News for hackers. Stuff that matters.
Re:Reimplementing the Coreutils in a modern language (Rust) by yjftsjthsd-h on Monday February 13, 2023 @05:39PM (#34781929)
> Looking at the source code, it is impressive that this re-implementation is essentially complete! Look at this: https://github.com/uutils/coreutils/tree/main/src/uu It's only missing some obscure and fringe things.

It's not just a matter of having all the individual tools, it's a matter of having them be functionally complete and behaving either the same as GNU or POSIX - note the graph part way down the readme that talks about having only slightly better than half of the tests passing. And the tests probably aren't truly comprehensive, especially for things that people didn't realize were implementation specific or non-guaranteed behaviors. It's good to see them making progress, I just want to someone temper the idea that they're close to done.

Re:Reimplementing the Coreutils in a modern language (Rust) by dho on Tuesday February 14, 2023 @10:44AM (#34790954)
You have to be careful with the interpretation of the graph. It tracks the number of files where all tests pass.
Re:Reimplementing the Coreutils in a modern language (Rust) by enriquto on Tuesday February 14, 2023 @03:52PM (#34795789)
Yes, it seems that they re-implemented essentially all coreutils, and nearly all command line options on all of them. That's the reason why there are so many failures in the test suite, because some obscure option here and there is still not yet implemented. The main functionality of all the tools is already there. This is a huge project, and really impressive.

(as a side note, I did not want to turn this thread into a licensing flamewar, even if I dislike the license change. It is more interesting to talk about the tools themselves! How does the performance compare to the GNU implementation? Are they faster? Do they use less memory? Are the executables smaller (e.g., when compiled statically)? Do they compile faster?)