I keep trying the "native" solutions every so often, but every time I quickly hit some snag that makes me question why I'm not just using the solution that actually works. As an example, I just generated a new project using create-vite & added two subpath imports:
The second one (#/*) is similar enough to what I usually use (@/*), and it's supported in Node since v25.4.0! Yet when I try to import the file at projectRoot/src/router/index.ts using:
import router from "#/router/index"
VS Code shows an error: "Cannot find module '#/router/index' or its corresponding type declarations."
Now, imports from e.g. "#assets/main.css" work, so I could work around this issue - but this is what I keep experiencing: the native variant usually kinda works except for the most common use case, which is made unnecessarily awkward. For a long time this is what ESM used to feel like, and IMO it still does in places (e.g. directory imports not working is a shame).
Possibly a double tap strike with airstrikes hitting shortly after the Tomahawks
> Middle East Eye, citing survivors and first responders, reported a possible "double-tap" strike — a second explosion hitting the area shortly after the first, striking people who had taken shelter.
This isn't surprising at all. We double tapped the boat survivors a few months back too. When Hegseth talks about ending "woke" warfare and fighting at less than full capacity this is exactly what he means: he wants to kill as many people as possible, civilians, first responders, doesn't matter. It's not an accident or aberration, it's official policy. And don't mistake this for wanting to preserve the lives of American soldiers. Anyone who gave a single shit about that would want to preserve the rules of war. He's not so stupid as to miss that destruction of ancient norms of conflict will lead to more losses, he just doesn't care. Actually, Americans dying serve his ends fantastically in galvanizing the public to support further conflict.
Mark my words, we're either going to attempt to occupy Iran or glass it in the coming months. That was the goal from the beginning, even if Trump seemed a bit unaware of that at first.
The article is talking about two things at once and trying to pretend they are the same thing. The us hit the military base with a tomahawk missile, and the school was hit by a missile from "nooneistakingaccountabilityville". They are trying to act like the US hitting the military base proves that they also hit a school like a week ago.
No, the school was hit by multiple missiles, not one, around the same time the military clinic was hit.
You can tell yourself that multiple air defense missile all failed the same way and fell at the same place, but saying that a single missile (especially air defense missile) erased the school is wrong.
This does sort of reveal the genius in Israel partnering with the US on this. At any point both parties can deny responsibility, leaving plausible deniability for both parties.
It's not genius at all. We all know that both the US and Israel are responsible for spending trillions of taxpayer's dollars on illegal wars of aggression and genocide. We weren't fooled by W, we weren't fooled by Obama, or Trump, or Biden.
Look at how the world has watched the US starving Cuba for decades, to take one example of many. Every year, every country except the US and Israel vote in the UN to condemn the sanctions. And every year, no one actually helps Cuba, because America threatens insane consequences.
Not genius. Just threats. Not very smart at all, if we want a liveable planet.
What does it take, exactly, for Donald Trump to be formally accused of war crimes and arrested on sight if he ever visits some actually civilized countries?
This strike is a fuck-up. Could be a mistake, could be a crime attributable to a person somewhere in the middle-ish of the chain of command, or even at the very bottom. You need a pattern of such strikes to move the needle firmly into "intentional government-wide war crime" territory.
Last I heard 16 hospitals have been damaged and 7 are no longer able to operate. Is that a pattern yet? They are also explicitly targeting residences where they believe officials live with their families (which is also a war crime).
Even if you think they are simply wreckless, it is well-established that wrecklessness still constitutes war crimes
Israel has a (recent) history of bombing hospitals, and committing warcrimes and I believe they are also engaged with Iran. This attack on Iran is wrong from both parties and all targets are unacceptable, but do you have any articles or evidence that the U.S. damaged these hospitals?
> At least 13 hospitals and other health facilities have been hit during the US-Israel attacks on Iran, global health chiefs have said.The World Health Organization (WHO) said it was checking reports that four medics had been killed and 25 others injured.
> The Iranian Red Crescent chief said that at least 3,090 homes, 528 commercial centres, 13 medical facilities and nine Red Crescent centres have been hit in Israeli-US strikes. Officials reported damage to major medical facilities, including Khatam Hospital, Gandhi Hospital, and various rehabilitation and welfare centres.
> Iran’s Ministry of Foreign Affairs claimed on Thursday that the US and Israeli strikes have targeted 33 civilian locations nationwide, including hospitals, schools, residential areas, the Tehran Grand Bazaar, and the historic Golestan Palace complex – a UNESCO World Heritage Site.
Why though? Do we really need a pattern of strikes or could we just hold the biggest military in the world to a slightly higher standard? Why equivocate away responsibility to 'oh it could have been private so and so who murdered 100+ school children. Shrug.'.
Power is a social construct. Our institutions are being dismantled and collapsing, but they retain a vestige of legitimacy owing to the fact that most Americans haven't experienced much change in their quality of life. Wait until the gas pumps run dry or people start missing meals, and "power" has a way of evaporating pretty quickly.
That and much worse happened in the ~1675-1775 era of English and chartered company and proprietors lording over the American people, including actions that lead to mass starvation and death. It still took 100 years to totally throw off that yoke, though there were a few failed rebellions (like Bacon's).
> I've long been irritated by the use of the term "server emulator" in gaming contexts. Technically these projects are just reimplementations of a proprietary networking protocol. Nobody calls Samba a "server emulator" because it reimplements the Windows file sharing protocol, because Samba isn't "emulating" anything from the perspective of the traditional definition of "emulator" in computer science.
I think the distinction is a lot greyer than the black/white you propose.
The very first popular online games used servers mostly to redistribute (and maybe time sync) packets from clients. There is no standard way to to do that. Player-created servers did their best to emulate the official servers logic but it was indeed impossible to replicate it perfectly.
e.g. when breaking up large maps into sectors, the official server might broadcast your location and projectiles X units away and emulators would broadcast it X + 500 units away, which could have an impact on gameplay.
