Add integration test for --fail-fast

[oehme]

No description provided.

[michael-o]

Will add and run it!

[michael-o]

I do not understand the test:

• You are not even passing --fail-fast. Did you forget that?
• You make sure that Module3 will never run because Module1 will fail, Module2 needs to be built first because Module3 relies on it?

[oehme]

You are not even passing --fail-fast. Did you forget that?

--fail-fast is the default

You make sure that Module3 will never run because Module1 will fail, Module2 needs to be built first because Module3 relies on it?

Yep

[michael-o]

@rfscholte . This looks good to me. It properly passes on my Windows 10 desktop, an old, low-perf 4-thread Atom CPU with FreeBSD 11.2-RELEASE and a 20-thread Xeon server with FreeBSD 12-STABLE. I'd ready to merge this into master along with the Maven change.

[michael-o]

I will now also try the negative case with Maven from master.

[michael-o]

Very good, with Maven from master:

mng6720FailFast(ItShouldWaitForConcurrentModulesToFinish)...FAILURE (3.1 s)

[INFO] Skipping parent
[INFO] This project has been banned from the build due to previous failures.
[INFO] module-1 ........................................... FAILURE [  2.905 s]
[INFO] ------------------------------------------------------------------------
[INFO] module-2 ........................................... SKIPPED
[INFO] module-3 ........................................... SKIPPED
[INFO] parent ............................................. SKIPPED
[INFO] ------------------------------------------------------------------------

[olamy]

IT test for apache/maven#272

[oehme]

Am I correct that this has been merged, so the PR can be closed now?

[olamy]

@oehme some changes have been requested :)

[oehme]

Yeah, but @michael-o already did those changes and committed here: 3ccccc4

[olamy]

merged via 3ccccc4

[michael-o]

I have now merged into master. @olamy the commit existed on a branch only. No issue. All is fine now.