X-Git-Url: https://git.saurik.com/apple/icu.git/blobdiff_plain/46f4442e9a5a4f3b98b7c1083586332f6a8a99a4..151279e3792e85d0417b499c229886b3af724f55:/icuSources/data/translit/Greek_Latin.txt diff --git a/icuSources/data/translit/Greek_Latin.txt b/icuSources/data/translit/Greek_Latin.txt index 53a25b76..9f626cee 100644 --- a/icuSources/data/translit/Greek_Latin.txt +++ b/icuSources/data/translit/Greek_Latin.txt @@ -1,6 +1,6 @@ # *************************************************************************** # * -# * Copyright (C) 2004-2008, International Business Machines +# * Copyright (C) 2004-2013, International Business Machines # * Corporation; Unicode, Inc.; and others. All Rights Reserved. # * # *************************************************************************** @@ -37,171 +37,171 @@ $beforeLetter = [[:M:]\']* [:L:] ; $beforeLower = $accent * $lower ; $notLetter = [^[:L:][:M:]] ; $under = \u0331; -\: <> \: $under ; -\? <> \? $under ; -\; <> \? ; -· <> \: ; -\u0342 <> \u0302 ; +\: ↔ \: $under ; +\? ↔ \? $under ; +\; ↔ \? ; +· ↔ \: ; +\u0342 ↔ \u0302 ; $accent_minus = [[$accent]-[$iotasub$macron]]; -Α } $accent_minus * $iotasub > | Α $macron ; -α } $accent_minus * $iotasub > | α $macron ; -$upper $accent * { $iotasub > I ; -$iotasub > i ; -| $1 $iotasub < ($evowel $macron $accentMinus *) i ; -| $1 $iotasub < ($evowel $macron $accentMinus *) I ; -Α ($macron?) $rough } $beforeLower > H | α $1; -Ε $rough } $beforeLower > H | ε; -Η $rough } $beforeLower > H | η ; -Ι ($ddot?) $rough } $beforeLower > H | ι $1; -Ο $rough } $beforeLower > H | ο ; -Υ $rough } $beforeLower > H | υ ; -Ω ($ddot?) $rough } $beforeLower > H | ω $1; -Α ($glower $macron?) $rough > H | α $1 ; -Ε ($glower) $rough > H | ε $1 ; -Η ($glower) $rough > H | η $1 ; -Ι ($glower $ddot?) $rough > H | ι $1 ; -Ο ($glower) $rough > H | ο $1 ; -Υ ($glower) $rough > H | υ $1 ; -Ω ($glower $ddot?) $rough > H | ω $1 ; -($lcgvowel + $ddotmac? ) $rough > h | $1 ; -($gvowel + $ddotmac? ) $rough > H | $1 ; -| $1 $rough < h ($evowel $macron $ddot? $evowel2_i $macron?) ; -| $1 $rough < h ($evowel $ddot? $evowel2 $macron?) ; -| $1 $rough < h ($evowel $macron? $ddot?) ; -| $1 $rough < H ([AEIOUY] $macron $ddot? $evowel2_i $macron?) ; -| $1 $rough < H ([AEIOUY] $ddot? $evowel2 $macron?) ; -| $1 $rough < H ([AEIOUY] $macron? $ddot?) ; -| A $1 $rough < H a ($macron $ddot? $evowel2_i $macron?) ; -| E $1 $rough < H e ($macron $ddot? $evowel2_i $macron?) ; -| I $1 $rough < H i ($macron $ddot? $evowel2_i $macron?) ; -| O $1 $rough < H o ($macron $ddot? $evowel2_i $macron?) ; -| U $1 $rough < H u ($macron $ddot? $evowel2_i $macron?) ; -| Y $1 $rough < H y ($macron $ddot? $evowel2_i $macron?) ; -| A $1 $rough < H a ($ddot? $evowel2 $macron?) ; -| E $1 $rough < H e ($ddot? $evowel2 $macron?) ; -| I $1 $rough < H i ($ddot? $evowel2 $macron?) ; -| O $1 $rough < H o ($ddot? $evowel2 $macron?) ; -| U $1 $rough < H u ($ddot? $evowel2 $macron?) ; -| Y $1 $rough < H y ($ddot? $evowel2 $macron?) ; -| A $1 $rough < H a ($macron? $ddot? ) ; -| E $1 $rough < H e ($macron? $ddot? ) ; -| I $1 $rough < H i ($macron? $ddot? ) ; -| O $1 $rough < H o ($macron? $ddot? ) ; -| U $1 $rough < H u ($macron? $ddot? ) ; -| Y $1 $rough < H y ($macron? $ddot? ) ; -$smooth > ; -| $1 $smooth < $notLetter { ([rR]) } [^hH$smooth$rough] ; -| $1 $smooth < $notLetter { ($evowel $macron? $evowel2 $macron?) } [^$smooth$rough] ; -| $1 $smooth < $notLetter { ($evowel $macron?) } [^$evowel2$smooth$rough] ; -Α < A $macron ; -α < a $macron ; -η <> e $macron ; -Η <> E $macron ; -φ <> ph ; -Ψ } $beforeLower <> Ps ; -Ψ <> PS ; -Φ } $beforeLower <> Ph ; -Φ <> PH ; -ψ <> ps ; -ω <> o $macron ; -Ω <> O $macron; -α <> a ; -Α <> A ; -β <> b ; -Β <> B ; -γ } $gammaLike <> n } $egammaLike ; -γ <> g ; -Γ } $gammaLike <> N } $egammaLike ; -Γ <> G ; -δ <> d ; -Δ <> D ; -ε <> e ; -Ε <> E ; -ζ <> z ; -Ζ <> Z ; -θ <> th ; -Θ } $beforeLower <> Th ; -Θ <> TH ; -ι <> i ; -Ι <> I ; -κ <> k ; -Κ <> K ; -λ <> l ; -Λ <> L ; -μ <> m ; -Μ <> M ; -ν } $gammaLike > n\' ; -ν <> n ; -Ν } $gammaLike <> N\' ; -Ν <> N ; -ξ <> x ; -Ξ <> X ; -ο <> o ; -Ο <> O ; -π <> p ; -Π <> P ; -ρ $rough <> rh; -Ρ $rough } $beforeLower <> Rh ; -Ρ $rough <> RH ; -ρ <> r ; -Ρ <> R ; -[Pp] { } [ςσΣϷϸϺϻ] > \' ; -Ϸ <> S\u030C ; # Ϸ GREEK CAPITAL LETTER SHO Uppercase_Letter Grek - L -ϸ <> s\u030C ; #ϸ GREEK SMALL LETTER SHO Lowercase_Letter Grek - L -Ϻ <> S\u0302 ; # Ϻ GREEK CAPITAL LETTER SAN Uppercase_Letter Grek - L -ϻ <> s\u0302 ; # ϻ GREEK SMALL LETTER SAN Lowercase_Letter Grek - L -ς } $beforeLetter <> s $underbar } $beforeLetter; -σ } $beforeLetter <> s } $beforeLetter; -$afterLetter { σ <> $afterLetter { s $underbar; -$afterLetter { ς <> $afterLetter { s ; -ς <> s $underbar; -σ <> s ; -Σ <> S ; -τ <> t ; -Τ <> T ; -$vowel {υ } <> u ; -υ <> y ; -$vowel { Υ <> U ; -Υ <> Y ; -χ <> ch ; -Χ } $beforeLower <> Ch ; -Χ <> CH ; +Α } $accent_minus * $iotasub → | Α $macron ; +α } $accent_minus * $iotasub → | α $macron ; +$upper $accent * { $iotasub → I ; +$iotasub → i ; +| $1 $iotasub ← ($evowel $macron $accentMinus *) i ; +| $1 $iotasub ← ($evowel $macron $accentMinus *) I ; +Α ($macron?) $rough } $beforeLower → H | α $1; +Ε $rough } $beforeLower → H | ε; +Η $rough } $beforeLower → H | η ; +Ι ($ddot?) $rough } $beforeLower → H | ι $1; +Ο $rough } $beforeLower → H | ο ; +Υ $rough } $beforeLower → H | υ ; +Ω ($ddot?) $rough } $beforeLower → H | ω $1; +Α ($glower $macron?) $rough → H | α $1 ; +Ε ($glower) $rough → H | ε $1 ; +Η ($glower) $rough → H | η $1 ; +Ι ($glower $ddot?) $rough → H | ι $1 ; +Ο ($glower) $rough → H | ο $1 ; +Υ ($glower) $rough → H | υ $1 ; +Ω ($glower $ddot?) $rough → H | ω $1 ; +($lcgvowel + $ddotmac? ) $rough → h | $1 ; +($gvowel + $ddotmac? ) $rough → H | $1 ; +| $1 $rough ← h ($evowel $macron $ddot? $evowel2_i $macron?) ; +| $1 $rough ← h ($evowel $ddot? $evowel2 $macron?) ; +| $1 $rough ← h ($evowel $macron? $ddot?) ; +| $1 $rough ← H ([AEIOUY] $macron $ddot? $evowel2_i $macron?) ; +| $1 $rough ← H ([AEIOUY] $ddot? $evowel2 $macron?) ; +| $1 $rough ← H ([AEIOUY] $macron? $ddot?) ; +| A $1 $rough ← H a ($macron $ddot? $evowel2_i $macron?) ; +| E $1 $rough ← H e ($macron $ddot? $evowel2_i $macron?) ; +| I $1 $rough ← H i ($macron $ddot? $evowel2_i $macron?) ; +| O $1 $rough ← H o ($macron $ddot? $evowel2_i $macron?) ; +| U $1 $rough ← H u ($macron $ddot? $evowel2_i $macron?) ; +| Y $1 $rough ← H y ($macron $ddot? $evowel2_i $macron?) ; +| A $1 $rough ← H a ($ddot? $evowel2 $macron?) ; +| E $1 $rough ← H e ($ddot? $evowel2 $macron?) ; +| I $1 $rough ← H i ($ddot? $evowel2 $macron?) ; +| O $1 $rough ← H o ($ddot? $evowel2 $macron?) ; +| U $1 $rough ← H u ($ddot? $evowel2 $macron?) ; +| Y $1 $rough ← H y ($ddot? $evowel2 $macron?) ; +| A $1 $rough ← H a ($macron? $ddot? ) ; +| E $1 $rough ← H e ($macron? $ddot? ) ; +| I $1 $rough ← H i ($macron? $ddot? ) ; +| O $1 $rough ← H o ($macron? $ddot? ) ; +| U $1 $rough ← H u ($macron? $ddot? ) ; +| Y $1 $rough ← H y ($macron? $ddot? ) ; +$smooth → ; +| $1 $smooth ← $notLetter { ([rR]) } [^hH$smooth$rough] ; +| $1 $smooth ← $notLetter { ($evowel $macron? $evowel2 $macron?) } [^$smooth$rough] ; +| $1 $smooth ← $notLetter { ($evowel $macron?) } [^$evowel2$smooth$rough] ; +Α ← A $macron ; +α ← a $macron ; +η ↔ e $macron ; +Η ↔ E $macron ; +φ ↔ ph ; +Ψ } $beforeLower ↔ Ps ; +Ψ ↔ PS ; +Φ } $beforeLower ↔ Ph ; +Φ ↔ PH ; +ψ ↔ ps ; +ω ↔ o $macron ; +Ω ↔ O $macron; +α ↔ a ; +Α ↔ A ; +β ↔ b ; +Β ↔ B ; +γ } $gammaLike ↔ n } $egammaLike ; +γ ↔ g ; +Γ } $gammaLike ↔ N } $egammaLike ; +Γ ↔ G ; +δ ↔ d ; +Δ ↔ D ; +ε ↔ e ; +Ε ↔ E ; +ζ ↔ z ; +Ζ ↔ Z ; +θ ↔ th ; +Θ } $beforeLower ↔ Th ; +Θ ↔ TH ; +ι ↔ i ; +Ι ↔ I ; +κ ↔ k ; +Κ ↔ K ; +λ ↔ l ; +Λ ↔ L ; +μ ↔ m ; +Μ ↔ M ; +ν } $gammaLike → n\' ; +ν ↔ n ; +Ν } $gammaLike ↔ N\' ; +Ν ↔ N ; +ξ ↔ x ; +Ξ ↔ X ; +ο ↔ o ; +Ο ↔ O ; +π ↔ p ; +Π ↔ P ; +ρ $rough ↔ rh; +Ρ $rough } $beforeLower ↔ Rh ; +Ρ $rough ↔ RH ; +ρ ↔ r ; +Ρ ↔ R ; +[Pp] { } [ςσΣϷϸϺϻ] → \' ; +Ϸ ↔ S\u030C ; # Ϸ GREEK CAPITAL LETTER SHO Uppercase_Letter Grek - L +ϸ ↔ s\u030C ; #ϸ GREEK SMALL LETTER SHO Lowercase_Letter Grek - L +Ϻ ↔ S\u0302 ; # Ϻ GREEK CAPITAL LETTER SAN Uppercase_Letter Grek - L +ϻ ↔ s\u0302 ; # ϻ GREEK SMALL LETTER SAN Lowercase_Letter Grek - L +ς } $beforeLetter ↔ s $underbar } $beforeLetter; +σ } $beforeLetter ↔ s } $beforeLetter; +$afterLetter { σ ↔ $afterLetter { s $underbar; +$afterLetter { ς ↔ $afterLetter { s ; +ς ↔ s $underbar; +σ ↔ s ; +Σ ↔ S ; +τ ↔ t ; +Τ ↔ T ; +$vowel {υ } ↔ u ; +υ ↔ y ; +$vowel { Υ ↔ U ; +Υ ↔ Y ; +χ ↔ ch ; +Χ } $beforeLower ↔ Ch ; +Χ ↔ CH ; $ignore = [[:Mark:]''] * ; -| k < c ; -| ph < f ; -| i < j ; -| k < q ; -| b < v } $vowel ; -| b < w } $vowel; -| u < v ; -| u < w; -| K < C ; -| Ph < F ; -| I < J ; -| K < Q ; -| B < V } $vowel ; -| B < W } $vowel ; -| U < V ; -| U < W ; -$rough } $ignore [:UppercaseLetter:] > H ; -$ignore [:UppercaseLetter:] { $rough > H ; -$rough < H ; -$rough <> h ; -ϐ > | β ; -ϑ > | θ ; -ϒ > | Υ ; -ϕ > | φ ; -ϖ > | π ; -ϰ > | κ ; -ϱ > | ρ ; -ϲ > | σ ; -Ϲ > | Σ; #U+03F9 GREEK CAPITAL LUNATE SIGMA SYMBOL -ϳ > j ; -ϴ > | Θ ; -ϵ > | ε ; -µ > | μ ; -ͺ > i; -< [Ππ] { \' } [Ss] ; -< [Νν] { \' } $egammaLike ; +| k ← c ; +| ph ← f ; +| i ← j ; +| k ← q ; +| b ← v } $vowel ; +| b ← w } $vowel; +| u ← v ; +| u ← w; +| K ← C ; +| Ph ← F ; +| I ← J ; +| K ← Q ; +| B ← V } $vowel ; +| B ← W } $vowel ; +| U ← V ; +| U ← W ; +$rough } $ignore [:UppercaseLetter:] → H ; +$ignore [:UppercaseLetter:] { $rough → H ; +$rough ← H ; +$rough ↔ h ; +ϐ → | β ; +ϑ → | θ ; +ϒ → | Υ ; +ϕ → | φ ; +ϖ → | π ; +ϰ → | κ ; +ϱ → | ρ ; +ϲ → | σ ; +Ϲ → | Σ; #U+03F9 GREEK CAPITAL LUNATE SIGMA SYMBOL +ϳ → j ; +ϴ → | Θ ; +ϵ → | ε ; +µ → | μ ; +ͺ → i; +← [Ππ] { \' } [Ss] ; +← [Νν] { \' } $egammaLike ; ::NFC (NFD) ; :: ( [':?A-Za-zÀ-ÅÇ-ÏÑ-ÖÙ-Ýà-åç-ïñ-öù-ýÿ-ďĒ-ĥĨ-İĴ-ķĹ-ľŃ-ňŌ-őŔ-ťŨ-žƠ-ơƯ-ưǍ-ǜǞ-ǣǦ-ǰǴ-ǵǸ-țȞ-ȟȦ-ȳ\u0300-\u0337\u0339-\u0345΅-ΆΈ-ΊΌΎ-ΐΪ-ΰϊ-ώϓ-ϔЀ-ЁЃЇЌ-ЎЙйѐ-ёѓїќ-ўѶ-ѷӁ-ӂӐ-ӓӖ-ӗӚ-ӟӢ-ӧӪ-ӵӸ-ӹḀ-ẙẛẠ-ỹἀ-ἕἘ-Ἕἠ-ὅὈ-Ὅὐ-ὗὙὛὝὟ-ώᾀ-ᾴᾶ-ᾼ῁-ῄῆ-ΐῖ-Ί῝-΅ῲ-ῴῶ-ῼK-Å] ) ;