Tomcat's WebApp Chinese project starts using anti-compiled methods to join the original string
String newstr = new string (Str.getbytes ("ISO-8859-1"), "GBK");
To solve, Java will have a double-byte of the Chinese characters as a separate two byte processing. But it is apparent that it is not a good way.
Later, Eclipse can use UTF-8 encoding mode to handle multilingual interface issues. (Of course, its search results are sometimes garbled). So the Chinese string file in the resource uses UTF-8 to convert (JDK Native2ASCII.exe), becoming / uhhhh / uhhhh;
Who knows or not, because the system also uses XML Parser to achieve resources, so the mode of / uhhhh becomes the XML escape character indicated:
A character reference references to a character in the ISO / IEC 10646 character set. For example, you cannot enter the character directly entered directly. Character reference [66] Charref :: = '& #' [0-9] ';' | '& # x' [0-9a-fa-f] ';' [WFC: legal character] So each Chinese characters are changed to & # xhhhh;
I finally all OK, I thought it didn't know how to restart. Who knows that the original Chinese becomes "??" After the machine started, JVM did not identify the correct locale, but I have specified it. Test, File .Encoding is a set-DFILE.Encoding = GBK in JavaService, and if the registry is changed directly, in addition to the addition of a parameter, all JVM system parameters All can be set by -d.
It may be because the system is in English for Windows2000, there is this problem.
