[Ada] AI12-0004 Normalization and allowed characters

gcc/ada/

	* scng.adb (Scan): Detect wide characters not in NFKC.
	* libgnat/a-chahan.adb, libgnat/a-chahan.ads,
	libgnat/a-wichha.adb, libgnat/a-wichha.ads,
	libgnat/a-wichun.adb, libgnat/a-wichun.ads,
	libgnat/a-zchhan.adb, libgnat/a-zchhan.ads,
	libgnat/a-zchuni.adb, libgnat/a-zchuni.ads (Is_NFKC): New.
	* libgnat/s-utf_32.ads, libgnat/s-utf_32.adb (Is_UTF_32_NFKC):
	New.

diff --git a/gcc/ada/libgnat/a-chahan.adb b/gcc/ada/libgnat/a-chahan.adb
index faee41b..de66846 100644 (file)
--- a/gcc/ada/libgnat/a-chahan.adb
+++ b/gcc/ada/libgnat/a-chahan.adb
@@ -399,6 +399,17 @@ package body Ada.Characters.Handling is
       return False;
    end Is_Mark;
 
+   -------------
+   -- Is_NFKC --
+   -------------
+
+   function Is_NFKC (Item : Character) return Boolean is
+   begin
+      return Character'Pos (Item) not in
+        160 | 168 | 170 | 175 | 178 | 179 | 180 | 181 | 184 | 185 | 186 |
+        188 | 189 | 190;
+   end Is_NFKC;
+
    ---------------------
    -- Is_Other_Format --
    ---------------------

diff --git a/gcc/ada/libgnat/a-chahan.ads b/gcc/ada/libgnat/a-chahan.ads
index 957d623..04f975c 100644 (file)
--- a/gcc/ada/libgnat/a-chahan.ads
+++ b/gcc/ada/libgnat/a-chahan.ads
@@ -58,6 +58,7 @@ package Ada.Characters.Handling is
    function Is_Other_Format          (Item : Character) return Boolean;
    function Is_Punctuation_Connector (Item : Character) return Boolean;
    function Is_Space                 (Item : Character) return Boolean;
+   function Is_NFKC                  (Item : Character) return Boolean;
 
    ---------------------------------------------------
    -- Conversion Functions for Character and String --

diff --git a/gcc/ada/libgnat/a-wichha.adb b/gcc/ada/libgnat/a-wichha.adb
index 7531ef6..feccc23 100644 (file)
--- a/gcc/ada/libgnat/a-wichha.adb
+++ b/gcc/ada/libgnat/a-wichha.adb
@@ -124,6 +124,13 @@ package body Ada.Wide_Characters.Handling is
    function Is_Mark (Item : Wide_Character) return Boolean
      renames Ada.Wide_Characters.Unicode.Is_Mark;
 
+   -------------
+   -- Is_NFKC --
+   -------------
+
+   function Is_NFKC (Item : Wide_Character) return Boolean
+     renames Ada.Wide_Characters.Unicode.Is_NFKC;
+
    ---------------------
    -- Is_Other_Format --
    ---------------------

diff --git a/gcc/ada/libgnat/a-wichha.ads b/gcc/ada/libgnat/a-wichha.ads
index bb9452f..23eb468 100644 (file)
--- a/gcc/ada/libgnat/a-wichha.ads
+++ b/gcc/ada/libgnat/a-wichha.ads
@@ -101,6 +101,12 @@ package Ada.Wide_Characters.Handling is
    --  Returns True if the Wide_Character designated by Item is categorized as
    --  separator_space, otherwise returns False.
 
+   function Is_NFKC (Item : Wide_Character) return Boolean;
+   pragma Inline (Is_NFKC);
+   --  Returns True if the Wide_Character designated by Item could be present
+   --  in a string normalized to Normalization Form KC (as defined by Clause
+   --  21 of ISO/IEC 10646:2017), otherwise returns False.
+
    function Is_Graphic (Item : Wide_Character) return Boolean;
    pragma Inline (Is_Graphic);
    --  Returns True if the Wide_Character designated by Item is categorized as

diff --git a/gcc/ada/libgnat/a-wichun.adb b/gcc/ada/libgnat/a-wichun.adb
index cfd84da..09cbad2 100644 (file)
--- a/gcc/ada/libgnat/a-wichun.adb
+++ b/gcc/ada/libgnat/a-wichun.adb
@@ -116,6 +116,15 @@ package body Ada.Wide_Characters.Unicode is
       return G.Is_UTF_32_Non_Graphic (G.Category (C));
    end Is_Non_Graphic;
 
+   -------------
+   -- Is_NFKC --
+   -------------
+
+   function Is_NFKC (U : Wide_Character) return Boolean is
+   begin
+      return G.Is_UTF_32_NFKC (Wide_Character'Pos (U));
+   end Is_NFKC;
+
    --------------
    -- Is_Other --
    --------------

diff --git a/gcc/ada/libgnat/a-wichun.ads b/gcc/ada/libgnat/a-wichun.ads
index c9eb938..9e42749 100644 (file)
--- a/gcc/ada/libgnat/a-wichun.ads
+++ b/gcc/ada/libgnat/a-wichun.ads
@@ -131,7 +131,7 @@ package Ada.Wide_Characters.Unicode is
    pragma Inline (Is_Other);
    --  Returns true iff U is an other format character, which means that it
    --  can be used to extend an identifier, but is ignored for the purposes of
-   --  matching of identiers, or if C is one of the corresponding categories,
+   --  matching of identifiers, or if C is one of the corresponding categories,
    --  which are the following:
    --    Other, Format (Cf)
 
@@ -150,6 +150,12 @@ package Ada.Wide_Characters.Unicode is
    --  of the corresponding categories, which are the following:
    --    Separator, Space (Zs)
 
+   function Is_NFKC (U : Wide_Character) return Boolean;
+   pragma Inline (Is_NFKC);
+   --  Returns True if the Wide_Character designated by U could be present
+   --  in a string normalized to Normalization Form KC (as defined by Clause
+   --  21 of ISO/IEC 10646:2017), otherwise returns False.
+
    function Is_Non_Graphic (U : Wide_Character) return Boolean;
    function Is_Non_Graphic (C : Category)       return Boolean;
    pragma Inline (Is_Non_Graphic);

diff --git a/gcc/ada/libgnat/a-zchhan.adb b/gcc/ada/libgnat/a-zchhan.adb
index 4fd7eba..6930121 100644 (file)
--- a/gcc/ada/libgnat/a-zchhan.adb
+++ b/gcc/ada/libgnat/a-zchhan.adb
@@ -108,6 +108,13 @@ package body Ada.Wide_Wide_Characters.Handling is
    function Is_Mark (Item : Wide_Wide_Character) return Boolean
      renames Ada.Wide_Wide_Characters.Unicode.Is_Mark;
 
+   -------------
+   -- Is_NFKC --
+   -------------
+
+   function Is_NFKC (Item : Wide_Wide_Character) return Boolean
+     renames Ada.Wide_Wide_Characters.Unicode.Is_NFKC;
+
    ---------------------
    -- Is_Other_Format --
    ---------------------

diff --git a/gcc/ada/libgnat/a-zchhan.ads b/gcc/ada/libgnat/a-zchhan.ads
index 354452b..74fab2a 100644 (file)
--- a/gcc/ada/libgnat/a-zchhan.ads
+++ b/gcc/ada/libgnat/a-zchhan.ads
@@ -98,6 +98,12 @@ package Ada.Wide_Wide_Characters.Handling is
    --  Returns True if the Wide_Wide_Character designated by Item is
    --  categorized as separator_space, otherwise returns false.
 
