I had a look at the C sources of the original coreutils, and I have to say that I seriously doubt the Rust version will add performance. Safety? Maybe, but these are mature programs that have been studied a lot. A new coreutils suite with hundreds of programs will have to go through the bug-fixing and compatibility issue-tracking-dance all over again.
I don't want my coreutils programs to have slight deviations from how they used to behave. What is the actual benefit of doing this? For example, with musl we had the benefit of easy access to static linking. Eventually they added "minimal dynamic linking," which is sort-of a feature in itself. It took a long time for feature parity, though, and it never reached the performance levels of glibc.
I don't want my coreutils programs to have slight deviations from how they used to behave. What is the actual benefit of doing this? For example, with musl we had the benefit of easy access to static linking. Eventually they added "minimal dynamic linking," which is sort-of a feature in itself. It took a long time for feature parity, though, and it never reached the performance levels of glibc.