sycommon-python-lib 0.2.10a13__tar.gz → 0.2.10a14__tar.gz

{sycommon_python_lib-0.2.10a13 → sycommon_python_lib-0.2.10a14}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sycommon-python-lib
-Version: 0.2.10a13
+Version: 0.2.10a14
 Summary: Add your description here
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown

{sycommon_python_lib-0.2.10a13 → sycommon_python_lib-0.2.10a14}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "sycommon-python-lib"
-version = "0.2.10a13"
+version = "0.2.10a14"
 description = "Add your description here"
 readme = "README.md"
 requires-python = ">=3.11"

{sycommon_python_lib-0.2.10a13 → sycommon_python_lib-0.2.10a14}/src/sycommon/logging/log_masker.py

@@ -383,19 +383,27 @@ class LogMasker(metaclass=SingletonMeta):
             return [self._decode_escaped_tree(v) for v in value]
         return value
 
-    _ESCAPED_UNICODE_RE = re.compile(r"\\u([0-9a-fA-F]{4})")
+    _ESCAPED_UNICODE_RE = re.compile(
+        r"\\u(d[89ab][0-9a-fA-F]{2})\\u(d[c-f][0-9a-fA-F]{2})"  # 代理对(emoji等)
+        r"|\\u([0-9a-fA-F]{4})"                                   # BMP 单对
+    )
 
     def _decode_escaped_unicode(self, value: Any) -> Any:
-        """字符串值内的 \\uXXXX 转义解码为可读中文(展示层修复)。
+        r"""字符串值内的 \uXXXX 转义解码为可读中文(展示层修复)。
 
         SQLAlchemy JSON 列绑定参数时 json.dumps 默认 ensure_ascii=True —
-        中文以 \\u660e\\u786e 形态入库/入参数;SQL 日志如实展示导致平台
+        中文以 明确 形态入库/入参数;SQL 日志如实展示导致平台
         不可读(2026-10-16 wxcoach ext_json 实测)。只解码【字符串叶子】。
 
         按对替换而非整串 unicode_escape: 真实参数里混有非法转义对
-        (如截断的 \\u4fd\\ 后跟引号)时,整串 decode 抛异常 → 原样返回,
+        (如截断的 \u4fd\ 后跟引号)时,整串 decode 抛异常 → 原样返回,
         合法对也全留下(17:59 ext_json 仍全编码的根因)。逐对处理:
-        合法 4 位十六进制对解码,非法序列(含字面 \\\\u)原样保留。
+        - 合法 BMP 对解码;但 \ud800-\udfff 区间(代理区)保留原文 —
+          解出 lone surrogate 会让 kafka producer 的 utf-8 encode 抛
+          UnicodeEncodeError,write() 吞异常 → 整条日志静默丢失。
+        - 高/低代理【成对】出现时(json.dumps 对 >BMP 字符的标准产物,
+          如 emoji 😀)合并为单个码点再解码。
+        - 非法序列(3 位 hex、字面 \\u 等)原样保留。
         """
         if not isinstance(value, str) or "\\u" not in value:
             return value
@@ -403,10 +411,17 @@ class LogMasker(metaclass=SingletonMeta):
             return value  # 已含可读中文,无需也不应再动
 
         def _sub(m: "re.Match[str]") -> str:
-            try:
-                return chr(int(m.group(1), 16))
-            except Exception:
-                return m.group(0)
+            hi, lo, single = m.group(1), m.group(2), m.group(3)
+            if hi is not None and lo is not None:
+                cp = 0x10000 + ((int(hi, 16) - 0xD800) << 10) \
+                    + (int(lo, 16) - 0xDC00)
+                return chr(cp)
+            if single is not None:
+                code = int(single, 16)
+                if 0xD800 <= code <= 0xDFFF:
+                    return m.group(0)  # lone surrogate: 保留原文防编码崩溃
+                return chr(code)
+            return m.group(0)
 
         return self._ESCAPED_UNICODE_RE.sub(_sub, value)
 

{sycommon_python_lib-0.2.10a13 → sycommon_python_lib-0.2.10a14}/src/sycommon_python_lib.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sycommon-python-lib
-Version: 0.2.10a13
+Version: 0.2.10a14
 Summary: Add your description here
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown

{sycommon_python_lib-0.2.10a13 → sycommon_python_lib-0.2.10a14}/tests/test_unicode_decode_regress.py

@@ -1,5 +1,6 @@
 # -*- coding: utf-8 -*-
 """复现 17:59 ext_json 残留编码的原始用例 — 逐对解码回归验证。"""
+import json
 import sys
 from pathlib import Path
 
@@ -42,3 +43,22 @@ assert "明确" in dec["ext"] and "明确" in dec["rows"][1]["inner"]
 print("树形递归 OK")
 
 print("\nALL PASS")
+
+# 7) 代理对合并(emoji): json.dumps 对 >BMP 字符的标准产物
+emoji_json = json.dumps({"m": "😀👍"}, ensure_ascii=True)  # 含 😀 对
+inner = json.loads(emoji_json)  # 还原回字符串形态用于模拟参数叶子
+esc = emoji_json.replace('{"m": "', '').replace('"}', '')
+out7 = m._decode_escaped_unicode(esc)
+assert "😀" in out7 and "👍" in out7, out7
+out7.encode("utf-8")  # kafka producer utf-8 编码不崩
+print("7 代理对合并 OK:", out7)
+
+# 8) lone surrogate 保留原文 — chr(0xd800) 会让 utf-8 encode 抛
+#    UnicodeEncodeError,write() except 吞掉 → 整条日志静默丢失
+lone = "bad \\ud800 end"
+out8 = m._decode_escaped_unicode(lone)
+assert "\\ud800" in out8, repr(out8)
+out8.encode("utf-8")  # 不抛
+print("8 lone surrogate 防崩溃 OK")
+
+print("SURROGATE PASS")