fix(编辑能力): 修复多版本字体裁剪与资源清理

ofdgo_editor_font_binding.go

@@ -963,7 +963,7 @@ func pruneUnreferencedFiles(ctx context.Context, reader *Reader, candidates ...s
 			if root == "" {
 				root = node.Name.Local
 			}
-			if classifyOFDNamespace(node.Name.Space) == ofdXMLUnknown || slices.Contains([]string{"CustomTags", "Extensions", "ExtendData", "Data"}, node.Name.Local) {
+			if classifyOFDNamespace(node.Name.Space) == ofdXMLUnknown || root == "CustomTags" || slices.Contains([]string{"Extensions", "ExtendData", "Data"}, node.Name.Local) {
 				uncertain = true
 				return fmt.Errorf("unknown file dependencies in %s", name)
 			}
@@ -990,7 +990,19 @@ func pruneUnreferencedFiles(ctx context.Context, reader *Reader, candidates ...s
 				versions[name] = append(versions[name], versionEditLink{node: node, target: target})
 			}
 		}
-		if err := reader.scanDocumentFileReferences(ctx, name, true, add, record, inspect); err != nil {
+		err := reader.scanDocumentFileReferences(ctx, name, true, add, record, inspect)
+		if uncertain && (root == "CustomTags" || root == "RichText") {
+			data, readErr := reader.readFileViewContext(ctx, name)
+			if readErr != nil {
+				return readErr
+			}
+			refs := &editorResourceRefs{ids: make(map[string]bool), files: make(map[string]bool), fonts: make(map[string]*editorFontUsage)}
+			if _, safe := refs.scanContext(ctx, bytes.NewReader(data), name); safe {
+				uncertain = false
+				err = reader.scanDocumentFileReferences(ctx, name, true, add, record, nil)
+			}
+		}
+		if err != nil {
 			var syntax *xml.SyntaxError
 			if uncertain || errors.As(err, &syntax) {
 				return ctx.Err()

ofdgo_editor_reference.go

@@ -53,6 +53,7 @@ type editorReferenceScan struct {
 	stack      []string
 	fonts      []*editorFontUsage
 	text       strings.Builder
+	richText   strings.Builder
 	resource   bool
 	capture    bool
 	safe       bool
@@ -140,14 +141,39 @@ func (s *editorReferenceScan) accept(token xml.Token) bool {
 	}
 	switch token := token.(type) {
 	case xml.StartElement:
-		if classifyOFDNamespace(token.Name.Space) == ofdXMLUnknown {
+		rich := token.Name.Space == richTextNamespace
+		known := classifyOFDNamespace(token.Name.Space) != ofdXMLUnknown
+		if rich {
+			known = len(s.stack) == 0 && token.Name.Local == "RichText" || len(s.stack) == 1 && s.stack[0] == "RichText" && token.Name.Local == "ObjectRef"
+		} else if len(s.stack) > 0 && s.stack[0] == "RichText" {
+			known = false
+		}
+		if len(s.stack) > 0 && s.stack[0] == "CustomTags" {
+			known = known && (len(s.stack) == 1 && token.Name.Local == "CustomTag" || len(s.stack) == 2 && s.stack[1] == "CustomTag" && (token.Name.Local == "FileLoc" || token.Name.Local == "SchemaLoc"))
+		}
+		if !known {
 			s.safe = false
 			return false
 		}
-		if token.Name.Local == "CustomTags" || token.Name.Local == "Extensions" || token.Name.Local == "ExtendData" || token.Name.Local == "Data" {
+		if token.Name.Local == "Extensions" || token.Name.Local == "ExtendData" || token.Name.Local == "Data" {
 			s.safe = false
 			return false
 		}
+		if token.Name.Local == "CustomTag" {
+			var kind, namespace string
+			for _, attr := range token.Attr {
+				if attr.Name.Space == "" && attr.Name.Local == "TypeID" {
+					kind = attr.Value
+				}
+				if attr.Name.Space == "" && attr.Name.Local == "NameSpace" {
+					namespace = attr.Value
+				}
+			}
+			if kind != richTextTagType || namespace != richTextNamespace {
+				s.safe = false
+				return false
+			}
+		}
 		if len(s.stack) == 0 {
 			if s.seen {
 				s.safe = false
@@ -184,7 +210,8 @@ func (s *editorReferenceScan) accept(token xml.Token) bool {
 			if attr.Name.Space == "xmlns" || attr.Name.Local == "xmlns" || attr.Name.Local == "xmlns:ofd" || attr.Name.Space == XMLNamespace {
 				continue
 			}
-			if attr.Name.Space != "" {
+			custom := len(s.stack) == 0 && token.Name.Local == "CustomTags" || len(s.stack) > 0 && s.stack[0] == "CustomTags"
+			if attr.Name.Space != "" || rich && (token.Name.Local != "ObjectRef" || attr.Name.Local != "PageRef") || custom && (token.Name.Local != "CustomTag" || attr.Name.Local != "TypeID" && attr.Name.Local != "NameSpace") {
 				s.safe = false
 				return false
 			}
@@ -225,6 +252,9 @@ func (s *editorReferenceScan) accept(token xml.Token) bool {
 			s.capture = true
 		}
 	case xml.CharData:
+		if len(s.stack) == 1 && s.stack[0] == "RichText" {
+			s.richText.Write(token)
+		}
 		if len(s.stack) == 0 && len(bytes.Trim(token, " \t\r\n")) != 0 {
 			s.safe = false
 			return false
@@ -233,6 +263,14 @@ func (s *editorReferenceScan) accept(token xml.Token) bool {
 			s.text.Write(token)
 		}
 	case xml.EndElement:
+		if len(s.stack) == 3 && s.stack[0] == "CustomTags" && token.Name.Local == "SchemaLoc" && strings.TrimSpace(s.text.String()) != "" {
+			s.safe = false
+			return false
+		}
+		if len(s.stack) == 1 && s.stack[0] == "RichText" && !richTextMetadataFontSafe(s.richText.String()) {
+			s.safe = false
+			return false
+		}
 		if len(s.stack) == 0 || s.stack[len(s.stack)-1] != token.Name.Local {
 			s.safe = false
 			return false

