Clojurians Log v2

What’s more idiomatic with core-async? 1. keeping track of open channels and making sure to close all of them when we are done with them 2. not closing channels at all when we are done with them and instead relying on them being garbage collected, and if some underlying data source still produces data for a while after we closed it, inputs will just get parked and eventually will be GC’d

I usually close from the producer side and let the closes cascade downstream

lots of core.async "mini-processes" will close their downstream when their upstream closes (see the close? option on a lot of fns)