Emulator feels fitting when there is no official server spec to reimplement.
edit: emulator also feels appropriate where servers are responsible for NPC activity or quest-like mechanics. This goes beyond implementing a network protocol. The gameplay is massively impacted.
Your reply did exactly what I complained about: expanding the definition of emulator to cover reimplementing a network protocol.
You're not wrong that "server emulator" is a generically correct use of the term emulation, in the same sense that it is a correct use of the word for someone to say they emulate a fashion sense of a celebrity they like in their own wardrobe.
But in computer science, strictly speaking, the original definition of emulator was more strict. It was about things like emulating processor architecture A so as to execute programs written for it on processor architecture B.
And part of why expanding the definition to include "server emulators" annoys me is why has this definition expansion occurred only in gaming contexts? If a free UO server is a "server emulator" then why is Samba not also a server emulator? The lack of consistency is irritating to me, and it only happened because gamers like the term emulator, not due to any kind of rigorous computer sciencey reason.
My reply is that, strictly speaking, it is not a reimplementation of a network protocol if you need to recreate parts of the game based on best guesses that _impact_ the gameplay if your guess is wrong/different. A game is more than a network protocol. It is the data within that makes it to you that is being emulated.
SMB3 is SMB3. I would probably classify SMB1 (proprietary, closed) implementations as emulators if the guesses resulted in differing functionality from client<>server to another client<>server.
Wow! This is no small feat... am I reading the contribution graph[0] correctly, you've done all this yourself?
This endeavour sounds a whole lot like a server emulator for Infantry Online that was started by an incredibly talented developed 16 years ago ("aaerox"). I found the original svn commit on Sourceforge [1]. It's since moved to GitHub but has been active for 16 years and it has much of the same functionality you've already built, but done by more than a dozen developers over a decade-and-a-half.
Kudos to you. You've gotta explain how you've managed to do so much all by yourself.
So: I took most of the infrastructure from the my first attempt at moongate (https://github.com/moongate-community/moongate, which failed miserably along with https://github.com/tgiachi/Prima). From there, I had a good starting point to quickly build the foundations. I had already done the Lua scripting part in another project (https://github.com/tgiachi/Lilly.Engine). Codex helped me with all the testing, implementing functionality and creating tests, so at least I have a good sparring pattern. For the data import part (which I called FileLoaders), I took the logic from ModernUO. For the items part, I created a script (scripts/dfn_*.sh) to import items from POL! Thanks for the compliments! The way I am, if I fixate on something, it becomes an obsession!
while i understand the motivation to have codex like, do this problem for you, that's fine. what is the ethos of corresponding with people on Hacker News through the chatbot too? like i get that this particular comment i am replying to, you authored, but ChatGPT authored your post, and your documentation, and some of your other comments.
the big picture question is, if you can mess around with the bot to do anything, why spend it on this game? why not make your own original game instead?
I do use ChatGPT sometimes as a tool while working on the project (similar to using documentation, Stack Overflow, or an IDE assistant), but the post and the project direction are my own. So what?
Did you have to massage/guide the UI quite a bit? I've had terrible luck with codex, claude, and gemini at doing frontend. It's always so close but so far at the same time
Useful and useless (or good and “less good”) aren’t easily mapped to big and small.
From a purely UX perspective, showing a red badge seems you’re conflating “less good” with size. Who is the target for this? Lots of useful codebases are large.
I do agree, however, that there’s value in splitting up domains into something a human can easily learn and keep in their head after, say, a few days of being deeply entrenched. Tokens could actually be a good proxy for this.
The idea is well articulated and comes across clear. What’s the issue? Taking a magnifying glass to the whole article to find sentence structure you think is “LLM-slop” is an odd way to dismiss the article entirely.
I’ve read my fair share of LLM slop. This doesn’t qualify.
Oh I’m curious. Love bash, and learning new things about it.
I can understand why [ is not ideal. Can you explain the rest to me? I use || true for custom error handling often (with the right set -euo pipefail of course)
I agree; I didn't want to editorialize too much as I think the writeup stands on its own.
My takeaway was that in this case, even an author with a clear and extreme bias against this sort of thing could find only unfortunately-common bad practices rather than deeply nefarious intent. Of course, this is just the front-end code, but this just looks like a KYC platform to me. Most of the secondary reports on this write-up seem to completely ignore section 0x13 and jump to the specific conclusions the author does not draw.
The fact that we've created a system where Discord need and want a KYC platform is a different and quite strange thing, but the KYC platform itself just looks like what it says on the tin.
Any time you interact with the financial services industry in a meaningful way, they are doing almost exactly all of these checks on you. It is mandated by law, and they're overseen by FINTRAC in Canada and FinCEN in US.
When you applied for a bank account for your freelancing business (or startup idea), some people googled you, looked for PEPs (politically exposed persons) in your family, stored photos of your IDs and probably even printed them off, and sent everything in a nice package to some "risk" department. Who knows how that department is handling your data.
The only difference is that Persona is trying to put a front-end on it and selling the process as a SaaS. Look up "KYC/KYB saas" and you'll find hundreds of businesses doing this (including, of course, Persona).
edit: I want to emphasize that this isn't restricted to just business banking. Poor wording on my part. Lots of industries are legally mandated to conduct KYC/IDV. Notaries do it in home sales, your stock brokerage is doing it, employers in regulated industries do it to everyone on payroll. The list is very long. Unfortunately...
The government should take on responsibility for KYC imo, instead of letting 100 vendors come up with their own solutions. But that would probably have some nasty externalities.
A great QoL change. One less place to duplicate (and potentially mistake) a config.
reply