News for hackers. Stuff that matters.
Re:Go is still not good by kragen on Friday August 22, 2025 @06:01PM (#44990389)
Isn't &[u8] what you should be using for command-line arguments and filenames and whatnot? In that case you'd want its name to be short, like &[u8], rather than long like &[bytes] or &[raw_uncut_byte] or something.
Re:Go is still not good by adastra22 on Friday August 22, 2025 @07:44PM (#44991326)
OsStr/OsString is what you would use in those circumstances. Path/PathBuf specifically for filenames or paths, which I think uses OsStr/OsString internally. I've never looked at OsStr's internals but I wouldn't be surprised if it is a wrapper around &[u8].

Note that &[u8] would allow things like null bytes, and maybe other edge cases.

Re:Go is still not good by kragen on Friday August 22, 2025 @08:20PM (#44991651)
You can't get null bytes from a command-line argument. And going by https://news.ycombinator.com/item?id=44991638 it's common to not use OsString when accepting command-line arguments, because std::env::args yields Strings, which means that probably most Rust programs that accept filenames on the command line have this bug.
Re:Go is still not good by adastra22 on Friday August 22, 2025 @09:03PM (#44991925)
Rust String can contain null bytes! Rust uses explicit string lengths. Agree though that most OS wouldn't be able to pass null bytes in arguments though.
Re:Go is still not good by kragen on Friday August 22, 2025 @09:09PM (#44991971)
Right, but it can't contain invalid UTF-8, which is valid in both command-line parameters and in filenames on Linux, FreeBSD, and other normal Unixes. See my link above for a demonstration of how this causes bugs in Rust programs.