So... kind of like a parrot, huh? /* ...or a mimic. Also... 'The Mimic (2017)' https://www.imdb.com/title/tt7046826 . Good movie. */

I made XC, it is how C++ should have been. It's actually C + Classes + Python Library Compatible APIs (dataset, asyncio, tempfile, pathlib, et cetera) and so on. It is an extension of native C. You are allowed to write classic C (just compiles) if not --strict mode is given.

Making your own programming language in this era, is kind of pointless. But Retoor has points, she got all the points, young man. I integrated AI development within the language. You should not manually write on projects and just use the customized make muse agent that will provide perfect XC code for you. And it is really, really, based and native.

Example:

class View(aiohttp::View) {
    async aiohttp::RequestHandler(aiohttp::Request res) get {
        // and so on...
    }
}

int main() {
    auto app = aiohttp::Application();
    app.routers.add_view("/", View);
    await app.run(7331);
}

Yes, that source code does have the exact same amount of lines as Python and matches the API very well. But it is NATIVELY compiled and it is extraordinary superior to Python when it comes to memory and CPU. Regarding performance, it does not matter that much if the view does nothing / only returns a string. It's the p99 area where the XC version of aiohttp really shines. We generate faster concurrent images that Python freaking a HTML template, young man. I actually made request handler using threading as async instead of epoll because having thread request handlers and epoll connection management is the best of both worlds.

Having all this said - a real web application that has a lot of mixed type of requests would actually do better on the aiohttp of XC.

But XC does actually win literally every benchmark, but epoll syscall is epoll system call and that kind of determines the speed of your event loop. Not the programming language is the bottleneck really.

But objectively, the XC aiohttp version is the best choice in any case. We did port literally whole Jinja2 to XC. I still have to review that, it is the first XC library that is included but not part of the transpiler aka native module. Almost forgot. Crossing fingers.