Add a distinct C++ parser

cmd/code-index/cmd/parse.go

@@ -86,10 +86,16 @@ func runParse(cmd *cobra.Command, args []string) error {
 				return fmt.Errorf("parsing R source %s: %w", source.Path, err)
 			}
 
-		case "c", "cpp":
+		case "c":
 			parser := indexer.NewCParser(srcPath, source.Exclude)
 			if err := parser.Parse(result); err != nil {
-				return fmt.Errorf("parsing C/C++ source %s: %w", source.Path, err)
+				return fmt.Errorf("parsing C source %s: %w", source.Path, err)
+			}
+
+		case "cpp":
+			parser := indexer.NewCPPParser(srcPath, source.Exclude)
+			if err := parser.Parse(result); err != nil {
+				return fmt.Errorf("parsing C++ source %s: %w", source.Path, err)
 			}
 
 		case "markdown":

docs/configuration.md

@@ -90,11 +90,13 @@ Each source object defines a directory to index:
 | `typescript` | `.ts`, `.tsx`, `.js`, `.jsx`, `.vue` | tree-sitter — functions, classes, interfaces, enums, JSDoc |
 | `javascript` | `.js`, `.jsx` | tree-sitter (same as typescript) |
 | `python` | `.py` | tree-sitter — functions, classes, decorators, docstrings |
-| `c` | `.c`, `.h`, `.cpp`, `.cc`, `.hpp` | tree-sitter — functions, structs, classes, enums, typedefs, Doxygen |
-| `cpp` | `.c`, `.h`, `.cpp`, `.cc`, `.hpp` | tree-sitter (same as c) |
+| `c` | `.c`, `.h` | tree-sitter — functions, structs, enums, typedefs, Doxygen |
+| `cpp` | `.cpp`, `.cc`, `.hpp`, `.cxx`, `.hxx` | tree-sitter — functions, classes, structs, namespaces, templates, enums, typedefs, Doxygen |
 | `r` | `.R`, `.r` | Native Rscript with regex fallback — functions, roxygen, S4/R6 classes |
 | `markdown` | `.md`, `.qmd` | Regex — headings as sections, YAML front matter |
 
+> **Note:** `.h` files are handled by the `c` parser, not `cpp`. This avoids double-parsing when a project configures both `c` and `cpp` sources over the same tree. C++ projects with `.h` headers should configure a `c` source pointing at the header directories.
+
 #### Vendor-aware Go indexing
 
 For Go projects, you can include specific vendored dependencies in the index:

indexer/parser_c.go

@@ -10,24 +10,21 @@ import (
 
 	sitter "github.com/smacker/go-tree-sitter"
 	treesitterc "github.com/smacker/go-tree-sitter/c"
-	"github.com/smacker/go-tree-sitter/cpp"
 )
 
-// CParser extracts structured information from C/C++ source files using tree-sitter.
+// CParser extracts structured information from C source files using tree-sitter.
 type CParser struct {
 	srcRoot  string
 	excludes []string
 	cLang    *sitter.Language
-	cppLang  *sitter.Language
 }
 
-// NewCParser creates a new C/C++ parser.
+// NewCParser creates a new C parser.
 func NewCParser(srcRoot string, excludes []string) *CParser {
 	return &CParser{
 		srcRoot:  srcRoot,
 		excludes: excludes,
 		cLang:    treesitterc.GetLanguage(),
-		cppLang:  cpp.GetLanguage(),
 	}
 }
 
@@ -47,9 +44,7 @@ func (p *CParser) Parse(result *ParseResult) error {
 		}
 
 		ext := filepath.Ext(path)
-		isCPP := ext == ".cpp" || ext == ".cc" || ext == ".hpp" || ext == ".cxx"
-		isC := ext == ".c" || ext == ".h"
-		if !isC && !isCPP {
+		if ext != ".c" && ext != ".h" {
 			return nil
 		}
 
@@ -67,22 +62,18 @@ func (p *CParser) Parse(result *ParseResult) error {
 			}
 		}
 
-		return p.parseFile(path, relPath, isCPP, result)
+		return p.parseFile(path, relPath, result)
 	})
 }
 
