Native RMT Authoring

Native RMT documents without legacy JSON as the preferred authoring path.

What it covers

Native authoring keeps an .rmt file as the editable source of truth. Legacy JSON and generated core files are comparison or runtime artifacts, not places for handwritten product logic.

Public building blocks

• tools/rmt-language/vnext-parser.js reads native syntax.
• tools/rmt-linter/cli.js returns local diagnostics.
• tools/rmt-language/vnext-compiler.js emits the core model.

Recommended workflow

Create RMT source, lint it, and compile only after a clean parser run. Inspect the core diff and commit source with its expected artifact when semantics intentionally change.

Editor helpers

RMT Linter and RMT Language Server cover the linter, LSP, Code Actions and agent report. Common snippets are rmt-component for component records and rmt-template-dom for DOM descriptor templates. Check tooling regressions with node scripts/run_xtend_tests.js rmt-language-regression --json.

Next steps

• XTendRMT overview
• RMT Authoring Guide
• RMT Linter
• RMT Language Server

Local verification

node scripts/run_xtend_tests.js rmt-tooling-docs rmt-language-regression --json

The first gate checks the documented authoring path; the second checks parser, diagnostic, and editor parity. On failure, change the .rmt source or responsible tool first, not generated core JSON.