libjava/scripts/unicode-blocks.pl

   1 #!/usr/bin/perl -w
   2 # unicode-blocks.pl -- Script to generate java.lang.Character.UnicodeBlock
   3 # Copyright (C) 2002 Free Software Foundation, Inc.
   4 #
   5 # This file is part of GNU Classpath.
   6 #
   7 # GNU Classpath is free software; you can redistribute it and/or modify
   8 # it under the terms of the GNU General Public License as published by
   9 # the Free Software Foundation; either version 2, or (at your option)
  10 # any later version.
  11 #
  12 # GNU Classpath is distributed in the hope that it will be useful, but
  13 # WITHOUT ANY WARRANTY; without even the implied warranty of
  14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  15 # General Public License for more details.
  16 #
  17 # You should have received a copy of the GNU General Public License
  18 # along with GNU Classpath; see the file COPYING.  If not, write to the
  19 # Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
  20 # 02110-1301 USA.
  21 #
  22 # Linking this library statically or dynamically with other modules is
  23 # making a combined work based on this library.  Thus, the terms and
  24 # conditions of the GNU General Public License cover the whole
  25 # combination.
  26 #
  27 # As a special exception, the copyright holders of this library give you
  28 # permission to link this library with independent modules to produce an
  29 # executable, regardless of the license terms of these independent
  30 # modules, and to copy and distribute the resulting executable under
  31 # terms of your choice, provided that you also meet, for each linked
  32 # independent module, the terms and conditions of the license of that
  33 # module.  An independent module is a module which is not derived from
  34 # or based on this library.  If you modify this library, you may extend
  35 # this exception to your version of the library, but you are not
  36 # obligated to do so.  If you do not wish to do so, delete this
  37 # exception statement from your version.
  38
  39
  40 # Code for reading Blocks.txt and generating (to standard out) the code for
  41 # java.lang.Character.UnicodeBlock, for pasting into java/lang/Character.java.
  42 # You should probably check that the results are accurate to the
  43 # specification, but I made sure it works OOB for Unicode 3.0.0 and JDK 1.4.
  44 # As the grammar for the Blocks.txt file is changing in Unicode 3.2.0, you
  45 # will have to tweak this some for future use.  For now, the relevant
  46 # Unicode definition files are found in libjava/gnu/gcj/convert/.
  47 #
  48 # author Eric Blake <ebb9@email.byu.edu>
  49 #
  50 # usage: unicode-blocks.pl <blocks.txt>
  51 #    where <blocks.txt> is obtained from www.unicode.org (named Blocks-3.txt
  52 #    for Unicode version 3.0.0).
  53
  54
  55 die "Usage: $0 <blocks.txt>" unless @ARGV == 1;
  56 open (BLOCKS, $ARGV[0]) || die "Can't open Unicode block file: $!\n";
  57
  58 # A hash of added fields and the JDK they were added in, to automatically
  59 # print @since tags.  Maintaining this is optional (and tedious), but nice.
  60 my %additions = ("SYRIAC" => "1.4",
  61                  "THAANA" => "1.4",
  62                  "SINHALA" => "1.4",
  63                  "MYANMAR" => "1.4",
  64                  "ETHIOPIC" => "1.4",
  65                  "CHEROKEE" => "1.4",
  66                  "UNIFIED_CANADIAN_ABORIGINAL_SYLLABICS" => "1.4",
  67                  "OGHAM" => "1.4",
  68                  "RUNIC" => "1.4",
  69                  "KHMER" => "1.4",
  70                  "MONGOLIAN" => "1.4",
  71                  "BRAILLE_PATTERNS" => "1.4",
  72                  "CJK_RADICALS_SUPPLEMENT" => "1.4",
  73                  "KANGXI_RADICALS" => "1.4",
  74                  "IDEOGRAPHIC_DESCRIPTION_CHARACTERS" => "1.4",
  75                  "BOPOMOFO_EXTENDED" => "1.4",
  76                  "CJK_UNIFIED_IDEOGRAPHS_EXTENSION_A" => "1.4",
  77                  "YI_SYLLABLES" => "1.4",
  78                  "YI_RADICALS" => "1.4",
  79                  );
  80
  81 print <<'EOF';
  82   /**
  83    * A family of character subsets in the Unicode specification. A character
  84    * is in at most one of these blocks.
