repl: handle buffered string logic on finish

[antsmartian]

Fixes : #24385

Looks like clearBufferedCommand will be called on almost all flows. Hence history was broken. Sorry about that, but I take this as learning and see why history test cases didn't catch up this in first place. I will spend some time this weekend to add few more test cases on to repl history which caused this regression.

Can some one ping repl team to make it land faster?

Checklist

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

[Trott]

@nodejs/repl

[antsmartian]

cc @vsemozhetbyt

[vsemozhetbyt]

Who do we need to cc to not forget including this fix in the security release that will be created soon?

[antsmartian]

@vsemozhetbyt Oh if this gets included in that release, it will be great :) Thanks for the follow up.

[Fishrock123]

CI: https://ci.nodejs.org/job/node-test-pull-request-lite-pipeline/1647/

[vsemozhetbyt]

CI: https://ci.nodejs.org/job/node-test-pull-request/18764/

[antsmartian]

@vsemozhetbyt You are super fast :)

[Fishrock123]

Bleh, some unrelated ARM failure.

Thanks, landed in eb42c1e.

(Hopefully I did that right... been an awfully long time since I landed anything...)

[Fishrock123]

Made an issue about the ARM failure (a crash): #24496

[Fishrock123]

Landed on 11.x in c11d345

[antsmartian]

Thanks @Fishrock123

[targos]

@Fishrock123 v11.x should only be updated when a release is merged. Can you force push to remove the commit?

[targos]

I fixed it, but had to temporarily unprotect the release branches to do so.

[targos]

@Fishrock123 No worries, btw :) I opened an issue in the Release repo to discuss a way to prevent this mistake in the future: nodejs/Release#392

[antsmartian]

@targos Thanks. Not sure if this commit bit (#22153) is backported to 10.x stages. If so, we need to backport this fix to 10.x as well.