src: prefer param function check over args length

[codebytere]

We have to override some fs methods in Electron, and as I looked at their original impls I believe that checking the type of the second argument as opposed to simply the number of arguments passed is more consistent and reliable.

Checklist

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

[nodejs-github-bot]

@codebytere build started: https://ci.nodejs.org/blue/organizations/jenkins/node-test-pull-request-lite-pipeline/detail/node-test-pull-request-lite-pipeline/1346/pipeline

[refack]

👍

I had a thought once about doing something like:

function fstat(fd, ...args) {
  const lastArg = args.pop();
  const callback = (typeof lastArg === 'function') ? lastArg : throw ..;
  const options = args.pop() || {};
  ...
}

Actually what we need is a generic arg validator :dream:

function fstat(...args) {
  const { fd, options, callback } = validateArgs(args, { some sort of contract });

[mscdex]

CI: https://ci.nodejs.org/job/node-test-pull-request/18086/

[mhdawson]

LGTM

[lpinca]

I think this will conflict with @bnoordhuis's work on #23767

[Trott]

Should we land #23767 first and then adapt this PR to that change? Seems do-able. /cc @codebytere @bnoordhuis

[Trott]

@codebytere This seems ready to go if we can get a rebase to resolve the merge conflict with #23767...

[codebytere]

@Trott i'll handle first thing tomorrow!

[codebytere]

@Trott should be set! i couldn't change this line becuase if i simply checked for flags being a function that didn't properly account for the error case where no callback was passed (i.e. only one arg passed) and would improperly fail tests with ERR_INVALID_OPT_VALUE instead of ERR_INVALID_CALLBACK.

[refack]

CI: https://ci.nodejs.org/job/node-test-pull-request/18378/

[Trott]

Landed in 7cf5679