Re: java/text/SimpleDateFormat.java (compileFormat)

[Ito Kazumitsu]

Hi,

>>>>> ":" == Michael Koch <address@hidden> writes:

>> I am afraid "Character.isLowerCase(char) || Character.isUpperCase(char)"
>> also allows too many characters,  including Greek or Slavic alphabet
>> or even Japanese Zenkaku alphabet.

:> Grrr! I thought its the same. All I found int the docs indicated this.
:> Can you please commit/provide a patch that fixes this ?

The API document of java.lang.Character says:

    The following are examples of lowercase characters:
   
         a b c d e f g h i j k l m n o p q r s t u v w x y z
         '\u00DF' '\u00E0' '\u00E1' '\u00E2' '\u00E3' '\u00E4' '\u00E5' '\u00E6'
         '\u00E7' '\u00E8' '\u00E9' '\u00EA' '\u00EB' '\u00EC' '\u00ED' '\u00EE'
         '\u00EF' '\u00F0' '\u00F1' '\u00F2' '\u00F3' '\u00F4' '\u00F5' '\u00F6'
         '\u00F8' '\u00F9' '\u00FA' '\u00FB' '\u00FC' '\u00FD' '\u00FE' '\u00FF'
         
    Many other Unicode characters are lowercase too.

I do not know of a smart way of describing characters between 'A' and 'Z'
and 'a' and 'z',  so my patch is:

--- java/text/SimpleDateFormat.java.orig        Mon Nov 24 23:59:09 2003
+++ java/text/SimpleDateFormat.java     Fri Nov 28 07:05:10 2003
@@ -117,8 +117,8 @@
       field = formatData.getLocalPatternChars().indexOf(thisChar);
       if (field == -1) {
        current = null;
-       if (Character.isLowerCase (thisChar)
-           || Character.isUpperCase (thisChar)) {
+       if ((thisChar >= 'A' && thisChar <= 'Z')
+           || (thisChar >= 'a' && thisChar <= 'z')) {
          // Not a valid letter
          tokens.add(new FieldSizePair(-1,0));
        } else if (thisChar == '\'') {

ChangeLog entry:

2003-11-28  Ito Kazumitsu  <address@hidden>

        * java/text/SimpleDateFormat.java (compileFormat): 
        isLowerCase and isUpperCase allow too many characters.
        Just use >= 'A' && <= 'Z' || >= 'a' && <= 'z'.