Re: strange problems with junit

> I tried to reproduce exactly the same command-line CC would produce by the
> ant builder
>

Did you get the command-line from the CC log?

Normally when there's a problem running JUnit tests under CC the problem is
triggered when you use the same -listener/-logger commands that CC is using
to create the xml log file.

Jtf


On Tue, Jun 9, 2009 at 4:10 AM, Sascha Ernst <[email protected]>wrote:

>   Hello,
>
>
>
> I’ll try your tips with the ports and the reduced buildscript.
>
>
>
> The successful builds ran on the same machine. Even in the same checkout.
>
> I tried to reproduce exactly the same command-line CC would produce by the
> ant builder
>
>
>
> Regards
>
>
>
> ///sascha ernst///
>
>
>
> *Von:* Knuplesch, Juergen [mailto:[email protected]]
> *Gesendet:* Dienstag, 9. Juni 2009 12:54
> *An:* [email protected]
> *Betreff:* Re: [Cruisecontrol-user] strange problems with junit
>
>
>
> Hello,
>
>
>
> just an idea:
>
> since this seems to be a sevlet problem: Are there the ports you need used
> by CC or sth. like that?
>
> I mean a conflict between your test and CC configuration. (Port 8080
> etc...)
>
> For a test change the config of your test.
>
>
>
> for debugging:
>
> write a buildfile that does only the one and only test (no checkout, no
> compile just use the last result of the big build) that fails and see what
> happens.
>
>
>
> Did you run the successfull build on the same machine?
>
> --
> Jürgen Knuplesch
>
>
>  ------------------------------
>
> *Von:* Sascha Ernst [mailto:[email protected]]
> *Gesendet:* Dienstag, 9. Juni 2009 12:02
> *An:* [email protected]
> *Betreff:* [Cruisecontrol-user] strange problems with junit
>
> Hello to all of you,
>
>
>
> I have some real strange problems in the constellation cruise-control óant
> ó junit test.
>
> From my point of view, It’s not the well known classpath problem (but I
> might be wrong)
>
>
>
> First some words about my environment:
>
> Win XP, java 1.6, cruise control 2.8.2, ant 1.7.1
>
>
>
> My way of software building may look a bit strange. I choosed this way to
> get maximal parallelism.
>
>
>
> The cruise control.xml calls an antscript by:
>
>     <schedule interval="3600" >
>
>       <ant antscript="C:/Programme/ANT/bin/ant.bat"
>
>                        buildfile="${_NBS.dir}/build.xml"
>
>                                   target="build_complete"
>
>                                   uselogger="true"
>
>                                   showProgress="true"
>
>                                   showAntOutput="true">
>
>                                   <property name="timestamp"
> value="${cctimestamp}" />
>
>     </ant>
>
>  </schedule>
>
>
>
> This antscript (let’s call it the superscript)
>
> -          Performs all needed checkouts
>
> -          Makes copies of the checkouts
>
> -          Calls different ant-buildscripts in these copies to build the
> different parts of our system
>
> I use this contruction to ensure a clean environment for the called
> buildscripts:
>
>     <parallel failonany="true">
>
>      <sequential>
>
>         <echo message="start irma buildscript in an external cmd" />
>
>         <exec executable="cmd "
>
>               dir="${sc.release.dir}/${sc.build.dir}">
>
>           <arg line="/c start cmd /c
>
>                                                             &quot;
>
>                                                                ant.bat
> ${build.target}
>
>                                                                 >
> ${basedir}/${release.dir}/3_build_irma.nbs-log
>
>                                                                 2>
> ${basedir}/${release.dir}/3_build_irma.err.nbs-log
>
>                                                                 &quot; "/>
>
>         </exec>
>
>       </sequential>
>
>
>
> One of these buildscripts calls junit to run the unit tests:
>
>   <target name="run.junit.tests"
