X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=lib%2FUnicode%2FCollate.pm;h=cd9b8e70e63124e4dcf48c5bb43445103d162146;hb=06c0cc96ebd866767a6d107ed78967600f7e0395;hp=fa0ef225fa37bf4504cd58bf4c4a95778e7c3016;hpb=ed423f7afb5038546a92d00ca689992f3e08bc61;p=p5sagit%2Fp5-mst-13.2.git diff --git a/lib/Unicode/Collate.pm b/lib/Unicode/Collate.pm index fa0ef22..cd9b8e7 100644 --- a/lib/Unicode/Collate.pm +++ b/lib/Unicode/Collate.pm @@ -1,8 +1,8 @@ package Unicode::Collate; BEGIN { - if (ord("A") == 193) { - die "Unicode::Collate not ported to EBCDIC\n"; + unless ("A" eq pack('U', 0x41)) { + die "Unicode::Collate cannot stringify a Unicode code point\n"; } } @@ -12,74 +12,137 @@ use warnings; use Carp; use File::Spec; -require Exporter; +no warnings 'utf8'; -our $VERSION = '0.20'; +our $VERSION = '0.52'; our $PACKAGE = __PACKAGE__; -our @ISA = qw(Exporter); - -our %EXPORT_TAGS = (); -our @EXPORT_OK = (); -our @EXPORT = (); - -(our $Path = $INC{'Unicode/Collate.pm'}) =~ s/\.pm$//; -our $KeyFile = "allkeys.txt"; - -our $UNICODE_VERSION; - -eval { require Unicode::UCD }; - -unless ($@) { - $UNICODE_VERSION = Unicode::UCD::UnicodeVersion(); -} -else { # XXX, Perl 5.6.1 - my($f, $fh); - foreach my $d (@INC) { - $f = File::Spec->catfile($d, "unicode", "Unicode.301"); - if (open($fh, $f)) { - $UNICODE_VERSION = '3.0.1'; - close $fh; - last; - } - } -} - -our $getCombinClass; # coderef for combining class from Unicode::Normalize - -use constant Min2 => 0x20; # minimum weight at level 2 -use constant Min3 => 0x02; # minimum weight at level 3 - -# format for pack -use constant VCE_FORMAT => 'Cn4'; # for variable + CE with 4 levels - -# values of variable +my @Path = qw(Unicode Collate); +my $KeyFile = "allkeys.txt"; + +# Perl's boolean +use constant TRUE => 1; +use constant FALSE => ""; +use constant NOMATCHPOS => -1; + +# A coderef to get combining class imported from Unicode::Normalize +# (i.e. \&Unicode::Normalize::getCombinClass). +# This is also used as a HAS_UNICODE_NORMALIZE flag. +my $CVgetCombinClass; + +# Supported Levels +use constant MinLevel => 1; +use constant MaxLevel => 4; + +# Minimum weights at level 2 and 3, respectively +use constant Min2Wt => 0x20; +use constant Min3Wt => 0x02; + +# Shifted weight at 4th level +use constant Shift4Wt => 0xFFFF; + +# A boolean for Variable and 16-bit weights at 4 levels of Collation Element +# PROBLEM: The Default Unicode Collation Element Table +# has weights over 0xFFFF at the 4th level. +# The tie-breaking in the variable weights +# other than "shift" (as well as "shift-trimmed") is unreliable. +use constant VCE_TEMPLATE => 'Cn4'; + +# A sort key: 16-bit weights +# See also the PROBLEM on VCE_TEMPLATE above. +use constant KEY_TEMPLATE => 'n*'; + +# Level separator in a sort key: +# i.e. pack(KEY_TEMPLATE, 0) +use constant LEVEL_SEP => "\0\0"; + +# As Unicode code point separator for hash keys. +# A joined code point string (denoted by JCPS below) +# like "65;768" is used for internal processing +# instead of Perl's Unicode string like "\x41\x{300}", +# as the native code point is different from the Unicode code point +# on EBCDIC platform. +# This character must not be included in any stringified +# representation of an integer. +use constant CODE_SEP => ';'; + +# boolean values of variable weights use constant NON_VAR => 0; # Non-Variable character use constant VAR => 1; # Variable character -our $DefaultRearrange = [ 0x0E40..0x0E44, 0x0EC0..0x0EC4 ]; +# specific code points +use constant Hangul_LBase => 0x1100; +use constant Hangul_LIni => 0x1100; +use constant Hangul_LFin => 0x1159; +use constant Hangul_LFill => 0x115F; +use constant Hangul_VBase => 0x1161; +use constant Hangul_VIni => 0x1160; # from Vowel Filler +use constant Hangul_VFin => 0x11A2; +use constant Hangul_TBase => 0x11A7; # from "no-final" codepoint +use constant Hangul_TIni => 0x11A8; +use constant Hangul_TFin => 0x11F9; +use constant Hangul_TCount => 28; +use constant Hangul_NCount => 588; +use constant Hangul_SBase => 0xAC00; +use constant Hangul_SIni => 0xAC00; +use constant Hangul_SFin => 0xD7A3; +use constant CJK_UidIni => 0x4E00; +use constant CJK_UidFin => 0x9FA5; +use constant CJK_UidF41 => 0x9FBB; +use constant CJK_ExtAIni => 0x3400; +use constant CJK_ExtAFin => 0x4DB5; +use constant CJK_ExtBIni => 0x20000; +use constant CJK_ExtBFin => 0x2A6D6; +use constant BMP_Max => 0xFFFF; + +# Logical_Order_Exception in PropList.txt +my $DefaultRearrange = [ 0x0E40..0x0E44, 0x0EC0..0x0EC4 ]; + +sub UCA_Version { "14" } + +sub Base_Unicode_Version { "4.1.0" } + +###### + +sub pack_U { + return pack('U*', @_); +} -sub UCA_Version { "9" } +sub unpack_U { + return unpack('U*', shift(@_).pack('U*')); +} -sub Base_Unicode_Version { $UNICODE_VERSION || 'unknown' } +###### -my (%AlternateOK); -@AlternateOK{ qw/ +my (%VariableOK); +@VariableOK{ qw/ blanked non-ignorable shifted shift-trimmed - / } = (); + / } = (); # keys lowercased our @ChangeOK = qw/ alternate backwards level normalization rearrange katakana_before_hiragana upper_before_lower overrideHangul overrideCJK preprocess UCA_Version + hangul_terminator variable /; our @ChangeNG = qw/ - entry entries table ignored combining maxlength - ignoreChar ignoreName undefChar undefName + entry mapping table maxlength + ignoreChar ignoreName undefChar undefName variableTable versionTable alternateTable backwardsTable forwardsTable rearrangeTable - derivCode normCode rearrangeHash isShift L3ignorable + derivCode normCode rearrangeHash + backwardsFlag /; +# The hash key 'ignored' is deleted at v 0.21. +# The hash key 'isShift' is deleted at v 0.23. +# The hash key 'combining' is deleted at v 0.24. +# The hash key 'entries' is deleted at v 0.30. +# The hash key 'L3_ignorable' is deleted at v 0.40. + +sub version { + my $self = shift; + return $self->{versionTable} || 'unknown'; +} my (%ChangeOK, %ChangeNG); @ChangeOK{ @ChangeOK } = (); @@ -89,6 +152,12 @@ sub change { my $self = shift; my %hash = @_; my %old; + if (exists $hash{variable} && exists $hash{alternate}) { + delete $hash{alternate}; + } + elsif (!exists $hash{variable} && exists $hash{alternate}) { + $hash{variable} = $hash{alternate}; + } foreach my $k (keys %hash) { if (exists $ChangeOK{$k}) { $old{$k} = $self->{$k}; @@ -99,39 +168,63 @@ sub change { } # else => ignored } - $self->checkCollator; + $self->checkCollator(); return wantarray ? %old : $self; } +sub _checkLevel { + my $level = shift; + my $key = shift; # 'level' or 'backwards' + MinLevel <= $level or croak sprintf + "Illegal level %d (in value for key '%s') lower than %d.", + $level, $key, MinLevel; + $level <= MaxLevel or croak sprintf + "Unsupported level %d (in value for key '%s') higher than %d.", + $level, $key, MaxLevel; +} + +my %DerivCode = ( + 8 => \&_derivCE_8, + 9 => \&_derivCE_9, + 11 => \&_derivCE_9, # 11 == 9 + 14 => \&_derivCE_14, +); + sub checkCollator { my $self = shift; - croak "Illegal level lower than 1 (passed $self->{level})." - if $self->{level} < 1; - croak "A level higher than 4 (passed $self->{level}) is not supported." - if 4 < $self->{level}; - - $self->{derivCode} = - $self->{UCA_Version} == -1 ? \&broken_derivCE : - $self->{UCA_Version} == 8 ? \&derivCE_8 : - $self->{UCA_Version} == 9 ? \&derivCE_9 : - croak "Illegal UCA version (passed $self->{UCA_Version})."; - - $self->{alternate} = lc($self->{alternate}); - croak "$PACKAGE unknown alternate tag name: $self->{alternate}" - unless exists $AlternateOK{ $self->{alternate} }; - - $self->{isShift} = $self->{alternate} eq 'shifted' || - $self->{alternate} eq 'shift-trimmed'; - - $self->{backwards} = [] - if ! defined $self->{backwards}; - $self->{backwards} = [ $self->{backwards} ] - if ! ref $self->{backwards}; - - $self->{rearrange} = [] - if ! defined $self->{rearrange}; - croak "$PACKAGE: A list for rearrangement must be store in an ARRAYREF" - if ! ref $self->{rearrange}; + _checkLevel($self->{level}, "level"); + + $self->{derivCode} = $DerivCode{ $self->{UCA_Version} } + or croak "Illegal UCA version (passed $self->{UCA_Version})."; + + $self->{variable} ||= $self->{alternate} || $self->{variableTable} || + $self->{alternateTable} || 'shifted'; + $self->{variable} = $self->{alternate} = lc($self->{variable}); + exists $VariableOK{ $self->{variable} } + or croak "$PACKAGE unknown variable parameter name: $self->{variable}"; + + if (! defined $self->{backwards}) { + $self->{backwardsFlag} = 0; + } + elsif (! ref $self->{backwards}) { + _checkLevel($self->{backwards}, "backwards"); + $self->{backwardsFlag} = 1 << $self->{backwards}; + } + else { + my %level; + $self->{backwardsFlag} = 0; + for my $b (@{ $self->{backwards} }) { + _checkLevel($b, "backwards"); + $level{$b} = 1; + } + for my $v (sort keys %level) { + $self->{backwardsFlag} += 1 << $v; + } + } + + defined $self->{rearrange} or $self->{rearrange} = []; + ref $self->{rearrange} + or croak "$PACKAGE: list for rearrangement must be store in ARRAYREF"; # keys of $self->{rearrangeHash} are $self->{rearrange}. $self->{rearrangeHash} = undef; @@ -144,20 +237,21 @@ sub checkCollator { if (defined $self->{normalization}) { eval { require Unicode::Normalize }; - croak "Unicode/Normalize.pm is required to normalize strings: $@" - if $@; - - Unicode::Normalize->import(); - $getCombinClass = \&Unicode::Normalize::getCombinClass - if ! $getCombinClass; - - $self->{normCode} = - $self->{normalization} =~ /^(?:NF)?C$/ ? \&NFC : - $self->{normalization} =~ /^(?:NF)?D$/ ? \&NFD : - $self->{normalization} =~ /^(?:NF)?KC$/ ? \&NFKC : - $self->{normalization} =~ /^(?:NF)?KD$/ ? \&NFKD : - croak "$PACKAGE unknown normalization form name: " - . $self->{normalization}; + $@ and croak "Unicode::Normalize is required to normalize strings"; + + $CVgetCombinClass ||= \&Unicode::Normalize::getCombinClass; + + if ($self->{normalization} =~ /^(?:NF)D\z/) { # tweak for default + $self->{normCode} = \&Unicode::Normalize::NFD; + } + elsif ($self->{normalization} ne 'prenormalized') { + my $norm = $self->{normalization}; + $self->{normCode} = sub { + Unicode::Normalize::normalize($norm, shift); + }; + eval { $self->{normCode}->("") }; # try + $@ and croak "$PACKAGE unknown normalization form name: $norm"; + } } return; } @@ -169,64 +263,76 @@ sub new # If undef is passed explicitly, no file is read. $self->{table} = $KeyFile if ! exists $self->{table}; - $self->read_table if defined $self->{table}; + $self->read_table() if defined $self->{table}; if ($self->{entry}) { - $self->parseEntry($_) foreach split /\n/, $self->{entry}; + while ($self->{entry} =~ /([^\n]+)/g) { + $self->parseEntry($1); + } } - $self->{level} ||= 4; + $self->{level} ||= MaxLevel; $self->{UCA_Version} ||= UCA_Version(); - $self->{overrideHangul} = '' + $self->{overrideHangul} = FALSE if ! exists $self->{overrideHangul}; - $self->{overrideCJK} = '' + $self->{overrideCJK} = FALSE if ! exists $self->{overrideCJK}; - $self->{normalization} = 'D' + $self->{normalization} = 'NFD' if ! exists $self->{normalization}; - $self->{alternate} = $self->{alternateTable} || 'shifted' - if ! exists $self->{alternate}; - $self->{rearrange} = $self->{rearrangeTable} || $DefaultRearrange + $self->{rearrange} = $self->{rearrangeTable} || + ($self->{UCA_Version} <= 11 ? $DefaultRearrange : []) if ! exists $self->{rearrange}; $self->{backwards} = $self->{backwardsTable} if ! exists $self->{backwards}; - $self->checkCollator; + $self->checkCollator(); return $self; } sub read_table { my $self = shift; - my $file = $self->{table} ne '' ? $self->{table} : $KeyFile; - my $filepath = File::Spec->catfile($Path, $file); - open my $fk, "<$filepath" - or croak "File does not exist at $filepath"; + my($f, $fh); + foreach my $d (@INC) { + $f = File::Spec->catfile($d, @Path, $self->{table}); + last if open($fh, $f); + $f = undef; + } + if (!defined $f) { + $f = File::Spec->catfile(@Path, $self->{table}); + croak("$PACKAGE: Can't locate $f in \@INC (\@INC contains: @INC)"); + } - while (<$fk>) { - next if /^\s*#/; - if (/^\s*\@/) { - if (/^\s*\@version\s*(\S*)/) { - $self->{versionTable} ||= $1; - } - elsif (/^\s*\@alternate\s+(\S*)/) { - $self->{alternateTable} ||= $1; - } - elsif (/^\s*\@backwards\s+(\S*)/) { - push @{ $self->{backwardsTable} }, $1; - } - elsif (/^\s*\@forwards\s+(\S*)/) { # parhaps no use - push @{ $self->{forwardsTable} }, $1; - } - elsif (/^\s*\@rearrange\s+(.*)/) { # (\S*) is NG - push @{ $self->{rearrangeTable} }, _getHexArray($1); - } + while (my $line = <$fh>) { + next if $line =~ /^\s*#/; + unless ($line =~ s/^\s*\@//) { + $self->parseEntry($line); next; } - $self->parseEntry($_); + + # matched ^\s*\@ + if ($line =~ /^version\s*(\S*)/) { + $self->{versionTable} ||= $1; + } + elsif ($line =~ /^variable\s+(\S*)/) { # since UTS #10-9 + $self->{variableTable} ||= $1; + } + elsif ($line =~ /^alternate\s+(\S*)/) { # till UTS #10-8 + $self->{alternateTable} ||= $1; + } + elsif ($line =~ /^backwards\s+(\S*)/) { + push @{ $self->{backwardsTable} }, $1; + } + elsif ($line =~ /^forwards\s+(\S*)/) { # parhaps no use + push @{ $self->{forwardsTable} }, $1; + } + elsif ($line =~ /^rearrange\s+(.*)/) { # (\S*) is NG + push @{ $self->{rearrangeTable} }, _getHexArray($1); + } } - close $fk; + close $fh; } @@ -237,7 +343,7 @@ sub parseEntry { my $self = shift; my $line = shift; - my($name, $ele, @key); + my($name, $entry, @uv, @key); return if $line !~ /^\s*[0-9A-Fa-f]/; @@ -251,67 +357,87 @@ sub parseEntry croak "Wrong Entry: must be separated by ';' from " if ! $k; - my @e = _getHexArray($e); - return if !@e; + @uv = _getHexArray($e); + return if !@uv; + + $entry = join(CODE_SEP, @uv); # in JCPS - $ele = pack('U*', @e); - return if defined $self->{undefChar} && $ele =~ /$self->{undefChar}/; + if (defined $self->{undefChar} || defined $self->{ignoreChar}) { + my $ele = pack_U(@uv); - # get sort key - if (defined $self->{ignoreName} && $name =~ /$self->{ignoreName}/ || - defined $self->{ignoreChar} && $ele =~ /$self->{ignoreChar}/) - { - $self->{entries}{$ele} = $self->{ignored}{$ele} = 1; + # regarded as if it were not entried in the table + return + if defined $self->{undefChar} && $ele =~ /$self->{undefChar}/; + + # replaced as completely ignorable + $k = '[.0000.0000.0000.0000]' + if defined $self->{ignoreChar} && $ele =~ /$self->{ignoreChar}/; } - else { - my $combining = 1; # primary = 0, secondary != 0; - my $level3ingore; - - foreach my $arr ($k =~ /\[([^\[\]]+)\]/g) { # SPACEs allowed - my $var = $arr =~ /\*/; # exactly /^\*/ but be lenient. - my @arr = _getHexArray($arr); - push @key, pack(VCE_FORMAT, $var, @arr); - $combining = 0 unless $arr[0] == 0 && $arr[1] != 0; - $level3ingore = 1 if $arr[0] == 0 && $arr[1] == 0 && $arr[2] == 0; - } - $self->{entries}{$ele} = \@key; - $self->{combining}{$ele} = 1 - if $combining; + # replaced as completely ignorable + $k = '[.0000.0000.0000.0000]' + if defined $self->{ignoreName} && $name =~ /$self->{ignoreName}/; + + my $is_L3_ignorable = TRUE; + + foreach my $arr ($k =~ /\[([^\[\]]+)\]/g) { # SPACEs allowed + my $var = $arr =~ /\*/; # exactly /^\*/ but be lenient. + my @wt = _getHexArray($arr); + push @key, pack(VCE_TEMPLATE, $var, @wt); + $is_L3_ignorable = FALSE + if $wt[0] || $wt[1] || $wt[2]; + # Conformance Test for 3.1.1 and 4.0.0 shows Level 3 ignorable + # is completely ignorable. + # For expansion, an entry $is_L3_ignorable + # if and only if "all" CEs are [.0000.0000.0000]. + } + + $self->{mapping}{$entry} = $is_L3_ignorable ? [] : \@key; - $self->{L3ignorable}{$e[0]} = 1 - if @e == 1 && $level3ingore; + if (@uv > 1) { + (!$self->{maxlength}{$uv[0]} || $self->{maxlength}{$uv[0]} < @uv) + and $self->{maxlength}{$uv[0]} = @uv; } - $self->{maxlength}{ord $ele} = scalar @e if @e > 1; } + ## -## arrayref CE = altCE(bool variable?, list[num] weights) +## VCE = _varCE(variable term, VCE) ## -sub altCE +sub _varCE { - my $self = shift; - my($var, @c) = unpack(VCE_FORMAT, shift); - - $self->{alternate} eq 'blanked' ? - $var ? [0,0,0,$c[3]] : \@c : - $self->{alternate} eq 'non-ignorable' ? - \@c : - $self->{alternate} eq 'shifted' ? - $var ? [0,0,0,$c[0] ] : [ @c[0..2], $c[0]+$c[1]+$c[2] ? 