+   function Is_NFKC (Item : Wide_Wide_Character) return Boolean;
+   pragma Inline (Is_NFKC);
+   --  Returns True if the Wide_Wide_Character designated by Item could be
+   --  present in a string normalized to Normalization Form KC (as defined by
+   --  Clause 21 of ISO/IEC 10646:2017), otherwise returns False.
+
    function Is_Graphic (Item : Wide_Wide_Character) return Boolean;
    pragma Inline (Is_Graphic);
    --  Returns True if the Wide_Wide_Character designated by Item is

diff --git a/gcc/ada/libgnat/a-zchuni.adb b/gcc/ada/libgnat/a-zchuni.adb
index b754af9..203c3aa 100644 (file)
--- a/gcc/ada/libgnat/a-zchuni.adb
+++ b/gcc/ada/libgnat/a-zchuni.adb
@@ -107,6 +107,15 @@ package body Ada.Wide_Wide_Characters.Unicode is
       return G.Is_UTF_32_Non_Graphic (G.Category (C));
    end Is_Non_Graphic;
 
+   -------------
+   -- Is_NFKC --
+   -------------
+
+   function Is_NFKC (U : Wide_Wide_Character) return Boolean is
+   begin
+      return G.Is_UTF_32_NFKC (Wide_Wide_Character'Pos (U));
+   end Is_NFKC;
+
    --------------
    -- Is_Other --
    --------------

diff --git a/gcc/ada/libgnat/a-zchuni.ads b/gcc/ada/libgnat/a-zchuni.ads
index 162d18d..7f4a30b 100644 (file)
--- a/gcc/ada/libgnat/a-zchuni.ads
+++ b/gcc/ada/libgnat/a-zchuni.ads
@@ -147,6 +147,12 @@ package Ada.Wide_Wide_Characters.Unicode is
    --  of the corresponding categories, which are the following:
    --    Separator, Space (Zs)
 
+   function Is_NFKC (U : Wide_Wide_Character) return Boolean;
+   pragma Inline (Is_NFKC);
+   --  Returns True if the Wide_Wide_Character designated by U could be present
+   --  in a string normalized to Normalization Form KC (as defined by Clause
+   --  21 of ISO/IEC 10646:2017), otherwise returns False.
+
    function Is_Non_Graphic (U : Wide_Wide_Character) return Boolean;
    function Is_Non_Graphic (C : Category)            return Boolean;
    pragma Inline (Is_Non_Graphic);

diff --git a/gcc/ada/libgnat/s-utf_32.adb b/gcc/ada/libgnat/s-utf_32.adb
index a722d62..a1346f3 100644 (file)
--- a/gcc/ada/libgnat/s-utf_32.adb
+++ b/gcc/ada/libgnat/s-utf_32.adb
@@ -49,7 +49,7 @@ package body System.UTF_32 is
    ----------------------
 
    --  Note these tables are derived from those given in AI-285. For details
-   --  see //www.ada-auth.org/cgi-bin/cvsweb.cgi/AIs/AI-00285.TXT?rev=1.22.
+   --  see www.ada-auth.org/cgi-bin/cvsweb.cgi/AIs/AI-00285.TXT?rev=1.22.
 
    type UTF_32_Range is record
       Lo : UTF_32;
@@ -6071,9 +6071,6 @@ package body System.UTF_32 is
         40,  -- DESERET CAPITAL LETTER LONG I .. DESERET CAPITAL LETTER EW
         32); -- TAG LATIN CAPITAL LETTER A .. TAG LATIN CAPITAL LETTER Z
 
-   pragma Warnings (On);
-   --  Temporary until pragma Warnings at start can be activated ???
-
    --  The following is a list of the 10646 names for CAPITAL LETTER entries
    --  that have no matching SMALL LETTER entry and are thus not folded
 
@@ -6117,6 +6114,403 @@ package body System.UTF_32 is
    --  GREEK CAPITAL LETTER ETA WITH PROSGEGRAMMENI
    --  GREEK CAPITAL LETTER OMEGA WITH PROSGEGRAMMENI
 
