While using ACE_Wide_To_Ascii string conversion class (and correspondingly ::WideCharToMultiByte) there must be expected that the length of the ouput multi-byte string is not equal to the length of the input unicode string. That's why we the variable iter used to iterate over the original unicode environment varaiables must be increased by the source string length (and not by the length of converted string).

Author: Viktor Chernenko

ACE/ace/Process.cpp

@@ -932,7 +932,7 @@ ACE_Process_Options::inherit_environment ()
           size_t const idx = temp_narrow_env.size ();
           temp_narrow_env.resize (idx + len + 1, 0);
           ACE_OS::strncpy (&temp_narrow_env[idx], wta.char_rep (), len);
-          iter += len;
+          iter += ACE_OS::strlen(iter);
         }
       temp_narrow_env.push_back (0);
       existing_environment = &temp_narrow_env[0];