  85    *
  86    * This inner class was generated automatically from
  87    * <code>$ARGV[0]</code>, by some perl scripts.
  88    * This Unicode definition file can be found on the
  89    * <a href="http://www.unicode.org">http://www.unicode.org</a> website.
  90    * JDK 1.4 uses Unicode version 3.0.0.
  91    *
  92    * @author scripts/unicode-blocks.pl (written by Eric Blake)
  93    * @since 1.2
  94    */
  95   public static final class UnicodeBlock extends Subset
  96   {
  97     /** The start of the subset. */
  98     private final char start;
  99
 100     /** The end of the subset. */
 101     private final char end;
 102
 103     /**
 104      * Constructor for strictly defined blocks.
 105      *
 106      * @param start the start character of the range
 107      * @param end the end character of the range
 108      * @param name the block name
 109      */
 110     private UnicodeBlock(char start, char end, String name)
 111     {
 112       super(name);
 113       this.start = start;
 114       this.end = end;
 115     }
 116
 117     /**
 118      * Returns the Unicode character block which a character belongs to.
 119      *
 120      * @param ch the character to look up
 121      * @return the set it belongs to, or null if it is not in one
 122      */
 123     public static UnicodeBlock of(char ch)
 124     {
 125       // Special case, since SPECIALS contains two ranges.
 126       if (ch == '\uFEFF')
 127         return SPECIALS;
 128       // Simple binary search for the correct block.
 129       int low = 0;
 130       int hi = sets.length - 1;
 131       while (low <= hi)
 132         {
 133           int mid = (low + hi) >> 1;
 134           UnicodeBlock b = sets[mid];
 135           if (ch < b.start)
 136             hi = mid - 1;
 137           else if (ch > b.end)
 138             low = mid + 1;
 139           else
 140             return b;
 141         }
 142       return null;
 143     }
 144 EOF
 145
 146 my $seenSpecials = 0;
 147 my $seenSurrogates = 0;
 148 my $surrogateStart = 0;
 149 my @names = ();
 150 while (<BLOCKS>) {
 151     next if /^\#/;
 152     my ($start, $end, $block) = split(/; /);
 153     next unless defined $block;
 154     chomp $block;
 155     $block =~ s/ *$//;
 156     if (! $seenSpecials and $block =~ /Specials/) {
 157         # Special case SPECIALS, since it is two disjoint ranges
 158         $seenSpecials = 1;
 159         next;
 160     }
 161     if ($block =~ /Surrogates/) {
 162         # Special case SURROGATES_AREA, since it one range, not three
 163         # consecutive, in Java
 164         $seenSurrogates++;
 165         if ($seenSurrogates == 1) {
 166             $surrogateStart = $start;
 167             next;
 168         } elsif ($seenSurrogates == 2) {
 169             next;
 170         } else {
 171             $start = $surrogateStart;
 172             $block = "Surrogates Area";
 173         }
 174     }
 175     # Special case the name of PRIVATE_USE_AREA.
 176     $block =~ s/(Private Use)/$1 Area/;
 177
 178     (my $name = $block) =~ tr/a-z -/A-Z__/;
 179     push @names, $name;
 180     my $since = (defined $additions{$name}
 181                  ? "\n     * \@since $additions{$name}" : "");
 182     my $extra = ($block =~ /Specials/ ? "'\\uFEFF', " : "");
 183     print <<EOF;
 184
 185     /**
 186      * $block.
 187      * $extra'\\u$start' - '\\u$end'.$since
 188      */
 189     public static final UnicodeBlock $name
 190       = new UnicodeBlock('\\u$start', '\\u$end',
 191                          "$name");
 192 EOF
 193 }
 194
 195 print <<EOF;
 196
 197     /**
 198      * The defined subsets.
 199      */
 200     private static final UnicodeBlock sets[] = {
 201 EOF
 202
 203 foreach (@names) {
 204     print "      $_,\n";
 205 }
 206
 207 print <<EOF;
 208     };
 209   } // class UnicodeBlock
 210 EOF