0xFFFF : 0 ] : - $self->{alternate} eq 'shift-trimmed' ? - $var ? [0,0,0,$c[0] ] : [ @c[0..2], 0 ] : - croak "$PACKAGE unknown alternate name: $self->{alternate}"; + my $vbl = shift; + my $vce = shift; + if ($vbl eq 'non-ignorable') { + return $vce; + } + my ($var, @wt) = unpack VCE_TEMPLATE, $vce; + + if ($var) { + return pack(VCE_TEMPLATE, $var, 0, 0, 0, + $vbl eq 'blanked' ? $wt[3] : $wt[0]); + } + elsif ($vbl eq 'blanked') { + return $vce; + } + else { + return pack(VCE_TEMPLATE, $var, @wt[0..2], + $vbl eq 'shifted' && $wt[0]+$wt[1]+$wt[2] ? Shift4Wt : 0); + } } sub viewSortKey { my $self = shift; - my $ver = $self->{UCA_Version}; + $self->visualizeSortKey($self->getSortKey(@_)); +} + +sub visualizeSortKey +{ + my $self = shift; + my $view = join " ", map sprintf("%04X", $_), unpack(KEY_TEMPLATE, shift); - my $key = $self->getSortKey(@_); - my $view = join " ", map sprintf("%04X", $_), unpack 'n*', $key; - if ($ver <= 8) { + if ($self->{UCA_Version} <= 8) { $view =~ s/ ?0000 ?/|/g; } else { $view =~ s/\b0000\b/|/g; @@ -321,27 +447,43 @@ sub viewSortKey ## -## list[strings] elements = splitCE(string arg) +## arrayref of JCPS = splitEnt(string to be collated) +## arrayref of arrayref[JCPS, ini_pos, fin_pos] = splitEnt(string, true) ## -sub splitCE +sub splitEnt { my $self = shift; + my $wLen = $_[1]; + my $code = $self->{preprocess}; my $norm = $self->{normCode}; - my $ent = $self->{entries}; + my $map = $self->{mapping}; my $max = $self->{maxlength}; my $reH = $self->{rearrangeHash}; - my $L3i = $self->{L3ignorable}; - my $ver9 = $self->{UCA_Version} > 8; + my $ver9 = $self->{UCA_Version} >= 9 && $self->{UCA_Version} <= 11; - my $str = ref $code ? &$code(shift) : shift; - $str = &$norm($str) if ref $norm; + my ($str, @buf); + + if ($wLen) { + $code and croak "Preprocess breaks character positions. " + . "Don't use with index(), match(), etc."; + $norm and croak "Normalization breaks character positions. " + . "Don't use with index(), match(), etc."; + $str = $_[0]; + } + else { + $str = $_[0]; + $str = &$code($str) if ref $code; + $str = &$norm($str) if ref $norm; + } - my @src = unpack('U*', $str); - my @buf; + # get array of Unicode code point of string. + my @src = unpack_U($str); - # rearrangement - if ($reH) { + # rearrangement: + # Character positions are not kept if rearranged, + # then neglected if $wLen is true. + if ($reH && ! $wLen) { for (my $i = 0; $i < @src; $i++) { if (exists $reH->{ $src[$i] } && $i + 1 < @src) { ($src[$i], $src[$i+1]) = ($src[$i+1], $src[$i]); @@ -350,239 +492,244 @@ sub splitCE } } - if ($ver9) { - @src = grep ! $L3i->{$_}, @src; + # remove a code point marked as a completely ignorable. + for (my $i = 0; $i < @src; $i++) { + $src[$i] = undef + if _isIllegal($src[$i]) || ($ver9 && + $map->{ $src[$i] } && @{ $map->{ $src[$i] } } == 0); } for (my $i = 0; $i < @src; $i++) { - my $ch; - my $u = $src[$i]; - - # non-characters - next unless defined $u; - next if $u < 0 || 0x10FFFF < $u # out of range - || (0xD800 <= $u && $u <= 0xDFFF) # unpaired surrogates - || (0xFDD0 <= $u && $u <= 0xFDEF) # non-character - ; - - my $four = $u & 0xFFFF; - next if $four == 0xFFFE || $four == 0xFFFF; - - if ($max->{$u}) { # contract - for (my $j = $max->{$u}; $j >= 1; $j--) { - next unless $i+$j-1 < @src; - $ch = pack 'U*', @src[$i .. $i+$j-1]; - $i += $j-1, last if $ent->{$ch}; + my $jcps = $src[$i]; + + # skip removed code point + if (! defined $jcps) { + if ($wLen && @buf) { + $buf[-1][2] = $i + 1; + } + next; + } + + my $i_orig = $i; + + # find contraction + if ($max->{$jcps}) { + my $temp_jcps = $jcps; + my $jcpsLen = 1; + my $maxLen = $max->{$jcps}; + + for (my $p = $i + 1; $jcpsLen < $maxLen && $p < @src; $p++) { + next if ! defined $src[$p]; + $temp_jcps .= CODE_SEP . $src[$p]; + $jcpsLen++; + if ($map->{$temp_jcps}) { + $jcps = $temp_jcps; + $i = $p; + } + } + + # not-contiguous contraction with Combining Char (cf. UTS#10, S2.1). + # This process requires Unicode::Normalize. + # If "normalization" is undef, here should be skipped *always* + # (in spite of bool value of $CVgetCombinClass), + # since canonical ordering cannot be expected. + # Blocked combining character should not be contracted. + + if ($self->{normalization}) + # $self->{normCode} is false in the case of "prenormalized". + { + my $preCC = 0; + my $curCC = 0; + + for (my $p = $i + 1; $p < @src; $p++) { + next if ! defined $src[$p]; + $curCC = $CVgetCombinClass->($src[$p]); + last unless $curCC; + my $tail = CODE_SEP . $src[$p]; + if ($preCC != $curCC && $map->{$jcps.$tail}) { + $jcps .= $tail; + $src[$p] = undef; + } else { + $preCC = $curCC; + } + } } - } else { - $ch = pack('U', $u); } - # with Combining Char (UTS#10, 4.2.1), here requires Unicode::Normalize. - if ($getCombinClass && defined $ch) { - for (my $j = $i+1; $j < @src; $j++) { - next unless defined $src[$j]; - last unless $getCombinClass->( $src[$j] ); - my $comb = pack 'U', $src[$j]; - next if ! $ent->{ $ch.$comb }; - $ch .= $comb; - $src[$j] = undef; + # skip completely ignorable + if ($map->{$jcps} && @{ $map->{$jcps} } == 0) { + if ($wLen && @buf) { + $buf[-1][2] = $i + 1; } + next; } - push @buf, $ch; + + push @buf, $wLen ? [$jcps, $i_orig, $i + 1] : $jcps; } - wantarray ? @buf : \@buf; + return \@buf; } ## -## list[arrayrefs] weight = getWt(string element) +## list of VCE = getWt(JCPS) ## sub getWt { my $self = shift; - my $ch = shift; - my $ent = $self->{entries}; - my $ign = $self->{ignored}; - my $cjk = $self->{overrideCJK}; - my $hang = $self->{overrideHangul}; + my $u = shift; + my $vbl = $self->{variable}; + my $map = $self->{mapping}; my $der = $self->{derivCode}; - return if !defined $ch || $ign->{$ch}; # ignored - return map($self->altCE($_), @{ $ent->{$ch} }) - if $ent->{$ch}; - - my $u = unpack('U', $ch); - - if (0xAC00 <= $u && $u <= 0xD7A3) { # is_Hangul - return map $self->altCE($_), - $hang - ? map(pack(VCE_FORMAT, NON_VAR, @$_), &$hang($u)) - : defined $hang - ? map({ - my $v = $_; - my $vCE = $ent->{pack('U', $v)}; - $vCE ? @$vCE : $der->($v); - } _decompHangul($u)) - : $der->($u); + return if !defined $u; + return map(_varCE($vbl, $_), @{ $map->{$u} }) + if $map->{$u}; + + # JCPS must not be a contraction, then it's a code point. + if (Hangul_SIni <= $u && $u <= Hangul_SFin) { + my $hang = $self->{overrideHangul}; + my @hangulCE; + if ($hang) { + @hangulCE = map(pack(VCE_TEMPLATE, NON_VAR, @$_), &$hang($u)); + } + elsif (!defined $hang) { + @hangulCE = $der->($u); + } + else { + my $max = $self->{maxlength}; + my @decH = _decompHangul($u); + + if (@decH == 2) { + my $contract = join(CODE_SEP, @decH); + @decH = ($contract) if $map->{$contract}; + } else { # must be <@decH == 3> + if ($max->{$decH[0]}) { + my $contract = join(CODE_SEP, @decH); + if ($map->{$contract}) { + @decH = ($contract); + } else { + $contract = join(CODE_SEP, @decH[0,1]); + $map->{$contract} and @decH = ($contract, $decH[2]); + } + # even if V's ignorable, LT contraction is not supported. + # If such a situatution were required, NFD should be used. + } + if (@decH == 3 && $max->{$decH[1]}) { + my $contract = join(CODE_SEP, @decH[1,2]); + $map->{$contract} and @decH = ($decH[0], $contract); + } + } + + @hangulCE = map({ + $map->{$_} ? @{ $map->{$_} } : $der->($_); + } @decH); + } + return map _varCE($vbl, $_), @hangulCE; } - elsif (0x3400 <= $u && $u <= 0x4DB5 || - 0x4E00 <= $u && $u <= 0x9FA5 || - 0x20000 <= $u && $u <= 0x2A6D6) { # CJK Ideograph - return map $self->altCE($_), + elsif (_isUIdeo($u, $self->{UCA_Version})) { + my $cjk = $self->{overrideCJK}; + return map _varCE($vbl, $_), $cjk - ? map(pack(VCE_FORMAT, NON_VAR, @$_), &$cjk($u)) + ? map(pack(VCE_TEMPLATE, NON_VAR, @$_), &$cjk($u)) : defined $cjk && $self->{UCA_Version} <= 8 && $u < 0x10000 - ? pack(VCE_FORMAT, NON_VAR, $u, 0x20, 0x02, $u) + ? _uideoCE_8($u) : $der->($u); } else { - return map $self->altCE($_), $der->($u); + return map _varCE($vbl, $_), $der->($u); } } + ## -## int = index(string, substring) +## string sortkey = getSortKey(string arg) ## -sub index +sub getSortKey { my $self = shift; my $lev = $self->{level}; - my $comb = $self->{combining}; - my $str = $self->splitCE(shift); - my $sub = $self->splitCE(shift); - - return wantarray ? (0,0) : 0 if ! @$sub; - return wantarray ? () : -1 if ! @$str; - - my @subWt = grep _ignorableAtLevel($_,$lev), - map $self->getWt($_), @$sub; - - my(@strWt,@strPt); - my $count = 0; - for (my $i = 0; $i < @$str; $i++) { - my $go_ahead = 0; - - my @tmp = grep _ignorableAtLevel($_,$lev), $self->getWt($str->[$i]); - $go_ahead += length $str->[$i]; - - # /*XXX*/ still broken. - # index("e\x{300}", "e") should be 'no match' at level 2 or higher - # as "e\x{300}" is a *single* grapheme cluster and not equal to "e". - - # go ahead as far as we find a combining character; - while ($i + 1 < @$str && - (! defined $str->[$i+1] || $comb->{ $str->[$i+1] }) ) { - $i++; - next if ! defined $str->[$i]; - $go_ahead += length $str->[$i]; - push @tmp, - grep _ignorableAtLevel($_,$lev), $self->getWt($str->[$i]); - } - - push @strWt, @tmp; - push @strPt, ($count) x @tmp; - $count += $go_ahead; + my $rEnt = $self->splitEnt(shift); # get an arrayref of JCPS + my $v2i = $self->{UCA_Version} >= 9 && + $self->{variable} ne 'non-ignorable'; + + my @buf; # weight arrays + if ($self->{hangul_terminator}) { + my $preHST = ''; + foreach my $jcps (@$rEnt) { + # weird things like VL, TL-contraction are not considered! + my $curHST = ''; + foreach my $u (split /;/, $jcps) { + $curHST .= getHST($u); + } + if ($preHST && !$curHST || # hangul before non-hangul + $preHST =~ /L\z/ && $curHST =~ /^T/ || + $preHST =~ /V\z/ && $curHST =~ /^L/ || + $preHST =~ /T\z/ && $curHST =~ /^[LV]/) { - while (@strWt >= @subWt) { - if (_eqArray(\@strWt, \@subWt, $lev)) { - my $pos = $strPt[0]; - return wantarray ? ($pos, $count-$pos) : $pos; + push @buf, $self->getWtHangulTerm(); } - shift @strWt; - shift @strPt; + $preHST = $curHST; + + push @buf, $self->getWt($jcps); } + $preHST # end at hangul + and push @buf, $self->getWtHangulTerm(); } - return wantarray ? () : -1; -} - -## -## bool _eqArray(arrayref, arrayref, level) -## -sub _eqArray($$$) -{ - my $a = shift; # length $a >= length $b; - my $b = shift; - my $lev = shift; - for my $v (0..$lev-1) { - for my $c (0..@$b-1){ - return if $a->[$c][$v] != $b->[$c][$v]; + else { + foreach my $jcps (@$rEnt) { + push @buf, $self->getWt($jcps); } } - return 1; -} - -## -## bool _ignorableAtLevel(CE, level) -## -sub _ignorableAtLevel($$) -{ - my $ce = shift; - return unless defined $ce; - my $lv = shift; - return ! grep { ! $ce->[$_] } 0..$lv-1; -} + # make sort key + my @ret = ([],[],[],[]); + my $last_is_variable; + foreach my $vwt (@buf) { + my($var, @wt) = unpack(VCE_TEMPLATE, $vwt); -## -## string sortkey = getSortKey(string arg) -## -sub getSortKey -{ - my $self = shift; - my $lev = $self->{level}; - my $rCE = $self->splitCE(shift); # get an arrayref - my $ver9 = $self->{UCA_Version} > 8; - my $sht = $self->{isShift}; - - # weight arrays - my (@buf, $last_is_variable); - - foreach my $ce (@$rCE) { - my @t = $self->getWt($ce); - if ($sht && $ver9) { - if (@t == 1 && $t[0][0] == 0) { - if ($t[0][1] == 0 && $t[0][2] == 0) { - $last_is_variable = 1; - } else { - next if $last_is_variable; - } - } else { - $last_is_variable = 0; + # "Ignorable (L1, L2) after Variable" since track. v. 9 + if ($v2i) { + if ($var) { + $last_is_variable = TRUE; + } + elsif (!$wt[0]) { # ignorable + next if $last_is_variable; + } + else { + $last_is_variable = FALSE; } } - push @buf, @t; - } - - # make sort key - my @ret = ([],[],[],[]); - foreach my $v (0..$lev-1) { - foreach my $b (@buf) { - push @{ $ret[$v] }, $b->[$v] if $b->[$v]; + foreach my $v (0..$lev-1) { + 0 < $wt[$v] and push @{ $ret[$v] }, $wt[$v]; } } - foreach (@{ $self->{backwards} }) { - my $v = $_ - 1; - @{ $ret[$v] } = reverse @{ $ret[$v] }; - } # modification of tertiary weights if ($self->{upper_before_lower}) { - foreach (@{ $ret[2] }) { - if (0x8 <= $_ && $_ <= 0xC) { $_ -= 6 } # lower - elsif (0x2 <= $_ && $_ <= 0x6) { $_ += 6 } # upper - elsif ($_ == 0x1C) { $_ += 1 } # square upper - elsif ($_ == 0x1D) { $_ -= 1 } # square lower + foreach my $w (@{ $ret[2] }) { + if (0x8 <= $w && $w <= 0xC) { $w -= 6 } # lower + elsif (0x2 <= $w && $w <= 0x6) { $w += 6 } # upper + elsif ($w == 0x1C) { $w += 1 } # square upper + elsif ($w == 0x1D) { $w -= 1 } # square lower } } if ($self->{katakana_before_hiragana}) { - foreach (@{ $ret[2] }) { - if (0x0F <= $_ && $_ <= 0x13) { $_ -= 2 } # katakana - elsif (0x0D <= $_ && $_ <= 0x0E) { $_ += 5 } # hiragana + foreach my $w (@{ $ret[2] }) { + if (0x0F <= $w && $w <= 0x13) { $w -= 2 } # katakana + elsif (0x0D <= $w && $w <= 0x0E) { $w += 5 } # hiragana + } + } + + if ($self->{backwardsFlag}) { + for (my $v = MinLevel; $v <= MaxLevel; $v++) { + if ($self->{backwardsFlag} & (1 << $v)) { + @{ $ret[$v-1] } = reverse @{ $ret[$v-1] }; + } } } - join "\0\0", map pack('n*', @$_), @ret; + + join LEVEL_SEP, map pack(KEY_TEMPLATE, @$_), @ret; } @@ -609,61 +756,360 @@ sub sort { } -sub derivCE_9 { +sub _derivCE_14 { my $u = shift; my $base = - (0x4E00 <= $u && $u <= 0x9FA5) # CJK - ? 0xFB40 : - (0x3400 <= $u && $u <= 0x4DB5 || 0x20000 <= $u && $u <= 0x2A6D6) - ? 0xFB80 : 0xFBC0; + (CJK_UidIni <= $u && $u <= CJK_UidF41) + ? 0xFB40 : # CJK + (CJK_ExtAIni <= $u && $u <= CJK_ExtAFin || + CJK_ExtBIni <= $u && $u <= CJK_ExtBFin) + ? 0xFB80 # CJK ext. + : 0xFBC0; # others my $aaaa = $base + ($u >> 15); my $bbbb = ($u & 0x7FFF) | 0x8000; return - pack(VCE_FORMAT, NON_VAR, $aaaa, Min2, Min3, $u), - pack(VCE_FORMAT, NON_VAR, $bbbb, 0, 0, $u); + pack(VCE_TEMPLATE, NON_VAR, $aaaa, Min2Wt, Min3Wt, $u), + pack(VCE_TEMPLATE, NON_VAR, $bbbb, 0, 0, $u); } -sub derivCE_8 { +sub _derivCE_9 { + my $u = shift; + my $base = + (CJK_UidIni <= $u && $u <= CJK_UidFin) + ? 0xFB40 : # CJK + (CJK_ExtAIni <= $u && $u <= CJK_ExtAFin || + CJK_ExtBIni <= $u && $u <= CJK_ExtBFin) + ? 0xFB80 # CJK ext. + : 0xFBC0; # others + + my $aaaa = $base + ($u >> 15); + my $bbbb = ($u & 0x7FFF) | 0x8000; + return + pack(VCE_TEMPLATE, NON_VAR, $aaaa, Min2Wt, Min3Wt, $u), + pack(VCE_TEMPLATE, NON_VAR, $bbbb, 0, 0, $u); +} + +sub _derivCE_8 { my $code = shift; my $aaaa = 0xFF80 + ($code >> 15); my $bbbb = ($code & 0x7FFF) | 0x8000; return - pack(VCE_FORMAT, NON_VAR, $aaaa, 2, 1, $code), - pack(VCE_FORMAT, NON_VAR, $bbbb, 0, 0, $code); + pack(VCE_TEMPLATE, NON_VAR, $aaaa, 2, 1, $code), + pack(VCE_TEMPLATE, NON_VAR, $bbbb, 0, 0, $code); } -sub broken_derivCE { # NG - my $code = shift; - my $aaaa = 0xFFC2 + ($code >> 15); - my $bbbb = $code & 0x7FFF | 0x1000; - return - pack(VCE_FORMAT, NON_VAR, $aaaa, 2, 1, $code), - pack(VCE_FORMAT, NON_VAR, $bbbb, 0, 0, $code); +sub _uideoCE_8 { + my $u = shift; + return pack(VCE_TEMPLATE, NON_VAR, $u, Min2Wt, Min3Wt, $u); } +sub _isUIdeo { + my ($u, $uca_vers) = @_; + return( + (CJK_UidIni <= $u && + ($uca_vers >= 14 ? ( $u <= CJK_UidF41) : ($u <= CJK_UidFin))) + || + (CJK_ExtAIni <= $u && $u <= CJK_ExtAFin) + || + (CJK_ExtBIni <= $u && $u <= CJK_ExtBFin) + ); +} + + +sub getWtHangulTerm { + my $self = shift; + return _varCE($self->{variable}, + pack(VCE_TEMPLATE, NON_VAR, $self->{hangul_terminator}, 0,0,0)); +} + + ## ## "hhhh hhhh hhhh" to (dddd, dddd, dddd) ## sub _getHexArray { map hex, $_[0] =~ /([0-9a-fA-F]+)/g } # -# $code must be in Hangul syllable. +# $code *must* be in Hangul syllable. # Check it before you enter here. # sub _decompHangul { my $code = shift; - my $SIndex = $code - 0xAC00; - my $LIndex = int( $SIndex / 588); - my $VIndex = int(($SIndex % 588) / 28); - my $TIndex = $SIndex % 28; + my $si = $code - Hangul_SBase; + my $li = int( $si / Hangul_NCount); + my $vi = int(($si % Hangul_NCount) / Hangul_TCount); + my $ti = $si % Hangul_TCount; return ( - 0x1100 + $LIndex, - 0x1161 + $VIndex, - $TIndex ? (0x11A7 + $TIndex) : (), + Hangul_LBase + $li, + Hangul_VBase + $vi, + $ti ? (Hangul_TBase + $ti) : (), ); } +sub _isIllegal { + my $code = shift; + return ! defined $code # removed + || ($code < 0 || 0x10FFFF < $code) # out of range + || (($code & 0xFFFE) == 0xFFFE) # ??FFF[EF] (cf. utf8.c) + || (0xD800 <= $code && $code <= 0xDFFF) # unpaired surrogates + || (0xFDD0 <= $code && $code <= 0xFDEF) # other non-characters + ; +} + +# Hangul Syllable Type +sub getHST { + my $u = shift; + return + Hangul_LIni <= $u && $u <= Hangul_LFin || $u == Hangul_LFill ? "L" : + Hangul_VIni <= $u && $u <= Hangul_VFin ? "V" : + Hangul_TIni <= $u && $u <= Hangul_TFin ? "T" : + Hangul_SIni <= $u && $u <= Hangul_SFin ? + ($u - Hangul_SBase) % Hangul_TCount ? "LVT" : "LV" : ""; +} + + +## +## bool _nonIgnorAtLevel(arrayref weights, int level) +## +sub _nonIgnorAtLevel($$) +{ + my $wt = shift; + return if ! defined $wt; + my $lv = shift; + return grep($wt->[$_-1] != 0, MinLevel..