Видео

1:20 I knew this would be a parody ;)

please turn down the background music bro. you speak so softly and soy-ley that I can't fucking hear a word

genius

Gettings segfaults is a skill issue tbh

Love this. The opening music reminds me of Monkey Island and the communism comment floored me 😆

Original is better! The segfault freeze is funny though.

🤣🤣🤣 nice prank 😂😂😆😃🤩😍👍

please tell me bgm...thanks

Guys I'm new to C what is segfault I'm scared 💀

This optimization stuff is absolute insanity! How anything works Is a miracle!

So wholesome 💜

Sure miss the good ol' days of lyrics being posted in the description or top comments, rather than dead links and needing to be manually copied from the subtitles in the video itself. Wish we could turn back time...

Dat ending tho

Here's the simple answer to the title: C and I didn't watch the whole video.

That's a 10x cat you got there.

Ok, so you don't really know c++ and you bash it because it was hard for you to use a trivial feature or two. E.g. operator overloading. Not very useful. The real differences starts where you get to memory management: rust has cute but stupid-limiting rules, c++ has more difficulty with pseudo-aliasing. You are not talking about either because... you are more interested in inconsequential semantic crap? Ok, lol

I'm not. I'm on of these who uses non x86, non arm and non Linux. So all I have is a custom gcc/llvm based toolchain. And I want multiplatform programs

@fasterthanlime couldn't you have made a proc macro at that you call at the crate level on `httpwg` that iterates through each module looking for tests matching your criteria and writes an impl block for some type that builds up your test suite? that way it recurses properly without being stateful and you only have to write one proc macro. I know it would still have to parse a lot of syntax but you could probably internally just feed syn individual modules instead of the whole crate

I couldn't disagree more with the video. Electron is a horrible technology, and one of the reasons why web pages today tend to work properly only in Chromium-based browsers, and facilitate Google's monopoly on being able to dictate its own rules. Electron, as you rightly mention, facilitates and democratizes the use of many advanced functions that would be difficult to implement with the libraries of an OS, so far I agree. But you forget to mention that each app written in Electron has its own instance, that is, each one uses a DIFFERENT embedded Chromium engine. If you have Spotify, Discord, Slack and the Xbox app open, you are literally using 4 non-optimized instances of the same technology. And my question is, why don't they share the same web engine? That's when I think you should talk about Tauri. Tauri has a smarter approach, because it uses the OS WebView in order to render the content, and thus uses the computer resources efficiently. Because although you don't mention it in the video, Electron wastes a lot of CPU cycles in the background, spends a lot of energy irresponsibly, and doesn't integrate properly wherever it runs. And I agree that many web APIs work better than their native OS counterpart. But that doesn't mean Electron is the ideal choice for building apps. In fact, it's the worst possible and least user-friendly option. You say that in Windows there are too many frameworks and a developer probably doesn't know which one to use. It's a reasonable argument, but it's funny that you use that pretext to push the use of Electron. Wouldn't it be fairer to require Microsoft to go with an official framework, instead of just using Electron for anything? And with Linux, you're lying, there are agnostic frameworks. Apart from this, Electron is very difficult to optimize. It supports too many unnecessary things that, even if you try to deny it, have a very negative impact, especially for users with low-resource PCs. This also has an impact on the size of the program. 500MB may be reasonable for a complex program, but it is unacceptable for simple apps, especially considering that if every developer opted for Electron, the average user would have 1GB occupied by only 2 applications. If you think 1GB is not enough, I remind you that Crash Bandicoot weighs 429MB, so you can't defend that a calculator in Electron weighs more than a videogame. I just hope you don't have to use an app written in Electron on a less powerful device.

