jakarta-oro-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Takashi Okamoto" <tokam...@rd.nttdata.co.jp>
Subject POSIX Rexpressioin
Date Sun, 01 Apr 2001 06:57:50 GMT
Hi!!

I investigated Perl5.6 [:punct:], [:graph:] and [:print:] more detail.
(ORO goal is compatibility with Perl5.6.)
Now ORO's matching characters are a little different from Perl5.6. 
I compared Perl5.6's matching result and  Java's Unicode Block. 

[:print:]
      - Following unicode blocks are NOT matching Perl5.6's [:print:]
            Character.CONTROL
            Character.FORMAT
            Character.SURROGATE
            Character.PRIVATE_USE
        Others are matching.

      - And some of Character.UNASSIGNED are matching [:print:]. But
        we can ignore them because they are not defined unicode block.
        (these characters are used for only special purpose.)

[:graph:]
        [:graph:] is excepted only U+0020 from [:print:] characters.

[:punct:]
      - Following unicode blocks are matching Perl5.6's [:punct:]
            Character.DASH_PUNCTUATION
            Character.START_PUNCTUATION
            Character.END_PUNCTUATION
            Character.CONNECTOR_PUNCTUATION
        
      - Some of Character.OTHER_PUNCTUATION characters are matching
 ,but followig characters are NOT matching it.
                U+0374, U+0375, U+0E2F, U+0EAF, U+3006

I attached patch including above result. However it makes code a
little tricky and user may not require such a detail compatibility.

Is it nitpick?;)

Regards.
---------------------
Takashi Okamoto


--- Perl5Matcher.java.orig Sat Mar 31 23:01:12 2001
+++ Perl5Matcher.java Sun Apr  1 11:33:55 2001
@@ -682,25 +682,58 @@
    if(Character.isUpperCase(code)) return isANYOF;
    break;
  case OpCode._PRINT:
-   if(Character.isSpaceChar(code)) return isANYOF;
-          // Fall through to check if the character is alphanumeric,
-   // or a punctuation mark.  Printable characters are either
-   // alphanumeric, punctuation marks, or spaces.
+   switch( Character.getType(code) ) {
+     // Following unicode blocks do NOT match [:print:].
+     case Character.UNASSIGNED:
+     case Character.CONTROL:
+     case Character.FORMAT:
+     case Character.SURROGATE:
+     case Character.PRIVATE_USE:
+       break;
+     default:
+       // Others match.
+       return isANYOF;
+   }
  case OpCode._GRAPH:
-   if(Character.isLetterOrDigit(code)) return isANYOF;
-          // Fall through to check if the character is a punctuation mark.
-          // Graph characters are either alphanumeric or punctuation.
+   switch ( Character.getType(code) ) {
+     // Following unicode blocks do NOT match [:graph:].
+     case Character.UNASSIGNED:
+     case Character.CONTROL:
+     case Character.FORMAT:
+     case Character.SURROGATE:
+     case Character.PRIVATE_USE:
+       break;
+     default:
+       // Others match except U+0020.
+       if ( code != 0x0020 )
+  return isANYOF;
+       break;
+   }
  case OpCode._PUNCT:
    switch ( Character.getType(code) ) {
+     // Following unicode blocks match [:punct:].
      case Character.DASH_PUNCTUATION:
      case Character.START_PUNCTUATION:
      case Character.END_PUNCTUATION:
      case Character.CONNECTOR_PUNCTUATION:
-     case Character.OTHER_PUNCTUATION:
        return isANYOF;
-     default:
-       break;
+     case Character.OTHER_PUNCTUATION:
+       switch ( code ) {
+  // following OTHER_PUNCTUATION characters don't match
+  // Perl5.6's [:punct:]
+         case 0x0374:
+         case 0x0375:
+         case 0x0e2f:
+         case 0x0eaf:
+         case 0x3006:
+    break;
+       default:
+  // other OTHER_PUNCTUATION characters match.
+  return isANYOF;
      }
+   default:
+     break;
+   }
    break;
  case OpCode._XDIGIT:
    if( (code >= '0' && code <= '9') ||



Mime
View raw message