v2-Neuerstellung

internal/article/load.go

@@ -1,8 +1,5 @@
-// internal/article/load.go
 package article
 
-// internal/article/load.go  (gekürzt)
-
 import (
 	"bytes"
 	"encoding/json"
@@ -19,15 +16,24 @@ import (
 	"github.com/gomarkdown/markdown/html"
 )
 
-// LoadDir liest alle *.html und *.md unter root/content,
-// parst Front‑Matter (JSON‑Kommentar in der 1. Zeile) und liefert []Article.
-func LoadDir(root string) ([]Article, error) {
-	var out []Article
-	seen := make(map[string]bool) // Duplikats‑Check
+type LoaderOptions struct {
+	// TrustedHTML erlaubt rohe .html-Inhalte aus deinem Content-Repository.
+	// Achtung: Nur aktivieren, wenn du dem Repository vertraust.
+	TrustedHTML bool
+}
 
-	mdRenderer := html.NewRenderer(html.RendererOptions{
-		Flags: html.CommonFlags | html.HrefTargetBlank,
-	})
+type frontMatter struct {
+	Title       string `json:"title"`
+	Date        string `json:"date"`
+	Slug        string `json:"slug"`
+	Cover       string `json:"cover"`
+	Description string `json:"description"`
+	Format      string `json:"format"`
+}
+
+func LoadDir(root string, opts LoaderOptions) ([]Article, error) {
+	var out []Article
+	seen := make(map[string]bool)
 
 	err := filepath.WalkDir(root, func(path string, d fs.DirEntry, walkErr error) error {
 		if walkErr != nil {
@@ -38,37 +44,15 @@ func LoadDir(root string) ([]Article, error) {
 		}
 
 		ext := strings.ToLower(filepath.Ext(path))
-		if ext != ".md" && ext != ".html" {
-			return nil // uninteressante Datei
+		if ext != ".md" && ext != ".markdown" && ext != ".html" && ext != ".htm" {
+			return nil
 		}
 
-		raw, err := os.ReadFile(path)
+		meta, body, format, err := readContentFile(path, opts)
 		if err != nil {
-			return fmt.Errorf("read %s: %w", path, err)
+			return err
 		}
 
-		parts := bytes.SplitN(raw, []byte("\n"), 2)
-		if len(parts) < 2 {
-			return fmt.Errorf("%s: missing front‑matter", path)
-		}
-
-		headerJSON, err := extractJSONHeader(parts[0])
-		if err != nil {
-			return fmt.Errorf("%s: %w", path, err)
-		}
-
-		var meta struct {
-			Title       string `json:"title"`
-			Date        string `json:"date"`
-			Slug        string `json:"slug"`
-			Cover       string `json:"cover"`
-			Description string `json:"description"`
-		}
-		if err := json.Unmarshal(headerJSON, &meta); err != nil {
-			return fmt.Errorf("%s: %w", path, err)
-		}
-
-		// Fallback‑Slug aus Dateinamen
 		if meta.Slug == "" {
 			meta.Slug = strings.TrimSuffix(filepath.Base(path), ext)
 		}
@@ -77,11 +61,6 @@ func LoadDir(root string) ([]Article, error) {
 		}
 		seen[meta.Slug] = true
 
-		body := parts[1]
-		if ext == ".md" {
-			body = md.ToHTML(body, nil, mdRenderer) // frischer Parser pro Call
-		}
-
 		date, err := time.Parse("2006-01-02", meta.Date)
 		if err != nil {
 			return fmt.Errorf("%s: parse date: %w", path, err)
@@ -93,7 +72,9 @@ func LoadDir(root string) ([]Article, error) {
 			Date:        date,
 			Cover:       meta.Cover,
 			Description: meta.Description,
-			Body:        template.HTML(body),
+			Body:        body,
+			Format:      format,
+			SourcePath:  path,
 		})
 		return nil
 	})
@@ -105,34 +86,7 @@ func LoadDir(root string) ([]Article, error) {
 	return out, nil
 }
 
-// ---------- NEU ----------
-type StaticPage struct {
-	Title       string
-	Slug        string
-	Description string
-	Body        template.HTML
-}
-
-// helper.go – einmal zentral verwenden
-func extractJSONHeader(line []byte) ([]byte, error) {
-	s := strings.TrimSpace(string(line))
-	s = strings.TrimPrefix(s, "<!--")
-
-	// alles bis zum ersten "-->" abschneiden
-	if idx := strings.Index(s, "-->"); idx != -1 {
-		s = s[:idx]
-	}
-	s = strings.TrimSpace(s)
-
-	// optional: falls doch kein JSON, besserer Fehler
-	if !strings.HasPrefix(s, "{") {
-		return nil, fmt.Errorf("no JSON front‑matter found")
-	}
-	return []byte(s), nil
-}
-
-// LoadStatic liest alle .md/.html‑Dateien unter dir und liefert sie als Map[slug]StaticPage.
-func LoadStatic(dir string) (map[string]StaticPage, error) {
+func LoadStatic(dir string, opts LoaderOptions) (map[string]StaticPage, error) {
 	pages := make(map[string]StaticPage)
 
