errors: alter and test ERR_INVALID_REPL_EVAL_CONFIG

[davidmarkclements]

changes the base instance for ERR_INVALID_REPL_EVAL_CONFIG
from Error to TypeError as a more accurate representation
of the error and adds a unit test for the repl options that
trigger this error.

Checklist

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

[BridgeAR]

CI https://ci.nodejs.org/job/node-test-pull-request/14270/

[jasnell]

CI failure is unrelated.

@nodejs/tsc .. any objections to this landing in 10.0.0?

[jasnell]

Landed in 8826cc9

[Trott]

This landed in violation of several of our rules.

• It has only one TSC approval and is semver-major. Our rules require a second TSC approval. (https://github.com/nodejs/node/blob/master/COLLABORATOR_GUIDE.md#consensus-seeking)
• It was open for less than 48 hours. (https://github.com/nodejs/node/blob/master/COLLABORATOR_GUIDE.md#waiting-for-approvals)
• It was open for less than 72 hours and it is a weekend everywhere in the world right now. (https://github.com/nodejs/node/blob/master/COLLABORATOR_GUIDE.md#waiting-for-approvals)
• CI was not green. The red CI task should have been re-run. (https://github.com/nodejs/node/blob/master/COLLABORATOR_GUIDE.md#testing-and-ci)

On the one hand, I'm inclined to cut @jasnell a ton of slack as he's on a PR-landing spree right now, probably in preparation for 10.0. On the other hand, though, I'm kind of worried about the example this sort of thing sets for other Collaborators.

[jasnell]

Think I missed this one up with another very similar one. And yes, I'm preparing for a big update for 10.0.0 on Monday so trying to get things that can land landed. My apologies for the oversight on this one

[Trott]

Ah well, just one more data point in favor of ever increased automation. :-D