In LSP, UTF-16 is not just for transport. Source locations are also referred to by their UTF-16 offset. That means applications are required to store a persistent mapping from UTF-16 offsets to whatever they use internally, keep that mapping updated as the document is edited, and map back and forth constantly.
Oh well. It was primarily designed to be a good protocol for VSCode, not for any other users. That's just the price you pay when using a corporate standard.
Oh you're right, except since LSP 3.17 position encoding can be negotiated with clients/servers that support it. I initially got confused because my work with LSP started with 3.17 and I always used UTF-8 encoding.
I'm not talking about HN in particular, since I know it's not a single give mind, but it's funny how quickly programmes as a group changed their mind on this when it became THEIR job that was replaced.
Without #[non_exhaustive], adding a variant is always a semver breaking change.
With #[non_exhaustive], adding a variant is not a breaking change, but only if that variant can safely be ignored. If it can't be, then it's still a breaking change, BUT you don't get a compiler error. That's a very big drawback IMO.
This makes me think of how the PNG format handles this. For context, PNG files consist of a list of chunks of different types (image data, color profile, text comment, etc). Each chunk has some flag bits that tell you whether it can be safely ignored when decoding the image, safely retained when the image data is changed, etc. That way the decoder can ignore unknown ignorable chunks and error on unknown non-ignorable ones.
Once the fix is released, people will reverse-engineer it and use the exploit on pages that haven't updated yet.
So they need to give time for infra people in all different time zones to wake up, learn about the announcement, and plan how to patch it.
The intended course of action is 1. take down website a few hours before the fix drops 2. apply the fix 3. only make website public again after it has been fixed
There's a big difference between suing people for prima facie illegal hiring practices and suing people for holding Grace Hopper celebrations. We probably shouldn't conflate them. Your links, as best as I can tell, are about the former sort of activity - hiring or paying people differently on the basis of protected characteristics.
You understand that none of these companies believe the merits of why they are settling?
They are settling because, like the law firms Trump suborned, they didn't want the MAGA rage; especially in the case of Texas, where Trump may (correctly as it goes) think Paxton is a disgusting slug of a man, but still vocally supports him.
The legal system was just the venue. All of this is bullshit corporate spinelessness in the face of threats without merit, especially in firms like Deloitte, who do not want their M&A interfered with.
OK. That is fine. It was literally a question so as to work out if we were on the same page for what I wrote subsequently. I will make a mental note that the phrasing doesn't come across that way.
As an aside I think my brain jumped the track and had me thinking I was replying to another commenter I'd replied to already -- who wrote a sibling comment to yours.
I personally found implementing an HTTP2 server from scratch quite difficult. The parsing and state machine was okay (and is also the easiest part to get from an external library), but correctly handling stream priority / bandwidth sharing and backpressure was really hard. Of course one could go HTTP1.1-only, but then you're losing out on the efficiency gains anyway :/
I'm a bit surprised the judges themselves didn't spot that it was AI. It immediately starts with things randomly appearing out of nothing.
AI-generated fake footage in medical science is one of the things that worries me a lot these days. Now there's practically a constant stream of papers getting retracted due to fake AI-generated microscopy, before/after photos, plots(!), ...
My thinking is the judges weren't familiar with what a footage like this would actually look like and they might be open to accepting minor video edits for making it interesting.
It's always a slippery slope to judge things as AI first and ask questions laters. I myself feel like I might have started to attribute human effort to AI, and that honestly is the truly sad part.
When writing GPU shaders it's not uncommon to use extensions like ".vert.glsl" = "vertex shader written in GLSL", ".frag.hlsl" = "fragment shader written in HLSL", etc.
Oh well. It was primarily designed to be a good protocol for VSCode, not for any other users. That's just the price you pay when using a corporate standard.
reply