libxml2-python3-unicode-errors.patch

--- libxml2-2.15.1/python/libxml.c~	2025-10-16 17:49:18.000000000 +0200
+++ libxml2-2.15.1/python/libxml.c	2025-10-18 12:06:23.895443826 +0200
@@ -1488,6 +1488,7 @@ libxml_xmlErrorFuncHandler(ATTRIBUTE_UNU
     PyObject *message;
     PyObject *result;
     char str[1024];
+    unsigned char *ptr = (unsigned char *)str;
 
     if (libxml_xmlPythonErrorFuncHandler == NULL) {
         va_start(ap, msg);
@@ -1498,12 +1499,20 @@ libxml_xmlErrorFuncHandler(ATTRIBUTE_UNU
         vsnprintf(str, sizeof(str), msg, ap);
         va_end(ap);
 
+#if PY_MAJOR_VERSION >= 3
+        /* Ensure the error string doesn't start at UTF8 continuation. */
+        while (*ptr && (*ptr & 0xc0) == 0x80)
+            ptr++;
+#endif
+
         list = PyTuple_New(2);
         PyTuple_SetItem(list, 0, libxml_xmlPythonErrorFuncCtxt);
         Py_XINCREF(libxml_xmlPythonErrorFuncCtxt);
-        message = libxml_charPtrConstWrap(str);
+        message = libxml_charPtrConstWrap(ptr);
         PyTuple_SetItem(list, 1, message);
         result = PyObject_CallObject(libxml_xmlPythonErrorFuncHandler, list);
+        /* Forget any errors caused in the error handler. */
+        PyErr_Clear();
         Py_XDECREF(list);
         Py_XDECREF(result);
     }