$lv) ? TRUE : FALSE; +} + +## +## bool _eqArray( +## arrayref of arrayref[weights] source, +## arrayref of arrayref[weights] substr, +## int level) +## * comparison of graphemes vs graphemes. +## @$source >= @$substr must be true (check it before call this); +## +sub _eqArray($$$) +{ + my $source = shift; + my $substr = shift; + my $lev = shift; + + for my $g (0..@$substr-1){ + # Do the $g'th graphemes have the same number of AV weigths? + return if @{ $source->[$g] } != @{ $substr->[$g] }; + + for my $w (0..@{ $substr->[$g] }-1) { + for my $v (0..$lev-1) { + return if $source->[$g][$w][$v] != $substr->[$g][$w][$v]; + } + } + } + return 1; +} + +## +## (int position, int length) +## int position = index(string, substring, position, [undoc'ed grobal]) +## +## With "grobal" (only for the list context), +## returns list of arrayref[position, length]. +## +sub index +{ + my $self = shift; + my $str = shift; + my $len = length($str); + my $subE = $self->splitEnt(shift); + my $pos = @_ ? shift : 0; + $pos = 0 if $pos < 0; + my $grob = shift; + + my $lev = $self->{level}; + my $v2i = $self->{UCA_Version} >= 9 && + $self->{variable} ne 'non-ignorable'; + + if (! @$subE) { + my $temp = $pos <= 0 ? 0 : $len <= $pos ? $len : $pos; + return $grob + ? map([$_, 0], $temp..$len) + : wantarray ? ($temp,0) : $temp; + } + $len < $pos + and return wantarray ? () : NOMATCHPOS; + my $strE = $self->splitEnt($pos ? substr($str, $pos) : $str, TRUE); + @$strE + or return wantarray ? () : NOMATCHPOS; + + my(@strWt, @iniPos, @finPos, @subWt, @g_ret); + + my $last_is_variable; + for my $vwt (map $self->getWt($_), @$subE) { + my($var, @wt) = unpack(VCE_TEMPLATE, $vwt); + my $to_be_pushed = _nonIgnorAtLevel(\@wt,$lev); + + # "Ignorable (L1, L2) after Variable" since track. v. 9 + if ($v2i) { + if ($var) { + $last_is_variable = TRUE; + } + elsif (!$wt[0]) { # ignorable + $to_be_pushed = FALSE if $last_is_variable; + } + else { + $last_is_variable = FALSE; + } + } + + if (@subWt && !$var && !$wt[0]) { + push @{ $subWt[-1] }, \@wt if $to_be_pushed; + } else { + push @subWt, [ \@wt ]; + } + } + + my $count = 0; + my $end = @$strE - 1; + + $last_is_variable = FALSE; # reuse + for (my $i = 0; $i <= $end; ) { # no $i++ + my $found_base = 0; + + # fetch a grapheme + while ($i <= $end && $found_base == 0) { + for my $vwt ($self->getWt($strE->[$i][0])) { + my($var, @wt) = unpack(VCE_TEMPLATE, $vwt); + my $to_be_pushed = _nonIgnorAtLevel(\@wt,$lev); + + # "Ignorable (L1, L2) after Variable" since track. v. 9 + if ($v2i) { + if ($var) { + $last_is_variable = TRUE; + } + elsif (!$wt[0]) { # ignorable + $to_be_pushed = FALSE if $last_is_variable; + } + else { + $last_is_variable = FALSE; + } + } + + if (@strWt && !$var && !$wt[0]) { + push @{ $strWt[-1] }, \@wt if $to_be_pushed; + $finPos[-1] = $strE->[$i][2]; + } elsif ($to_be_pushed) { + push @strWt, [ \@wt ]; + push @iniPos, $found_base ? NOMATCHPOS : $strE->[$i][1]; + $finPos[-1] = NOMATCHPOS if $found_base; + push @finPos, $strE->[$i][2]; + $found_base++; + } + # else ===> no-op + } + $i++; + } + + # try to match + while ( @strWt > @subWt || (@strWt == @subWt && $i > $end) ) { + if ($iniPos[0] != NOMATCHPOS && + $finPos[$#subWt] != NOMATCHPOS && + _eqArray(\@strWt, \@subWt, $lev)) { + my $temp = $iniPos[0] + $pos; + + if ($grob) { + push @g_ret, [$temp, $finPos[$#subWt] - $iniPos[0]]; + splice @strWt, 0, $#subWt; + splice @iniPos, 0, $#subWt; + splice @finPos, 0, $#subWt; + } + else { + return wantarray + ? ($temp, $finPos[$#subWt] - $iniPos[0]) + : $temp; + } + } + shift @strWt; + shift @iniPos; + shift @finPos; + } + } + + return $grob + ? @g_ret + : wantarray ? () : NOMATCHPOS; +} + +## +## scalarref to matching part = match(string, substring) +## +sub match +{ + my $self = shift; + if (my($pos,$len) = $self->index($_[0], $_[1])) { + my $temp = substr($_[0], $pos, $len); + return wantarray ? $temp : \$temp; + # An lvalue ref \substr should be avoided, + # since its value is affected by modification of its referent. + } + else { + return; + } +} + +## +## arrayref matching parts = gmatch(string, substring) +## +sub gmatch +{ + my $self = shift; + my $str = shift; + my $sub = shift; + return map substr($str, $_->[0], $_->[1]), + $self->index($str, $sub, 0, 'g'); +} + +## +## bool subst'ed = subst(string, substring, replace) +## +sub subst +{ + my $self = shift; + my $code = ref $_[2] eq 'CODE' ? $_[2] : FALSE; + + if (my($pos,$len) = $self->index($_[0], $_[1])) { + if ($code) { + my $mat = substr($_[0], $pos, $len); + substr($_[0], $pos, $len, $code->($mat)); + } else { + substr($_[0], $pos, $len, $_[2]); + } + return TRUE; + } + else { + return FALSE; + } +} + +## +## int count = gsubst(string, substring, replace) +## +sub gsubst +{ + my $self = shift; + my $code = ref $_[2] eq 'CODE' ? $_[2] : FALSE; + my $cnt = 0; + + # Replacement is carried out from the end, then use reverse. + for my $pos_len (reverse $self->index($_[0], $_[1], 0, 'g')) { + if ($code) { + my $mat = substr($_[0], $pos_len->[0], $pos_len->[1]); + substr($_[0], $pos_len->[0], $pos_len->[1], $code->($mat)); + } else { + substr($_[0], $pos_len->[0], $pos_len->[1], $_[2]); + } + $cnt++; + } + return $cnt; +} + 1; __END__ @@ -684,22 +1130,29 @@ Unicode::Collate - Unicode Collation Algorithm #compare $result = $Collator->cmp($a, $b); # returns 1, 0, or -1. + # If %tailoring is false (i.e. empty), + # $Collator should do the default collation. + =head1 DESCRIPTION +This module is an implementation of Unicode Technical Standard #10 +(a.k.a. UTS #10) - Unicode Collation Algorithm (a.k.a. UCA). + =head2 Constructor and Tailoring The C method returns a collator object. $Collator = Unicode::Collate->new( UCA_Version => $UCA_Version, - alternate => $alternate, + alternate => $alternate, # deprecated: use of 'variable' is recommended. backwards => $levelNumber, # or \@levelNumbers entry => $element, - normalization => $normalization_form, + hangul_terminator => $term_primary_weight, ignoreName => qr/$ignoreName/, ignoreChar => qr/$ignoreChar/, katakana_before_hiragana => $bool, level => $collationLevel, + normalization => $normalization_form, overrideCJK => \&overrideCJK, overrideHangul => \&overrideHangul, preprocess => \&preprocess, @@ -708,51 +1161,39 @@ The C method returns a collator object. undefName => qr/$undefName/, undefChar => qr/$undefChar/, upper_before_lower => $bool, + variable => $variable, ); - # if %tailoring is false (i.e. empty), - # $Collator should do the default collation. =over 4 =item UCA_Version -If the version number of the older UCA is given, -the older behavior of that version is emulated on collating. +If the tracking version number of UCA is given, +behavior of that tracking version is emulated on collating. If omitted, the return value of C is used. +C should return the latest tracking version supported. -The supported version: 8 or 9. - -B - -=item alternate - --- see 3.2.2 Alternate Weighting, UTR #10. +The supported tracking version: 8, 9, 11, or 14. -This key allows to alternate weighting for variable collation elements, -which are marked with an ASTERISK in the table -(NOTE: Many punction marks and symbols are variable in F). - - alternate => 'blanked', 'non-ignorable', 'shifted', or 'shift-trimmed'. + UCA