feat(#678): migrate C++ indexing to Rust-owned engine

CHANGELOG.md

@@ -9,6 +9,16 @@ and adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
 
+### New Features
+
+- `rust-hybrid` indexing now handles C++ baseline symbols through the Rust-owned path, so C++ functions, classes, structs, namespaces, enums, type aliases, includes, and calls no longer depend on TypeScript fallback extraction. (#678)
+- `.h` header files are now classified as C, C++, or Objective-C by content sniffing on both the Rust and TypeScript sides, using aligned regex patterns for consistent routing. (#678)
+- `rust-hybrid` metadata now includes `cpp` in `rustOwnedLanguages`, so C++ files are correctly assigned to the Rust engine with no TypeScript fallback. (#678)
+
+### Changes
+
+- Removed the TypeScript-owned C++ extractor (`c-cpp.ts`) and `tree-sitter-cpp.wasm` grammar dependency, as C++ extraction is now fully handled by the Rust core via `tree-sitter-cpp` crate. (#678)
+
 
 ## [0.11.0] - 2026-07-14
 

README.md

@@ -587,8 +587,8 @@ zcodegraph index --engine typescript
 
 | State | Meaning | Current languages / files |
 |---|---|---|
-| Rust-owned | Indexed by the Rust core on the default `rust-hybrid` path. | JavaScript, JSX, TypeScript, TSX, Go, Python, Rust, C, Java. |
-| TS-indexed | Indexed by the TypeScript indexer as the mature multi-language path. | C#, PHP, Ruby, C++, Objective-C, Swift, Kotlin, Scala, Dart, Svelte, Vue, Liquid, Pascal/Delphi, Lua, Luau, and other supported non-Rust-owned sources. |
+| Rust-owned | Indexed by the Rust core on the default `rust-hybrid` path. | JavaScript, JSX, TypeScript, TSX, Go, Python, Rust, C, C++, Java. |
+| TS-indexed | Indexed by the TypeScript indexer as the mature multi-language path. | C#, PHP, Ruby, Objective-C, Swift, Kotlin, Scala, Dart, Svelte, Vue, Liquid, Pascal/Delphi, Lua, Luau, and other supported non-Rust-owned sources. |
 | Hybrid fallback | `rust-hybrid` uses TypeScript fallback for a file or language and reports it in status/doctor. | Expected for non-Rust-owned supported files, and for Rust-owned parse gaps when recoverable. |
 | Not covered | Not indexed as source symbols/edges. | Unsupported extensions, ignored paths, default-excluded dependency/build/cache directories, and files over the size limit. |
 

__tests__/rust-index-engine-cli-fallback.test.ts

@@ -77,10 +77,10 @@ describe('zcodegraph rust-hybrid fallback degraded status and doctor output', ()
     const plan = planRustHybridAssignments(tempDir);
 
     expect(plan.rustOwnedFiles).toContain('plain.h');
-    expect(plan.fallbackFiles).toContain('widget.h');
+    expect(plan.rustOwnedFiles).toContain('widget.h');
     expect(plan.fallbackFiles).toContain('View.h');
-    expect(plan.engineByLanguage).toMatchObject({ c: 'rust', cpp: 'typescript', objc: 'typescript' });
-    expect(plan.fallbackByLanguage).toMatchObject({ cpp: 1, objc: 1 });
+    expect(plan.engineByLanguage).toMatchObject({ c: 'rust', cpp: 'rust', objc: 'typescript' });
+    expect(plan.fallbackByLanguage).toMatchObject({ objc: 1 });
   });
 
   it('records Rust-owned per-file gap diagnostics without same-language TypeScript fallback append', () => {

__tests__/rust-index-engine-cli-language-smoke.test.ts

@@ -363,6 +363,87 @@ describe('zcodegraph rust index language framework and MCP smoke behavior', () =
     }
   }, 30_000);
 
