Skip to content

Latest commit

 

History

History
64 lines (58 loc) · 4.44 KB

2022-01-19.md

File metadata and controls

64 lines (58 loc) · 4.44 KB

Meeting from: January 19th, 2022

Open RFC Meeting (npm)

Attendees

  • Darcy Clarke (@darcyclarke)
  • Ruy Adorno (@ruyadorno)
  • Owen Buckley (@thescientist13)
  • Nathan Fritz (@fritzy)
  • Wes Garland (@wesgarland4)

Previously...

Agenda

  1. Housekeeping
    1. Introduction(s)
    2. Code of Conduct Acknowledgement
    3. Outline Intentions & Desired Outcomes
    4. Announcements
  2. Issue: #511 [RRFC] remove `npm-shrinkwrap.json` from the list of unignorable files - @ljharb
  • @ljharb was using shrinkwrap in a specific project and found out that npm-shrinkwrap.json is part of the list of files that can't be ignored
  • @ljharb would like for it to be possible to ignore the file in case it's in either .gitignore or .npmignore
  • @darcyclarke it is likely that npm-shrinkwrap.json still has some usage in the ecosystem
  • @ljharb it would be an opt-in behavior to stop shipping npm-shrinkwrap.json in case it's in an ignore file
  • @mylesborins recalls a recent episode in which npm-shrinkwrap.json helped unblock the node core team to handle a broken transitive dependency and being able to rely on it for globally-installed users by shipping that file, just to add that there's still lots of value in that feature
  • @darcyclarke would still prefer to ship this in a major version in case some users are inadvertly ignoring something that they would not, trying to stay on the safe side
  • @mylesborins would like to bring in the conversation of having some sort of validation prior to publishing a package
  • @ljharb would like for any files extraneous to that package to be highlighted at the moment of publishing
  • @ruyadorno related approved RFC publish prompt
  • Actions:
    • @darcyclarke backlog a ticket to add messaging about diff of files include in a package (comparing .npmignore & .gitignore to files)
    • @darcyclarke backlog a ticket for next major, dropping npm-packlist opt-in of npm-shrinkwrap.json files
  1. Discussion: #508 Show which tag will be used in the output of npm publish ~ @fvictorio
  • @ruyadorno sounds like a good idea, we should just backlog it
  • @ljharb also likes the idea of improving the emails with more relevant info, can be useful to highlight any security threat, in case access gets compromised
  • Actions:
    • @darcyclarke take feedback on improved publish email information back to registry team under the assumption its a security improvement
    • @darcyclarke backlog ticket to address this feature work (ie. improve npm publish output)
  1. Discussion: #440 Peer dependency groups ~ @Jamesernator
  • @darcyclarke
    • seems like this could be supported by a future "package distributions" RFC
  • @bradleyfarias
    • node.js recently shipped a change which caused native C modules to not be importable, this could be related
  • @ruyadorno
    • sounds a little bit different from the distributions RFC idea in which it means to give a more flexible way for transitive packages to consume alternative packages given that any of the alternatives is provided as a peer dependency
    • it sounds like an interesting idea worth exploring, maybe turn this discussion into an actual RFC?
  • Action:
    • @darcyclarke to create RFC for "package distributions" by next week
  1. Discussion: #501 npm cli tool should report name similarity problems in --dry-run ~ @msikma
  • @ljharb
    • 👍
  • Actions:
    • @darcyclarke to circle back with registry team on getting an endpoint for package name validation
  1. Discussion: #496 Mention workspace in run-script output header ~ @smhg
  • @fritzy
    • should have brackets or some kind of visual indication this output is specific to the workspace name & not being logged from the command/script
  • Actions:
    • @darcyclarke to backlog adding this output to --workspace aware commands (ex. run, exec ...)