Introducing tmux-rs

Original Article ↗ Hacker News Discussion ↗

Hobby Motivation and Reception

  • Many commenters appreciate the “for fun” motivation and compare it to other hobby rewrites (e.g., fzf clones) used to learn Rust or algorithms.
  • There’s broad support for experimentation without a business case; some argue this kind of tinkering is how real innovation often appears later.

Porting Strategy: 100% Unsafe Rust First

  • The project is essentially a transliteration of tmux’s C code into “C written in Rust,” using raw pointers and many unsafe blocks.
  • Several people note this is a common two‑step pattern:
    • Step 1: get a faithful, mostly mechanical port working (often largely unsafe).
    • Step 2: progressively refactor into safe, idiomatic Rust.
  • Others criticize the approach: code is ~20–25% larger, still unsafe, and currently less stable than the battle‑tested C version.

Rust vs C (and Go/Zig): Safety, Portability, Value

  • Pro‑Rust side:
    • Even unsafe Rust is safer than C because all “dangerous” regions are explicitly marked and easier to audit.
    • Rewriting in a memory‑safe language is seen as a long‑term win for extensibility, maintainability, and reducing whole classes of bugs, including security issues.
  • Skeptical side:
    • tmux is already extremely stable with few CVEs; some see little practical gain from a risky rewrite.
    • Concerns about Rust’s portability (especially on some OpenBSD targets or obscure platforms) vs C’s ubiquity.
    • Some argue a garbage‑collected language (e.g., Go) would be perfectly adequate given tmux’s IO‑bound nature.
    • A few feel Rust hype is driving interest more than concrete benefits here.

Automated Translation: c2rust and LLMs

  • The author’s experience with c2rust: fast but produced bloated, unidiomatic, hard‑to‑maintain code; eventually discarded in favor of manual porting.
  • Discussion suggests c2rust might improve (e.g., preserving constants), but currently isn’t good enough for clean, maintainable Rust.
  • LLMs and tools like Cursor were tried late in the process:
    • They reduced typing fatigue but still inserted subtle bugs, requiring as much review as manual coding.
    • Opinions split: some see automated C→Rust translation as a “killer app” for future AI; others are deeply skeptical that current models can handle a non‑trivial codebase reliably.

Tmux Usage, Issues, and Alternatives

  • Users reaffirm tmux as “life in the terminal”: session managers (tmuxinator/rmuxinator), long histories, multiplexing across projects.
  • Reported issues: memory use with large scrollback, mouse behavior, keybinding ergonomics, and desires for features like better Windows support or remote backends.
  • Comparisons:
    • GNU screen vs tmux: defaults (status bar, keybindings) and splits cited as reasons tmux won mindshare.
    • zellij (Rust multiplexer) is praised but seen as still missing some tmux features and keybinding flexibility.
  • Some doubt tmux maintainers would ever adopt this port; in its current “C-in-Rust” state it’s seen more as an educational fork than a drop‑in successor.