What has happened

• We got everything done, bar finalizing the merge request to add shell as an output language for codegen. That one lacks a small modification to be done.
  • Avoid Pandoc::add_filter
  • Do lints in one place
  • Bug fix: first scenario step uses alias
  • Simple patterns
  • Check embedded files have unique names

Discussion

• We reviewed and updated all open merge requests and issues, and picked issues to work on for the next iteration. We picked five issues to work on, two to discuss, and to as a stretch goal if there’s extra time.

• We discussed whether we should require Subplot to be buildable with the Rust toolchain in Debian 10 (buster). The changes to support shell trigger a panic in that rustc (the same upstream version doesn’t). We decided that since Subplot will not go into buster, and we don’t use that rustc for our own work, we don’t need to be buildable with that rustc.

  We’d like the rustc in the next Debian stable version to work to build Subplot, though.

• We discussed using the “converntional commits” specification for Subplot. We decided that since Subplot is such a small project, it’s not useful enough at this time. We may reconsider later.

• Reviewing all open issues in these meetings is useful, but also time consuming. We agreed to only do it every other time from now on.

Actions

• Lars to make the remaining changes to the shell support merge request. Daniel to re-review.

• Lars to set his personal CI to use rustup for installing Rust, instead of buster’s rustc.

• Lars to provide Daniel with a reproducible test case, crash output, and minimum commit range to reproduce the rustc crash. Daniel will have a look if he can see what causes the crash.