ofdgo_editor_rich_text.go

@@ -909,22 +909,7 @@ func richTextTagsFontSafe(ctx context.Context, reader *Reader, index int) (bool,
 		if root.Name != "RichText" || root.Namespace != richTextNamespace || len(root.Attributes) != 0 {
 			return false, nil
 		}
-		var meta richTextMetadata
+		if !richTextMetadataFontSafe(root.Text) {
-		decoder := json.NewDecoder(bytes.NewBufferString(root.Text))
-		decoder.DisallowUnknownFields()
-		if err := decoder.Decode(&meta); err != nil {
-			return false, nil
-		}
-		if decoder.Decode(new(any)) != io.EOF || meta.Version != 1 {
-			return false, nil
-		}
-		for i := range meta.Value.Runs {
-			if meta.Value.Runs[i].Format.Font != "" {
-				return false, nil
-			}
-			meta.Value.Runs[i].Format.Font = "1"
-		}
-		if err := meta.Value.Validate(); err != nil {
 			return false, nil
 		}
 		for _, child := range root.Children {
@@ -935,3 +920,28 @@ func richTextTagsFontSafe(ctx context.Context, reader *Reader, index int) (bool,
 	}
 	return true, ctx.Err()
 }
+
+// richTextMetadataFontSafe 核对本库富文本元数据不包含字体资源编号或未知字段
+// 入参: value 元数据JSON
+// 返回: bool 是否可由标准图元完整统计字体引用
+func richTextMetadataFontSafe(value string) bool {
+	var meta richTextMetadata
+	decoder := json.NewDecoder(bytes.NewBufferString(value))
+	decoder.DisallowUnknownFields()
+	if err := decoder.Decode(&meta); err != nil {
+		return false
+	}
+	if decoder.Decode(new(any)) != io.EOF || meta.Version != 1 {
+		return false
+	}
+	for i := range meta.Value.Runs {
+		if meta.Value.Runs[i].Format.Font != "" {
+			return false
+		}
+		meta.Value.Runs[i].Format.Font = "1"
+	}
+	if err := meta.Value.Validate(); err != nil {
+		return false
+	}
+	return true
+}