Nushell core team meeting 2022-10-12

Attendees

  • Darren
  • JT
  • Stefan
  • Michael
  • Reilly
  • Jakub
  • Andrés

Agenda

  • How to not burn out the core team ;)
    • Q: Release cycle length 3wk vs 4wk?
    • Onboarding additional people
    • Implementing structures to remove some of the burden for the core teams or engage more active contributors (without having to have full core team responsibilities)
  • What's the plan for the parser?
    • Are we using treesitter?
    • Can we also include the new syntax?
  • Dataframe plans
  • https://github.com/nushell/nushell/pull/6649 Variable scoping changes?

Discussed Topics

Avoid burnout - Sustainable project development

We have several contributors who might be able to help out on the core team

Action item: list potential core team candidates in #core-team

Release cycle length:

  • JT not a big fan of 4 weeks
  • Are releases polished/featureful enough after 3 weeks? JT thinks yes
  • Darren: would be good to have fewer emergency fixes

Treesitter as a grammar backend

  • Darren spend exploring it, seems productive
  • Lhkipp has treesitter grammar under unfortunate license (just for ide stuff)

Way forward: crate with the treesitter parser
Benefit: not having to rip everything out or having to do an engine-alphabet rewrite

Dataframe plans

Don't have too many dataframe users right now. Is it worth keeping around?
- users who ask the most questions about df: kira and eggcaker

Darren: there are situations where users end up bumping against the limits of standard Nu tables, DF is good for that

datafusion with SQL support pathway -> arrow-rs
polars currently known to support a range of ops -> arrow2.

Discussion among the core userbase in "biggish" data to decide?
Source of friction command differences with the base data structures

Select a repo