how do you learn about this concepts

As a C dev, I concur. Peops who don't need control should use rust, those who do should use C. I always need control, on the other hand I've recently figure out how to do thread safety without ever forcing the dev to call lock/unlock on something. requires va_list, enumed calls, global indices and hidden semaphores/mutices. Any index assigned is by default owned by the thread that requested it. Any index that does not have a semaphore/mutex assigned is only accessible by the owning thread - which is the default case for any index assigned, just have to manually request a semaphore/mutex be assigned at the time of acquiring the index. The rest is handled by a mid-wrapper for va_list (the top level being the main one called by devs, the bottom being the handler for the function enums). The midwrapper denies access to any thread but the owning thread if no semaphore/mutex is assigned, otherwise it tries to lock the index for the thread and if successful passes onto the bottom level handler. If unsuccessful it just tells the caller it failed to gain synchronisation access. As for buffers you should be using read/write like functions for the actual writing and use global indices for referencing the buffer rather than pass the buffer address around. By using indices and read/write like functions it becomes impossible to do a buffer overflow and you should only use malloc/free for temporary buffers to read/write to/from. Sure the temporary buffers can still have buffer overflows but they're a heck of a lot easier to catch the source of the segfault for than ones with longer lifespans. The ones with longer lifespans should always use mmap/VirtualAlloc to avoid being near the malloc page/s

what an incredible amount of effort to put into ... a video essay about github actions.

Always when I got a segment fault I come back to this video. So basically I am coming back here a lot

Awesome

for some reason I thought it just loaded it's own glibc wrappers with LD_PRELOAD but that would not work if you were to call the syscalls manually instead of with glibc

I FOUND IT, this was used as a base to troll vedal and now im stoked i found where it came from

i heard a whole lot of nothing

Hey, it's probably a bit odd, but do you upload your music anywhere ? I had the end credit song stuck in my head for the past couple of days.

I have cats I find outside. They are extremely friendly and demand attention.

iosevka font mentioned(/used (same thing)) 🎉🎉🎉

You're very irresponsible and dumb in your support of this, but that's your entire personality too, so...

People also might not want to run "trust me bro" binaries on their dev environment...

Internet audio is the final frontier

Where did you get your shirt?

AS a 1 year later comment(the c++ syntax i will talk about was intrdouced in c++20 so it still would have worked) in c++ you can also automatically let the compiler implement operator<=>, which also implements automatically all the classic comparison operators

The short answer: it doesn't really matter. These types of micro-optimizations are only relevant for very, very specific use cases. People are writing production code in Java, Python, hell even JavaScript. Sometimes the C++ compiler will generate better code and sometimes Rust. I would say C++ has a slight advantage because of how long it's been out there but on the other side Rust can potentially surpass it because of how explicit it is: the compiler has just more information and safety guarantees and can thus find more shortcuts.

Amazing channel

Thanks for validating our use of code generation @fasterthanlime 😆 Our use case is exposing a generic Rust API to Python through pyo3 which does not support generics (and probably never will). On the Python side we use enums that provide variants for all known implementers of a trait and our code generator generates the hundreds of required match arms. We generate the code once and ship it.

The result would be different on ARM64, so try that on Raspberry Pi 5 later

Big fan of using intermediate data formats for crimes. Beautiful work!

“A series of unfortunate events” is such an understatement

That was a great ad

That little pink ear at 6:56 made me happy. I once had a lovely white cat that ALWAYS wanted to sit on top of people :)

Due that outro was first class!

The purity of proc macros is a bit unfortunate even if understandable. A while ago I wanted to do some code injection into an asmjs app. I wanted to do this semi-automatically by annotating the rust library with proc macros and then parsing both my lib as well as the asmjs source and patching it based on metadata generated from the proc macros. In the end I ended up using a feature flag in the proc macro that would either modify the injected function and pass in some extra values or generate calls to the inventory crate. This would then be collected into a json by using the example(?) binary type in cargo.

I swear that "Life is too short for long builds" was a slogan for some product but for the life of me I can't find it.

"I decided port it to rust" then "don't repeat yourself" Okey okey

You could also use libtest-mimic or similar to build a custom test harness that has test cases generated in runtime (from the central test registry you could have collected)

The outro really caught me by surprise, really informative and interesting video

re: git hooks @ 14:11, aren't git hooks "disabled" by default? I thought that they are not committed to the repo and you have to copy them over manually (or use something like the `pre-commit` tool). Which makes sense, because they can run code without you knowing.