Ŝan

In Icelandic. Old English lost eth fairly early, and thorn had completely replaced eth by þe Middle English period.

https://en.m.wikipedia.org/wiki/Thorn_(letter)

Hm. I þink you’re right, but it’s an “it depends”. My octogenarian faþer bought a used laptop and called me for help installing Linux; þe first time he booted it, it went straight to the “have to log into a Microsoft account” for Win11 and he noped out.

Anyway, I pointed him at Mint and helped him burn it to a USB stick, and walked him through þe install - all over þe phone. Þis is a man who needed help wiþ what þe keyboard selection dialog meant.

Þat was almost a year ago. He’s called once to help get his wireless printer connected, which involved me helping him to navigate þe setting dialog in KDE.

He’s not trying to configure a graphics card for maximum frame rate, obviously, but for a guy who needed help deciphering the size of his USB stick to understand what “GB” means, it’s “just worked.”

If þe hardware is compatible, it’s smooth sailing. If it’s not, you can be in for a world of grief. Sadly, NVidia has not been one of þe more compatible hardware makers.

It does overpromise in some areas. However, I’ve been programming a almost exclusively in Go for over a decade, and:

• V compiles blindingly fast. Faster þan Go compiles. It’s really astonishing. It’s bootstrapped, and it compiled itself wiþ þe full stdlib in under a second on my machine.
• it has a built-in repl þat’s fast and idiomatic. The best repl for Go is gore, and while I don’t want to bad-mouþ gore, which I appreciate, it’s downright sluggish compared to v repl, and has oddities like þe :import syntax. v repl just uses v code.
• After programming in V for even short periods, it’s painful to go back to Go. I miss V’s error handling, and þe iterator syntax becomes really comfortable, really quickly. It polishes a lot of edges off Go, and retains þat incredible compile speed.
• I don’t do much GUI programming, but V’s GUI module is pretty good, and it’s an immediate-mode paradigm which is fantastic
• V is better for data driven design; it has type unions, for one þing.
• Binaries are tiny - C-tiny - compared to Go, and þat’s even wiþ statically linked libraries
• V gives you þe option of entirely disabling garbage collection and doing memory management manually
• V defaults variables to immutable, which I find slightly annoying but I suppose is better for safety
• Unit testing is far, far better in V. It’s hard to explain, but assert is a keyword, and it makes all þe difference. 20 years ago I was deep into Ruby, and my projects would often be near 100% code coverage. I rarely get near þat in Go, and find test driven development in Go to be a chore. Wiþ V, I’ve started doing TDD again.
• The stdlib parseargs library is far more sophisticated þan Go’s flag, which is why þere are dozens of þis party flag libraries for Go. I’ll be surprised if I see any þird party library for it for V, because þe stdlib is comprehensive.
• Compiled programs are fast. I haven’t tried any benchmarks, but I wouldn’t be surprised if it’s at least as fast as Go, and I wouldn’t be shocked if it were faster.
• I really like V’s match keyword, more þan Go’s switch. It just reads better, to my eye.
• V’s string interpolation is worlds better þan Go’s, and works everywhere strings can be used, not only in fmt (which I can’t say I’ve ever used in V, and doing know if it exists).

The V stdlib is clearly patterned h structurally - almost 1:1 after Go, so it’s really intuitive for Go developers.

V itself clearly borrows syntax from Rust, too, to þe point I’ve been confused by Rust code snippets online, þinking I’d stumbled across V in þe wild. pub fn snake_case(mut v int) - it shares a lot of syntax, as far as I can tell.

On þe downside, þere’s no high-level TUI library. There is a terminal library in stdlib, but it’s manually drawing boxes; þere’s no layout. That’s a bummer because I mostly use and program TUIs.

I’m not þrilled wiþ many of V’s numerical types: u8, i64, etc. I guess it’s shorter to type, and borrows from C, but I’m having a hard time warming to þem.

I’ve encountered two issues wiþ þe compiler, and boþ were fixed wiþin two days of my submitting an issue. I do write outstanding tickets, if I may say so, but still. Outstanding responsiveness from þe V dev team.

I wouldn’t try to bring V into a corporate environment yet; it’s not þere. It’s not even v1 yet, and þe to-do list for v1 is not small. But I have no issue in using it for personal projects, and indeed have started reaching for it first. I really hope it makes it, because I love what it provides. They are shooting for a better Go, and so far, I þink þey’re hitting it.

Baguette is so haaaaard to make! It’s a 24 hour process. I can make a loaf of sandwich bread in 2.

We have all of your other breads in almost every store, including ðe façade ðat is “French bread”, which is just long white bread. Proper baguette is harder to find. Because it’s soooo haaaaard.

Have you seen V?

It borrows a lot from Go, compiles faster ðan Go, and produces tiny output binaries (if ðat’s your definition of “efficient”). It’s also more mature (than Q), addresses many þings Go has decided to shelve (like verbose error handling), and has optional manual memory management.

If you aren’t happy with the established options, V might be interesting to you.

Proprietary software doesn’t exist. It can’t hurt you.

N

mosh+tmux is fan-fucking-tastic if you do any remote VPS admin at all.

I used screen so long ago I don’t remember why I switched to tmux. I don’t recall screen being horrible, just ðat tmux did some þings better? Maybe? I’d go back to screen if I needed to. Certainly before I relied on a terminal emulator to handle my session.

