Handle duplicate names in "Contributors" list

[lukemassa]

Feature request

What problem does this feature solve?

Right now, if there are two git commits for the same with the same name but different email, it will show up as two different "contributors". It is relatively common for the same person to commit with different emails, for example this instance from https://github.com/runatlantis/atlantis.

atlantis % git log runatlantis.io/docs/README.md | grep Author
Author: Luke Kysow <[email protected]>
Author: Luke Kysow <[email protected]>
Author: Luke Kysow <[email protected]>
Author: Luke Kysow <[email protected]>
Author: Luke Kysow <[email protected]>
Author: Luke Kysow <[email protected]>
Author: Luke Kysow <[email protected]>
Author: Luke Kysow <[email protected]>

This results in https://www.runatlantis.io/docs/ saying "Contributors: Luke Kysow, Luke Kysow" at the bottom.

What does the proposed API look like?

Not sure, see below.

How should this be implemented in your opinion?

I'm actually not sure. The issue is that these are rendered as spans, which are in fact different (see attached), but only in metadata not in any way visible on the page.

Here are a few ideas:

1. The simplest answer is "These are technically two different contributors, the code is working as expected". If this is the decision, that's fine, feel free to close this issue.
2. The deduping could happen at the git plugin level here: https://github.com/vuepress/vuepress-next/blob/9f21432e578c61fedfd4acc3466a08a58a0976a0/ecosystem/plugin-git/src/node/utils/getContributors.ts#L4, by adding a filter() and eliminating entries with the same name (picking I guess the first email?)
3. It could also be done at rendering time here: https://github.com/vuepress/vuepress-next/blob/9f21432e578c61fedfd4acc3466a08a58a0976a0/ecosystem/theme-default/src/client/components/PageMeta.vue#L115
4. Some sort of configuration option that allows users to either toggle this behavior, or to provide code that said "these two users are the same" or "run this filter on contributors". This seems kind of heavy-handed, and might not useful enough to belong in a public configuration, but 🤷

Are you willing to work on this yourself?

Yup!

[Mister-Hope]

Nice catch, actually I am not sure what's the best too.

I think it's reasonable for the git plugin to output as is, and maybe the theme should filter duplicate names before displaying it?

Edited: for the git plugin, maybe it should try to drop any email containing no-reply if a contributor with same name already exisits.

[lukemassa]

Yeah I like your suggestion: basically omit anything that's obviously a duplicate, but nothing else. That way in the (rare) case that two people with the same name but different emails we include them both. Then later at the theme stage theoretically something could be done to distinguish them, but that could be a later improvement if it happens frequently.

It also opens the door to later refining what it means to be "obviously a duplicate"; this is the only specific case I know about now, so best to start conservatively, and users who see this problem will have a clear place to extend the code.

[Mister-Hope]

@meteorlxy Any suggestions?

[meteorlxy]

Maybe we can add an option for git plugin to manipulate the contributors list

[lukemassa]

@Mister-Hope this what you were thinking? lukemassa@c327018 If so I'll turn this into an "issue" and write up a PR.

As for an option to manipulate the contributors list, I'm happy to try that out. I'm just not sure what else people might want to add other than this, and it might be kind of awkward to have people copy and paste this particular filter around. I recognize this represents a change in functionality over the existing situation, but it seems innocuous and unlikely to cause regressions.

[Mister-Hope]

I think we should widen the email to just contain noreply or no-reply

[lukemassa]

I incorporated this, and turned it into a PR #1364