Language LSP parity — findings (2026)
Language LSP parity — findings (2026)
Section titled “Language LSP parity — findings (2026)”Probe charter
Section titled “Probe charter”Compare what authors need (completion, hover, goto-def, rename, code actions, diagnostics freshness) against what vox-lsp negotiates and implements.
Evidence (2026-05-11)
Section titled “Evidence (2026-05-11)”- Validation path reuses lexer/parser/typechecker (+ optional HIR) — strong parity for diagnostics when the same flags apply.
- Web App Archetype Coverage Map explicitly lists LSP completion/hover/goto-def gaps for several archetypes — treat as authoritative backlog input.
Gaps (initial)
Section titled “Gaps (initial)”| Area | Risk |
|---|---|
| Cross-file symbols | Goto-def may miss generated or workspace-relative paths. |
| Completions vs SSOT builtins | Until Phase 1 codegen completes, completions may drift from typechecker manifest. |
| Code actions | Diagnostic embeds fixes; editor may not advertise full action kinds. |
Next steps
Section titled “Next steps”- Export capability list from
crates/vox-lsp/src/main.rsinitialize handler (manual table until scripted). - Add integration tests mirroring
crates/vox-integration-tests/tests/lsp_test.rspatterns for each advertised capability. - Refresh vox-lsp-capabilities-ssot-2026.md after measurements.