GStreamer Now Supports Multi-Threaded Scaling/Conversion For Big Performance Win

[TheBlackCat]

It seems pretty obvious doesn't it? I don't actually use gstreamer. I'm lucky enough that Gentoo devs have worked real hard to make certain gstreamer can be removed. Thank god for that. Because it suffers from a whole slew of bugs and escalation flaws that are not possible in Rust programs. In todays world Gstreamer is one of the most blatent attack surfaces on most linux distro's. Basically all of them would be better off completely removing it. The second option is rewrite it in a better language like rust.

Yet you readily admit that you don't actually know that much about rust or gstreamer, so what makes you so sure rust would actually help with gstreamer's particular problems?

[duby229]

Yet you readily admit that you don't actually know that much about rust or gstreamer, so what makes you so sure rust would actually help with gstreamer's particular problems?

Because I can read and I can understand what I read.

[mmstick]

Yet you readily admit that you don't actually know that much about rust or gstreamer, so what makes you so sure rust would actually help with gstreamer's particular problems?

Plenty of zero days have been uncovered lying around in Gstreamer that wouldn't have been possible had it been written in Rust. The same is true of basically every C program in existence. Rust's memory safety invalidates 99% of these vulnerabilities on the spot by making it a compile-time requirement. C is littered with undefined behavior, memory unsafety, and pointer aliasing. Rust does not allow undefined behavior, the borrow checker and lifetime mechanisms protects memory safety, and pointer aliasing can be entirely avoided (once MIR starts optimizing it). There's good reason why the GNOME team has been transitioning to Rust and experimenting with it in their projects.