Unicode Standard DUCET (@version) + --------------------------------------------------- + 8 3.1 3.0.1 (3.0.1d9) + 9 3.1 with Corrigendum 3 3.1.1 (3.1.1) + 11 4.0 4.0.0 (4.0.0) + 14 4.1.0 4.1.0 (4.1.0) -These names are case-insensitive. -By default (if specification is omitted), 'shifted' is adopted. +Note: Recent UTS #10 renames "Tracking Version" to "Revision." - 'Blanked' Variable elements are ignorable at levels 1 through 3; - considered at the 4th level. - - 'Non-ignorable' Variable elements are not reset to ignorable. +=item alternate - 'Shifted' Variable elements are ignorable at levels 1 through 3 - their level 4 weight is replaced by the old level 1 weight. - Level 4 weight for Non-Variable elements is 0xFFFF. +-- see 3.2.2 Alternate Weighting, version 8 of UTS #10 - 'Shift-Trimmed' Same as 'shifted', but all FFFF's at the 4th level - are trimmed. +For backward compatibility, C (old name) can be used +as an alias for C. =item backwards --- see 3.1.2 French Accents, UTR #10. +-- see 3.1.2 French Accents, UTS #10. backwards => $levelNumber or \@levelNumbers @@ -761,32 +1202,113 @@ If omitted, forwards at all the levels. =item entry --- see 3.1 Linguistic Features; 3.2.1 File Format, UTR #10. +-- see 3.1 Linguistic Features; 3.2.1 File Format, UTS #10. + +If the same character (or a sequence of characters) exists +in the collation element table through C, +mapping to collation elements is overrided. +If it does not exist, the mapping is defined additionally. + + entry => <<'ENTRY', # for DUCET v4.0.0 (allkeys-4.0.0.txt) +0063 0068 ; [.0E6A.0020.0002.0063] # ch +0043 0068 ; [.0E6A.0020.0007.0043] # Ch +0043 0048 ; [.0E6A.0020.0008.0043] # CH +006C 006C ; [.0F4C.0020.0002.006C] # ll +004C 006C ; [.0F4C.0020.0007.004C] # Ll +004C 004C ; [.0F4C.0020.0008.004C] # LL +00F1 ; [.0F7B.0020.0002.00F1] # n-tilde +006E 0303 ; [.0F7B.0020.0002.00F1] # n-tilde +00D1 ; [.0F7B.0020.0008.00D1] # N-tilde +004E 0303 ; [.0F7B.0020.0008.00D1] # N-tilde +ENTRY + + entry => <<'ENTRY', # for DUCET v4.0.0 (allkeys-4.0.0.txt) +00E6 ; [.0E33.0020.0002.00E6][.0E8B.0020.0002.00E6] # ae ligature as +00C6 ; [.0E33.0020.0008.00C6][.0E8B.0020.0008.00C6] # AE ligature as +ENTRY + +B The code point in the UCA file format (before C<';'>) +B be a Unicode code point (defined as hexadecimal), +but not a native code point. +So C<0063> must always denote C, +but not a character of C<"\x63">. + +Weighting may vary depending on collation element table. +So ensure the weights defined in C will be consistent with +those in the collation element table loaded via C
. + +In DUCET v4.0.0, primary weight of C is C<0E60> +and that of C is C<0E6D>. So setting primary weight of C to C<0E6A> +(as a value between C<0E60> and C<0E6D>) +makes ordering as C CH E D>. +Exactly speaking DUCET already has some characters between C and C: +C (C) with primary weight C<0E64>, +C (C) with C<0E65>, +and C (C) with C<0E69>. +Then primary weight C<0E6A> for C makes C +ordered between C and C. + +=item hangul_terminator + +-- see 7.1.4 Trailing Weights, UTS #10. + +If a true value is given (non-zero but should be positive), +it will be added as a terminator primary weight to the end of +every standard Hangul syllable. Secondary and any higher weights +for terminator are set to zero. +If the value is false or C key does not exist, +insertion of terminator weights will not be performed. + +Boundaries of Hangul syllables are determined +according to conjoining Jamo behavior in F +and F. + +B +(1) For expansion mapping (Unicode character mapped +to a sequence of collation elements), a terminator will not be added +between collation elements, even if Hangul syllable boundary exists there. +Addition of terminator is restricted to the next position +to the last collation element. + +(2) Non-conjoining Hangul letters +(Compatibility Jamo, halfwidth Jamo, and enclosed letters) are not +automatically terminated with a terminator primary weight. +These characters may need terminator included in a collation element +table beforehand. -Overrides a default order or defines additional collation elements - - entry => <<'ENTRIES', # use the UCA file format -00E6 ; [.0861.0020.0002.00E6] [.08B1.0020.0002.00E6] # ligature as -0063 0068 ; [.0893.0020.0002.0063] # "ch" in traditional Spanish -0043 0068 ; [.0893.0020.0008.0043] # "Ch" in traditional Spanish -ENTRIES +=item ignoreChar =item ignoreName -=item ignoreChar +-- see 3.2.2 Variable Weighting, UTS #10. --- see Completely Ignorable, 3.2.2 Alternate Weighting, UTR #10. +Makes the entry in the table completely ignorable; +i.e. as if the weights were zero at all level. -Makes the entry in the table ignorable. -If a collation element is ignorable, -it is ignored as if the element had been deleted from there. +Through C, any character matching C +will be ignored. Through C, any character whose name +(given in the C
file as a comment) matches C +will be ignored. E.g. when 'a' and 'e' are ignorable, 'element' is equal to 'lament' (or 'lmnt'). +=item katakana_before_hiragana + +-- see 7.3.1 Tertiary Weight Table, UTS #10. + +By default, hiragana is before katakana. +If the parameter is made true, this is reversed. + +B: This parameter simplemindedly assumes that any hiragana/katakana +distinctions must occur in level 3, and their weights at level 3 must be +same as those mentioned in 7.3.1, UTS #10. +If you define your collation elements which violate this requirement, +this parameter does not work validly. + =item level --- see 4.3 Form a sort key for each string, UTR #10. +-- see 4.3 Form Sort Key, UTS #10. Set the maximum level. Any higher levels than the specified one are ignored. @@ -794,7 +1316,7 @@ Any higher levels than the specified one are ignored. Level 1: alphabetic ordering Level 2: diacritic ordering Level 3: case ordering - Level 4: tie-breaking (e.g. in the case when alternate is 'shifted') + Level 4: tie-breaking (e.g. in the case when variable is 'shifted') ex.level => 2, @@ -802,33 +1324,56 @@ If omitted, the maximum is the 4th. =item normalization --- see 4.1 Normalize each input string, UTR #10. +-- see 4.1 Normalize, UTS #10. If specified, strings are normalized before preparation of sort keys (the normalization is executed after preprocess). -As a form name, one of the following names must be used. +A form name C accepts will be applied +as C<$normalization_form>. +Acceptable names include C<'NFD'>, C<'NFC'>, C<'NFKD'>, and C<'NFKC'>. +See C for detail. +If omitted, C<'NFD'> is used. - 'C' or 'NFC' for Normalization Form C - 'D' or 'NFD' for Normalization Form D - 'KC' or 'NFKC' for Normalization Form KC - 'KD' or 'NFKD' for Normalization Form KD +C is performed after C (if defined). -If omitted, the string is put into Normalization Form D. +Furthermore, special values, C and C<"prenormalized">, can be used, +though they are not concerned with C. -If C is passed explicitly as the value for this key, +If C (not a string C<"undef">) is passed explicitly +as the value for this key, any normalization is not carried out (this may make tailoring easier -if any normalization is not desired). - -see B. +if any normalization is not desired). Under C<(normalization =E undef)>, +only contiguous contractions are resolved; +e.g. even if C (and C) is ordered after C, +C would be primary equal to C. +In this point, +C<(normalization =E undef, preprocess =E sub { NFD(shift) })> +B equivalent to C<(normalization =E 'NFD')>. + +In the case of C<(normalization =E "prenormalized")>, +any normalization is not performed, but +non-contiguous contractions with combining characters are performed. +Therefore +C<(normalization =E 'prenormalized', preprocess =E sub { NFD(shift) })> +B equivalent to C<(normalization =E 'NFD')>. +If source strings are finely prenormalized, +C<(normalization =E 'prenormalized')> may save time for normalization. + +Except C<(normalization =E undef)>, +B is required (see also B). =item overrideCJK --- see 7.1 Derived Collation Elements, UTR #10. +-- see 7.1 Derived Collation