311 karma · joined October 30, 2019
https://tmpod.dev
tom@tmpod.dev
But again, could be wrong about Ktor specifically.
Even then, everywhere you go, unless they're really small or old places that still don't accept anything other than cash, you can pay with MBWAY, through a QR code or NFC. All POS terminals support that nowadays. You can even add an email address and VAT number and newer terminals automatically skip printing the customer receipt, send it to your email and give the VAT number to the merchant (not entirely sure how that last one works, but it's there).
The language ends up being more complex, but I find it a joy to use; going "back" to Java projects always leaving me wishing I could use Kotlin instead :P
Yet it's true, Java has improved quite a lot in the last decade. It is ahead of Kotlin in a few things, like pattern matching (though not incredible, it's still better than the nigh nonexistent support in Kotlin).
They should take your velocity vector into account, or something along those lines, to minimize rotations, even when not following a route. It just leads to a ton of unnecessary screen redraws and jerky movement.
Furthermore, as I've been driving in a region with a lot of tunnels, I wish they implemented a better system for when you pass through one while following a route. It animated the arrow for a bit but then stops and only jumps forward when you get out. I wish it would keep going with a speed based on your entry speed (and maybe the tunnel's speed limit, if available), taking the routed exit if there are any forks inside. This would make the happy path much better and would not really worsen accuracy, since it isn't accurate in any way as is right now. You could tweak the arrow's color/shape to indicate it is showing a predicted route.
I've been an active lurker in the self-hosting community and I'm definitely not alone. Nearly everyone hosting public facing websites, particularly those whose form is rather juicy for LLMs, have been facing these issues. It costs more time and money to deal with this, when applying a simple User-Agent block would be much cheaper and trivial to do and maintain.
sigh
Those toggles actually mimic real hardware that used to be fairly common. I find those should be preferred over checkboxes for anything that takes immediate effect. If they don't, and you're collecting a bunch of options at once, in a form, then use checkboxes.
It's very minimal while having useful features: - exit codes, even for pipelines - git branch, status (displayed as a dot if your tree is dirty) and ahead/behind counts - command execution time (if above some configurable threshold) - truncated/minified $CWD, always maintaining the git root's name (I sometimes like it, sometimes don't; fortunately, it's very easy to change) - current vi-like mode (I don't use that)
It's very fast and async (prompt repaints don't block your input or running commands), and totals 132 lines of fish (according to cloc[1]). It's also very customisable through variables, which can be declared as universal to instantly change on all sessions you have open.
If you're on fish and like this feature set, definitely give it a shot, or at least look at the code as a base for a bespoke prompt :P
[0]: https://github.com/jorgebucaran/hydro [1]: https://github.com/AlDanial/cloc
Has anyone used fish + autuin? What are your thoughts? I'm curious to know if it is worth the extra tool, maybe I should just try it :)
Briar looks really nice for people that really need the extra security and privacy, though.
I've been interested in SimpleX lately, but I don't really have anyone to properly test it with.
I do a web search and get a list of results for the English version, open it and it automatically changes to PT-br as soon as JS loads, but then, after a few seconds, once the page is fully loaded, it jumps back to English, while keeping the /pt-br/ slug in the URL...
What the heck??
And yeah, the translation is very obviously not human made or reviewed. Furthermore, I'm PT-pt, so the differences to PT-br just make the experience even more annoying X)
The core office suite is very good and people are very used to it. It also seems to be the hardest to truly replace, in my opinion.
LibreOffice and OnlyOffice are a good step forward but they're not replacements yet, for a lot of people, due to a simple fact (already mentioned somewhere here): people have been using this software for decades and are very used to the way they work, and replicating that, especially nicher things, is not trivial and takes a lot of work. Still, I'm hopeful.
Outlook, Acrobat and Teams may be easier to switch, particularly the last two. Outlook (and Exchange) has some extra nicities compared to plain standard email (from the top of my head, I recall read receipts and automatic responses for when you're away being important ones), but there are some nice projects tackling both the client and server side parts of those features. Acrobat could maybe be replaced by a reader like Okular, given a bit more polish, and Teams is so bad and often used in such basic ways that it could trivially be replaced by something like Mattermost (though I personally much prefer Zulip's model).
Again, all this is based on my experience and certainly won't apply everywhere.
You also have Pumpkin [1] and Ferrum [2] that are trying to implement a vanilla-like server, but are a little ways off still. The former has an interesting plugin system already; works with shared libraries.
One of the neat things of these Rust implementations, apart from good base performance (though Java's JIT can be pretty good on this type of software), is the memory footprint, one of the main painpoints of hosting large server networks. The official server implementation is extremely memory hungry, with lots of allocations everywhere; even with optimization forks like Paper, you can't really run away from it too much. Minestom is also nicer in that regard, but not on the same level as Pumpkin or Ferrum. The Minestom people are also waiting on Project Valhalla to unlock some optimizations on that front.
Another interesting approach would be Erlang or some other BEAM language, but there isn't much activity on that front either, as far as I know. McEx [3] was pretty neat, but it has long since been quiet.
[0]: https://valence.rs [1]: https://pumpkinmc.org [2]: https://github.com/ferrumc-rs/ferrumc [3]: https://github.com/McEx/McEx
Minestom isn't a plug-and-play server, but rather a framework for building custom servers. It's geared towards cases where performance is important (think big servers) and implementing your features is faster than pruning an existing implementation. There are several libraries that implement vanilla behaviour and that you can just plug onto your server, but the main idea is that you tailor the server to your specific needs.
Minestom is an open-source library that enables developers to create their own Minecraft server software, without any code from Mojang.
The main difference between Mojang's vanilla server and a minestom-based server, is that ours does not contain any features by default! However, we have a complete API which is designed to allow you to make anything possible, with ease.
This is a developer API not meant to be used by end-users.
---
I believe the target audience is people developing non-vanilla/survival multiplayer servers, like minigames, PvP duels, etc.
[1]: https://lsposed.org [2]: https://github.com/M66B/XPrivacyLua / https://github.com/0bbedCode/XPL-EX [3]: https://appops.rikka.app