+   --  The following array includes all characters in the Unicode table with
+   --  the category NFKC_Quick_Check=No, taken from
+   --  www.unicode.org/Public/UCD/latest/ucd/DerivedNormalizationProps.txt
+
+   UTF_32_NFKC_QC_No : constant UTF_32_Ranges := (
+     (16#00A0#, 16#00A0#),     -- NO-BREAK SPACE
+     (16#00A8#, 16#00A8#),     -- DIAERESIS
+     (16#00AA#, 16#00AA#),     -- FEMININE ORDINAL INDICATOR
+     (16#00AF#, 16#00AF#),     -- MACRON
+     (16#00B2#, 16#00B3#),     -- SUPERSCRIPT TWO..SUPERSCRIPT THREE
+     (16#00B4#, 16#00B4#),     -- ACUTE ACCENT
+     (16#00B5#, 16#00B5#),     -- MICRO SIGN
+     (16#00B8#, 16#00B8#),     -- CEDILLA
+     (16#00B9#, 16#00B9#),     -- SUPERSCRIPT ONE
+     (16#00BA#, 16#00BA#),     -- MASCULINE ORDINAL INDICATOR
+     (16#00BC#, 16#00BE#),     -- VULGAR FRACTION ONE QUARTER..VULGAR FRACTION THREE QUARTERS
+     (16#0132#, 16#0133#),     -- LATIN CAPITAL LIGATURE IJ..LATIN SMALL LIGATURE IJ
+     (16#013F#, 16#0140#),     -- LATIN CAPITAL LETTER L WITH MIDDLE DOT..LATIN SMALL LETTER L WITH MIDDLE DOT
+     (16#0149#, 16#0149#),     -- LATIN SMALL LETTER N PRECEDED BY APOSTROPHE
+     (16#017F#, 16#017F#),     -- LATIN SMALL LETTER LONG S
+     (16#01C4#, 16#01CC#),     -- LATIN CAPITAL LETTER DZ WITH CARON..LATIN SMALL LETTER NJ
+     (16#01F1#, 16#01F3#),     -- LATIN CAPITAL LETTER DZ..LATIN SMALL LETTER DZ
+     (16#02B0#, 16#02B8#),     -- MODIFIER LETTER SMALL H..MODIFIER LETTER SMALL Y
+     (16#02D8#, 16#02DD#),     -- BREVE..DOUBLE ACUTE ACCENT
+     (16#02E0#, 16#02E4#),     -- MODIFIER LETTER SMALL GAMMA..MODIFIER LETTER SMALL REVERSED GLOTTAL STOP
+     (16#0340#, 16#0341#),     -- COMBINING GRAVE TONE MARK..COMBINING ACUTE TONE MARK
+     (16#0343#, 16#0344#),     -- COMBINING GREEK KORONIS..COMBINING GREEK DIALYTIKA TONOS
+     (16#0374#, 16#0374#),     -- GREEK NUMERAL SIGN
+     (16#037A#, 16#037A#),     -- GREEK YPOGEGRAMMENI
+     (16#037E#, 16#037E#),     -- GREEK QUESTION MARK
+     (16#0384#, 16#0385#),     -- GREEK TONOS..GREEK DIALYTIKA TONOS
+     (16#0387#, 16#0387#),     -- GREEK ANO TELEIA
+     (16#03D0#, 16#03D6#),     -- GREEK BETA SYMBOL..GREEK PI SYMBOL
+     (16#03F0#, 16#03F2#),     -- GREEK KAPPA SYMBOL..GREEK LUNATE SIGMA SYMBOL
+     (16#03F4#, 16#03F5#),     -- GREEK CAPITAL THETA SYMBOL..GREEK LUNATE EPSILON SYMBOL
+     (16#03F9#, 16#03F9#),     -- GREEK CAPITAL LUNATE SIGMA SYMBOL
+     (16#0587#, 16#0587#),     -- ARMENIAN SMALL LIGATURE ECH YIWN
+     (16#0675#, 16#0678#),     -- ARABIC LETTER HIGH HAMZA ALEF..ARABIC LETTER HIGH HAMZA YEH
+     (16#0958#, 16#095F#),     -- DEVANAGARI LETTER QA..DEVANAGARI LETTER YYA
+     (16#09DC#, 16#09DD#),     -- BENGALI LETTER RRA..BENGALI LETTER RHA
+     (16#09DF#, 16#09DF#),     -- BENGALI LETTER YYA
+     (16#0A33#, 16#0A33#),     -- GURMUKHI LETTER LLA
+     (16#0A36#, 16#0A36#),     -- GURMUKHI LETTER SHA
+     (16#0A59#, 16#0A5B#),     -- GURMUKHI LETTER KHHA..GURMUKHI LETTER ZA
+     (16#0A5E#, 16#0A5E#),     -- GURMUKHI LETTER FA
+     (16#0B5C#, 16#0B5D#),     -- ORIYA LETTER RRA..ORIYA LETTER RHA
+     (16#0E33#, 16#0E33#),     -- THAI CHARACTER SARA AM
+     (16#0EB3#, 16#0EB3#),     -- LAO VOWEL SIGN AM
+     (16#0EDC#, 16#0EDD#),     -- LAO HO NO..LAO HO MO
+     (16#0F0C#, 16#0F0C#),     -- TIBETAN MARK DELIMITER TSHEG BSTAR
+     (16#0F43#, 16#0F43#),     -- TIBETAN LETTER GHA
+     (16#0F4D#, 16#0F4D#),     -- TIBETAN LETTER DDHA
+     (16#0F52#, 16#0F52#),     -- TIBETAN LETTER DHA
+     (16#0F57#, 16#0F57#),     -- TIBETAN LETTER BHA
+     (16#0F5C#, 16#0F5C#),     -- TIBETAN LETTER DZHA
+     (16#0F69#, 16#0F69#),     -- TIBETAN LETTER KSSA
+     (16#0F73#, 16#0F73#),     -- TIBETAN VOWEL SIGN II
+     (16#0F75#, 16#0F79#),     -- TIBETAN VOWEL SIGN UU..TIBETAN VOWEL SIGN VOCALIC LL
+     (16#0F81#, 16#0F81#),     -- TIBETAN VOWEL SIGN REVERSED II
+     (16#0F93#, 16#0F93#),     -- TIBETAN SUBJOINED LETTER GHA
+     (16#0F9D#, 16#0F9D#),     -- TIBETAN SUBJOINED LETTER DDHA
+     (16#0FA2#, 16#0FA2#),     -- TIBETAN SUBJOINED LETTER DHA
+     (16#0FA7#, 16#0FA7#),     -- TIBETAN SUBJOINED LETTER BHA
+     (16#0FAC#, 16#0FAC#),     -- TIBETAN SUBJOINED LETTER DZHA
+     (16#0FB9#, 16#0FB9#),     -- TIBETAN SUBJOINED LETTER KSSA
+     (16#10FC#, 16#10FC#),     -- MODIFIER LETTER GEORGIAN NAR
+     (16#1D2C#, 16#1D2E#),     -- MODIFIER LETTER CAPITAL A..MODIFIER LETTER CAPITAL B
+     (16#1D30#, 16#1D3A#),     -- MODIFIER LETTER CAPITAL D..MODIFIER LETTER CAPITAL N
+     (16#1D3C#, 16#1D4D#),     -- MODIFIER LETTER CAPITAL O..MODIFIER LETTER SMALL G
+     (16#1D4F#, 16#1D6A#),     -- MODIFIER LETTER SMALL K..GREEK SUBSCRIPT SMALL LETTER CHI
