Build and test mergely on JDK 16 (on Travis)

[lrytz]

No description provided.

[dwijnand]

Mostly LGTM. Just a tiny request, question, and a "let's not do it this way" pushback. 😄

[som-snytt]

I have a couple of test failures on jdk16... I mentioned it on 3613.

test/files/neg/macro-invalidret-neg.log
test/files/run/t3613-run.log

[SethTisue]

is the idea here that if we test on 16, we don't also need to test on 11?

[lrytz]

is the idea here that if we test on 16, we don't also need to test on 11?

I'd say that's fine, since we have a community build on 11?

[SethTisue]

I'd say that's fine, since we have a community build on 11?

I don't have a strong feeling either way.

(In the community build, I chose to have both 11 and 16 builds, but I chose that mainly because several important upstream maintainers don't support 15 or 16, so the 11 build tests noticeably more stuff. Here in this repo, that doesn't really apply, so 16 only is probably fine. On the other hand, doing 11 too wouldn't be especially costly. 🤷 )

[som-snytt]

On my lunch break, I am PRing removal of use of ThreadGroup in testkit, which is day job adjacent.

My other innovative contribution is to pronounce PR as purr.

[lrytz]

The JDK 16 build went green too: https://travis-ci.com/github/scala/scala/builds/223846952

(I pushed to a scala/scala branch for testing: https://github.com/scala/scala/tree/travis-jdk16)

re-review by @dwijnand

[lrytz]

(the github page summarizing the travis build incorrectly says ADOPTOPENJDK=8, https://github.com/scala/scala/runs/2408503535. if you look at the build log you see that the env is ADOPTOPENJDK=16, and java -version is correct)

[SethTisue]

@som-snytt https://twitter.com/SethTisue/status/1386797658988187648

[som-snytt]

@SethTisue whew, when I saw the notification while waiting at child's piano lesson, I wondered if something broke under -Xsource:3 in the CB.

Thanks for reminding me that I still have to add the ThreadGroup commit to my PR for REPL pretty printing.