Unicode data updated to be the latest beta of the Unicode 3.0.

[p5sagit/p5-mst-13.2.git] / lib / unicode / mktables.PL

diff --git a/lib/unicode/mktables.PL b/lib/unicode/mktables.PL
index 184e5ca..f54ea69 100755 (executable)
--- a/lib/unicode/mktables.PL
+++ b/lib/unicode/mktables.PL
@@ -1,7 +1,5 @@
 #!../../miniperl
 
-# Create the equivalence mappings.
-
 $UnicodeData = "UnicodeData-Latest.txt";
 
 # Note: we try to keep filenames unique within first 8 chars.  Using
@@ -154,6 +152,21 @@ mkdir "Eq", 0777;
 # Jamo
 
     ['JamoShort',      '1',            '$short'],
+
+# Syllables
+
+    ['IsSylV', '$syl eq "V"',          ''],
+    ['IsSylU', '$syl eq "U"',          ''],
+    ['IsSylI', '$syl eq "I"',          ''],
+    ['IsSylA', '$syl eq "A"',          ''],
+    ['IsSylE', '$syl eq "E"',          ''],
+    ['IsSylC', '$syl eq "C"',          ''],
+    ['IsSylO', '$syl eq "O"',          ''],
+    ['IsSylWV',        '$syl eq "V"',          ''],
+    ['IsSylWI',        '$syl eq "I"',          ''],
+    ['IsSylWA',        '$syl eq "A"',          ''],
+    ['IsSylWE',        '$syl eq "E"',          ''],
+    ['IsSylWC',        '$syl eq "C"',          ''],
 );
 
 # This is not written for speed...
@@ -225,6 +238,11 @@ sub proplist {
 
        $split = '($code, $short, $name) = split(/; */); $code =~ s/^U\+//;';
     }
+    elsif ($table =~ /^IsSyl/) {
+       open(UD, "syllables.txt") or warn "Can't open $table: $!";
+
+       $split = '($code, $short, $syl) = split(/; */); $code =~ s/^U\+//;';
+    }
     else {
        open(UD, $UnicodeData) or warn "Can't open $UnicodeData: $!";
 
@@ -304,6 +322,8 @@ END
     $out;
 }
 
+# Create the equivalence mappings.
+
 open(UNICODEDATA, $UnicodeData) || die "$0: $UnicodeData: $!\n";
 
 while (<UNICODEDATA>) {
@@ -312,48 +332,44 @@ while (<UNICODEDATA>) {
     $code{$name} = $code;
     $name{$code} = $name;
 
-    if ($name =~ /^((?:LATIN|GREEK|CYRILLIC|HEBREW|BENGALI) .+? LETTER .+?) WITH /) {
-       push @base, [ $code, $1 ];
-    } elsif ($name =~ /^(ARABIC LETTER \w+?) WITH .+ (\w+ FORM)$/) {
-       push @base, [ $code, "$1 $2" ];
-    } elsif ($name =~ /^(ARABIC LETTER \w+?) WITH /) {
-       push @base, [ $code, $1 ];
-# Is the concept of turning ligatures into character classes sound?
-    } elsif ($name =~ /^(ARABIC) LIGATURE (.+?) (WITH .+ )+(\w+ FORM)$/) {
-       my $script = $1;
-       my $base   = $2;
-       my $with   = $3;
-       my $form   = $4;
-       push @base, [ $code, "$script LETTER $base" ];
-       push @base, [ $code, "$script LETTER $base $form" ];
-       my @with = split(/\bWITH\s+/, $with);
-       shift @with;
-       @with = grep { ! /^ (?:ABOVE|BELOW)/ } @with;
-       foreach my $base (@with) {
-           push @base, [ $code, "$script LETTER $base" ];
-           push @base, [ $code, "$script LETTER $base $form" ];
-       }
-    } elsif ($name =~ /^((?:ARMENIAN|CYRILLIC) .+) LIGATURE (\w+) (\w+)$/) {
-       push @base, [ $code, "$1 LETTER $2" ];
-       push @base, [ $code, "$1 LETTER $3" ];
-# Latin ligatures (ae, oe, ij, ff, fi, fl, ffi, ffl, long st, st) ignored.
-# Hebrew Yiddish ligatures (double vav, vav yod, double yod, yod yod patah,
-# alef lamed) ignored.
-    } else {
-       next;
-    }
-
+    next unless $name =~ /^(.+? LETTER .+?) WITH .+( \w+ FORM)?$/;
+
+    push @base, [ $code, $1 ];
+    push @base, [ $code, $1.$2 ] if $2 ne '';
+
+    # Before this "diacritics stripping" phase (and for Arabic, also
+    # "form stripping" phase) all ligatures could be decomposed into
+    # their constituent letters.
+    #
+    # For example the ligature
+    # ARABIC LIGATURE YEH WITH HAMZA ABOVE WITH ALEF ISOLATED FORM
+    # would go first through ligature decomposition producing the two letters
+    # ARABIC LETTER YEH WITH HAMZA ABOVE ISOLATED FORM
+    # ARABIC LETTER ALEF WITH HAMZA ABOVE ISOLATED FORM
+    # and those with diacritics stripping
+    # ARABIC LETTER YEH ISOLATED FORM
+    # ARABIC LETTER ALEF ISOLATED FORM
+    # and those with the Arabic form stripping
+    # ARABIC LETTER YEH
+    # ARABIC LETTER ALEF ISOLATED FORM
+    # ARABIC LETTER YEH
+    # ARABIC LETTER ALEF ISOLATED FORM
+    #
+    # Similarly for ligatures from other scripts.
+    # Effectively this would mean that ligatures turn into categories
+    # (Unicodese for character classes).
 }
 
 foreach my $b (@base) {
     ($code, $base) = @$b;
     next unless exists $code{$base};
     push @{$unicode{$code{$base}}}, $code;
-    print "$code: $name{$code} -> $base\n",
+#   print "$code: $name{$code} -> $base\n",
 }
 
 @unicode = sort keys %unicode;
 
+print "Eq/Unicode\n";
 if (open(EQ_UNICODE, ">Eq/Unicode")) {
     foreach my $c (@unicode) {
        print EQ_UNICODE "$c @{$unicode{$c}}\n";
@@ -363,6 +379,7 @@ if (open(EQ_UNICODE, ">Eq/Unicode")) {
     die "$0: failed to open Eq/Unicode for writing: $!\n";
 }
 
+print "Eq/Latin1\n";
 if (open(EQ_LATIN1, ">Eq/Latin1")) {
     foreach my $c (@unicode) {
        last if hex($c) > 255;