Repository navigation
Store list of collaborators and TSC in a JSON file #23623
Description
Is your feature request related to a problem? Please describe.
As mentioned by @joyeecheung during CollabSummit, node-core-utils reads list of collaborators and TSC members from README.md
Describe the solution you'd like
It would be helpful if we have the list stored in JSON, and is read from all places where it's required.
Activity
joyeecheung commentedon Oct 12, 2018
This can also help other projects to reuse node-core-utils if they just have a similar list of collaborators
Refs: nodejs/llnode#242
ChALkeR commentedon Oct 12, 2018
Is it supposed to be auto-generated from README.md, or is it supposed to be maintained manually?
I would prefer the former.
joyeecheung commentedon Oct 12, 2018
@ChALkeR I guess it wouldn't be too hard to auto-generate this considering we have the parser already in https://github.com/nodejs/node-core-utils/blob/master/lib/collaborators.js and the JSON.stringify basically just looks like https://github.com/nodejs/node-core-utils/blob/master/test/fixtures/collaborators.json
joyeecheung commentedon Oct 12, 2018
The problem is...who runs the generator? Should we make it part of the onboarding process...?
joyeecheung commentedon Oct 12, 2018
Or, should we generate things the other way around? (JSON -> markdown)? At least I personally find it easier to fill a JSON object than to fill a markdown row with a bunch of markers, and put it into the right place alphabetically..
refack commentedon Oct 12, 2018
I'd say store the json in an HTML comment inside
README.mdthen at onboarding, update the object, then run the generator, then commit.richardlau commentedon Oct 13, 2018
Is there a reason ncu can't be pointed at the relevant GitHub teams and retrieve via the API?
refack commentedon Oct 13, 2018
I can think of two minor reasons.
sagitsofan commentedon Oct 13, 2018
Does this JSON file needs to be part of the
node-core-utilsproject or it can be under thenodeproject?Why not running this collaborators generator whenever the api document generator run?
joyeecheung commentedon Oct 13, 2018
It's not necessarily true that the name and the email from people's GitHub profile match the name and the email that they want to use in this project. Also there's no pronouns in GitHub profiles.
joyeecheung commentedon Oct 13, 2018
I'd say it should be in the project otherwise you'll need to submit PRs to ncu to add yourself to the list of collaborators, which kind of misses the point of onboarding - landing a PR in core.
richardlau commentedon Oct 13, 2018
.mailmap? And why does ncu need to know pronouns?joyeecheung commentedon Oct 13, 2018
@richardlau
That probably works as well if we regenerate the current
.mailmapto include people's github handles (not everyone makes their email available in their GitHub profile so you can't get emails from handles, it could be some no-reply email or null, and there is no guarantee that you can get handles from emails either)Was talking about generating markdown from JSON (or another file like
.mailmap) and make that file the source of truth.Trott commentedon Nov 16, 2018
Is there a clear first step to take here?
joyeecheung commentedon Nov 19, 2018
@Trott I personally think it would be more efficient if we just try to solve #24117 - and the "list of collaborators" would just be a subset of the data requested there, if, say, we add the concept of
committersto the metadata, so there will bejoyeecheung commentedon Nov 19, 2018
Merged into #24117