Organizing the rust-memory-model repo


#1

I’ve been trying to organize the rust-memory-model repo a bit. Here is the scheme I attempted:

  • Right now, all the activity is in the issues, not the files themselves. Issues are basically being used to catalog interesting models, code-examples, optimizations, and collect conversation on that topic.
    • I plan to move the examples out of the files in the repo and into issues; some of them are already covered
  • K-tags indicate the “kind” of an issue:
    • K-code-example – some code or a pattern that may be legal or not
    • K-optimization – something the compiler would like to be able to do
    • K-model – a proposed model for how to think about whether unsafe code examples are legal or not
    • K-task – something we have to consider whether to do or not
  • T-tags indicate the “team” this most affects:
    • T-lang – this is a rule for the language itself
    • T-libs – this is primarily a library issue

Sound good?


#2

maybe with some form of export into the git repo as a backup so you’re not relying on github issues exclusively forever?


#3

Ultimately, once we’ve decided the answers, the plan is to move the decisions into files in the repository, which should summarize the discussion as well.

I agree that it would be nice to capture GH issue history as well, of course.