pass-fail-pass-fail, etc. correctly passing with 2/3 
> views but then failing with 3/4 views.)
>
> Can anyone replicate this?
>
> Thanks,
> Glen
>


Mime
View raw message