Elements, UTS #10. + +By default, CJK Unified Ideographs are ordered in Unicode codepoint order +but C (if C is 8 to 11, its range is +C; if C is 14, its range is C) +are lesser than C (its range is +C and C). -By default, mapping of CJK Unified Ideographs -uses the Unicode codepoint order. -But the mapping of CJK Unified Ideographs may be overrided. +Through C, ordering of CJK Unified Ideographs can be overrided. ex. CJK Unified Ideographs in the JIS code point order. @@ -850,30 +1395,31 @@ ex. ignores all CJK Unified Ideographs. If C is passed explicitly as the value for this key, weights for CJK Unified Ideographs are treated as undefined. But assignment of weight for CJK Unified Ideographs -in table or L is still valid. +in table or C is still valid. =item overrideHangul --- see 7.1 Derived Collation Elements, UTR #10. +-- see 7.1 Derived Collation Elements, UTS #10. -By default, Hangul Syllables are decomposed into Hangul Jamo. +By default, Hangul Syllables are decomposed into Hangul Jamo, +even if C<(normalization =E undef)>. But the mapping of Hangul Syllables may be overrided. -This tag works like L, so see there for examples. +This parameter works like C, so see there for examples. If you want to override the mapping of Hangul Syllables, -the Normalization Forms D and KD are not appropriate -(they will be decomposed before overriding). +NFD, NFKD, and FCD are not appropriate, +since they will decompose Hangul Syllables before overriding. If C is passed explicitly as the value for this key, weight for Hangul Syllables is treated as undefined without decomposition into Hangul Jamo. But definition of weight for Hangul Syllables -in table or L is still valid. +in table or C is still valid. =item preprocess --- see 5.1 Preprocessing, UTR #10. +-- see 5.1 Preprocessing, UTS #10. If specified, the coderef is used to preprocess before the formation of sort keys. @@ -884,36 +1430,44 @@ Then, "the pen" is before "a pencil". preprocess => sub { my $str = shift; $str =~ s/\b(?:an?|the)\s+//gi; - $str; + return $str; }, +C is performed before C (if defined). + =item rearrange --- see 3.1.3 Rearrangement, UTR #10. +-- see 3.1.3 Rearrangement, UTS #10. Characters that are not coded in logical order and to be rearranged. -By default, +If C is equal to or lesser than 11, default is: rearrange => [ 0x0E40..0x0E44, 0x0EC0..0x0EC4 ], -If you want to disallow any rearrangement, -pass C or C<[]> (a reference to an empty list) -as the value for this key. +If you want to disallow any rearrangement, pass C or C<[]> +(a reference to empty list) as the value for this key. + +If C is equal to 14, default is C<[]> (i.e. no rearrangement). B =item table --- see 3.2 Default Unicode Collation Element Table, UTR #10. +-- see 3.2 Default Unicode Collation Element Table, UTS #10. + +You can use another collation element table if desired. -You can use another element table if desired. -The table file must be in your C directory. +The table file should locate in the F directory +on C<@INC>. Say, if the filename is F, +the table file is searched as F in C<@INC>. -By default, the file C is used. +By default, F (as the filename of DUCET) is used. +If you will prepare your own table file, any name other than F +may be better to avoid namespace conflict. If C is passed explicitly as the value for this key, -no file is read (but you can define collation elements via L). +no file is read (but you can define collation elements via C). A typical way to define a collation element table without any file of table: @@ -930,11 +1484,14 @@ without any file of table: ENTRIES ); -=item undefName +If C or C is used, character names should be +specified as a comment (following C<#>) on each line. =item undefChar --- see 6.3.4 Reducing the Repertoire, UTR #10. +=item undefName + +-- see 6.3.4 Reducing the Repertoire, UTS #10. Undefines the collation element as if it were unassigned in the table. This reduces the size of the table. @@ -946,23 +1503,52 @@ as it is greater than any other assigned collation elements But, it'd be better to ignore characters unfamiliar to you and maybe never used. -=item katakana_before_hiragana +Through C, any character matching C +will be undefined. Through C, any character whose name +(given in the C
file as a comment) matches C +will be undefined. + +ex. Collation weights for beyond-BMP characters are not stored in object: + + undefChar => qr/[^\0-\x{fffd}]/, =item upper_before_lower --- see 6.6 Case Comparisons; 7.3.1 Tertiary Weight Table, UTR #10. +-- see 6.6 Case Comparisons, UTS #10. + +By default, lowercase is before uppercase. +If the parameter is made true, this is reversed. + +B: This parameter simplemindedly assumes that any lowercase/uppercase +distinctions must occur in level 3, and their weights at level 3 must be +same as those mentioned in 7.3.1, UTS #10. +If you define your collation elements which differs from this requirement, +this parameter doesn't work validly. + +=item variable + +-- see 3.2.2 Variable Weighting, UTS #10. + +This key allows to variable weighting for variable collation elements, +which are marked with an ASTERISK in the table +(NOTE: Many punction marks and symbols are variable in F). + + variable => 'blanked', 'non-ignorable', 'shifted', or 'shift-trimmed'. + +These names are case-insensitive. +By default (if specification is omitted), 'shifted' is adopted. + + 'Blanked' Variable elements are made ignorable at levels 1 through 3; + considered at the 4th level. -By default, lowercase is before uppercase -and hiragana is before katakana. + 'Non-Ignorable' Variable elements are not reset to ignorable. -If the tag is made true, this is reversed. + 'Shifted' Variable elements are made ignorable at levels 1 through 3 + their level 4 weight is replaced by the old level 1 weight. + Level 4 weight for Non-Variable elements is 0xFFFF. -B: These tags simplemindedly assume -any lowercase/uppercase or hiragana/katakana distinctions -should occur in level 3, and their weights at level 3 -should be same as those mentioned in 7.3.1, UTR #10. -If you define your collation elements which violates this, -these tags doesn't work validly. + 'Shift-Trimmed' Same as 'shifted', but all FFFF's at the 4th level + are trimmed. =back @@ -1003,7 +1589,7 @@ They works like the same name operators as theirs. =item C<$sortKey = $Collator-EgetSortKey($string)> --- see 4.3 Form a sort key for each string, UTR #10. +-- see 4.3 Form Sort Key, UTS #10. Returns a sort key. @@ -1018,6 +1604,9 @@ and get the result of the comparison of the strings using UCA. =item C<$sortKeyForm = $Collator-EviewSortKey($string)> +Converts a sorting key into its representation form. +If C is 8, the output is slightly different. + use Unicode::Collate; my $c = Unicode::Collate->new(); print $c->viewSortKey("Perl"),"\n"; @@ -1026,29 +1615,32 @@ and get the result of the comparison of the strings using UCA. # [0B67 0A65 0B7F 0B03 | 0020 0020 0020 0020 | 0008 0002 0002 0002 | FFFF FFFF FFFF FFFF] # Level 1 Level 2 Level 3 Level 4 - (If C is 8, the output is slightly different.) +=back + +=head2 Methods for Searching -=item C<$position = $Collator-Eindex($string, $substring)> +B If C or C parameter is true +for C<$Collator>, calling these methods (C, C, C, +C, C) is croaked, +as the position and the length might differ +from those on the specified string. +(And C and C parameters are neglected.) -=item C<($position, $length) = $Collator-Eindex($string, $substring)> +The C, C, C, C methods work +like C, C, C, C, respectively, +but they are not aware of any pattern, but only a literal substring. --- see 6.8 Searching, UTR #10. +=over 4 + +=item C<$position = $Collator-Eindex($string, $substring[, $position])> + +=item C<($position, $length) = $Collator-Eindex($string, $substring[, $position])> If C<$substring> matches a part of C<$string>, returns the position of the first occurrence of the matching part in scalar context; in list context, returns a two-element list of the position and the length of the matching