moh-eulith
moh-eulith
According to [this](https://github.com/rust-lang/rustup/issues/2417), the solution is >The solution is to remove the entire toolchain and reinstall it with all the desired components. There is no need to uninstall all of...
Pushed new code that addresses both the existing unicode issues and adds left truncation.
Diff was just trying too hard to make sense of a simple delete/add. Moved the add and now diff is a lot more readable.
Is the diff ok now?
I had a look at this. unicode encoding makes this diabolically difficult. The existing code is already running into this issue, so maybe fixing the unicode issues is the first...
I'm not clear on what the `buf` contents are that are passed to the `write` function. Is the content guaranteed to be on byte, unicode code point, or grapheme boundary?
So a full fix would require some manner of buffering. Max encoding length for utf8 is 4 bytes, another 2 for combining, so that's a 6 byte buffer. I'm by...
Agreed that grapheme counting is the right way to go. However, because we don't have a guarantee on the `buf` segmentation, it leads to failures: ```rust #[test] fn char_starts_gs() {...
Sorry, should've made the example simpler. I think the minimal test case that fails `test_all_splits` is: ```rust let buf = "e\u{0331}"; test_all_splits(buf, 1); ```
This issue is easily reproducible. Any chance we could get a fix?