improved frontmatter processing, added tests

internal/builder/build_page.go

@@ -50,32 +50,6 @@ func processWithYaml(f []byte) (Metadata, []byte, error) {
 	return meta, []byte(split[2]), nil
 }
 
-func processFrontmatter(p string) (Metadata, error) {
-	// TODO: Also save index of the line where the actual document starts?
-	f, err := util.ReadFile(p)
-	if err != nil {
-		return nil, err
-	}
-	// Check if the file has valid metadata
-	trimmed := bytes.TrimSpace(f)
-	normalized := strings.ReplaceAll(string(trimmed), "\r\n", "\n")
-	if !strings.HasPrefix(normalized, ("---\n")) {
-		// No frontmatter, return nil -- handled by caller
-		return nil, nil
-	}
-	// Separate YAML from rest of document
-	split := strings.SplitN(normalized, "---\n", 3)
-	if len(split) < 3 {
-		return nil, fmt.Errorf("invalid frontmatter format")
-	}
-	var meta Metadata
-	// Parse YAML
-	if err := yaml.Unmarshal([]byte(split[1]), &meta); err != nil {
-		return nil, err
-	}
-	return meta, nil
-}
-
 func buildPageData(m Metadata, in string, out string, settings *Settings) *PageData {
 	p := &PageData{}
 	if title, ok := m["title"].(string); ok {

internal/builder/frontmatter.go

@@ -0,0 +1,81 @@
+package builder
+
+import (
+	"bufio"
+	"bytes"
+	"errors"
+	"fmt"
+	"os"
+
+	"gopkg.in/yaml.v3"
+)
+
+func processFrontmatter(p string) (Metadata, int, error) {
+	f, l, err := readFrontmatter(p)
+	if err != nil {
+		return nil, l, err
+	}
+	var meta Metadata
+	// Parse YAML
+	if err := yaml.Unmarshal(f, &meta); err != nil {
+		return nil, l, fmt.Errorf("yaml frontmatter could not be parsed: %w", err)
+	}
+	return meta, l, nil
+}
+
+// readFrontmatter reads the file at `path` and scans
+// it for --- delimited frontmatter. It does not attempt
+// to parse the data, it only scans for the delimiters.
+// It returns the frontmatter contents as a byte array
+// and its length in lines.
+func readFrontmatter(path string) ([]byte, int, error) {
+	file, err := os.Open(path)
+	if err != nil {
+		return nil, 0, err
+	}
+	defer file.Close()
+
+	lines := make([]string, 0, 10)
+	s := bufio.NewScanner(file)
+	i := 0
+	delims := 0
+	for s.Scan() {
+		l := s.Text()
+		if l == `---` {
+			if i > 0 && delims == 0 {
+				// if --- is not the first line, we
+				// assume the file does not contain frontmatter
+				return nil, 0, nil
+			}
+			delims += 1
+			i += 1
+			if delims == 2 {
+				break
+			}
+		} else {
+			lines = append(lines, l)
+			i += 1
+		}
+	}
+	// check whether any errors occurred while scanning
+	if err := s.Err(); err != nil {
+		return nil, 0, err
+	}
+	if delims == 2 {
+		l := len(lines)
+		if l == 0 {
+			// no valid frontmatter
+			return nil, 0, errors.New("frontmatter cannot be empty")
+		}
+		// convert to byte array
+		var b bytes.Buffer
+		for _, line := range lines {
+			b.WriteString(line + "\n")
+		}
+		return b.Bytes(), l, nil
+	} else {
+		// not enough delimiters, don't
+		// treat as frontmatter
+		return nil, 0, errors.New("frontmatter is missing closing delimiter")
+	}
+}

internal/builder/frontmatter_test.go

@@ -28,10 +28,13 @@ This is the body of the document.`
 	}
 
 	// Test the processFrontmatter function with valid frontmatter
-	meta, err := processFrontmatter(tmpfile.Name())
+	meta, l, err := processFrontmatter(tmpfile.Name())
 	if err != nil {
 		t.Fatalf("processFrontmatter failed: %v", err)
 	}
+	if l != 2 {
+		t.Errorf("Expected length 2, got %d", l)
+	}
 
 	if meta["title"] != "Test Title" || meta["description"] != "Test Description" {
 		t.Errorf("Expected title 'Test Title' and description 'Test Description', got title '%s' and description '%s'", meta["title"], meta["description"])
@@ -41,6 +44,7 @@ This is the body of the document.`
 	invalidContent := `---
 title: "Test Title"
 description: "Test Description"
+There is no closing delimiter???
 This is the body of the document.`
 
 	tmpfile, err = os.CreateTemp("", "testfile")
@@ -57,7 +61,30 @@ This is the body of the document.`
 	}
 
 	// Test the processFrontmatter function with invalid frontmatter
-	_, err = processFrontmatter(tmpfile.Name())
+	_, _, err = processFrontmatter(tmpfile.Name())
+	if err == nil {
+		t.Fatalf("Expected error for invalid frontmatter, got nil")
+	}
+	// Create a temporary file with invalid frontmatter
+	invalidContent = `---
+---
+This is the body of the document.`
+
+	tmpfile, err = os.CreateTemp("", "testfile")
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer os.Remove(tmpfile.Name()) // clean up
+
+	if _, err := tmpfile.Write([]byte(invalidContent)); err != nil {
+		t.Fatal(err)
+	}
+	if err := tmpfile.Close(); err != nil {
+		t.Fatal(err)
+	}
+
+	// Test the processFrontmatter function with invalid frontmatter
+	_, _, err = processFrontmatter(tmpfile.Name())
 	if err == nil {
 		t.Fatalf("Expected error for invalid frontmatter, got nil")
 	}

internal/builder/process.go

@@ -70,7 +70,7 @@ func processFile(inPath string, entry fs.DirEntry, err error, outRoot string, se
 	hasFrontmatter := false
 	if toProcess {
 		// process its frontmatter here
-		m, err := processFrontmatter(inPath)
+		m, _, err := processFrontmatter(inPath)
 		if err != nil {
 			return err
 		}