The Servo Browser Engine Has Been Making Great Progress In 2023

[Anux]

Memory leaks are needed for any language that wants to communicate with other programs/languages/libraries, if you want to send data away you need to be able to drop all references to it.
That said, it's really hard to accidentally leak memory in rust.

[Artim]

Memory leaks are needed for any language that wants to communicate with other programs/languages/libraries, if you want to send data away you need to be able to drop all references to it.
That said, it's really hard to accidentally leak memory in rust.

That is rubbish. That's not how memory leaks work.

[Jabberwocky]

Once compiled it will just be a binary. And hopefully a small and efficient one. I would love Servo to replace Webkit in electron for example. All my software are electron now, even if I don't like it... Visual Code / Atom, Docker Desktop, Signal, Mattermost, Spotify... I think Thunderbird and Firefox are only two left that I use on a daily basis and are not electron.

It would be nice indeed. We do need multiple backends for these apps. If it was open source then it wouldn't be as bad but many of these apps are proprietary (VS Code, Spotify, etc).

Chrome/Chromium still suffers from massive vulnerabilities even though it's being used everywhere. Here's a very recent example: https://nvd.nist.gov/vuln/detail/CVE-2023-4863 - Heap buffer overflow in libwebp in Google Chrome prior to 116.0.5845.187 and libwebp 1.3.2 allowed a remote attacker to perform an out of bounds memory write via a crafted HTML page

[Artim]

It would be nice indeed. We do need multiple backends for these apps. If it was open source then it wouldn't be as bad but many of these apps are proprietary (VS Code, Spotify, etc).

Chrome/Chromium still suffers from massive vulnerabilities even though it's being used everywhere. Here's a very recent example: https://nvd.nist.gov/vuln/detail/CVE-2023-4863 - Heap buffer overflow in libwebp in Google Chrome prior to 116.0.5845.187 and libwebp 1.3.2 allowed a remote attacker to perform an out of bounds memory write via a crafted HTML page

All true, but the biggest problem with that is developers using Electron need to keep the Chromium base up to date. Just look at Discord to see how miserably you can fail at that task. Right now it's based on v108.0.5359.215 (at least that's what I get from the dev tools console after executing >>navigator.appVersion.match(/.*Chrome\/([0-9\.]+)/)[1]<<). It's about a year old and probably nobody ports back any fixes. So while a rust written browser would limit the amount of security issues that can occur, it's not a solution for that problem. And I'd argue that I have less of a problem with dozens of apps wasting memory than dozens of apps being a security nightmare.

[Anux]

That is rubbish. That's not how memory leaks work.

Oh this will be funny, please enlighten us how memory leaks work if not by losing the reference without freeing the memory.

[Artim]

Oh this will be funny, please enlighten us how memory leaks work if not by losing the reference without freeing the memory.

There's a difference between sending information to another process to ingest and a garbage (or you yourself in the absence of a garbage collecotor) collector neglecting freeing up space. A memory leak isn't the process of removing reference to memory but the faulty behavior of not releasing memory when not needed anymore.

[moltonel]

No need to call others names and be aggressive when you don't share their opinions.

This is good general advice, but Weasel has held an aggressive anti-Rust stance for a long time, with some outlandish views and zero interest in counter-arguments. Friendly constructive replies don't last forever.

Besides, the polls are most likely biased, even if not on purpose, and it is easy to see why. It is easy for a programming language to be popular in polls when most of its users are using it voluntarily, not prescribed by legacy codebases or employers. This basically implies that all its users will like it to some degree, because otherwise they wouldn't be using it. The polls though will look different in 10-20 years, not because there will necessarily be better alternatives, but because Rust needs time to accumulate history like other "less-popular" languages. Until then, such polls aren't meaningful.

And yet, there are a few biases in your analysis too. The premise that Rust use is mostly voluntary rather than prescribed by an employer. The assumption that you necessarily like a tech that you freely choose for a project (the best tool for the job can still suck). The intuition that niche/upcoming languages will naturally have more happy users (how do you explain the low "admired" score of Zig/Nim/Groovy/Julia compared to Rust ?). The belief that Rust is still a niche / early-adopters language (SO ranks it as commonly used as Go/Powershell, not far behind C/PHP, well ahead of Kotlin/Ruby/Dart/Swift...).

I'll be the first to say that all popularity measures are full of biases and should not change your credence by much. But explaining Rust's high user satisfaction only by it being a new language doesn't hold up to scrutiny. FWIW, after 19 years and nearly as many languages used professionally, Rust is the first one that feels so right in so many dimensions. I'm happy to see it used more and more.

[Weasel]

Just because a moron like you doesn't want it to be true doesn't make it any less true.

Just because morons like you says it's true doesn't make it any true, in fact it says the exact opposite since it comes from you specimens.

[Weasel]

And yet it keeps getting adopted by project after project by experts in their fields. I think that counts as practice.

Nope, because it doesn't. You can keep claiming stuff, it won't make it true.

"Experts" lmao. If they use Rust they're already off that tag.

[Weasel]

He has proven to be an absolute moron over and over. So nothing wrong with that.

Coming from a rusted brain that sure means a lot. You want some help unrusting those neurons?

Keep wondering why Servo is dead and not adopted by most apps, though. I'm sure you'll figure it out... eventually...