Heya, author here - sorry if my module names make you feel uneasy. It’s totally understandable that at a glance naming might seem confusing; many people write JS, and equally many approaches to naming things. This is generally how I name my things:

• General tooling has plain, boring names, e.g. assert-html, on-idle, scroll-to-anchor
• High level frameworks have fancy names, more akin to brands. E.g. choo, merry, bankai
• Packages that are parts of the framework’s ecosystem have a prefix to group them together:
  • nano* for browser packages, e.g. nanobus, nanorouter, nanoraf
  • server-* for HTTP server packages, e.g. server-router
  • hyper* for p2p things, e.g. hypercore, hyperdrive
  • *ify for browserify things, e.g. sheetify, uglifyify

Writing this out makes me realize it might indeed be hard to get into if you’re not already used to it. I agree discoverability can be much improved upon; both by tooling (e.g. GitHub allowing us to group things outside of orgs), and humans (your comment made me pause and reflect on this, thank you).

I hope this somewhat explains how I (and perhaps some others) approach naming JS things (:

I feel this is a toxic part of the JS community. I recently had to patch some nodejs code and searched “how to lower case a string”. The top suggestion was the fancy ChangeCase dependency when a simple String.ToLowerCase from the standard functions did the trick with no downloads required.

Actually I believe it was originally designed to support the normal use case of ccache and recently extended to support Rust. So it would be more accurate to describe it as “like ccache but can use an s3/redis store instead of local disk and has a windows binary, and supports caching Rust compilations.”

In addition to keepalive, you know when you accidentally cat a gigantic log file over ssh and have to wait hours for it to scroll by or kill the session? mosh prevents this and jumps to the end of the out. Granted, you lose scrollback due to this, so it’s a tradeoff.

Mosh keeps the session alive when my laptop sleeps or I close the lid, so I don’t have to re-authenticate. If I lose the session over regular ssh, then I believe the tmux instance would die instead of simply detaching, so I wouldn’t be able to resume anyway.

tmux is for N-to-N sharing of a terminal: run multiple programs on the server at the same time, see them from some number of clients, isolate them from client-server connection termination.

mosh is for making client-server connections more robust. If you close your computer in one place and reopen it in another, it keeps going instead of needing you to hit up and enter twice to reconnect and tmux attach. This is invaluable if some part of your link is flaky, like oversubscribed conference wifi, or the connection between a moving vehicle and a terrestrial network.

It literally does.

Perhaps I’m mistaken, but I believe the claim is “single header”, not “single file”.

It’s like a monorepo for your program! /s

I would have preferred something like sql amalgamation, but if the author can manage to develop while things are all in one file and make it work I Would say it’s fine.