net: remove usage of require('util')

[dnlup]

Use require('internal/util/inspect').inspect and
require('internal/util/debuglog').debuglog instead of
require('util').debuglog and require('util').inspect.

Refs: #26546

Checklist

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

[BridgeAR]

CI https://ci.nodejs.org/job/node-test-pull-request/21764/

[BridgeAR]

Landed in e112fb4 🎉

[BridgeAR]

This broke master (likely due to a different change that landed in the meanwhile that conflicts with this one) and I had to pull it out again.

@dnlup would you be so kind and rebase this and to take another look? Thanks!

[Trott]

This commit is still in master and the test is still broken. I'll open a fast-track revert unless somebody else is already doing that. And if we get a fix in before we revert, cool. And if not, we can re-land later.

[dnlup]

@BridgeAR Yes, no problem

[dnlup]

@BridgeAR Rebasing I get this error in node/test/parallel/test-net-access-byteswritten.js:15:

assert.strictEqual(tls.TLSSocket.super_.prototype.bytesWritten, undefined);

TypeError: Cannot read property 'prototype' of undefined

which I think it is caused by using

Object.setPrototypeOf(Socket.prototype, stream.Duplex.prototype);
Object.setPrototypeOf(Socket, stream.Duplex);

instead of

util.inherits(Socket, stream.Duplex);

which sets the super_ property:

https://github.com/nodejs/node/blob/master/lib/util.js#L145

[BridgeAR]

@dnlup you can change the test as I did in another PR that I closed: 43fa556

[dnlup]

@BridgeAR I would change this too if you're ok with it

const internalUtil = require('internal/util');

to

const { deprecate } = require('internal/util');

since deprecate is the only function used in this module.

[BridgeAR]

@dnlup sure, go ahead.

[dnlup]

@BridgeAR So I rebased master again, modified the test and it looks good. Only thing is that I should force push again. Would that be ok?

[BridgeAR]

@dnlup thanks. Normally it would absolutely be fine to force push but I would rather have a new PR for this since I messed up removing the original commit (I pushed to my fork instead of to upstream) and it was instead reverted. So this PR is still referenced by the commit and it should actually be closed and ideally another PR would be opened.

[dnlup]

Closing in favour of #26920