+  it('indexes C++ as Rust-owned under rust-hybrid', () => {
+    fs.writeFileSync(
+      path.join(tempDir, 'widget.h'),
+      [
+        '#pragma once',
+        '',
+        'namespace app {',
+        'class Widget {',
+        'public:',
+        '  Widget();',
+        '  ~Widget();',
+        '  int render();',
+        '};',
+        '}',
+        '',
+        'using WidgetPtr = Widget*;',
+      ].join('\n') + '\n',
+    );
+    fs.writeFileSync(
+      path.join(tempDir, 'main.cpp'),
+      [
+        '#include <cstdio>',
+        '#include "widget.h"',
+        '',
+        'namespace app {',
+        'Widget::Widget() {}',
+        'Widget::~Widget() {}',
+        'int Widget::render() { return 42; }',
+        '}',
+        '',
+        'int main() {',
+        '  Widget w;',
+        '  return w.render();',
+        '}',
+      ].join('\n') + '\n',
+    );
+
+    const result = runZcodegraphCli(tempDir, ['index', '--quiet'], {
+      ZCODEGRAPH_RUST_CORE_BINARY: RUST_CORE_BIN,
+    });
+
+    expect(result.status, `stdout:\n${result.stdout}\nstderr:\n${result.stderr}`).toBe(0);
+    const cg = CodeGraph.openSync(tempDir);
+    try {
+      expect(cg.getStats().filesByLanguage.cpp).toBe(2);
+      const expectations = [
+        ['cstdio', 'import'],
+        ['widget.h', 'import'],
+        ['Widget', 'class'],
+        ['app', 'namespace'],
+        ['WidgetPtr', 'type_alias'],
+        ['main', 'function'],
+      ] as const;
+      for (const [name, kind] of expectations) {
+        expect(
+          cg.searchNodes(name).some((match) => match.node.name === name && match.node.kind === kind && match.node.language === 'cpp'),
+          `${name} (${kind}) should be indexed as C++`,
+        ).toBe(true);
+      }
+
+      const mainFn = cg.searchNodes('main').find((match) => match.node.kind === 'function' && match.node.language === 'cpp')?.node;
+      const renderFn = cg.searchNodes('render').find((match) => match.node.kind === 'function' && match.node.language === 'cpp')?.node;
+      expect(mainFn).toBeDefined();
+      expect(renderFn).toBeDefined();
+      const calls = cg.getOutgoingEdges(mainFn!.id).filter((edge) => edge.kind === 'calls');
+      expect(calls.some((edge) => edge.target === renderFn!.id)).toBe(true);
+
+      const buildInfo = cg.getIndexBuildInfo();
+      expect(buildInfo.engine).toBe('rust-hybrid');
+      expect(buildInfo.hybrid).toMatchObject({
+        rustOwnedLanguages: expect.arrayContaining(['cpp']),
+        engineByLanguage: { cpp: 'rust' },
+        fallbackByLanguage: {},
+        fallbackFileCount: 0,
+        fallbackReasonTaxonomy: {},
+      });
+    } finally {
+      cg.close();
+    }
+  }, 30_000);
+
   it('reports Rust index-engine metadata through MCP status', async () => {
     const indexResult = runZcodegraphCli(tempDir, ['index', '--engine', 'rust', '--quiet'], {
       ZCODEGRAPH_RUST_CORE_BINARY: RUST_CORE_BIN,

crates/zcodegraph-core/Cargo.toml

@@ -7,12 +7,14 @@ publish = false
 [dependencies]
 dhat = { version = "0.3", optional = true }
 libc = "0.2"
+regex = "1"
 rusqlite = { version = "0.32", features = ["bundled"] }
 serde = { version = "1", features = ["derive"] }
 serde_json = "1"
 sha2 = "0.10"
 tree-sitter = "0.24"
 tree-sitter-c = "0.23"
+tree-sitter-cpp = "0.23"
 tree-sitter-go = "0.23"
 tree-sitter-java = "0.23"
 tree-sitter-javascript = "0.23"