Some simple information syncing

includes/ast/parser.py

@@ -7,7 +7,7 @@ from .base import (
     AST, ASTCtx, ASTKind, OpKind, ASTFlag,
     _set_pos, _binop_from_op, _augop_from_op, _cmpop_from_op, _copy_str, _inherit_pos,
     _init_ast,
-    CONST_INT, CONST_FLOAT, CONST_STR, CONST_BOOL, CONST_NONE,
+    CONST_INT, CONST_FLOAT, CONST_STR, CONST_BOOL, CONST_NONE, CONST_CHAR,
     FLAG_IS_ASYNC,
 )
 from .exprs import (
@@ -169,7 +169,30 @@ def _new_const_float(pool: memhub.MemManager | t.CPtr, val: t.CDouble,
 
 
 def _new_const_str(pool: memhub.MemManager | t.CPtr, val: str,
-                   lineno: t.CInt, col: t.CInt) -> AST | t.CPtr:
+                   lineno: t.CInt, col: t.CInt,
+                   is_dquote: t.CInt = 1,
+                   is_triple: t.CInt = 0,
+                   is_char: t.CInt = 0) -> AST | t.CPtr:
+    """创建字符串/字符常量节点
+
+    is_dquote: 0=单引号, 非0=双引号（默认双引号）
+    is_triple: 非0=三引号字符串
+    is_char: 非0=单字符候选（lexer 已确认转义后正好 1 字节, 含 '\\0'）
+    语义区分:
+      - 'x'  (单引号单字符, 非三引号) → CONST_CHAR (i8 值)
+      - "x"  (双引号, 含\\0终止符)      → CONST_STR (i8* 指针, i8 x 2)
+      - 'ab' (单引号多字符)            → CONST_STR (i8* 指针)
+      - '''x''' / \"\"\"x\"\"\" (三引号) → CONST_STR (i8* 指针)
+      - '' / \"\" (空字符串)            → CONST_STR (i8* 指针)
+    """
+    # 单引号、非三引号、lexer 标记的单字符候选 → CONST_CHAR
+    # 用 is_char 标记而非 strlen(val)，因为 '\0' 转义后是 NUL 字节，strlen 会返回 0
+    if is_dquote == 0 and is_triple == 0 and is_char != 0:
+        # 单字符 → CONST_CHAR，int_val 存储字符 ASCII 值
+        # val[0] 即使是 NUL 也是合法值 0
+        ch_val: t.CInt64T = val[0]
+        return Constant(pool, CONST_CHAR, ch_val, 0.0, val, lineno, col)
+    # 其他情况 → CONST_STR
     return Constant(pool, CONST_STR, 0, 0.0, val, lineno, col)
 
 
@@ -751,10 +774,13 @@ def _parse_atom(ps: Parser | t.CPtr) -> AST | t.CPtr:
     if ttype == TokenType.String:
         sval: str = _cur_str(ps)
         is_fstr: t.CInt = ps.cur.kw_subtype & 4
+        is_triple: t.CInt = ps.cur.kw_subtype & 8
+        is_dquote: t.CInt = ps.cur.kw_subtype & 16
+        is_char: t.CInt = ps.cur.kw_subtype & 32
         _advance(ps)
         if is_fstr:
             return _parse_fstring(ps, sval, lineno, col)
-        return _new_const_str(ps.pool, sval, lineno, col)
+        return _new_const_str(ps.pool, sval, lineno, col, is_dquote, is_triple, is_char)
 
     if _match_op(ps, TokOp.Ellipsis):
         _advance(ps)
@@ -1551,12 +1577,14 @@ def _parse_from_import(ps: Parser | t.CPtr) -> AST | t.CPtr:
         level += 3
         _advance(ps)
     module_name: str = None
-    if _cur_type(ps) == TokenType.Name:
+    # 关键字（如 import）的 token type 也是 Name，必须通过 kw_subtype 排除，
+    # 否则 `from . import` 中的 `import` 会被误认为 module_name
+    if _cur_type(ps) == TokenType.Name and _cur_kw(ps) == Keyword.NotKeyword:
         module_name = _cur_str(ps)
         _advance(ps)
         while _match_op(ps, TokOp.Dot):
             _advance(ps)
-            if _cur_type(ps) == TokenType.Name:
+            if _cur_type(ps) == TokenType.Name and _cur_kw(ps) == Keyword.NotKeyword:
                 next_name: str = _cur_str(ps)
                 mlen: t.CInt = string.strlen(module_name)
                 nlen: t.CInt = string.strlen(next_name)
@@ -1571,6 +1599,7 @@ def _parse_from_import(ps: Parser | t.CPtr) -> AST | t.CPtr:
         ps.error_count += 1
         return None
     _advance(ps)
+    _skip_nl(ps)
     # import *
     if _match_op(ps, TokOp.Star):
         _advance(ps)
@@ -2058,8 +2087,9 @@ def _parse_closed_pattern(ps: Parser | t.CPtr) -> AST | t.CPtr:
     # 字符串字面量 pattern
     if ttype == TokenType.String:
         sval = _cur_str(ps)
+        is_dquote_m: t.CInt = ps.cur.kw_subtype & 16
         _advance(ps)
-        val = _new_const_str(ps.pool, sval, lineno, col)
+        val = _new_const_str(ps.pool, sval, lineno, col, is_dquote_m)
         node = MatchValue(ps.pool, val)
         _set_pos(node, lineno, col, 0, 0)
         return node