cobol_testgen/grammar.lark

@@ -1,7 +1,7 @@
 start: data_div_content
 data_div_content: (file_section | working_storage | linkage)*
 file_section: "FILE" "SECTION" DOT (fd | sd)+
-fd: "FD" NAME FD_SUFFIX data_item+
+fd: "FD" NAME FD_SUFFIX data_item*
 sd: "SD" NAME FD_SUFFIX data_item*
 FD_SUFFIX: /(?:"[^"]*"|'[^']*'|[^.])*\./
 working_storage: "WORKING-STORAGE" "SECTION" DOT data_item*
@@ -22,13 +22,13 @@ value_literal: INT | SIGNED_NUMBER | STRING | SQSTRING
              | "LOW-VALUE" | "LOW-VALUES"
 SQSTRING: /'[^']*'/
 redefines_clause: "REDEFINES" NAME
-occurs_clause: "OCCURS" INT "TIMES"? ("DEPENDING" "ON" NAME)? key_clause? indexed_clause?
+occurs_clause: "OCCURS" INT ("TO" INT)? "TIMES"? ("DEPENDING" "ON" NAME)? key_clause? indexed_clause?
 key_clause: ("ASCENDING" | "DESCENDING") "KEY" "IS"? NAME (","? NAME)*
 indexed_clause: "INDEXED" "BY" NAME (","? NAME)*
 usage_clause: USAGE_VAL
 USAGE_VAL: "COMP" | "COMP-3" | "COMP-5" | "BINARY" | "PACKED-DECIMAL" | "DISPLAY"
 LEVEL: /0[1-9]|[1-4][0-9]|49|77|88/
-NAME: /[A-Z][A-Z0-9-]*/
+NAME: /[A-Z][A-Z0-9-]*/i
 PICTURE_STRING: /[0-9A-Z()+,\-*\/V]+/i
 INT: /[0-9]+/
 DOT: /\./

cobol_testgen/read.py

@@ -27,6 +27,10 @@ def _is_fixed_format(source: str) -> bool:
 
 
 def preprocess(source: str) -> str:
+    # COPY 预处理：展开或移除 COPY 语句
+    # Lark 语法不支持 COPY（这是预处理指令），必须在解析前处理
+    source = resolve_copybooks(source, '.')
+
     fixed = _is_fixed_format(source)
     lines = []
     for raw_line in source.splitlines():
@@ -110,10 +114,12 @@ def resolve_copybooks(source: str, source_dir: str) -> str:
                             re.escape(old.strip()), new.strip(),
                             cb, flags=re.IGNORECASE
                         )
-                result.append(f'      * COPY {name}')
+                # 展开 COPYBOOK 内容，不添加注释行（避免 Lark 在 FD 块内看到注释）
                 result.append(cb)
             else:
-                result.append(line)
+                # COPY 未找到时完全跳过（预处理指令，Lark 不应处理）
+                # 该行可能在 FD/SD 块内，保留会破坏 Lark 解析
+                pass
         else:
             result.append(line)
     return '\n'.join(result)

tests/cobol_testgen/test_read.py

@@ -92,11 +92,12 @@ def test_resolve_copybooks_found():
 
 
 def test_resolve_copybooks_not_found():
-    """COPY 文件不存在时返回含 NOT FOUND 或 NOTEXIST 的文本"""
+    """COPY 不可解析时移除该行（预处理器指令，Lark 不应处理）"""
     with tempfile.TemporaryDirectory() as tmp:
         src = "       COPY NOTEXIST.\n"
         result = resolve_copybooks(src, tmp)
-        assert "NOT FOUND" in result or "NOTEXIST" in result.upper()
+        # COPY 被移除（无残留）
+        assert "NOTEXIST" not in result.upper()
 
 
 def test_resolve_copybooks_no_copy():