Merge pull request #37 from levineuwirth/worktree-tui-tree-sitter-cpp-dual-authority

tree-sitter-cpp/c + dual-authority styling (real C++ colors)

Cargo.lock

@@ -1264,6 +1264,8 @@ dependencies = [
  "thiserror 2.0.18",
  "toml",
  "tree-sitter",
+ "tree-sitter-c",
+ "tree-sitter-cpp",
  "tree-sitter-lua",
  "tree-sitter-md",
  "tree-sitter-rust",
@@ -2039,6 +2041,26 @@ dependencies = [
  "tree-sitter-language",
 ]
 
+[[package]]
+name = "tree-sitter-c"
+version = "0.24.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a9b2eb57a55fed6b00812912e730b7a275cf4fe98bfd6a5d76263d4438371728"
+dependencies = [
+ "cc",
+ "tree-sitter-language",
+]
+
+[[package]]
+name = "tree-sitter-cpp"
+version = "0.23.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "df2196ea9d47b4ab4a31b9297eaa5a5d19a0b121dceb9f118f6790ad0ab94743"
+dependencies = [
+ "cc",
+ "tree-sitter-language",
+]
+
 [[package]]
 name = "tree-sitter-language"
 version = "0.1.7"

Cargo.toml

@@ -101,6 +101,14 @@ tree-sitter = "0.26"
 # `crate::syntax::BUILTIN_LANGUAGES`.
 tree-sitter-rust = "0.24"
 tree-sitter-lua = "0.5"
+# T M_B3 — C / C++. Bundled so opening a `.c`/`.cpp`/header file in
+# the grid TUI gets lexical highlighting (keywords/strings/operators)
+# via tree-sitter, with `LspStyleView` layered on top for semantic
+# refinement (functions/types/macros via clangd's semantic tokens).
+# `.h` is ambiguous C / C++; the `c` entry below claims it by default,
+# matching the LSP filetype map in `builtin/runtime/lsp.lua`.
+tree-sitter-c = "0.24"
+tree-sitter-cpp = "0.23"
 # T M9.7: markdown grammar so prompt result buffers with
 # `_meta.format = "markdown"` get structured highlighting through the
 # same M4 path as rust/lua — no special-case painter in Lua.

builtin/runtime/lsp.lua

@@ -347,16 +347,17 @@ local function attach_buffer(buf)
   -- did_open is a notification; the manager queues it cleanly even
   -- while the server is in `starting` / `initializing`.
   pcall(pmacs.lsp.did_open, sid, uri, rec.version, active_buffer_text())
-  -- M_B1: policy A — when this buffer has no bundled tree-sitter
-  -- grammar (C/C++, …), push the LSP-driven style overlay onto the
-  -- active window so the grid TUI gets cell coloring from LSP
-  -- semantic tokens. Grammar-backed languages already have
-  -- `SyntaxHighlightView`; never both authorities on one buffer.
-  -- `language_for_path` returning nil is the grammar-less signal
-  -- (the LSP filetype map is consulted independently above).
-  if not styled_buffers[key]
-      and not pmacs.parse.language_for_path(path)
-  then
+  -- M_B3: dual-authority styling. Always push the LSP style overlay
+  -- when an LSP server is up — whether or not the buffer has a
+  -- bundled tree-sitter grammar. When the grammar exists too,
+  -- `SyntaxHighlightView` paints first (lexical: keywords, strings,
+  -- operators) and `LspStyleView` paints after (semantic: function /
+  -- type / macro / namespace identifiers from clangd's tokens); their
+  -- styles compose via `crate::overlay::merge_styles`, so the final
+  -- cell carries both authorities' contributions. This replaces
+  -- M_B1's policy-A exclusivity, which left grammar-backed languages
+  -- (Rust, C, C++) without LSP semantic refinement.
+  if not styled_buffers[key] then
     local ok, attached = pcall(pmacs.lsp._attach_style, buf)
     if ok and attached then styled_buffers[key] = true end
   end

src/highlight.rs

@@ -464,13 +464,28 @@ pub fn is_default_style(style: Style) -> bool {
 // LspStyleView
 // ---------------------------------------------------------------------------
 
