[FEAT] System exit

[danglotb]

We call now a System.exit() at the end of the main methods of the runner classes. I recall that these classes should never be used directly, but only called by the EntryPoint, which will use a new java command line to run the main of the runner classes.

@martinezmatias @andre15silva
This should solve the issue on Math2, since the process was not ending properly. I guess that the System.exit() is a bit rough but IMO, it is a valid workaround.

The tests is executing directly the main methods of the runner classes, this is why we must now use a special flag when running the test: mvn clean test -DnoExitForTesting=true.

[danglotb]

Merging and releasing.

[martinezmatias]

Thanks a lot @danglotb !
Which is the version of the release? 3.1.1? In that case, we will update the dependency from Flacoco (it uses 3.1.0)
Regards
Matias

[danglotb]

Hum, I have issue releasing, since the release goal of maven launches an inside maven command, without the option for testing and avoiding the System.exit().

I'll investigate to find a solution but might have revert this commit

[martinezmatias]

ok!

[andre15silva]

Hum, I have issue releasing, since the release goal of maven launches an inside maven command, without the option for testing and avoiding the System.exit().

You can use https://github.com/stefanbirkner/system-rules.

https://stefanbirkner.github.io/system-rules/apidocs/org/junit/contrib/java/lang/system/ExpectedSystemExit.html

We use it in Flacoco:

https://github.com/SpoonLabs/flacoco/blob/6f15f8162ce59c66af9f378691a5c9dcc603af7a/src/test/java/fr/spoonlabs/flacoco/cli/FlacocoMainTest.java#L56-L76

[danglotb]

Hi @martinezmatias

In the meantime, you can try on the SNAPSHOT, which is deployed on Nexus Repository.

[martinezmatias]

great, thanks, I will try 3.1.1-SNAPSHOT.

[danglotb]

@andre15silva Awesome! Thank you very much for the resources. I'll fix the test and release, thank you! 😄

[danglotb]

test-runner-3.1.1 is out 🎊

[andre15silva]

Thanks @danglotb !

I'll merge ASSERT-KTH/flacoco#117 once it is green @martinezmatias

[andre15silva]

Hi @danglotb @martinezmatias

I confirm Math5 generates fault localization results on my local setup! Should work form Math2 and Math8 too. Good job @danglotb !

[martinezmatias]

Great job @danglotb @andre15silva !!

I confirm that the issue is solved on the CI for Math-5, Math-8 and Math-28!! see here.
(for M2 we don't have the memory issue but still no result, to be discussed in other issue suspicions logging from TestRunner)

[monperrus]

nice!