What has happened

  • Lars has discussed using Subplot at work, but hasn’t started using it yet.

  • Lars has added pointers to GitLab to the Subplot website.

  • Lars has files some issues on the GitLab issue tracker, and created a milestone for the iteration that starts today.

  • Lars has started using Subplot to sketch out ideas for hobby projects.

  • Daniel has written a crate, marked_yaml, for parsing a subset of YAML, with location information for values, for better error messages. He will write a similar one for Markdown, and he and Lars will discuss it then.

Discussion

  • We discussed scalability / stress testing for Subplot. We decided to write a tool to generare “stupidly big” input files so that we have an idea if they work at all. We’re more interested in checking that RAM requriements are insane than that things are fast. For example, a stupidly many scenarios, with stupidly many steps, using stupidly many bindings, and both a PDF can ge benerated, and a test program generated and run.

  • We discussed various strategies for supporting writing large documents with Subplot. We pondered different “include file” options, but decided that, for now, an external tool is the way to go. We’ll revisit this later if it seems worth it.

  • We decided to add plantuml support to Subplot. Further, later on it seems like this kind of support should be configurable: given a fenced code block of this type, use that shell command to convert it to an SVG image.

  • We decided to add support for the citations Pandoc extension.

  • We decided that scenario keywords should always be case insensitive, and so should the regular expressions matching the rest of the scenario step. However, we further decided that the binding should be able to specify a case sensitive regex, though we couldn’t think of a use case yet.

  • We decided to wait until Subplot supports test programs in Rust before we start a rewrite or the jt in Rust.

  • We discussed merge strategies in GitLab. Daniel configured the project to always use a standard merge commit. We may want to revisit this if we get more contributors.

  • [https://gitlab.com/larswirzenius/subplot/issues/8] (Scenario steps can’t be continue to the next physical line): we discussed various ways of indicating line continuation. The discussion was inconclusive except that whatever syntax the implementation supports should be easy to change.

  • We may start drafting a blog post to introduce Subplot and try to attract one or two volunteers to use it for real. However, we won’t post it until after the current iteration is finished and issues marked for it resolved.

  • Lars will ask a co-worker about helping with website structure. Failing that, we’ll approach other friends.

  • We reviewed the roadmap and decided on how to update it.

Actions

  • Lars to file issues based on decisions above.

  • Daniel to continue work on marked_markdown.