+     (16#1D78#, 16#1D78#),     -- MODIFIER LETTER CYRILLIC EN
+     (16#1D9B#, 16#1DBF#),     -- MODIFIER LETTER SMALL TURNED ALPHA..MODIFIER LETTER SMALL THETA
+     (16#1E9A#, 16#1E9B#),     -- LATIN SMALL LETTER A WITH RIGHT HALF RING..LATIN SMALL LETTER LONG S WITH DOT ABOVE
+     (16#1F71#, 16#1F71#),     -- GREEK SMALL LETTER ALPHA WITH OXIA
+     (16#1F73#, 16#1F73#),     -- GREEK SMALL LETTER EPSILON WITH OXIA
+     (16#1F75#, 16#1F75#),     -- GREEK SMALL LETTER ETA WITH OXIA
+     (16#1F77#, 16#1F77#),     -- GREEK SMALL LETTER IOTA WITH OXIA
+     (16#1F79#, 16#1F79#),     -- GREEK SMALL LETTER OMICRON WITH OXIA
+     (16#1F7B#, 16#1F7B#),     -- GREEK SMALL LETTER UPSILON WITH OXIA
+     (16#1F7D#, 16#1F7D#),     -- GREEK SMALL LETTER OMEGA WITH OXIA
+     (16#1FBB#, 16#1FBB#),     -- GREEK CAPITAL LETTER ALPHA WITH OXIA
+     (16#1FBD#, 16#1FBD#),     -- GREEK KORONIS
+     (16#1FBE#, 16#1FBE#),     -- GREEK PROSGEGRAMMENI
+     (16#1FBF#, 16#1FC1#),     -- GREEK PSILI..GREEK DIALYTIKA AND PERISPOMENI
+     (16#1FC9#, 16#1FC9#),     -- GREEK CAPITAL LETTER EPSILON WITH OXIA
+     (16#1FCB#, 16#1FCB#),     -- GREEK CAPITAL LETTER ETA WITH OXIA
+     (16#1FCD#, 16#1FCF#),     -- GREEK PSILI AND VARIA..GREEK PSILI AND PERISPOMENI
+     (16#1FD3#, 16#1FD3#),     -- GREEK SMALL LETTER IOTA WITH DIALYTIKA AND OXIA
+     (16#1FDB#, 16#1FDB#),     -- GREEK CAPITAL LETTER IOTA WITH OXIA
+     (16#1FDD#, 16#1FDF#),     -- GREEK DASIA AND VARIA..GREEK DASIA AND PERISPOMENI
+     (16#1FE3#, 16#1FE3#),     -- GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND OXIA
+     (16#1FEB#, 16#1FEB#),     -- GREEK CAPITAL LETTER UPSILON WITH OXIA
+     (16#1FED#, 16#1FEF#),     -- GREEK DIALYTIKA AND VARIA..GREEK VARIA
+     (16#1FF9#, 16#1FF9#),     -- GREEK CAPITAL LETTER OMICRON WITH OXIA
+     (16#1FFB#, 16#1FFB#),     -- GREEK CAPITAL LETTER OMEGA WITH OXIA
+     (16#1FFD#, 16#1FFE#),     -- GREEK OXIA..GREEK DASIA
+     (16#2000#, 16#200A#),     -- EN QUAD..HAIR SPACE
+     (16#2011#, 16#2011#),     -- NON-BREAKING HYPHEN
+     (16#2017#, 16#2017#),     -- DOUBLE LOW LINE
+     (16#2024#, 16#2026#),     -- ONE DOT LEADER..HORIZONTAL ELLIPSIS
+     (16#202F#, 16#202F#),     -- NARROW NO-BREAK SPACE
+     (16#2033#, 16#2034#),     -- DOUBLE PRIME..TRIPLE PRIME
+     (16#2036#, 16#2037#),     -- REVERSED DOUBLE PRIME..REVERSED TRIPLE PRIME
+     (16#203C#, 16#203C#),     -- DOUBLE EXCLAMATION MARK
+     (16#203E#, 16#203E#),     -- OVERLINE
+     (16#2047#, 16#2049#),     -- DOUBLE QUESTION MARK..EXCLAMATION QUESTION MARK
+     (16#2057#, 16#2057#),     -- QUADRUPLE PRIME
+     (16#205F#, 16#205F#),     -- MEDIUM MATHEMATICAL SPACE
+     (16#2070#, 16#2070#),     -- SUPERSCRIPT ZERO
+     (16#2071#, 16#2071#),     -- SUPERSCRIPT LATIN SMALL LETTER I
+     (16#2074#, 16#2079#),     -- SUPERSCRIPT FOUR..SUPERSCRIPT NINE
+     (16#207A#, 16#207C#),     -- SUPERSCRIPT PLUS SIGN..SUPERSCRIPT EQUALS SIGN
+     (16#207D#, 16#207D#),     -- SUPERSCRIPT LEFT PARENTHESIS
+     (16#207E#, 16#207E#),     -- SUPERSCRIPT RIGHT PARENTHESIS
+     (16#207F#, 16#207F#),     -- SUPERSCRIPT LATIN SMALL LETTER N
+     (16#2080#, 16#2089#),     -- SUBSCRIPT ZERO..SUBSCRIPT NINE
+     (16#208A#, 16#208C#),     -- SUBSCRIPT PLUS SIGN..SUBSCRIPT EQUALS SIGN
+     (16#208D#, 16#208D#),     -- SUBSCRIPT LEFT PARENTHESIS
+     (16#208E#, 16#208E#),     -- SUBSCRIPT RIGHT PARENTHESIS
+     (16#2090#, 16#209C#),     -- LATIN SUBSCRIPT SMALL LETTER A..LATIN SUBSCRIPT SMALL LETTER T
+     (16#20A8#, 16#20A8#),     -- RUPEE SIGN
+     (16#2100#, 16#2101#),     -- ACCOUNT OF..ADDRESSED TO THE SUBJECT
+     (16#2102#, 16#2102#),     -- DOUBLE-STRUCK CAPITAL C
+     (16#2103#, 16#2103#),     -- DEGREE CELSIUS
+     (16#2105#, 16#2106#),     -- CARE OF..CADA UNA
+     (16#2107#, 16#2107#),     -- EULER CONSTANT
+     (16#2109#, 16#2109#),     -- DEGREE FAHRENHEIT
+     (16#210A#, 16#2113#),     -- SCRIPT SMALL G..SCRIPT SMALL L
+     (16#2115#, 16#2115#),     -- DOUBLE-STRUCK CAPITAL N
+     (16#2116#, 16#2116#),     -- NUMERO SIGN
+     (16#2119#, 16#211D#),     -- DOUBLE-STRUCK CAPITAL P..DOUBLE-STRUCK CAPITAL R
+     (16#2120#, 16#2122#),     -- SERVICE MARK..TRADE MARK SIGN
+     (16#2124#, 16#2124#),     -- DOUBLE-STRUCK CAPITAL Z
+     (16#2126#, 16#2126#),     -- OHM SIGN
+     (16#2128#, 16#2128#),     -- BLACK-LETTER CAPITAL Z
+     (16#212A#, 16#212D#),     -- KELVIN SIGN..BLACK-LETTER CAPITAL C
+     (16#212F#, 16#2131#),     -- SCRIPT SMALL E..SCRIPT CAPITAL F
+     (16#2133#, 16#2134#),     -- SCRIPT CAPITAL M..SCRIPT SMALL O
+     (16#2135#, 16#2138#),     -- ALEF SYMBOL..DALET SYMBOL
+     (16#2139#, 16#2139#),     -- INFORMATION SOURCE
