atlassian-labs
Runtime performance optimizations
Current runtime code is, ironically, quite heavy during lots of re-renders. We need to optimize the HOT paths to make things faster.
ax()
This function does a lot. Ideally we want to memoize invocations so if it is called with the same args, it returns immediately. It already short circuits, but we need to handle the worst case better as well. There needs to be a balance between memozing everything, and just enough. We don't want to consume a ton of memory.
Style component
Related to #720. If it re-renders it will do a bit of work that isn't needed. Easy answer it to add memo back that checks the children.
This is also tangently related to #719 - we need to be able to measure before we change things. They should run in CI, and report on every PR.
