errgroup is the thing to use for concurrencyBeen a while since I checked in with tunable worker pools and it appears errgroup is the thing to use in this new context.Context world. I couldn’t find my own worker pool abstraction that I did some years ago, but probably just as well. Anything without Context these days is really, um, bad. I did find neilotoole/errgroup1 which is a nice implementation that includes being able to throttle up and down the number of concurrent jobs in the group.
I won’t really need this until I get the other framework for keg grep done, which I can do in a linear way for now and layer in the errgroup later.