 	entries, err := os.ReadDir(dir)
@@ -144,38 +98,17 @@ func LoadStatic(dir string) (map[string]StaticPage, error) {
 		if e.IsDir() {
 			continue
 		}
-		ext := filepath.Ext(e.Name())
-		if ext != ".md" && ext != ".html" {
-			continue // unbekanntes Format
+		ext := strings.ToLower(filepath.Ext(e.Name()))
+		if ext != ".md" && ext != ".markdown" && ext != ".html" && ext != ".htm" {
+			continue
 		}
 
 		path := filepath.Join(dir, e.Name())
-		raw, err := os.ReadFile(path)
+		meta, body, format, err := readContentFile(path, opts)
 		if err != nil {
-			return nil, fmt.Errorf("read %s: %w", path, err)
+			return nil, err
 		}
 
-		// Front‑Matter (erste Zeile) herauslösen
-		parts := bytes.SplitN(raw, []byte("\n"), 2)
-		if len(parts) < 2 {
-			return nil, fmt.Errorf("%s: missing front‑matter", path)
-		}
-
-		headerJSON, err := extractJSONHeader(parts[0])
-		if err != nil {
-			return nil, fmt.Errorf("%s: %w", path, err)
-		}
-
-		var meta struct {
-			Title       string `json:"title"`
-			Slug        string `json:"slug"`
-			Description string `json:"description"`
-		}
-		if err := json.Unmarshal(headerJSON, &meta); err != nil {
-			return nil, fmt.Errorf("%s: %w", path, err)
-		}
-
-		// Fallback: Slug aus Dateinamen ableiten, falls im Header leer
 		if meta.Slug == "" {
 			meta.Slug = strings.TrimSuffix(e.Name(), ext)
 		}
@@ -183,17 +116,88 @@ func LoadStatic(dir string) (map[string]StaticPage, error) {
 			return nil, fmt.Errorf("%s: duplicate slug %q", path, meta.Slug)
 		}
 
-		body := parts[1]
-		if ext == ".md" {
-			body = md.ToHTML(body, nil, nil) // eigener Parser je Aufruf
-		}
-
 		pages[meta.Slug] = StaticPage{
 			Title:       meta.Title,
 			Slug:        meta.Slug,
 			Description: meta.Description,
-			Body:        template.HTML(body),
+			Body:        body,
+			Format:      format,
+			SourcePath:  path,
 		}
 	}
+
 	return pages, nil
 }
+
+func readContentFile(path string, opts LoaderOptions) (frontMatter, template.HTML, Format, error) {
+	raw, err := os.ReadFile(path)
+	if err != nil {
+		return frontMatter{}, "", "", fmt.Errorf("read %s: %w", path, err)
+	}
+
+	parts := bytes.SplitN(raw, []byte("\n"), 2)
+	if len(parts) < 2 {
+		return frontMatter{}, "", "", fmt.Errorf("%s: missing front-matter", path)
+	}
+
+	headerJSON, err := extractJSONHeader(parts[0])
+	if err != nil {
+		return frontMatter{}, "", "", fmt.Errorf("%s: %w", path, err)
+	}
+
+	var meta frontMatter
+	if err := json.Unmarshal(headerJSON, &meta); err != nil {
+		return frontMatter{}, "", "", fmt.Errorf("%s: %w", path, err)
+	}
+
+	ext := strings.ToLower(filepath.Ext(path))
+	format := detectFormat(ext, meta.Format)
+	body := renderBody(parts[1], format, opts)
+
+	return meta, body, format, nil
+}
+
+func detectFormat(ext, explicit string) Format {
+	switch strings.ToLower(strings.TrimSpace(explicit)) {
+	case "html":
+		return FormatHTML
+	case "md", "markdown":
+		return FormatMarkdown
+	}
+
+	switch ext {
+	case ".html", ".htm":
+		return FormatHTML
+	default:
+		return FormatMarkdown
+	}
+}
+
+func renderBody(body []byte, format Format, opts LoaderOptions) template.HTML {
+	if format == FormatMarkdown {
+		renderer := html.NewRenderer(html.RendererOptions{
+			Flags: html.CommonFlags | html.HrefTargetBlank,
+		})
+		return template.HTML(md.ToHTML(body, nil, renderer))
+	}
+
+	if opts.TrustedHTML {
+		return template.HTML(body)
+	}
+
+	// Sicherheits-Fallback: HTML-Dateien werden escaped, solange BLOG_TRUSTED_HTML=false ist.
+	return template.HTML(template.HTMLEscapeString(string(body)))
+}
+
+func extractJSONHeader(line []byte) ([]byte, error) {
+	s := strings.TrimSpace(string(line))
+	s = strings.TrimPrefix(s, "<!--")
+	if idx := strings.Index(s, "-->"); idx != -1 {
+		s = s[:idx]
+	}
+	s = strings.TrimSpace(s)
+	if !strings.HasPrefix(s, "{") {
+		return nil, fmt.Errorf("no JSON front-matter found")
+	}
+	return []byte(s), nil
+}