-func (p *CParser) parseFile(path, relPath string, isCPP bool, result *ParseResult) error {
+func (p *CParser) parseFile(path, relPath string, result *ParseResult) error {
 	content, err := os.ReadFile(path)
 	if err != nil {
 		return nil
 	}
 
 	parser := sitter.NewParser()
-	if isCPP {
-		parser.SetLanguage(p.cppLang)
-	} else {
-		parser.SetLanguage(p.cLang)
-	}
+	parser.SetLanguage(p.cLang)
 
 	tree, err := parser.ParseCtx(context.Background(), nil, content)
 	if err != nil {
@@ -102,7 +93,7 @@ func (p *CParser) parseFile(path, relPath string, isCPP bool, result *ParseResul
 		ImportPath: importPath,
 	}
 
-	p.extractDeclarations(root, content, fileInfo, isCPP)
+	p.extractDeclarations(root, content, fileInfo)
 
 	fileInfo.ASTHash = hashString(string(content))
 
@@ -121,7 +112,7 @@ func (p *CParser) parseFile(path, relPath string, isCPP bool, result *ParseResul
 	return nil
 }
 
-func (p *CParser) extractDeclarations(node *sitter.Node, content []byte, fileInfo *FileInfo, isCPP bool) {
+func (p *CParser) extractDeclarations(node *sitter.Node, content []byte, fileInfo *FileInfo) {
 	for i := 0; i < int(node.ChildCount()); i++ {
 		child := node.Child(i)
 		nodeType := child.Type()
@@ -153,26 +144,10 @@ func (p *CParser) extractDeclarations(node *sitter.Node, content []byte, fileInf
 				fileInfo.Types = append(fileInfo.Types, *t)
 			}
 
-		// C++ specific
-		case "class_specifier":
-			if isCPP {
-				if t := p.extractClass(child, content, fileInfo.Path); t != nil {
-					fileInfo.Types = append(fileInfo.Types, *t)
-				}
-			}
-
-		case "namespace_definition":
-			if isCPP {
-				// Recurse into namespace body.
-				if body := child.ChildByFieldName("body"); body != nil {
-					p.extractDeclarations(body, content, fileInfo, isCPP)
-				}
-			}
-
 		// Preprocessor blocks — recurse into their bodies to find declarations
 		// inside #ifndef/#ifdef/#if guards (common in header files).
 		case "preproc_ifdef", "preproc_if", "preproc_else", "preproc_elif":
-			p.extractDeclarations(child, content, fileInfo, isCPP)
+			p.extractDeclarations(child, content, fileInfo)
 		}
 	}
 }
@@ -319,26 +294,6 @@ func (p *CParser) extractTypedef(node *sitter.Node, content []byte, filePath str
 	}
 }
 
-func (p *CParser) extractClass(node *sitter.Node, content []byte, filePath string) *TypeInfo {
-	nameNode := node.ChildByFieldName("name")
-	if nameNode == nil {
-		return nil
-	}
-
-	name := nameNode.Content(content)
-	doc := extractPrecedingComment(node, content)
-
-	return &TypeInfo{
-		Name:     name,
-		Kind:     "class",
-		Doc:      doc,
-		File:     filePath,
-		Line:     int(node.StartPoint().Row) + 1,
-		Exported: true,
-		ASTHash:  hashString(node.Content(content)),
-	}
-}
-
 // extractCName finds the function/variable name from a declarator node.
 func extractCName(declarator *sitter.Node, content []byte) string {
 	// Walk down through pointer_declarator, function_declarator, etc.
@@ -371,19 +326,12 @@ func extractCSignature(node *sitter.Node, content []byte) string {
 	// Take everything before the function body.
 	body := node.ChildByFieldName("body")
 	if body != nil {
-		sig := strings.TrimSpace(string(content[node.StartByte():body.StartByte()]))
-		if len(sig) > 200 {
-			sig = sig[:200] + "..."
-		}
-		return sig
+		return truncateSignature(strings.TrimSpace(string(content[node.StartByte():body.StartByte()])))
 	}
 	// Fallback: first line.
 	text := node.Content(content)
 	if idx := strings.Index(text, "{"); idx > 0 {
 		text = strings.TrimSpace(text[:idx])
 	}
-	if len(text) > 200 {
-		text = text[:200] + "..."
-	}
-	return text
+	return truncateSignature(text)
 }

indexer/parser_common.go

@@ -0,0 +1,11 @@
+// Copyright (C) 2026 by Posit Software, PBC
+package indexer
+
+const maxSignatureLen = 200
+
+func truncateSignature(sig string) string {
+	if len(sig) <= maxSignatureLen {
+		return sig
+	}
+	return sig[:maxSignatureLen] + "..."
+}