@hyperscript-tools/language-server 7.1.0 → 7.3.0

package/CHANGELOG.md

@@ -7,6 +7,72 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [7.3.0] - 2026-10-21
+
+`@hyperfixi/mcp-server` lists its 14 core tools by default and keeps the other 70 in toolsets
+that `LOKASCRIPT_MCP_TOOLSETS` lists, and five of its tools that were broken work. A minor, as
+when the sampling tools went opt-in (2.11.0): every tool still answers a call by name, and the
+variable lists any of them again.
+
+### Changed
+
+- **`@hyperfixi/mcp-server` lists 14 core tools by default; the other 70 are in opt-in
+  toolsets.** A connection listed all 84: about 13k tokens of definitions in each client that
+  loads them all, and a month of sessions called 7 of them. The default list is the agent loop
+  and what it leans on (`validate_and_compile`, `get_diagnostics`, `get_code_fixes`,
+  `get_command_docs`, `get_expression_docs`, `search_patterns`), translation (`translate_code`,
+  `translate_to_english`, `get_keyword_translations`, `list_supported_languages`), proof that
+  a change kept its meaning (`score_fidelity`, `diff_behaviors`), and `detect_domain` /
+  `compile_auto`: about 2.5k tokens. Set `LOKASCRIPT_MCP_TOOLSETS` to list more: `domains`
+  (26), `lse` (13), `grail` (5), `debug` (4), `editor` (3), `analysis` (9), `profiles` (5),
+  `project` (5), comma-separated, or `all`. Only the list changes: every tool still answers a
+  call by name. A client that picks tools from the list (Claude Code, Claude Desktop) needs
+  the variable to see a tool outside the core; the README lists each toolset's tools (#1607).
+
+### Fixed
+
+- **`lse_from_hyperscript` returned `[get patient:null]` for every input**, with `ok: true` and
+  confidence 1: it read the semantic parse as the AST builder's tree. It now renders the parse,
+  and refuses (`ok: false`, the rendered LSE under `partial`) a parse that left input unread or
+  a shape whose LSE does not parse back, such as a `then` chain inside a handler (#1607).
+- **`extract_routes` and `generate_server_routes` threw on every input** ("Cannot read
+  properties of undefined"), from the day they were added: they read `.routes` from the array
+  `scanRoutes` returns. They report server-bridge's route conflicts where they promised an
+  `errors` field it never had. The import that hid this from the typecheck (an untyped ambient
+  module) is gone (#1607).
+- **`get_code_fixes` had no fixes for the codes `validate_and_compile` emits**
+  (`ENGINE_PARSE_ERROR`, `UNCONSUMED_INPUT`, …), though the server's instructions send an agent
+  there from its diagnostics; `get_diagnostics`' `parse-error` and `low-confidence` mapped to
+  keys that did not exist. Every code compilation-service emits now has fixes, and a test holds
+  that (#1607).
+- **`explain_code` listed commands by kind, not in the order they run** (toggle → wait → add
+  came out "toggles … adds … waits") and dropped their targets. It explains the parse now: each
+  step rendered back to English with what it does, a block's body and an `if`'s two branches (#1607).
+- **`get_diagnostics`' typo fix was titled `Fix typo: ${suggestion}`**, unfilled, with no edit.
+  It now reads "Fix typo: toggle" with an edit that replaces the misspelled word (#1607).
+
+## [7.2.0] - 2026-10-20
+
+`@hyperfixi/mcp-server` moves to `@lokascript/domains` 7.1.0: learn left the domains for
+lokascript-learn, its only user, and the server's four learn tools went with it. A minor, though
+it removes tools: there are no known users of them, and `translate_code` writes what they wrote.
+
+### Changed
+
+- **`@hyperfixi/mcp-server` is on `@lokascript/domains` 7.1.0.** The domain DSLs take each
+  language's grammar from `@lokascript/framework/languages` (framework 7.1); the six that remain
+  read and write exactly as on 5.0.4 (the domains' round-trip gate: 0 worse, 0 better, 0 parses
+  moved). Domains 7.1.0 peers framework and intent `^7.1.0`; semantic is no longer a peer
+  (#1604, #1605).
+
+### Removed
+
+- **`@hyperfixi/mcp-server`'s four learn tools** (`parse_learn`, `compile_learn`,
+  `validate_learn`, `translate_learn`), and learn in `detect_domain` and the composite tools.
+  learn left `@lokascript/domains` in 7.1.0 for lokascript-learn, its only user. Its commanding
+  sentence was `@lokascript/semantic`'s translation, which `translate_code` writes. A default
+  connection lists 84 tools (88 before) (#1605).
+
 ## [7.1.0] - 2026-10-20
 
 Each language's grammar is framework's data now (`@lokascript/framework/languages/<code>`), and
@@ -2194,7 +2260,8 @@ _Synchronized version release. See git history for details._
 - npm access token stored in GitHub Secrets
 - 2FA recommended for npm organization
 
-[Unreleased]: https://github.com/codetalcott/hyperfixi/compare/v7.1.0...HEAD
+[Unreleased]: https://github.com/codetalcott/hyperfixi/compare/v7.2.0...HEAD
+[7.2.0]: https://github.com/codetalcott/hyperfixi/compare/v7.1.0...v7.2.0
 [7.1.0]: https://github.com/codetalcott/hyperfixi/compare/v7.0.0...v7.1.0
 [7.0.0]: https://github.com/codetalcott/hyperfixi/compare/v6.3.0...v7.0.0
 [6.3.0]: https://github.com/codetalcott/hyperfixi/compare/v6.2.0...v6.3.0

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@hyperscript-tools/language-server",
-  "version": "7.1.0",
+  "version": "7.3.0",
   "description": "Language Server Protocol implementation for original _hyperscript — completions, diagnostics, hover docs",
   "type": "module",
   "main": "dist/server.js",
@@ -25,7 +25,7 @@
   "author": "HyperFixi Contributors",
   "license": "MIT",
   "dependencies": {
-    "@lokascript/language-server": "^7.1.0"
+    "@lokascript/language-server": "^7.3.0"
   },
   "devDependencies": {
     "@types/node": "^26.6.5",