1521288 - Pontoon should commit with multiple authors (Co-authored-by:)

Status: RESOLVED FIXED

(Webtools Graveyard :: Pontoon, enhancement)

Description

[Tomer Cohen :tomer]

While I dislike the way Pontoon commit translations on behalf of one of the translation authors instead of creating multiple commits one for each author, there is a Co-authored-by which will cause Github to credit multiple users for the contribution. Since headers in Git/Mercurial are written as the commit message, this change require only to change the commit_message template file.

https://help.github.com/articles/creating-a-commit-with-multiple-authors/

Comment 1

[GitHub Bugzilla PR Linker]

Attachment: Link to GitHub pull-request: https://github.com/mozilla/pontoon/pull/1173

Comment 2

[Axel Hecht [:Pike]]

We should do for attribution what serves us best. There are a couple of different approaches, and I wonder how they relate to the studies we're doing with OI.

I'm wondering how co-authored-by compares to multiple commits to multiple commit authors.

One question I have about co-authored-by in particular, is if that's already loaded with connotations in our studies?

For example, I could have seen me using this feature on a patch I wrote the other day, where one function in that patch was actually written by stas and not me. I wrote the tests for that function, though.

In l10n contributions, that actual collaboration isn't what we're doing, and I'd not want us to use a feature on gh that's misinterpreted, or makes comparing l10n contributions to other contributions harder still.

For the sake of the argument, I think that multi-author commits are not that much harder to implement than a different format in the commit message. Separate commits are probably the hardest investment, but come with the benefit of actually working on hg commits, too.

Comment 3

[Rubén Martín [:Nukeador] ❌ [away till Aug 31st]]

For the studies that we have done, as long as it was clear each user number of contributions we were fine.

Comment 4

[Matjaz Horvat [:mathjazz]]

Pike, do you suggest that we:

• wontfix this bug,
• as a short-term solution implement multi-author commits,
• as a long-term solution implement separate commmits (one for each author)?

Comment 5

[Axel Hecht [:Pike]]

Yeah, I think multi-author commits are the better semantic for now.

Comment 6

[Matjaz Horvat [:mathjazz]]

OK, closing. Sorry about that Tomer. :/

Bug 1522112 should also work better across different VCS systems.

Comment 7

[Matjaz Horvat [:mathjazz]]

Fixed in https://github.com/mozilla/pontoon/pull/1601.