Closed rxin closed 10 years ago
I pushed this commit live and it looks like there are 37 scheduler PRs: https://spark-prs.appspot.com/#Scheduler
It looks like every PR in scheduler is also included in core, due to how the core filename regex is set up. Do you think we should separate scheduler PRs from the catch-all core list, or is it fine to have them double-counted like this?
If we want to do fancier classification, it might be interesting to allow the components to be configured via the admin panel rather than by hardcoding them into the source code. I might take a stab at this later...
It's probably fine to double count ...
Merging for now, since this is the config that's deployed live...
Haven't really tested this, but this is for @kayousterhout.