My system has a midrange amd cpu with 6 cores. I have an ssd.

I think the issue has to do with procedural macros in maud. The project is about 1000 dependencies and 10k loc.

Combined with tailwind classes and a complex ui, the templates get bloated. I have to test this theory.

Are you sure? I have tried using vim and it was pretty much the same.

I have also confirmed the cause using task manager to be rust analyzer.

8GB RAM isn't a small amount (though by no means a lot). As far as RAM usage goes, the amount you need will scale with project+dependencies size, so for smaller projects, it shouldn't be a problem at all.

8GB RAM doesn't tell us about the rest of your system though. What CPU do you have? Is your storage slow? Performance is affected by a lot of factors. A slow CPU will naturally run programs slower, fewer hardware threads means less running in parallel, and slower storage means that reading incremental build data and writing it could be a bottleneck.

This will work if you use emacs or vim. The problem is helix.

what are you compiling?

slice::element_offset() looks really nice. But I don't know yet where I would use it.

The has_abba example is neat!

I really want to see what happens to the web at large if given a way to create websites using wasm instead of js.

I'm also looking forward to seeing if any new frameworks (with hopefully better APIs, robustness or just better practices) will be developed for the web, I'm tired of inaccessible React (or React-like) applications that break easily.

The majority of actual rustaceans don't care about these polls (or any "official community" activity for that matter).

If you want actually relevant (and objective) stats, look here.

Let's take Lemmy UIs as an example. In a world where this "RCE" is removed, all API calls and returned data would have to go through a "server client" first. I hope this won't take you long to ponder if that's an improvement or not 😉

The web is indeed shit. But dumber web means more "clouding", or if it's not "clouding", and to borrow from your reductionist fatalism: Dumber web replaces a potential RCE with a definite MITM.

wasm and js are by definition remote code execution

"oh but is sandboxed" how many sandbox bypass and sanbox escape CVEs have we had? incountably many

beyond that, that is code using your cpu cycles often inefficiently and for useless purposes or outright malicious purposes such as tracking

I'm guessing, they mean JavaScript and WebAssembly in general...

All the RCE vulnerabilities that Apple introduced as "features"

This is such an exciting proposal for the web. From what I understand, this would be one of the final hurdles to having WASM "just work" in browsers the way people always expected it to (i.e. having standardized support for directly interacting with Web APIs/the DOM, without any janky glue code and skipping the huge inefficiencies of intermediary JS).

The end result would be developers having a much easier path to implementing huge parts of their applications with native-like speeds in any language they want. Some people will want to do their whole application this way, and others will have specific modules that can benefit hugely from this while still making considered use of the rest of the web platform.

As they say in the article, the current status quo is that this is possible but that ad-hoc/custom WASM modules are out of reach for all but the largest developers because of the cliff of complexity involved in moving beyond the most well-trod paths.

What RCE are you talking about?

make RCE a second class feature instead

Please get rid of JS.

Bah, very funny, but this shit is real. I can't find any paid projects lately, going back to C/C++ or TS or even Go and Kotlin feels so disgusting now. I had gigs for last 6 years before 2025 or so but I suppose that data slide with Rust positions is... somewhat misleading.

AI can't figure out most of Rust tasks indeed, I keep trying to start doing that vibecoding thing and failing, especially in blockchain and baremetal, so what, who needs those anyway? AI does great in areas like JS (as far as I can tell, not worse than regular devs), yet job boards are still full of python and JS positions.

Nothing can beat Rust at mission critical areas, but people would rather save pennies on development and then have their butts on fire than do things right.

They just abandoned the previous project of moving to Swift.

There’s also the issue of trying to manage a project rewrite while the project itself is under continuing feature development. Many projects have failed at that in the past because the rewriting effort is so much of a moving target and because a team split between the sides can develop a lot of friction and conflict points.

I really want them to succeed. I think it’s pretty amazing what they’ve achieved so far with Ladybird, just as a proof that it is indeed possible to write a brand new browser engine from scratch on web standards. I also think it’s extremely important that we have some mitigation against abusive behaviour by Google & Apple, however small it may be with a hobbyist browser like Ladybird.

He also thought that Rust integrates poorly into project with a deep C++ OOP hierarchy. That is probably still true as well.

Is there any language that can do that? As far as I know there isn't. You can use SWIG or whatever but it's just as awful as any Rust/C++ interop. There's Carbon, but that's a work in progress.

IMO if you need integration with a deep C++ OOP hierarchy your options are a) give up and just use C++, or b) pain, no matter what language you target.

I made this point below but it is telling that the first project was a total rewrite of the entire JavaScript engine. There is no shared hierarchy and a well defined boundary between the C++ and Rust code. That may be a sign of things to come.

The Ladybird founder initially rejected Rust because he found it hard to integrate into their deep C++ OOP hierarchy. That problem likely remains.

We may see entire modules that are either all Rust or all C++ rather than more fine grained mixing.

Andreas did not like the Rust community. It sounds like that is still true.

He also thought that Rust integrates poorly into project with a deep C++ OOP hierarchy. That is probably still true as well.

It is telling that the first project was a total rewrite of the entire JavaScript engine. There is no shared hierarchy and a well defined boundary between the C++ and Rust code. That may be a sign of things to come.

We may see entire modules that are either all Rust or all C++ rather than more fine grained mixing.

I think making the standard library larger in a composable way would be great. If we actually uplifted "blessed crates" from a loose concept to an actual thing in the build system, it would make Rust much easier to adopt for those who are not already well versed in the ecosystem.

i think use it an appropriate amount. im not sure how to quantify that. i use different AI models on different tasks in the code as well as the documentation.

its worth repeating its far from finished and i hope with feedback i can make it better. i have put efforts towards directing it towards unit-tests, an audit and formal-proofs. none of that is good-enough, but i hope it can act as a starting point for verifying the implementation is correct.

i get the whole semantic versioning rhetoric and branching strategies, etc. this project is a while from being promoted as "perfect". this is still a work-in-progress.

im sure people have better things to do with their time than review unstable and unfinished code. as a solo dev on this, there isnt anyone reviewing my code. if i dont share it like this, no one with come across it. i hope you can understand i get pushback when i promote my messaging app is "secure", so this transparency is nessesary.