part. -B that the length of the matching part may differ from -the length of C<$substring>. - -B that the position and the length are counted on the string -after the process of preprocess, normalization, and rearrangement. -Therefore, in case the specified string is not binary equal to -the preprocessed/normalized/rearranged string, the position and the length -may differ form those on the specified string. But it is guaranteed -that, if matched, it returns a non-negative value as C<$position>. - If C<$substring> does not match any part of C<$string>, returns C<-1> in scalar context and an empty list in list context. @@ -1056,15 +1648,86 @@ an empty list in list context. e.g. you say my $Collator = Unicode::Collate->new( normalization => undef, level => 1 ); - my $str = "Ich mu\x{00DF} studieren."; - my $sub = "m\x{00FC}ss"; + # (normalization => undef) is REQUIRED. + my $str = "Ich muß studieren Perl."; + my $sub = "MÜSS"; my $match; if (my($pos,$len) = $Collator->index($str, $sub)) { $match = substr($str, $pos, $len); } -and get C<"mu\x{00DF}"> in C<$match> since C<"mu>E<223>C<"> -is primary equal to C<"m>E<252>C. +and get C<"muß"> in C<$match> since C<"muß"> +is primary equal to C<"MÜSS">. + +=item C<$match_ref = $Collator-Ematch($string, $substring)> + +=item C<($match) = $Collator-Ematch($string, $substring)> + +If C<$substring> matches a part of C<$string>, in scalar context, returns +B the first occurrence of the matching part +(C<$match_ref> is always true if matches, +since every reference is B); +in list context, returns the first occurrence of the matching part. + +If C<$substring> does not match any part of C<$string>, +returns C in scalar context and +an empty list in list context. + +e.g. + + if ($match_ref = $Collator->match($str, $sub)) { # scalar context + print "matches [$$match_ref].\n"; + } else { + print "doesn't match.\n"; + } + + or + + if (($match) = $Collator->match($str, $sub)) { # list context + print "matches [$match].\n"; + } else { + print "doesn't match.\n"; + } + +=item C<@match = $Collator-Egmatch($string, $substring)> + +If C<$substring> matches a part of C<$string>, returns +all the matching parts (or matching count in scalar context). + +If C<$substring> does not match any part of C<$string>, +returns an empty list. + +=item C<$count = $Collator-Esubst($string, $substring, $replacement)> + +If C<$substring> matches a part of C<$string>, +the first occurrence of the matching part is replaced by C<$replacement> +(C<$string> is modified) and return C<$count> (always equals to C<1>). + +C<$replacement> can be a C, +taking the matching part as an argument, +and returning a string to replace the matching part +(a bit similar to C($1)/e>). + +=item C<$count = $Collator-Egsubst($string, $substring, $replacement)> + +If C<$substring> matches a part of C<$string>, +all the occurrences of the matching part is replaced by C<$replacement> +(C<$string> is modified) and return C<$count>. + +C<$replacement> can be a C, +taking the matching part as an argument, +and returning a string to replace the matching part +(a bit similar to C($1)/eg>). + +e.g. + + my $Collator = Unicode::Collate->new( normalization => undef, level => 1 ); + # (normalization => undef) is REQUIRED. + my $str = "Camel donkey zebra came\x{301}l CAMEL horse cAm\0E\0L..."; + $Collator->gsubst($str, "camel", sub { "$_[0]" }); + + # now $str is "Camel donkey zebra came\x{301}l CAMEL horse cAm\0E\0L..."; + # i.e., all the camels are made bold-faced. =back @@ -1100,104 +1763,114 @@ In the scalar context, returns the modified collator $Collator->change(level => 4)->eq("perl", "PERL"); # false -=item UCA_Version +=item C<$version = $Collator-Eversion()> + +Returns the version number (a string) of the Unicode Standard +which the C
file used by the collator object is based on. +If the table does not include a version line (starting with C<@version>), +returns C<"unknown">. + +=item C -Returns the version number of Unicode Technical Standard 10 -this module consults. +Returns the tracking version number of UTS #10 this module consults. -=item Base_Unicode_Version +=item C -Returns the version number of the Unicode Standard -this module is based on. +Returns the version number of UTS #10 this module consults. =back -=head2 EXPORT +=head1 EXPORT -None by default. +No method will be exported. -=head2 TODO +=head1 INSTALL -Unicode::Collate has not been ported to EBCDIC. The code mostly would -work just fine but a decision needs to be made: how the module should -work in EBCDIC? Should the low 256 characters be understood as -Unicode or as EBCDIC code points? Should one be chosen or should -there be a way to do either? Or should such translation be left -outside the module for the user to do, for example by using -Encode::from_to()? -(or utf8::unicode_to_native()/utf8::native_to_unicode()?) +Though this module can be used without any C
file, +to use this module easily, it is recommended to install a table file +in the UCA format, by copying it under the directory +/Unicode/Collate. -=head2 CAVEAT +The most preferable one is "The Default Unicode Collation Element Table" +(aka DUCET), available from the Unicode Consortium's website: -Use of the C parameter requires -the B module. + http://www.unicode.org/Public/UCA/ + + http://www.unicode.org/Public/UCA/latest/allkeys.txt (latest version) + +If DUCET is not installed, it is recommended to copy the file +from http://www.unicode.org/Public/UCA/latest/allkeys.txt +to /Unicode/Collate/allkeys.txt +manually. + +=head1 CAVEATS + +=over 4 + +=item Normalization + +Use of the C parameter requires the B +module (see L). If you need not it (say, in the case when you need not handle any combining characters), assign C undef> explicitly. --- see 6.5 Avoiding Normalization, UTR #10. +-- see 6.5 Avoiding Normalization, UTS #10. -=head2 Conformance Test +=item Conformance Test -The Conformance Test for the UCA is provided -in L -and L +The Conformance Test for the UCA is available +under L. For F, a collator via Cnew( )> should be used; for F, a collator via -Cnew(alternate =E "non-ignorable", level =E 3)>. - -B - -=head2 BUGS +Cnew(variable =E "non-ignorable", level =E 3)>. -C is an experimental method and -its return value may be unreliable. -The correct implementation for C must be based -on Locale-Sensitive Support: Level 3 in UTR #18, -F. +B -See also 4.2 Locale-Dependent Graphemes in UTR #18. - -=head1 AUTHOR +=back -SADAHIRO Tomoyuki, ESADAHIRO@cpan.orgE +=head1 AUTHOR, COPYRIGHT AND LICENSE - http://homepage1.nifty.com/nomenclator/perl/ +The Unicode::Collate module for perl was written by SADAHIRO Tomoyuki, +. This module is Copyright(C) 2001-2005, +SADAHIRO Tomoyuki. Japan. All rights reserved. - Copyright(C) 2001-2002, SADAHIRO Tomoyuki. Japan. All rights reserved. +This module is free software; you can redistribute it and/or +modify it under the same terms as Perl itself. - This library is free software; you can redistribute it - and/or modify it under the same terms as Perl itself. +The file Unicode/Collate/allkeys.txt was copied directly +from L. +This file is Copyright (c) 1991-2005 Unicode, Inc. All rights reserved. +Distributed under the Terms of Use in L. =head1 SEE ALSO =over 4 -=item http://www.unicode.org/reports/tr10/ +=item Unicode Collation Algorithm - UTS #10 -Unicode Collation Algorithm - UTR #10 +L -=item http://www.unicode.org/reports/tr10/allkeys.txt +=item The Default Unicode Collation Element Table (DUCET) -The Default Unicode Collation Element Table +L -=item http://www.unicode.org/reports/tr10/CollationTest.html -http://www.unicode.org/reports/tr10/CollationTest.zip +=item The conformance test for the UCA -The latest versions of the conformance test for the UCA +L -=item http://www.unicode.org/reports/tr15/ +L -Unicode Normalization Forms - UAX #15 +=item Hangul Syllable Type -=item http://www.unicode.org/reports/tr18 +L -Unicode Regular Expression Guidelines - UTR #18 +=item Unicode Normalization Forms - UAX #15 -=item L +L =back