Context
10 home-os kernel modules use `switch` statements with match-style `=>` arrows:
```home switch (event) { BRCMF_E_LINK => { handle_link() } BRCMF_E_DISCONNECT => { handle_disconnect() } } ```
The parser expects `case`/`default` keywords inside `switch` and rejects `=>` arrows. Meanwhile, `match` already uses `=>` arrows — so the kernel writers (reasonably) reached for the same form in `switch`.
Affected files
`kernel/src/drivers/wifi/broadcom.home`, `kernel/src/drivers/wifi/iwlwifi.home`, `kernel/src/i18n/bidi.home`, plus 7 more (full list in `~/Desktop/home-os-kernel-audit-2026-05-04-pm.tsv`, filter `first_error` = `Expected 'case' or 'default' in switch statement`).
Question for the language design
Two options:
- Unify `switch` and `match` — accept `=>` arms in `switch` too. They become syntactic siblings; `match` is the more general form (with destructuring, guards, etc.) and `switch` is the simple-equality flavor. Or even drop `switch` as a keyword in favor of just `match` everywhere.
- Keep them distinct — `switch` requires `case Pattern: ...` C-style, `match` keeps `=>` arrows. Mass-migrate the home-os kernel to `case`/`default` style.
Option 1 is simpler for users and matches what the kernel code already assumes. Option 2 is more conventional (Zig has `switch` with `=>`; Rust has `match` with `=>`; C/Go/JS have `switch` with `case`).
Recommended: Option 1, with `switch` accepting `=>` arms. Aligns with what kernel writers naturally wrote and removes a syntactic gotcha. `match` and `switch` differ in features (destructuring, exhaustiveness checking) but share arm syntax.
Why this matters
10 modules unblocked once decided. #2 of 5 ranked follow-ups in the 2026-05-04-pm audit. Listed as `question` label until the design call is made.
Reference
- Audit: `~/Desktop/home-os-kernel-audit-2026-05-04-pm.tsv`