Rust-GCC

Posted Oct 19, 2024 0:54 UTC (Sat) by jsakkine (subscriber, #80603)
Parent article: On Rust in enterprise kernels

Looking at https://github.com/Rust-GCC/Reporting/blob/main/2024/2024..., GCC is making steady progress. IMHO much more important than anything that happens in-tree. When one can take just good old GCC and compile Rust-enabled kernel that should open doors for being non-experimental and being part of arch defconfig-files. The problem sort of starts to solving itself then.

Rust-GCC

Posted Oct 19, 2024 1:07 UTC (Sat) by intelfx (subscriber, #130118) [Link] (2 responses)

> When one can take just good old GCC and compile Rust-enabled kernel that should open doors for being non-experimental

That shouldn't be a prerequisite.

Rust-GCC

Posted Oct 19, 2024 7:07 UTC (Sat) by smurf (subscriber, #17840) [Link] (1 responses)

> > When one can take just good old GCC and compile Rust-enabled kernel that should open doors for being non-experimental

> That shouldn't be a prerequisite.

I can only assume that a few kernel maintainers more-or-less-strongly disagree.

IMHO a strong and demonstrated commitment for the GCC people to "get there", in the not-too-distant future, should go a long way towards assuaging their concerns.

NB: There are standards organizations that require at least two, sometimes three, independent implementations before they even start thinking about standardizing something.

Rust-GCC

Posted Oct 21, 2024 8:35 UTC (Mon) by taladar (subscriber, #68407) [Link]

And all languages standardized by those organizations are a horrible mess of design by committee compromises between the authors of the various implementations. No thanks.