I'll be taking notes in this topic while I work on drafting the agenda items for tonight's kickoff meeting
here are high-level items that I have written on my whiteboard in my office (not all are suitable for meeting topics):
scheduling a 2nd meeting in alternative timeslot (for e.g.
@davidtwco ) review cgu policy/granularity (e.g. one-per-fn vs one-per-mod; cgu merge strategy)
review/revise benchmarking on perf
instrumentation (e.g. are there additional details we should expose to
problems with rustc bootstrap (?)
disk usage from incremental build artifacts
structure of group. (e.g. I'm working under assumption that we are a self-selected domain working group, and that we will spin up individual project groups, presumably only a few at a time, for each large effort we take on)
how do we prioritize which projects to work on? (We may need to just take some leaps of faith rather than try to devise concrete experiments to guide such decisions.)
To the last point: Can we find projects independent enough they can be worked on in parallel? Can we expose more (unstable) knobs to make experimenting easier?
Last updated: Oct 21 2021 at 21:02 UTC