>
>           depends="init"
>
>           unless="release"
>
>                                  if="junit" >
>
>
>
>                 <mkdir dir="${temp.dir}/junit"/>
>
>
>
>                 <junit printsummary="withOutAndErr"
>
>                        haltonfailure="false"
>
>                                   haltonerror="false"
>
>                                   showoutput="true"
>
>                                   timeout="60000"
>
>                                   tempdir="${temp.dir}" >
>
>                   <classpath refid="irma.build.classpath" />
>
>
>
>       <formatter type="xml"/>
>
>
>
>       <batchtest fork="yes" todir="${temp.dir}/junit">
>
>         <fileset dir="${base.src.dir}">
>
>           <include name="**/test/*.java"/>
>
>         </fileset>
>
>         <fileset dir="${irma.src.dir}">
>
>           <include name="**/test/*.java"/>
>
>         </fileset>
>
>       </batchtest>
>
>
>
>     </junit>
>
>
>
>
>
> When I start the superscript by hand, all junit tests were performed and
> reportet.
>
> When cruise control starts the superscript, the junit tests stop always at
> the 64. test with this error
>
>     [junit] Test com.xtramind.common.http.test.TestService FAILED
>
>     [junit] Test com.xtramind.common.http.test.TestServlet FAILED
>
>
>
> BUILD FAILED
>
> F:\NightlyBuildSystem\NBS-run-20090609061149\developer\irma\buildfiles\irma\build.xml:3394:
> The following error occurred while executing this line:
>
> F:\NightlyBuildSystem\NBS-run-20090609061149\developer\irma\buildfiles\irma\build.xml:1392:
> org.apache.tools.ant.taskdefs.optional.junit.XMLJUnitResultFormatter is not
> a JUnitResultFormatter
>
>
>
> Up to this point, 63 junit reports in xml format have been written (using
> this XMLJUnitResultFormatter).
>
>
>
> I checked the FAQs, I googled around, I searched the mailing list.
>
> But I did not find something that helped me out of this mess.
>
> I’m puzzled because I simply do  not understand what’s going on there L
>
>
>
> Has anyone an idea?
>
>
>
> If you need more information: just ask.
>
>
>
> Ein Gruss von der Saar
>
> --
> Sascha Ernst
> Dipl. Inform.
> Head of Software Quality Management
>   ------------------------------
>
>
>
> Telefon: + 49 (0)681 - 302 – 5133
>
> Telefax: + 49 (0)681 - 302 – 5109
>
> E-Mail: [email protected]
>
> Website: http://www.living-e.com
>   ------------------------------
>
>
>
> Living-e AG
>
> Campus D3 2
>
> 66123 Saarbrücken
>   ------------------------------
>
> * *
>
> *Besuchen Sie uns bei der Call Center Innovations Tour 2009*
>
> CCIT’09
>
> *05. Mai  **–** Duisburg*
> *06. Mai  **–** Hannover*
> *12. Mai  **–** Ingolstadt*
> *14. Mai  **–** Le Méridien, Wien*
> *09. Juni **–** Lake Side, Zürich*
>
> CALL CENTER INNOVATIONS TOUR
>
> *
> Für nähere Informationen besuchen Sie **www.ccit-online.de*<http://www.ccit-online.de/>
>   ------------------------------
>
>
>
> Vorstand: Guido Polko
>
> Aufsichtsratsvorsitzender: Uwe Feuersenger
>
> Amtsgericht: Mannheim, HRB 111323
>
>
>
>
>
>
> ------------------------------------------------------------------------------
> Crystal Reports - New Free Runtime and 30 Day Trial
> Check out the new simplified licensing option that enables unlimited
> royalty-free distribution of the report engine for externally facing
> server and web deployment.
> http://p.sf.net/sfu/businessobjects
> _______________________________________________
> Cruisecontrol-user mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/cruisecontrol-user
>
>


-- 
Build engineer training: http://ci-guys.com

------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables unlimited
royalty-free distribution of the report engine for externally facing 
server and web deployment.
http://p.sf.net/sfu/businessobjects

_______________________________________________
Cruisecontrol-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/cruisecontrol-user