skullt

SDF font rendering has been around 20+ years though? Valve really popularized in their 2007 SIGGRAPH paper and Chlumský developed MSDF font rendering in a 2015 thesis.

If you're paying the cost of serializing/deserializing the whole game state around every reload of the worker, why bother with the unsafe type-erased PersistWrapper (which still requires you to restart the host any time the game state struct changes)? Just keep the game state entirely on the worker side, and have the host instead pass, say, a &mut dyn Write to the before/after_reload functions. The old worker can write whatever it feels like there, JSON or whatever else, on the assumption that the new worker will decode it somehow. The new worker then stashes the deserialized game state in a static Mutex<GameState> during the after_reload hook and the actual hot_update function goes back to taking just the host trait object. The big advantage here is you can change the game state struct during a hot reload as long as the deserialization can handle it. With JSON, for example, adding new fields to the game state struct would be perfectly fine: old fields still get deserialized and new fields get default initialized.

He's also got Cleartype on and set to RGB stripe even though the OLED is not RGB stripe (though to be fair, Windows doesn't really make it clear what each page of the ClearType tuner does).

I doubt it's Cleartype, the close up photo of the U3223QE show all subpixels uniformly dimmed on the fringes. The post also says the monitor is attached to a Mac mini and a previous post about OpenSCAD has a screenshot with MacOS window decorations.

I would think leaders sometimes not following their own codes of conduct is the strongest argument in favor of having them: yes, they are obvious to everyone but they are also evidently easy to forget in the heat of the moment. It's a standard of behavior to strive for not one statically attainable. Reminders are needed and that's the purpose their deliberate codifying serves.

In fairness, Emacs has long had cua-mode for rebinding C-c, C-v, C-x, and C-z to copy, paste, cut, and undo, so at least those changes are not too radical.

There's a bit of a trick in that solution: n is assumed to have the lower two bits clear so for an arbitrary n the array would really be:

[(n & ~3), 1, (n & ~3) + 3, 0][n % 4]

where the (n & ~3) makes sure those lower 2 bits are cleared. But note that we only ever can look at the first element when n % 4 == 0. In that case, (n & ~3) == n already. And further, we only ever can look at the third element when n % 4 == 2. In that case (n & ~3) == n - 2, so (n & ~3) + 3 == n + 1. Hence the array can be simplified to the one given in the other comment.

It's a common extension but it's not standard. Both C23 and C++23 only allow identifiers to start with a Unicode XID_START character, which excludes $. And with older standards, it was just Latin letters, the underscore, and explicit \u escapes.

If you look in the project's pedantic.cmake, you'll find a add_c_and_cxx_compile_options("-Wno-dollar-in-identifier-extension") just to suppress warnings about this.

That doesn't quite fit either. Slotin did the screwdriver trick a bunch of times before the accident. He was showing off.

SSDs using Host Memory Buffer for cache instead of dedicated DRAM are getting pretty common on PCs too.

C11s Atomics will never be implemented on Windows because they cant be!

They are implementing them though. See https://devblogs.microsoft.com/cppblog/c11-atomics-in-visual...

Sure but you can't ignore a 20% obesity rate all throughout Europe just because the US is worse.

Points 3 & 4 could be combined: like .plan, each user has an o+r .status file. Its contents are a user's status message, but its modified time is used specially to indicate when that user was last active. Anytime the usuc client writes a message to a channel, it would simply touch the .status file. And whenever an expanded usuc or some other tool lists the users in a channel (that is, all users in the group that owns the channel file), it would mark as idle any who had a .status mtime older than 30 minutes, say.

Maybe they compute in tate mode.

Why not just fix it? It's a trivial fix and has the added benefit that a zero-initialized DYN_ARR_OF(x) struct would be in a valid state, which is always nice. A struct with several dynamic arrays can then much simpler to initialize, for example.

a journalist intending no malice

He outed her to investors and likely to business connections through Gary McCord. He could have stuck to the holes in her professional history, which was all that was relevant, but he didn't. He was absolutely muckraking and about a woman he knew had previously attempted suicide.

Does that not contradict the Zig principle of no hidden allocations?

principal game engine engineer on The Witness

Is that true? He is listed in the credits under "Additional contributions in programming by", and in his blog posts about the "Walk Monster" problem, his role sounds like that of an independent contractor brought on late in the project.

There's also a comment on that alternative interpretation here: https://twitter.com/LinManuelRwanda/status/15058362781090611...

The article doesn't have the full quote from the video. After mentioning community-based transmission, he goes on to say that "You can start to think of it like seasonal flu," which does suggest recurrence.

The graph is measuring in six month intervals, so the March 2009 release of IE 8 needs to also be considered.

Strictly taking everything people say at face value is great way to become somebody who "can't take a hint." The problem is that there are many things people want to communicate but would rather not say because they're likely to lead to uncomfortable situations.

Suppose a coworker offers you mints or gum every time you speak to them. If you only take their words literally, you'll think, "Gee, what a generous person," and miss entirely their true meaning: "Your breath stinks and it's bothering me but I don't want to hurt your feelings, so I'm giving you out where we can both pretend you fixed the problem before anyone noticed."

Because in many places, standard time is the exceptional case. In the US, at least, we spend almost twice as much of the year in daylight saving time than in standard time. For example, standard time during the 2017 to 2018 winter lasted 127 days. Daylight saving time this spring through fall lasts 238 days.

Now just let us change the tab width.