node dev piping broken

[jordanh]

Issue - Bug

@mattkrick:

piping is broken & i'm not sure why. When it auto-restarts, i get EADDRESSINUSE almost always. Also, it only worked in development, and we should be debugging server code in production.

@jordanh:

This has been driving me nuts. Let's create an issue. I suspect it's because the SC worker process isn't exiting.

I want to track this down either via a git bisect and a test script or adding a hook on exit that looks at node's run queue.

• OS: OSX
• Estimated effort: 2 points (see CONTRIBUTING.md)

[jordanh]

So this is goofy. I tried to reproduce this on the airplane over to Germany and now I couldn't get it to break. Before I began I toasted an reinstalled node_modules. Is this still failing reliably for you @mattkrick?

[jordanh]

@mattkrick you pulled piping out of the codebase right? I bet it was related to why npm run build:server was failing. I wonder if it would work again if you re-added it?

[mattkrick]

Yup. TBH I don't miss it, but we can add in back in. If we do, we should turn it on only if a DEPLOY environment variable is false since it's only really useful when we're in __PRODUCTION__ (if we're in dev mode that means waiting 15+ seconds for the server to restart & recompile the entire client webpack in memory).