[saghm]

Sure, not caring about contributing to Linux in practice is valid, but at that point it would make more sense just not to argue about the aptness of C++ for integrating into Linux at all then if no one ever legitimately tried. My point is that the parent comment is trying to have it both ways by arguing that C++ totally could be integrated into the Linux kernel as a dismissal of the real-world case of Rust being integrated but then ignoring the actual real-world case of C++ not getting integrated into the kernel.

At a higher level, I've noticed a long-term trend where dismissals of Rust from C++ proponents have continued to grow increasingly abstract; the line of reasoning basically seems to be that how Rust is used in practice can be ignored as long as there's a sufficient theoretical argument for C++ being able to do the same thing. This is a remarkable retreat from the previous objections to Rust being impractical compared to C++ which I'd argue is itself evidence of the ground that Rust has gained in the past several years. It seems to have happened without the C++ community noticing though, which at least in part seems to come from some cognitive dissonance; in your two replies to my comments in this thread, you demonstrated this yourself by arguing on pragmatic grounds about how C++ is used in practice in this comment but then rejecting criticism of C++ in your other comment by asserting that a "recipe" for memory safety is enough to dismiss concerns about how pretty much no real-world project actually achieves this.

My (obviously biased) perception is that the shortcomings of Rust tend to be pretty widely agreed upon by the experts in the community and mostly dismissed by people who honestly aren't nearly as experienced in the language, whereas with C++ the most expert members of the community seem prone to handwaving away concerns that deserve a lot more attention.