tsc is (very) slow and there are also no convenient ways to interact with it from Rust.

So it saves a lot development and CI time to roll our own. The downside is that our inference still isn’t as good as tsc of course, but we’re hopeful the community can help us get very close at least.

Heh, I agree with everything you said, but I’m afraid such a framework is impossible to create, let alone implement. It’s impossible to foresee the infinite possibilities for people to screw themselves through bad decisions, so all you’d create is a lot of bureaucracy to still end up in the same place.

That’s still a very major achievement! Do I understand correctly this means all target architectures supported by GCC are now unlocked for Rust too?

It’s that the compiler doesn’t help you with preventing race conditions. This makes some problems so hard to solve in C that C programmers simply stay away from attempting it, because they fear the complexity involved.

It’s a variation of the same theme: Maybe a C programmer could do it too, given infinite time and skill. But in practice it’s often not feasible.

I’m not saying you can’t, but it’s a lot more work to use such solutions, to say nothing about their quality compared to std solutions in other languages.

And it’s also just one example. If we bring multi-threading into it, we’re opening another can of worms where C doesn’t particularly shine.

I would argue that because C is so hard to program in, even the claim to machine efficiency is arguable. Yes, if you have infinite time for implementation, then C is among the most efficient, but then the same applies to C++, Rust and Zig too, because with infinite time any artificial hurdle can be cleared by the programmer.

In practice however, programmers have limited time. That means they need to use the tools of the language to save themselves time. Languages with higher levels of abstraction make it easier, not harder, to reach high performance, assuming the abstractions don’t provide too much overhead. C++, Rust and Zig all apply in this domain.

An example is the situation where you need a hash map or B-Tree map to implement efficient lookups. The languages with higher abstraction give you reusable, high performance options. The C programmer will need to either roll his own, which may not be an option if time Is limited, or choose a lower-performance alternative.

Of course, but it needn’t be black and white. You can also diversify, make yourself less reliant on a single platform. And by doing so, enable your audience to follow you elsewhere. Or diversify into different activities altogether. And when it’s no longer half your income on the line, then switch.

But doing nothing and saying, “but half my income!”? That’s not only a choice, but also complacency.

Great points, except:

People can’t leave for anything smaller.

They can and some do. It’s still a choice.

Ah yes, then we are in agreement. I thought we were talking about unintentionally arriving at the same implementation after looking at the original, which is where the discussion started.

Because coming up with the same implementation independently is legal, while copying someone else’s implementation isn’t. Which method you used to arrive at your implementation can be difficult to prove either way, which is why it’s important for implementors to be able to say they never looked at the original. It’s a legal defence, in case you ever need to stand in front of a judge or jury who will question how you arrived at yours.

Maybe not the best example then, but not the only example. If you unintentionally create something that resembles the original too much you may still become liable. It’s hard to draw the line, which is why many in such a position would prefer to be safe rather than sorry.

It’s a gray area, legally. What you say is theoretically correct, but there’s practical issues once you’ve looked at the code that will open you up to legal liability anyway.

For instance, what if you need a utility function during your reimplementation for which there is really only one obvious implementation? You can no longer claim to have come up with it by yourself.

I doubt the FSF would sue over it, but companies are known to avoid the risk.

I think it’s the latter. I once had to take care of a sick friend who was pretty much puking her guts out. Her moans sounded arousing. Of course she wasn’t intentionally doing that, it’s just our own male brains playing tricks on us.

I’m making a case for custom codes, not for using a 200 status code with it. My reply said the 200 didn’t make sense.

Of course once you use custom codes, the actual HTTP status codes do become less important, because there’s some redundancy there. That’s not an argument to do it wrong, but it is an argument that accurate HTTP status codes are less of a priority. So understandably some people will take shortcuts.

Apparently you find this very frustrating, but in the end it’s just an implementation detail. But it also sounds like you’re more frustrated with the service API as a whole than the fact it uses custom error codes specifically, so I’m just going to leave it at that.

I’m not arguing against that. Merely providing some counterweight to the idea that the author was “flinging shit in the trenches” 😅

I found the title of that section slightly triggering too, but the argument they lay down actually makes sense. Consistency helps you to achieve correctness in large codebases, because it means you don’t have to reinvent what is correct over and over in separate pockets of the codebase. Such pockets also make incremental improvements to the codebase harder and harder, so they do come back to bite you.

Your example of vendors doesn’t relate to that, because you don’t control your vendor’s code. But you do control your organisation’s.

Well, looking at your example, I think a good case can even be made for it.

“s23” doesn’t look like an HTTP status code, so including it can make total sense. After all, there’s plenty of reasons why you could want custom error codes that don’t really align with HTTP codes, and customised error messages are also a sensible use case for that.

Of course duplicating the actual HTTP status code in your body is just silly. And if you use custom error codes, it often still makes sense to use the closest matching HTTP status code in addition to it (so yeah, I agree the 200 in your example doesn’t make a lot of sense). But neither of those preclude good reasons for custom codes.

Oh, and then I pair it with a very boring Dell mouse for extra style.

Ha, I better not tell you about the Apple Keyboard I use with my Linux laptop then. Don’t like macOS much, but I love their flat keyboards.

Another data structure that you can consider is the red green tree: https://willspeak.me/2021/11/24/red-green-syntax-trees-an-overview.html

We use it in Biome too, and it’s great for building trees that are immutable and yet still need frequent updates, as well as traversal in all directions. Its implementation contains quite a bit of unsafe to make it fast, though as a consumer you’re not really exposed to that.