test: replace s_client in test-https-ci-reneg-attack

[Trott]

Replace s_client in test-https-ci-reneg-attack with built-in
client calling tls.renegotiate(). This also fixes the currently-broken
test. (It is broken due to a change in behavior in a
recently-updated-in-core version of s_client.)

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• commit message follows commit guidelines

[nodejs-github-bot]

@Trott build started: https://ci.nodejs.org/blue/organizations/jenkins/node-test-pull-request-lite-pipeline/detail/node-test-pull-request-lite-pipeline/2398/pipeline

[Trott]

Pummel tests are not run in regular CI so this combination should be sufficient:

Lite CI: https://ci.nodejs.org/job/node-test-pull-request-lite-pipeline/2398/

Repurposed custom-suites job from node-daily-master that runs the pummel tests, but set to only run this one (because there is one other broken pummel test fixed in another PR that hasn't landed yet): https://ci.nodejs.org/job/node-test-commit-custom-suites/841/

[bnoordhuis]

Maybe add a trailing comma?

[Trott]

Landed in 43c2a13`

[refack]

BTW, a run with the full suite fails
https://ci.nodejs.org/job/node-test-commit-custom-suites/846/default 🤔

[Trott]

BTW, a run with the full suite fails

Interesting. it's a different test that fails (test-tls rather than test-https) but it's still TLS renegotiation...

[Trott]

@refack Is it possible something funky is up with the git checkout on that run? Here's output from that run:

08:25:56     AssertionError [ERR_ASSERTION]: The expression evaluated to a falsy value:
08:25:56     
08:25:56       assert(/TLS session renegotiation attack/.test(err))
08:25:56     
08:25:56         at TLSSocket.<anonymous> (/home/iojs/build/workspace/node-test-commit-custom-suites/default/test/pummel/test-tls-ci-reneg-attack.js:59:7)

But the assert() in question is not on line 59 of test-tls-ci-reneg-attack.js. It's on line 56.

[Trott]

Yeah, in fact, I'm going to call "funky checkout" on this one. It seems that somehow c421619 is missing. That change fixed that test for this sort of thing. And before that change, the assert() was online 59, as reported in CI. But now it's on line 56. So it sure seems like it's running without that change.

[refack]

something funky is up with the git checkout on that run?

I used the head of this PR - b2ad179:

11:20:08  > git fetch --no-tags --progress git@github.com:nodejs/node.git +refs/heads/*:refs/remotes/origin/* +refs/pull/25720/head:refs/remotes/origin/_jenkins_local_branch # timeout=20
11:20:15 Checking out Revision b2ad1797bde3da028813b073e68dc5ae2d3c9541 (refs/remotes/origin/_jenkins_local_branch)
11:20:15  > git config core.sparsecheckout # timeout=10
11:20:15  > git checkout -f b2ad1797bde3da028813b073e68dc5ae2d3c9541 # timeout=10

[Trott]

I used the head of this PR - b2ad179:

Ah! That'll do it. This branch didn't have the fix for the other test in it. But it had landed on master.