-/// View that paints LSP semantic-token styling into cells for buffers
-/// with **no bundled tree-sitter grammar** (C/C++, …). Sibling of
-/// [`SyntaxHighlightView`] for the same policy-A "one styling
-/// authority per language" rule that `SemanticRenderState` applies on
-/// the semantic-frontend side — but feeding the grid TUI's existing
-/// cell-painter pipeline instead. Closes the visible "no syntax
-/// coloring for grammar-less languages in the TUI" gap.
+/// View that paints LSP semantic-token styling into cells, layered
+/// **alongside** [`SyntaxHighlightView`] when both exist for a
+/// buffer. The grid TUI's `View::render` composition stack means
+/// tree-sitter paints first (lexical: keywords, strings, operators,
+/// the things a parser identifies from the source's *shape*) and
+/// `LspStyleView` paints after (semantic: function vs variable,
+/// macro vs constant, the things a language server identifies from
+/// program *meaning*). Their styles compose through
+/// [`crate::overlay::merge_styles`], so the final cell carries both
+/// authorities' contributions — the `VSCode` / `Zed` "`TextMate` +
+/// LSP semantic tokens" model, on a terminal grid.
+///
+/// For languages with no bundled tree-sitter grammar (Python, Go,
+/// languages added later via LSP only), `LspStyleView` is the *sole*
+/// styling source and still works — the merge with an absent
+/// tree-sitter overlay is identity.
+///
+/// `M_B3` (this) dropped the earlier policy-A exclusivity gate in
+/// `lsp.lua` that restricted attachment to grammar-less buffers. The
+/// previous gate left grammar-backed languages (Rust, C, C++) without
+/// LSP semantic refinement; dual-authority delivers richer coloring
+/// without conflict.
 ///
 /// Holds shared handles only (`SharedLspManager` + `ThemeHandle`), so
 /// it survives buffer renames and LSP server restarts: every `render`
@@ -763,9 +778,11 @@ mod tests {
         let state = EditorState::new();
         let buffer_id = state.core.borrow().active_window().buffer_id;
 
-        // Seed the buffer: one line, `.cpp` path (no bundled grammar →
-        // SyntaxHighlightView never attaches, the LSP path owns the
-        // styling per policy A).
+        // Seed the buffer: one line, `.cpp` path. The test renders
+        // `LspStyleView` directly (no `SyntaxHighlightView` attached
+        // alongside in this fixture), so the asserted cells reflect
+        // the LSP authority alone — M_B3's dual-authority composition
+        // is exercised end-to-end by Lua's `attach_buffer`, not here.
         {
             let mut core = state.core.borrow_mut();
             core.registry

src/syntax.rs

@@ -377,6 +377,33 @@ pub const BUILTIN_LANGUAGES: &[LanguageEntry] = &[
         loader: || tree_sitter_md::LANGUAGE.into(),
         highlights_query: tree_sitter_md::HIGHLIGHT_QUERY_BLOCK,
     },
+    // T M_B3 — C / C++. Lexical highlighting (keywords / strings /
+    // operators) so the grid TUI shows code-shaped C++ on first open.
+    // `LspStyleView` layers on top with semantic refinement
+    // (functions / types / macros / namespaces) from clangd's
+    // semantic tokens — the two views' styles merge via
+    // `crate::overlay::merge_styles`.
+    //
+    // `.h` is ambiguous C / C++; the `c` entry below claims it
+    // (matches the LSP filetype map's default in `lsp.lua`). Users
+    // who want `.h` parsed as C++ can override via Lua.
+    // Note the const names: `tree-sitter-c` and `tree-sitter-cpp`
+    // expose `HIGHLIGHT_QUERY` (singular), matching the `tree-sitter-md`
+    // crate's `HIGHLIGHT_QUERY_BLOCK` style; `tree-sitter-rust` and
+    // `tree-sitter-lua` use `HIGHLIGHTS_QUERY` (plural). No semantic
+    // difference — same bundled `highlights.scm` either way.
+    LanguageEntry {
+        name: "c",
+        extensions: &["c", "h"],
+        loader: || tree_sitter_c::LANGUAGE.into(),
+        highlights_query: tree_sitter_c::HIGHLIGHT_QUERY,
+    },
+    LanguageEntry {
+        name: "cpp",
+        extensions: &["cpp", "cc", "cxx", "hpp", "hh", "hxx", "ipp", "inl", "cppm"],
+        loader: || tree_sitter_cpp::LANGUAGE.into(),
+        highlights_query: tree_sitter_cpp::HIGHLIGHT_QUERY,
+    },
 ];
 
 /// Registry that the Lua surface ([`crate::lua_bindings::install_parse`])
@@ -712,6 +739,38 @@ mod tests {
         bundle
     }
 
+    #[test]
+    fn builtin_languages_include_c_and_cpp() {
+        // M_B3 regression guard: a future refactor must not silently
+        // drop C / C++ tree-sitter coverage — the dual-authority
+        // styling in the TUI depends on these entries existing and
+        // claiming their canonical extensions.
+        let c = BUILTIN_LANGUAGES
+            .iter()
+            .find(|l| l.name == "c")
+            .expect("`c` language entry must be present");
+        assert!(c.extensions.contains(&"c"), "`c` claims `.c`");
+        assert!(
+            c.extensions.contains(&"h"),
+            "`c` claims `.h` (matches the LSP filetype map's default)"
+        );
+        assert!(
+            !c.highlights_query.is_empty(),
+            "`c` ships a non-empty highlights query"
+        );
+        let cpp = BUILTIN_LANGUAGES
+            .iter()
+            .find(|l| l.name == "cpp")
+            .expect("`cpp` language entry must be present");
+        for ext in ["cpp", "cc", "cxx", "hpp", "hh", "hxx"] {
+            assert!(cpp.extensions.contains(&ext), "`cpp` claims `.{ext}`");
+        }
+        assert!(
+            !cpp.highlights_query.is_empty(),
+            "`cpp` ships a non-empty highlights query"
+        );
+    }
+
     #[test]
     fn byte_to_point_handles_first_line() {
         let src = b"hello world";