It is pretty weird, that is my exact setup too. But instead of alacritty I use iterm2, and instead of tmux I use screen, and instead of neovim I use emacs. But regardless, spooky!
I'm trying to tell from your comment as well as the others' comments whether you're running screen locally, and what benefits that brings above and beyond iTerm for local shell sessions.
For anyone else using tmux locally, I'd be interested as well.
Genuinely curious. I'd consider doing likewise if the reason is compelling.
> what benefits that brings above and beyond iTerm for local shell sessions
> For anyone else using tmux locally, I'd be interested as well.
I run tmux in iTerm2 on macOS; tmux is my window-manager of choice for terminal applications; shells, nvim, etc. I usually have 2–10 tmux “windows” (more like tabs) with 2–4 split panes in each (more like tiled windows). I generally never detach from my local tmux session.
Most things I use and love tmux for could probably be done with iTerm2 tabs and splits, or a modern text editor and its terminal integrations.
But, I can use tmux in iTerm2 or Terminal.app or Kitty or Gnome Terminal or urxvt, on macOS or Linux or FreeBSD and it works the same everywhere. Perhaps that's a key feature; decoupling your “terminal window manager” or “terminal desktop environment” from your terminal emulator (I guess that's like the X server in this analogy).
Also, tmux feels more keyboard-native… I can do everything from the keyboard, including navigating scrollbacks, finding/selecting/copying/pasting text, etc. Again, you can probably achieve at least 90% of this using a richly-featured terminal emulator like iTerm2, but tmux does it well, and it's portable.
Also, tmux feels more Unix… it has config files that can be git-managed, it has man pages, it is scriptable from bash/anything, etc.
That easy scriptability leads to some nice integrations, like I can configure key bindings in nvim that run commands (e.g. the test I have open) in a tmux split alongside nvim. Again, that's nothing that things like VS Code can't do too with their built-in terminal emulators etc. But it's nice.
(Despite loving tmux and iTerm2, the one thing I don't like is iTerm2's native tmux integration, where it uses tmux as the “engine” but replaces the UI with OS-native windows/splits etc. It's technically impressive, but mostly negates all the reasons I run tmux inside iTerm2 in the first place)
> (Despite loving tmux and iTerm2, the one thing I don't like is iTerm2's native tmux integration, where it uses tmux as the “engine” but replaces the UI with OS-native windows/splits etc. It's technically impressive, but mostly negates all the reasons I run tmux inside iTerm2 in the first place)
I completely agree. I'm sure there are people who love it but it seems a strange implementation since it takes the best bits of tmux and replaces that with the worst bits of iTerm.
The great thing about this stuff though, is that everyone can personalise and run the set up they prefer. Whatever that might be
I’m mostly the same, but I can’t leave the native smooth infinite scrolling of the Terminal.app
iTerm2’s tmux integration seems all I have ever wanted, but somehow I keep returning to the two extremes. It maybe just habit, or there maybe something subtle missing.
Kitty performs better and renders font more beautifully than Alacritty, and Alacritty's devs are openly hostile to macOS users, so I've stuck with Kitty. I really want a native Rust terminal emulator to be good, though.
> Alacritty's devs are openly hostile to macOS users
What do you mean by that? I noticed some bindings randomly stop working even though I'm explicitly configuring them -- ALT and left / right arrow are the most prominent example -- but I'm curious if your observations are rooted in other problems?
It likely broke because the dev won’t touch macOS with a 39.5' pole, so they obviously don’t test changes.
I tried offering to help get the macOS build notarized (financially and logistically) so that it would work without disabling gatekeeper and was laughed out of the issue tracker for using a stupid OS. I tried explaining that I don’t always control which OS I use (not that it matters) and that since I use all of them one of the major advantages of a terminal emulator is that it works consistently across platforms. They basically told me that I’m an idiot for using macOS and put me in the middle of their ideological battle with Apple.
The best part is the maintainer wouldn’t even approve a PR to add one sentence to the readme explaining the expectation that users need to manually add a gatekeeper exception, because helping macOS users onboard “has no value” to the project (to him).
i legit thought this was an incredibly popular setup, because i have three coworkers (+ myself) with the exact same config (alacritty + tmux + neovim).
Where does the undercurl come from? TFA credits kitty, which I understand to mean supporting undercurl in a terminal. My first experience of it was with the spell checker in Microsoft word circa 1993, does anyone have an earlier sighting?
if you read HN for more than a month you will probably see more than one post discussing them as well as their alternatives, or you can use search to see those posts. If you find that you like each tool for what it offers (and doesn't offer) then you will probably like them in combination. If not then not. Why would you expect someone to know you enough to say it's worth the hassle for you?
HN's policy about dupes has been consistent for many years (it goes back to pg, who started the site and ran it for the first 8 years or so), so I hope it's ok if I explain it a bit.
On HN, a post only counts as a dupe if the same story has had significant attention in the last year or so. If it hasn't had significant attention yet, or if it has, but more than a year or so ago, then reposts are fine. The reason is that we want to give good stories multiple cracks at the bat, i.e. multiple chances at getting attention. Otherwise it's too random—randomness is a strong factor in what does or doesn't get noticed on the /newest page and we want to even that out over time.
Pointing out dupes in HN comments is great, but only if you link to a previous post that has had significant attention in the last year or so. Otherwise HN users will get irritated when they click on a link and are taken to a previous submission that didn't get any activity. I hope that makes sense!
By the way, some of the posts you've flagged as dupes were actually invited reposts—we comb through old submissions looking for good articles that didn't get noticed the first (or second, or third) time around, and invite one of the previous submitters to submit it again. Invited reposts go into the second-chance pool (https://news.ycombinator.com/pool, explained at https://news.ycombinator.com/item?id=26998308), so they get a random placement on HN's front page. This is another device we came up with for trying to mitigate the randomness of /newest. We want to maximize the opportunity for the best (for HN) articles to get in front of the community. At that point, readers can make up their own minds if the article is of interest or not.
... as I frequently point out both dupes and previous discussions, I'll make a point of noting that the latter are permitted, especially where those are near the threshold of recency, or where previous submissions didn't cross the "significant discussion" threshold. I take that at about 20 comments, though I'm not aware of an official stance.
If you continue antagonizing people with your fake dupe claims and have a large proportion of your posts flashed, there's a chance that the entire account gets shadowbanned. And then most people won't see the subset of posts where you have a legit point either.
A small number of reposts is not only permitted by the rules of the doors, it is very often explicitly encouraged by the mods. Railing against that is really not productive, especially since those reposts are often some of the very best posts to the site.
