test: run 'abort' suite on Windows

[refack]

• For now a C++ node::Abort() failure exits with code 3 on Windows

Fixes: #14012
Refs: #14013

CI: https://ci.nodejs.org/job/node-test-commit/12055/

/cc @nodejs/platform-windows @nodejs/testing

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• tests and/or benchmarks are included
• commit message follows commit guidelines

Affected core subsystem(s)

test,process

[refack]

@nodejs/build PTAL do you think aborting on windows will somehow create core dumps (via the unpredictable diagsvc)?

[refack]

New CI: https://ci.nodejs.org/job/node-test-commit/12056/

[Trott]

Looks like abort/test-process-abort-exitcode needs a change (failing on Windows in CI).

[addaleax]

typo: process

[Trott]

Looks like there's a relevant test failure. Also very interested in whether we can turn off core files in test.py for Windows or not.

[refack]

Also very interested in whether we can turn off core files in test.py for Windows or not.

Windows does not generate core files by default. So it's actually a matter to finding a way to turn them on (if we find a way to validate them)...

[Trott]

Windows does not generate core files by default.

Awesome. Then it's just the failing test...

[refack]

@Trott done
Windows CI: https://ci.nodejs.org/job/node-test-commit-windows-fanned/11461/

[Trott]

LGTM if CI is green

[MylesBorins]

This was relying on a semver major change. If possible in future please appropriately tag dont-land, thanks.

[richardlau]

It isn't relying per se. #13947 should have changed test-process-abort-exitcode.js on Windows to match the new behaviour -- Presumably this was missed because prior to this PR the abort tests weren't included in the CI runs.

Perhaps this PR should have been two commits -- one to unbreak the tests (i.e. the bit that is dependent on the semver-major change) and a separate commit to enable the abort tests in the list of testsuites to run on Windows (which should be backported if #14013 is backported otherwise there'll be a disparity between the tests that are run on Windows vs. everywhere else).

[MylesBorins]

@richardlau please feel free to submit a backport if you think a partial change makes sense on 8.x

[refack]

@MylesBorins Backport PR #15460 (since it's build/test you could land it irrespective of a new release)

[MylesBorins]

Should this be backported to v6.x-staging? If yes please follow the guide and raise a backport PR, if not let me know or add the dont-land-on label.

[refack]

Backport PR: #16442