+     (16#213B#, 16#213B#),     -- FACSIMILE SIGN
+     (16#213C#, 16#213F#),     -- DOUBLE-STRUCK SMALL PI..DOUBLE-STRUCK CAPITAL PI
+     (16#2140#, 16#2140#),     -- DOUBLE-STRUCK N-ARY SUMMATION
+     (16#2145#, 16#2149#),     -- DOUBLE-STRUCK ITALIC CAPITAL D..DOUBLE-STRUCK ITALIC SMALL J
+     (16#2150#, 16#215F#),     -- VULGAR FRACTION ONE SEVENTH..FRACTION NUMERATOR ONE
+     (16#2160#, 16#217F#),     -- ROMAN NUMERAL ONE..SMALL ROMAN NUMERAL ONE THOUSAND
+     (16#2189#, 16#2189#),     -- VULGAR FRACTION ZERO THIRDS
+     (16#222C#, 16#222D#),     -- DOUBLE INTEGRAL..TRIPLE INTEGRAL
+     (16#222F#, 16#2230#),     -- SURFACE INTEGRAL..VOLUME INTEGRAL
+     (16#2329#, 16#2329#),     -- LEFT-POINTING ANGLE BRACKET
+     (16#232A#, 16#232A#),     -- RIGHT-POINTING ANGLE BRACKET
+     (16#2460#, 16#249B#),     -- CIRCLED DIGIT ONE..NUMBER TWENTY FULL STOP
+     (16#249C#, 16#24E9#),     -- PARENTHESIZED LATIN SMALL LETTER A..CIRCLED LATIN SMALL LETTER Z
+     (16#24EA#, 16#24EA#),     -- CIRCLED DIGIT ZERO
+     (16#2A0C#, 16#2A0C#),     -- QUADRUPLE INTEGRAL OPERATOR
+     (16#2A74#, 16#2A76#),     -- DOUBLE COLON EQUAL..THREE CONSECUTIVE EQUALS SIGNS
+     (16#2ADC#, 16#2ADC#),     -- FORKING
+     (16#2C7C#, 16#2C7D#),     -- LATIN SUBSCRIPT SMALL LETTER J..MODIFIER LETTER CAPITAL V
+     (16#2D6F#, 16#2D6F#),     -- TIFINAGH MODIFIER LETTER LABIALIZATION MARK
+     (16#2E9F#, 16#2E9F#),     -- CJK RADICAL MOTHER
+     (16#2EF3#, 16#2EF3#),     -- CJK RADICAL C-SIMPLIFIED TURTLE
+     (16#2F00#, 16#2FD5#),     -- KANGXI RADICAL ONE..KANGXI RADICAL FLUTE
+     (16#3000#, 16#3000#),     -- IDEOGRAPHIC SPACE
+     (16#3036#, 16#3036#),     -- CIRCLED POSTAL MARK
+     (16#3038#, 16#303A#),     -- HANGZHOU NUMERAL TEN..HANGZHOU NUMERAL THIRTY
+     (16#309B#, 16#309C#),     -- KATAKANA-HIRAGANA VOICED SOUND MARK..KATAKANA-HIRAGANA SEMI-VOICED SOUND MARK
+     (16#309F#, 16#309F#),     -- HIRAGANA DIGRAPH YORI
+     (16#30FF#, 16#30FF#),     -- KATAKANA DIGRAPH KOTO
+     (16#3131#, 16#318E#),     -- HANGUL LETTER KIYEOK..HANGUL LETTER ARAEAE
+     (16#3192#, 16#3195#),     -- IDEOGRAPHIC ANNOTATION ONE MARK..IDEOGRAPHIC ANNOTATION FOUR MARK
+     (16#3196#, 16#319F#),     -- IDEOGRAPHIC ANNOTATION TOP MARK..IDEOGRAPHIC ANNOTATION MAN MARK
+     (16#3200#, 16#321E#),     -- PARENTHESIZED HANGUL KIYEOK..PARENTHESIZED KOREAN CHARACTER O HU
+     (16#3220#, 16#3229#),     -- PARENTHESIZED IDEOGRAPH ONE..PARENTHESIZED IDEOGRAPH TEN
+     (16#322A#, 16#3247#),     -- PARENTHESIZED IDEOGRAPH MOON..CIRCLED IDEOGRAPH KOTO
+     (16#3250#, 16#3250#),     -- PARTNERSHIP SIGN
+     (16#3251#, 16#325F#),     -- CIRCLED NUMBER TWENTY ONE..CIRCLED NUMBER THIRTY FIVE
+     (16#3260#, 16#327E#),     -- CIRCLED HANGUL KIYEOK..CIRCLED HANGUL IEUNG U
+     (16#3280#, 16#3289#),     -- CIRCLED IDEOGRAPH ONE..CIRCLED IDEOGRAPH TEN
+     (16#328A#, 16#32B0#),     -- CIRCLED IDEOGRAPH MOON..CIRCLED IDEOGRAPH NIGHT
+     (16#32B1#, 16#32BF#),     -- CIRCLED NUMBER THIRTY SIX..CIRCLED NUMBER FIFTY
+     (16#32C0#, 16#33FF#),     -- IDEOGRAPHIC TELEGRAPH SYMBOL FOR JANUARY..SQUARE GAL
+     (16#A69C#, 16#A69D#),     -- MODIFIER LETTER CYRILLIC HARD SIGN..MODIFIER LETTER CYRILLIC SOFT SIGN
+     (16#A770#, 16#A770#),     -- MODIFIER LETTER US
+     (16#A7F8#, 16#A7F9#),     -- MODIFIER LETTER CAPITAL H WITH STROKE..MODIFIER LETTER SMALL LIGATURE OE
+     (16#AB5C#, 16#AB5F#),     -- MODIFIER LETTER SMALL HENG..MODIFIER LETTER SMALL U WITH LEFT HOOK
+     (16#AB69#, 16#AB69#),     -- MODIFIER LETTER SMALL TURNED W
+     (16#F900#, 16#FA0D#),     -- CJK COMPATIBILITY IDEOGRAPH-F900..CJK COMPATIBILITY IDEOGRAPH-FA0D
+     (16#FA10#, 16#FA10#),     -- CJK COMPATIBILITY IDEOGRAPH-FA10
+     (16#FA12#, 16#FA12#),     -- CJK COMPATIBILITY IDEOGRAPH-FA12
+     (16#FA15#, 16#FA1E#),     -- CJK COMPATIBILITY IDEOGRAPH-FA15..CJK COMPATIBILITY IDEOGRAPH-FA1E
+     (16#FA20#, 16#FA20#),     -- CJK COMPATIBILITY IDEOGRAPH-FA20
+     (16#FA22#, 16#FA22#),     -- CJK COMPATIBILITY IDEOGRAPH-FA22
+     (16#FA25#, 16#FA26#),     -- CJK COMPATIBILITY IDEOGRAPH-FA25..CJK COMPATIBILITY IDEOGRAPH-FA26
+     (16#FA2A#, 16#FA6D#),     -- CJK COMPATIBILITY IDEOGRAPH-FA2A..CJK COMPATIBILITY IDEOGRAPH-FA6D
+     (16#FA70#, 16#FAD9#),     -- CJK COMPATIBILITY IDEOGRAPH-FA70..CJK COMPATIBILITY IDEOGRAPH-FAD9
+     (16#FB00#, 16#FB06#),     -- LATIN SMALL LIGATURE FF..LATIN SMALL LIGATURE ST
+     (16#FB13#, 16#FB17#),     -- ARMENIAN SMALL LIGATURE MEN NOW..ARMENIAN SMALL LIGATURE MEN XEH
+     (16#FB1D#, 16#FB1D#),     -- HEBREW LETTER YOD WITH HIRIQ
+     (16#FB1F#, 16#FB28#),     -- HEBREW LIGATURE YIDDISH YOD YOD PATAH..HEBREW LETTER WIDE TAV
