util: include reference anchor for circular structures

[BridgeAR]

This adds a reference anchor to circular structures when using
util.inspect. That way it's possible to identify with what object
the circular reference corresponds too.

// cc @nodejs/util

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

[nodejs-github-bot]

Lite-CI: https://ci.nodejs.org/job/node-test-pull-request-lite-pipeline/3513

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/23106/

[addaleax]

I think it’s sufficient to either pick the asterisk or ref as a symbol, using both seems to make the output a bit verbose, but either way 👍

[BridgeAR]

@addaleax I played around with it and this just seemed the most obvious and hopefully straight forward to understand. It does also look nice to just use *1 as anchor but I fear someone might miss that while looking at some output? In the end, I am fine both ways. @nodejs/util what do you think?

I also thought about using &1 as anchor. I just fear that having a separate sign for the anchor might actually confuse some people instead of clarifying that it stands for the anchor.

[BridgeAR]

@nodejs/tsc @nodejs/releasers I currently wonder if it would make sense to postpone landing this PR shortly before the next semver-major release date instead of in the next few days. The reason for this consideration is that this PR is not possible to be backported in a semver-patch manner, even with a compatibility patch (almost all changed code lines are test changes and most other changed lines would also have to be reverted). That would reduce the number of conflicts during the regular releases and that's something we try to achieve in the release working group at the moment. We try to take new measures against conflicts (as e.g., backporting semver-major releases with compatibility patches).
To postpone this, it'll likely require the PR to be rebased and updated again to land cleanly at a specific time before the next semver-major. It's not really possible to tell if the PR author (my point is a generic one, so it's not only about me and this specific PR) has time to do that on their own and it would be bad if the postponing would therefore result in not landing the PR after all. Thus, it would require multiple persons to be willing to potentially step up and to rebase and land the PR when it's time.

I am not sure if this is something we might want to give a try or not but I thought it's worth bringing it up and it would be great to get some feedback!

[BridgeAR]

It's might be just as much work as resolving the conflicts to land the semver-major PRs later (including a minimal new review). So it might not be worth it.

[antsmartian]

Nice! I like the idea.

[addaleax]

@BridgeAR I think it’s fine to land this. If you want to wait, I think it’s best to switch the labels author ready → blocked?

[BridgeAR]

Landed in 9f71dbc 🎉