Issue #1170: std: lossy UTF-8 decode of bytes and a raw byte write to stdout

Open and answer this issue on GitHub.

ready · opened 2026-11-02 · updated 2026-11-02 · 0 comments

Found during the M8 move (Part of #125). Moving the rest of project.rs run_tests is blocked on this: the failures report prints a test’s captured stdout as lossy UTF-8, and std has no form for that.

Supervisor decision (an obvious extension by the language principles): std gets

• a lossy decode from a byte list to String, where invalid sequences become U+FFFD, the same as Rust’s String::from_utf8_lossy;
• a write of raw bytes to stdout.

A strict decode stays the default. The lossy one is named so that a reader sees it replaces bytes.

To build

1. Add both to std (the .nika side, plus the runtime hook if one is needed), each with a tests/compiler/ program and its .stdout. One test must give invalid UTF-8 and show U+FFFD.
2. Move the failures report of project.rs (ProjectTestsHost’s report side) into project_tests.nika.

Back to Open Issues