+     (16#FB29#, 16#FB29#),     -- HEBREW LETTER ALTERNATIVE PLUS SIGN
+     (16#FB2A#, 16#FB36#),     -- HEBREW LETTER SHIN WITH SHIN DOT..HEBREW LETTER ZAYIN WITH DAGESH
+     (16#FB38#, 16#FB3C#),     -- HEBREW LETTER TET WITH DAGESH..HEBREW LETTER LAMED WITH DAGESH
+     (16#FB3E#, 16#FB3E#),     -- HEBREW LETTER MEM WITH DAGESH
+     (16#FB40#, 16#FB41#),     -- HEBREW LETTER NUN WITH DAGESH..HEBREW LETTER SAMEKH WITH DAGESH
+     (16#FB43#, 16#FB44#),     -- HEBREW LETTER FINAL PE WITH DAGESH..HEBREW LETTER PE WITH DAGESH
+     (16#FB46#, 16#FBB1#),     -- HEBREW LETTER TSADI WITH DAGESH..ARABIC LETTER YEH BARREE WITH HAMZA ABOVE FINAL FORM
+     (16#FBD3#, 16#FD3D#),     -- ARABIC LETTER NG ISOLATED FORM..ARABIC LIGATURE ALEF WITH FATHATAN ISOLATED FORM
+     (16#FD50#, 16#FD8F#),     -- ARABIC LIGATURE TEH WITH JEEM WITH MEEM INITIAL FORM..ARABIC LIGATURE MEEM WITH KHAH WITH MEEM INITIAL FORM
+     (16#FD92#, 16#FDC7#),     -- ARABIC LIGATURE MEEM WITH JEEM WITH KHAH INITIAL FORM..ARABIC LIGATURE NOON WITH JEEM WITH YEH FINAL FORM
+     (16#FDF0#, 16#FDFB#),     -- ARABIC LIGATURE SALLA USED AS KORANIC STOP SIGN ISOLATED FORM..ARABIC LIGATURE JALLAJALALOUHOU
+     (16#FDFC#, 16#FDFC#),     -- RIAL SIGN
+     (16#FE10#, 16#FE16#),     -- PRESENTATION FORM FOR VERTICAL COMMA..PRESENTATION FORM FOR VERTICAL QUESTION MARK
+     (16#FE17#, 16#FE17#),     -- PRESENTATION FORM FOR VERTICAL LEFT WHITE LENTICULAR BRACKET
+     (16#FE18#, 16#FE18#),     -- PRESENTATION FORM FOR VERTICAL RIGHT WHITE LENTICULAR BRAKCET
+     (16#FE19#, 16#FE19#),     -- PRESENTATION FORM FOR VERTICAL HORIZONTAL ELLIPSIS
+     (16#FE30#, 16#FE30#),     -- PRESENTATION FORM FOR VERTICAL TWO DOT LEADER
+     (16#FE31#, 16#FE32#),     -- PRESENTATION FORM FOR VERTICAL EM DASH..PRESENTATION FORM FOR VERTICAL EN DASH
+     (16#FE33#, 16#FE34#),     -- PRESENTATION FORM FOR VERTICAL LOW LINE..PRESENTATION FORM FOR VERTICAL WAVY LOW LINE
+     (16#FE35#, 16#FE35#),     -- PRESENTATION FORM FOR VERTICAL LEFT PARENTHESIS
+     (16#FE36#, 16#FE36#),     -- PRESENTATION FORM FOR VERTICAL RIGHT PARENTHESIS
+     (16#FE37#, 16#FE37#),     -- PRESENTATION FORM FOR VERTICAL LEFT CURLY BRACKET
+     (16#FE38#, 16#FE38#),     -- PRESENTATION FORM FOR VERTICAL RIGHT CURLY BRACKET
+     (16#FE39#, 16#FE39#),     -- PRESENTATION FORM FOR VERTICAL LEFT TORTOISE SHELL BRACKET
+     (16#FE3A#, 16#FE3A#),     -- PRESENTATION FORM FOR VERTICAL RIGHT TORTOISE SHELL BRACKET
+     (16#FE3B#, 16#FE3B#),     -- PRESENTATION FORM FOR VERTICAL LEFT BLACK LENTICULAR BRACKET
+     (16#FE3C#, 16#FE3C#),     -- PRESENTATION FORM FOR VERTICAL RIGHT BLACK LENTICULAR BRACKET
+     (16#FE3D#, 16#FE3D#),     -- PRESENTATION FORM FOR VERTICAL LEFT DOUBLE ANGLE BRACKET
+     (16#FE3E#, 16#FE3E#),     -- PRESENTATION FORM FOR VERTICAL RIGHT DOUBLE ANGLE BRACKET
+     (16#FE3F#, 16#FE3F#),     -- PRESENTATION FORM FOR VERTICAL LEFT ANGLE BRACKET
+     (16#FE40#, 16#FE40#),     -- PRESENTATION FORM FOR VERTICAL RIGHT ANGLE BRACKET
+     (16#FE41#, 16#FE41#),     -- PRESENTATION FORM FOR VERTICAL LEFT CORNER BRACKET
+     (16#FE42#, 16#FE42#),     -- PRESENTATION FORM FOR VERTICAL RIGHT CORNER BRACKET
+     (16#FE43#, 16#FE43#),     -- PRESENTATION FORM FOR VERTICAL LEFT WHITE CORNER BRACKET
+     (16#FE44#, 16#FE44#),     -- PRESENTATION FORM FOR VERTICAL RIGHT WHITE CORNER BRACKET
+     (16#FE47#, 16#FE47#),     -- PRESENTATION FORM FOR VERTICAL LEFT SQUARE BRACKET
+     (16#FE48#, 16#FE48#),     -- PRESENTATION FORM FOR VERTICAL RIGHT SQUARE BRACKET
+     (16#FE49#, 16#FE4C#),     -- DASHED OVERLINE..DOUBLE WAVY OVERLINE
+     (16#FE4D#, 16#FE4F#),     -- DASHED LOW LINE..WAVY LOW LINE
+     (16#FE50#, 16#FE52#),     -- SMALL COMMA..SMALL FULL STOP
+     (16#FE54#, 16#FE57#),     -- SMALL SEMICOLON..SMALL EXCLAMATION MARK
+     (16#FE58#, 16#FE58#),     -- SMALL EM DASH
+     (16#FE59#, 16#FE59#),     -- SMALL LEFT PARENTHESIS
+     (16#FE5A#, 16#FE5A#),     -- SMALL RIGHT PARENTHESIS
+     (16#FE5B#, 16#FE5B#),     -- SMALL LEFT CURLY BRACKET
+     (16#FE5C#, 16#FE5C#),     -- SMALL RIGHT CURLY BRACKET
+     (16#FE5D#, 16#FE5D#),     -- SMALL LEFT TORTOISE SHELL BRACKET
+     (16#FE5E#, 16#FE5E#),     -- SMALL RIGHT TORTOISE SHELL BRACKET
+     (16#FE5F#, 16#FE61#),     -- SMALL NUMBER SIGN..SMALL ASTERISK
+     (16#FE62#, 16#FE62#),     -- SMALL PLUS SIGN
+     (16#FE63#, 16#FE63#),     -- SMALL HYPHEN-MINUS
+     (16#FE64#, 16#FE66#),     -- SMALL LESS-THAN SIGN..SMALL EQUALS SIGN
+     (16#FE68#, 16#FE68#),     -- SMALL REVERSE SOLIDUS
+     (16#FE69#, 16#FE69#),     -- SMALL DOLLAR SIGN
+     (16#FE6A#, 16#FE6B#),     -- SMALL PERCENT SIGN..SMALL COMMERCIAL AT
