Hacker Newsnew | past | comments | ask | show | jobs | submit | WhyNotHugo's commentslogin

A matter of taste: when I use wine on Linux I prefer to confine all window to a single "real" window. In truth: I suspect that your suggestions is more work to pull off properly.

Anyone with an older toolchain can’t build that library of anything that depends on it.

Some environments might not even have the newer version available.


Anyone with an older toolchain is free to fork it on github, test with the older version, and CI to the project that tests with the older version, and submit a patch, too!

This may not get the project as many users, but not everyone who writes a 50 line project is trying to figure out which versions it supports and setting up full test matrices either.


Not a Go dev, but I typically set up a CI with the oldest toolchains I support (usually a debian release), and only bump those versions when I really need something from the latest versions. Locally I build with the most recent tools. This ensures good enough coverage for very little work, as I notice when I start using something that's newer and can bump the toolchain accordingly.

Sure, but if you start a new small project and throw it on GitHub, it's not totally insane to just put the version you tested. Just because someone put up their tiny library doesn't mean they've put in the effort to figure out which version they need.

Are you sure you replied to the right comment? I'm not sure how this relates to the question being asked.

I did.

If you have an older tool chain, it is on you to fix the library to build with the older tool chain, that's what open source is about!


kqueue is quite portable and works across all the BSDs.

The OpenBSD documentation for it is top notch, as usual. No idea about the rest (I suspect they’ve all converged at this point).


Kqueue originated in FreeBSD, and like most all of FreeBSD, it is very well documented.

I find that (neo)vim enable code navigation to be much faster than any GUI as well, once past the learning curve. If you’re going to work with code long term (eg: years), the learning curve pays off quickly.

Intuitively, I’ve always had an impression that using an analogue circuit would be feasible for neural networks (they just matrix multiplication!). These should provide instantaneous output.

Isn’t this kind of approach feasible for something so purpose-built?


You might wanna look at https://taalas.com/

They aren't using analog circuits, are they?


PSR (panel self-refresh) lets you send a single frame from software and tell the display to keep using that.

You don’t need to render 60 times the same frame in software just to keep that visible on screen.


How often is that used? Is there a way to check?

With the amount of bullshit animations all OSes come with these days, enabled by default, and most applications being webapp with their own secondary layer of animations, and with the typical developer's near-zero familiarity with how floating point numbers behave, I imagine there's nearly always some animation somewhere, almost but not quite eased to a stop, that's making subtle color changes across some chunk of the screen - not enough to notice, enough to change some pixel values several times per second.

I wonder what existing mitigations are at play to prevent redisplay churn? It probably wouldn't matter on Windows today, but will matter with those low-refresh-rate screens.


Android has a debug tool that flashes colors when any composed layer changes. It's probably an easy optimization for them to not re-render when nothing changes.

I never thought about it but you've made me realise that a lot of people in our industry have been so enthusiastically working on random "creative" things that at best no one even asked for and it turns out to hurt the end users in ways no one even knows.

I used to be a front end dev and I always hated that animation was coded per element. There should be just a global graphics API that does all the morphing and magic moves that user can turn off on the OS.


Normally, your posts are very coherent, but this one flies on the rails. (Half joking: Did someone hack your account!?) I don't understand your rant here:

    > With the amount of bullshit animations all OSes come with these days, enabled by default, and most applications being webapp with their own secondary layer of animations, and with the typical developer's near-zero familiarity with how floating point numbers behave
I use KDE/GNU/Linux, and I don't see a lot of unnecessary animations. Even at work where I use Win11, it seems fine. "[M]ost applications being webapp": This is a pretty wild claim. Again, I don't think any apps that I use on Linux are webapps, and most at work (on Win11) are not.

Seriously? What is _this_ comment? TeMPOraL makes perfect sense.

LLMs learned that users have post histories? /s

> Passing legislation takes about as much effort as repealing.

While true, those trying to pass this legislation get paid to do so, while those against it have work hard and pay taxes to fund the former.


> those trying to pass this legislation get paid to do so

Chat Control has paid lobbyists on both sides. Also, paying lobbyists is still sinking resources. And the people taking their meetings are still sinking political capital into a fight that has–to date–yielded zilch.

> while those against it have work hard and pay taxes to fund the former

The principal moneyed interests in this fight are the tech companies. Your taxes aren't funding their fight. The police lobby is less effective if filtered through paid lobbyists versus having a police chief personally pitch lawmakers.


You have “centralised democracy”, a form of democracy where decisions, once debated and adopted, are implemented uniformly throughout an organisation. They are not debated a second time, and there’s no room for dissenting against decisions already made.

It’s a double-edge sword though: if something you dislike gets votes, it’s never going away.


> They are not debated a second time, and there’s no room for dissenting against decisions already made

Of course they are and of course there is. The "EU passed a temporary derogation" to the ePrivacy Directive in 2021 "called Chat Control 1.0 by critics" [1]. That is now dead [2].

> if something you dislike gets votes, it’s never going away

Weird to be saying precedent is infintely binding in 2026 of all years.

[1] https://en.wikipedia.org/wiki/Chat_Control#Legislative_proce...

[2] https://x.com/NoToDigitalID/status/2037213272131203339


The sentence which you are quoting is referencing the concept described in the previous sentence of the same paragraph. It is not describing the EU’s form of democracy.

The EU parliament can't retract existing laws if the EC doesn't agree and proposes a law doing it.

Yes, if I don't like something, I can't just ignore it. That is called democracy, and rule of law. Democracy is often interpreted to mean only things I like get passed, but that is incorrect.

> it's simply no longer reasonable to BYO CI or accept one that can't natively build for a common set of end-user architectures.

GitHub’s hosted runners support a grant total of two architectures.

The only forges which I’ve seen with more variety are distributios’ forges usually hosting their own runners.


I probably should have said platform tuples, I guess. I'm not aware of a free alternative to GitHub that gives me (AMD64, ARM64) x (macOS, Windows, Linux).

> that's where the community is

The part of the FOSS community that embraces proprietary dependencies are there, but there’s a lot of the community outside of it.

Fortunately, GitHub is pushing hard for folks to want to move away.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: