[cdt-dev] Best practices to run and analyze CDT JUnit tests?

[Raphael Zulliger <zulliger@xxxxxxxx>]

Hi

I'm playing around with running the CDT JUnit tests. I followed the 
instructions here http://wiki.eclipse.org/CDT/release_engineering in 
order to build CDT on a 64Bit Ubuntu 11.10 machine. Building (which 
includes running all tests) succeeded. I also figured out that a 
subsequent 'mvn install' does (among others) re-run the tests. Moreover, 
I found the test results, which are spread all over the filesystem:

    ./codan/org.eclipse.cdt.codan.core.test/target/surefire-reports
    ./xlc/org.eclipse.cdt.core.lrparser.xlc.tests/target/surefire-reports
    ./core/org.eclipse.cdt.ui.tests/target/surefire-reports
    ./core/org.eclipse.cdt.core.tests/target/surefire-reports
    
./build/org.eclipse.cdt.managedbuilder.core.tests/target/surefire-reports
    ./build/org.eclipse.cdt.managedbuilder.ui.tests/target/surefire-reports
    ./lrparser/org.eclipse.cdt.core.lrparser.tests/target/surefire-reports
    ./upc/org.eclipse.cdt.core.parser.upc.tests/target/surefire-reports

My main question is: How do I best analyze these reports? Is there a 
(simple) way to get an overview of all tests similar what hudson offers 
(e.g. 
https://hudson.eclipse.org/hudson/job/cdt-nightly/934/testReport/)? One 
of the main reasons for me to run these tests is to compare several runs 
with each other (after having applied patches) to verify that nothing 
got damaged. I guess that this is a common use case and therefore I ask 
myself how others perform this task? (just using a simple diff tool?!)

Other questions would be:
 - Is 'mvn install' the best available option if I only want to run all 
unit tests? Or is there another command that simply executes all tests?
 - How do you deal with the huge output of maven/tycho? Is there some 
kind of tool that gets rid of 'unimportant information', or do you 
simply do a 'grep -v "[INFO]" or similar?

Thanks
Raphael