+     (16#FE70#, 16#FE72#),     -- ARABIC FATHATAN ISOLATED FORM..ARABIC DAMMATAN ISOLATED FORM
+     (16#FE74#, 16#FE74#),     -- ARABIC KASRATAN ISOLATED FORM
+     (16#FE76#, 16#FEFC#),     -- ARABIC FATHA ISOLATED FORM..ARABIC LIGATURE LAM WITH ALEF FINAL FORM
+     (16#FF01#, 16#FF03#),     -- FULLWIDTH EXCLAMATION MARK..FULLWIDTH NUMBER SIGN
+     (16#FF04#, 16#FF04#),        -- FULLWIDTH DOLLAR SIGN
+     (16#FF05#, 16#FF07#),     -- FULLWIDTH PERCENT SIGN..FULLWIDTH APOSTROPHE
+     (16#FF08#, 16#FF08#),     -- FULLWIDTH LEFT PARENTHESIS
+     (16#FF09#, 16#FF09#),     -- FULLWIDTH RIGHT PARENTHESIS
+     (16#FF0A#, 16#FF0A#),     -- FULLWIDTH ASTERISK
+     (16#FF0B#, 16#FF0B#),     -- FULLWIDTH PLUS SIGN
+     (16#FF0C#, 16#FF0C#),     -- FULLWIDTH COMMA
+     (16#FF0D#, 16#FF0D#),     -- FULLWIDTH HYPHEN-MINUS
+     (16#FF0E#, 16#FF0F#),     -- FULLWIDTH FULL STOP..FULLWIDTH SOLIDUS
+     (16#FF10#, 16#FF19#),     -- FULLWIDTH DIGIT ZERO..FULLWIDTH DIGIT NINE
+     (16#FF1A#, 16#FF1B#),     -- FULLWIDTH COLON..FULLWIDTH SEMICOLON
+     (16#FF1C#, 16#FF1E#),     -- FULLWIDTH LESS-THAN SIGN..FULLWIDTH GREATER-THAN SIGN
+     (16#FF1F#, 16#FF20#),     -- FULLWIDTH QUESTION MARK..FULLWIDTH COMMERCIAL AT
+     (16#FF21#, 16#FF3A#),     -- FULLWIDTH LATIN CAPITAL LETTER A..FULLWIDTH LATIN CAPITAL LETTER Z
+     (16#FF3B#, 16#FF3B#),     -- FULLWIDTH LEFT SQUARE BRACKET
+     (16#FF3C#, 16#FF3C#),     -- FULLWIDTH REVERSE SOLIDUS
+     (16#FF3D#, 16#FF3D#),     -- FULLWIDTH RIGHT SQUARE BRACKET
+     (16#FF3E#, 16#FF3E#),     -- FULLWIDTH CIRCUMFLEX ACCENT
+     (16#FF3F#, 16#FF3F#),     -- FULLWIDTH LOW LINE
+     (16#FF40#, 16#FF40#),     -- FULLWIDTH GRAVE ACCENT
+     (16#FF41#, 16#FF5A#),     -- FULLWIDTH LATIN SMALL LETTER A..FULLWIDTH LATIN SMALL LETTER Z
+     (16#FF5B#, 16#FF5B#),     -- FULLWIDTH LEFT CURLY BRACKET
+     (16#FF5C#, 16#FF5C#),     -- FULLWIDTH VERTICAL LINE
+     (16#FF5D#, 16#FF5D#),     -- FULLWIDTH RIGHT CURLY BRACKET
+     (16#FF5E#, 16#FF5E#),     -- FULLWIDTH TILDE
+     (16#FF5F#, 16#FF5F#),     -- FULLWIDTH LEFT WHITE PARENTHESIS
+     (16#FF60#, 16#FF60#),     -- FULLWIDTH RIGHT WHITE PARENTHESIS
+     (16#FF61#, 16#FF61#),     -- HALFWIDTH IDEOGRAPHIC FULL STOP
+     (16#FF62#, 16#FF62#),     -- HALFWIDTH LEFT CORNER BRACKET
+     (16#FF63#, 16#FF63#),     -- HALFWIDTH RIGHT CORNER BRACKET
+     (16#FF64#, 16#FF65#),     -- HALFWIDTH IDEOGRAPHIC COMMA..HALFWIDTH KATAKANA MIDDLE DOT
+     (16#FF66#, 16#FF6F#),     -- HALFWIDTH KATAKANA LETTER WO..HALFWIDTH KATAKANA LETTER SMALL TU
+     (16#FF70#, 16#FF70#),     -- HALFWIDTH KATAKANA-HIRAGANA PROLONGED SOUND MARK
+     (16#FF71#, 16#FF9D#),     -- HALFWIDTH KATAKANA LETTER A..HALFWIDTH KATAKANA LETTER N
+     (16#FF9E#, 16#FF9F#),     -- HALFWIDTH KATAKANA VOICED SOUND MARK..HALFWIDTH KATAKANA SEMI-VOICED SOUND MARK
+     (16#FFA0#, 16#FFBE#),     -- HALFWIDTH HANGUL FILLER..HALFWIDTH HANGUL LETTER HIEUH
+     (16#FFC2#, 16#FFC7#),     -- HALFWIDTH HANGUL LETTER A..HALFWIDTH HANGUL LETTER E
+     (16#FFCA#, 16#FFCF#),     -- HALFWIDTH HANGUL LETTER YEO..HALFWIDTH HANGUL LETTER OE
+     (16#FFD2#, 16#FFD7#),     -- HALFWIDTH HANGUL LETTER YO..HALFWIDTH HANGUL LETTER YU
+     (16#FFDA#, 16#FFDC#),     -- HALFWIDTH HANGUL LETTER EU..HALFWIDTH HANGUL LETTER I
+     (16#FFE0#, 16#FFE1#),     -- FULLWIDTH CENT SIGN..FULLWIDTH POUND SIGN
+     (16#FFE2#, 16#FFE2#),     -- FULLWIDTH NOT SIGN
+     (16#FFE3#, 16#FFE3#),     -- FULLWIDTH MACRON
+     (16#FFE4#, 16#FFE4#),     -- FULLWIDTH BROKEN BAR
+     (16#FFE5#, 16#FFE6#),     -- FULLWIDTH YEN SIGN..FULLWIDTH WON SIGN
+     (16#FFE8#, 16#FFE8#),     -- HALFWIDTH FORMS LIGHT VERTICAL
+     (16#FFE9#, 16#FFEC#),     -- HALFWIDTH LEFTWARDS ARROW..HALFWIDTH DOWNWARDS ARROW
+     (16#FFED#, 16#FFEE#),     -- HALFWIDTH BLACK SQUARE..HALFWIDTH WHITE CIRCLE
+     (16#1D15E#, 16#1D164#),   -- MUSICAL SYMBOL HALF NOTE..MUSICAL SYMBOL ONE HUNDRED TWENTY-EIGHTH NOTE
+     (16#1D1BB#, 16#1D1C0#),   -- MUSICAL SYMBOL MINIMA..MUSICAL SYMBOL FUSA BLACK
+     (16#1D400#, 16#1D454#),   -- MATHEMATICAL BOLD CAPITAL A..MATHEMATICAL ITALIC SMALL G
+     (16#1D456#, 16#1D49C#),   -- MATHEMATICAL ITALIC SMALL I..MATHEMATICAL SCRIPT CAPITAL A
+     (16#1D49E#, 16#1D49F#),   -- MATHEMATICAL SCRIPT CAPITAL C..MATHEMATICAL SCRIPT CAPITAL D
+     (16#1D4A2#, 16#1D4A2#),   -- MATHEMATICAL SCRIPT CAPITAL G
+     (16#1D4A5#, 16#1D4A6#),   -- MATHEMATICAL SCRIPT CAPITAL J..MATHEMATICAL SCRIPT CAPITAL K
+     (16#1D4A9#, 16#1D4AC#),   -- MATHEMATICAL SCRIPT CAPITAL N..MATHEMATICAL SCRIPT CAPITAL Q
+     (16#1D4AE#, 16#1D4B9#),   -- MATHEMATICAL SCRIPT CAPITAL S..MATHEMATICAL SCRIPT SMALL D
+     (16#1D4BB#, 16#1D4BB#),   -- MATHEMATICAL SCRIPT SMALL F