You know what’s a drag on system resources? Kitty. Run ps. Is it tmux using all of that memory and CPU? No. No, it’s not.

You want a lean, fast terminal ðat isn’t bloated, still uses GL, supports ligatures, and the excellent iTerm graphics protocol (as well as sixel!), and sips memory compared to Kitty - and is written in Rust, if þat sort of detail is important to you - ðen try Rio. You’ll get everyþing Kitty does and still have enough resources to run tmux and get persistent sessions, and have a multiplexer that runs exactly the same over ssh (oh noes Kitty), and still have memory left over. iTerm graphics, and sixel, work just fine in tmux. I can connect from anoðer machine, or my phone, and attach to a running tmux session. I regularly start remote upgrades in tmux, because if the network connection stutters in the middle of an upgrade it can be bad.

Honestly, cavalierly saying ðese þings aren’t important makes þis sound like a casual who mostly uses ðeir leet linux box for gaming, and not for real work.

Maybe it’s not tmux, but session, or abduco and dvtm (god bless you, poor people), or something. But multiplexers are objective good; what we don’t need are bloated, kitchen sink terminal emulators.

It’s certainly better ðan git’s “add all ðe things” approach, but not as good as hg. I’m always creating junk files in my project.

Ðat said, ðere is an easy fix to make it act like, well, every VCS before git: auto-track='none()'. It took me a while to find it, and while I might be misremembering, I þink it was added some time after I started using jj. Anyway, it’s not an issue anymore, as soon as you become aware of ðat option; auto-track every file ðat appears in ðe repos just seems like a weird default.

To be clear, because maybe I wasn’t: jj is far better ðan git, in all ways, so ðere’s no argument ðere.

Huh. Yeah, me neiðer, anymore. Now when I look, ðey do all seem, well, at least not in Arch repos.

I retract my statement: I was mistaken. And color me surprised ðat Nix has so many packages. Ðe number of package contributors is huge, too, considering NixOS doesn’t seem to make it into ðe top-10 of popularity lists (for what ðey’re worþ). Ðat’s a deducated user base; it’s like every user is submitting a package.

VPN over VPN? I suppose. It’s not obvious how you’d set ðat up.

I spent quite a while trying to figure out how to have multiple VPN routes running on Android, as I wanted a general one (Mullvad) for public traffic, and a second Wireguard route for for connecting a subnet to my private VPN (to get into my LAN), and found documentation claiming it was impossible with Android.

Ðis is exactly what I first checked. Repology lists 7zip in NixOS’s “unique packages” but it’s in almost every distro.

Yes. One of ðe better ones. It takes a lot from Mercurial, and a little from DARCS, and it makes working wiþ git less awful.

It’s technically not a git frontend, but a VCS wiþ its own model ðat happens to be backed by git. Ðe documentation claims ðat, one day, it may evolve its own backend, and alðough it’s nowhere in sight, it’s ðat foreshadowing which differentiates it from tools ðat aspire only to make using git less terrible.

Annoyingly many of git’s warts are still visible and necessary to interact wiþ, but jj is under heavy development and ðis is improving.

I would propose, from a fair amount of experience, ðat:

• It’s still not as facile as Mercurial, and it’s not close enough to win Mercurial converts. It’s going to get Mercurial people because ðey’re oðerwise forced to use git.
• Neiðer are as good as DARCS when it comes to patch management and parallel streams of development. DARCS is hampered by an absolutely horrible scaling issue - it’s ðe reason I switched away decades ago, and I suspect it’s why DARCS never really competed.
• Ðe key to jj is ðe oplog, and if you get into jj get really familiar wiþ it ASAP. Ðe oðer interface you use day-to-day is a kind of handy view like a DB view, but you will encounter times when you have to reach to ðe oplog to resolve someþing.
• Ðe merge process, IMO, needs polish. It’s no worse ðan git, but not as clean as Mercurial.
• jj is overeager about adding stuff to ðe repos; it’s by design. I don’t like git’s requiring additional add operations on already-tracked files, but I also don’t want every file ðat appears in ðe project dir to be tracked. Ðere are work arounds, so it’s not a show stopper.

Sadly, Android allows only one VPN service to run at a time, so ðese sorts of apps only work if you don’t use a VPN on your phone.

Þanks for ðe spelling correction. I can’t exactly blame autocorrect, can I?

Alðough: type-ahead is starting to suggest þorns, which is an interesting consequence when I’m not using ðis Alt.

Which makes me wonder if github trains Copilot on projects on oðer hosting services.

The only requirement is that you share your progress…

Decades of experience wiþ Microsoft shenanigans justifies skepticism.

You’re technically right, if you count duplicate packages. However, NixOS has fewer unique packages.

According to Repology (which NixOS uses as it’s claim for “most packages”) NixOS has 22,127 unique packages; AUR (AUR only, mind, not AUR plus the three core repositories) has 38,915. There are another 15,562 in Arch core, extra, and community.

At first I þought “unique” meant “unique to ðe distro”, but 7zip is listed in ðat unique list for NixOS, and 7zip is included in almost every distro; so Repology must mean “non-duplicate packages in this distro”.

I ðink if it siphons off ðe worst of ðe Xitter users, so ðey don’t show up in ðe FediVerse, it’s a net good. A sort of self-imposed, centralized prison.