+     (16#1D4BD#, 16#1D4C3#),   -- MATHEMATICAL SCRIPT SMALL H..MATHEMATICAL SCRIPT SMALL N
+     (16#1D4C5#, 16#1D505#),   -- MATHEMATICAL SCRIPT SMALL P..MATHEMATICAL FRAKTUR CAPITAL B
+     (16#1D507#, 16#1D50A#),   -- MATHEMATICAL FRAKTUR CAPITAL D..MATHEMATICAL FRAKTUR CAPITAL G
+     (16#1D50D#, 16#1D514#),   -- MATHEMATICAL FRAKTUR CAPITAL J..MATHEMATICAL FRAKTUR CAPITAL Q
+     (16#1D516#, 16#1D51C#),   -- MATHEMATICAL FRAKTUR CAPITAL S..MATHEMATICAL FRAKTUR CAPITAL Y
+     (16#1D51E#, 16#1D539#),   -- MATHEMATICAL FRAKTUR SMALL A..MATHEMATICAL DOUBLE-STRUCK CAPITAL B
+     (16#1D53B#, 16#1D53E#),   -- MATHEMATICAL DOUBLE-STRUCK CAPITAL D..MATHEMATICAL DOUBLE-STRUCK CAPITAL G
+     (16#1D540#, 16#1D544#),   -- MATHEMATICAL DOUBLE-STRUCK CAPITAL I..MATHEMATICAL DOUBLE-STRUCK CAPITAL M
+     (16#1D546#, 16#1D546#),   -- MATHEMATICAL DOUBLE-STRUCK CAPITAL O
+     (16#1D54A#, 16#1D550#),   -- MATHEMATICAL DOUBLE-STRUCK CAPITAL S..MATHEMATICAL DOUBLE-STRUCK CAPITAL Y
+     (16#1D552#, 16#1D6A5#),   -- MATHEMATICAL DOUBLE-STRUCK SMALL A..MATHEMATICAL ITALIC SMALL DOTLESS J
+     (16#1D6A8#, 16#1D6C0#),   -- MATHEMATICAL BOLD CAPITAL ALPHA..MATHEMATICAL BOLD CAPITAL OMEGA
+     (16#1D6C1#, 16#1D6C1#),   -- MATHEMATICAL BOLD NABLA
+     (16#1D6C2#, 16#1D6DA#),   -- MATHEMATICAL BOLD SMALL ALPHA..MATHEMATICAL BOLD SMALL OMEGA
+     (16#1D6DB#, 16#1D6DB#),   -- MATHEMATICAL BOLD PARTIAL DIFFERENTIAL
+     (16#1D6DC#, 16#1D6FA#),   -- MATHEMATICAL BOLD EPSILON SYMBOL..MATHEMATICAL ITALIC CAPITAL OMEGA
+     (16#1D6FB#, 16#1D6FB#),   -- MATHEMATICAL ITALIC NABLA
+     (16#1D6FC#, 16#1D714#),   -- MATHEMATICAL ITALIC SMALL ALPHA..MATHEMATICAL ITALIC SMALL OMEGA
+     (16#1D715#, 16#1D715#),   -- MATHEMATICAL ITALIC PARTIAL DIFFERENTIAL
+     (16#1D716#, 16#1D734#),   -- MATHEMATICAL ITALIC EPSILON SYMBOL..MATHEMATICAL BOLD ITALIC CAPITAL OMEGA
+     (16#1D735#, 16#1D735#),   -- MATHEMATICAL BOLD ITALIC NABLA
+     (16#1D736#, 16#1D74E#),   -- MATHEMATICAL BOLD ITALIC SMALL ALPHA..MATHEMATICAL BOLD ITALIC SMALL OMEGA
+     (16#1D74F#, 16#1D74F#),   -- MATHEMATICAL BOLD ITALIC PARTIAL DIFFERENTIAL
+     (16#1D750#, 16#1D76E#),   -- MATHEMATICAL BOLD ITALIC EPSILON SYMBOL..MATHEMATICAL SANS-SERIF BOLD CAPITAL OMEGA
+     (16#1D76F#, 16#1D76F#),   -- MATHEMATICAL SANS-SERIF BOLD NABLA
+     (16#1D770#, 16#1D788#),   -- MATHEMATICAL SANS-SERIF BOLD SMALL ALPHA..MATHEMATICAL SANS-SERIF BOLD SMALL OMEGA
+     (16#1D789#, 16#1D789#),   -- MATHEMATICAL SANS-SERIF BOLD PARTIAL DIFFERENTIAL
+     (16#1D78A#, 16#1D7A8#),   -- MATHEMATICAL SANS-SERIF BOLD EPSILON SYMBOL..MATHEMATICAL SANS-SERIF BOLD ITALIC CAPITAL OMEGA
+     (16#1D7A9#, 16#1D7A9#),   -- MATHEMATICAL SANS-SERIF BOLD ITALIC NABLA
+     (16#1D7AA#, 16#1D7C2#),   -- MATHEMATICAL SANS-SERIF BOLD ITALIC SMALL ALPHA..MATHEMATICAL SANS-SERIF BOLD ITALIC SMALL OMEGA
+     (16#1D7C3#, 16#1D7C3#),   -- MATHEMATICAL SANS-SERIF BOLD ITALIC PARTIAL DIFFERENTIAL
+     (16#1D7C4#, 16#1D7CB#),   -- MATHEMATICAL SANS-SERIF BOLD ITALIC EPSILON SYMBOL..MATHEMATICAL BOLD SMALL DIGAMMA
+     (16#1D7CE#, 16#1D7FF#),   -- MATHEMATICAL BOLD DIGIT ZERO..MATHEMATICAL MONOSPACE DIGIT NINE
+     (16#1EE00#, 16#1EE03#),   -- ARABIC MATHEMATICAL ALEF..ARABIC MATHEMATICAL DAL
+     (16#1EE05#, 16#1EE1F#),   -- ARABIC MATHEMATICAL WAW..ARABIC MATHEMATICAL DOTLESS QAF
+     (16#1EE21#, 16#1EE22#),   -- ARABIC MATHEMATICAL INITIAL BEH..ARABIC MATHEMATICAL INITIAL JEEM
+     (16#1EE24#, 16#1EE24#),   -- ARABIC MATHEMATICAL INITIAL HEH
+     (16#1EE27#, 16#1EE27#),   -- ARABIC MATHEMATICAL INITIAL HAH
+     (16#1EE29#, 16#1EE32#),   -- ARABIC MATHEMATICAL INITIAL YEH..ARABIC MATHEMATICAL INITIAL QAF
+     (16#1EE34#, 16#1EE37#),   -- ARABIC MATHEMATICAL INITIAL SHEEN..ARABIC MATHEMATICAL INITIAL KHAH
+     (16#1EE39#, 16#1EE39#),   -- ARABIC MATHEMATICAL INITIAL DAD
+     (16#1EE3B#, 16#1EE3B#),   -- ARABIC MATHEMATICAL INITIAL GHAIN
+     (16#1EE42#, 16#1EE42#),   -- ARABIC MATHEMATICAL TAILED JEEM
+     (16#1EE47#, 16#1EE47#),   -- ARABIC MATHEMATICAL TAILED HAH
+     (16#1EE49#, 16#1EE49#),   -- ARABIC MATHEMATICAL TAILED YEH
+     (16#1EE4B#, 16#1EE4B#),   -- ARABIC MATHEMATICAL TAILED LAM
+     (16#1EE4D#, 16#1EE4F#),   -- ARABIC MATHEMATICAL TAILED NOON..ARABIC MATHEMATICAL TAILED AIN
+     (16#1EE51#, 16#1EE52#),   -- ARABIC MATHEMATICAL TAILED SAD..ARABIC MATHEMATICAL TAILED QAF
+     (16#1EE54#, 16#1EE54#),   -- ARABIC MATHEMATICAL TAILED SHEEN
+     (16#1EE57#, 16#1EE57#),   -- ARABIC MATHEMATICAL TAILED KHAH
+     (16#1EE59#, 16#1EE59#),   -- ARABIC MATHEMATICAL TAILED DAD
+     (16#1EE5B#, 16#1EE5B#),   -- ARABIC MATHEMATICAL TAILED GHAIN
+     (16#1EE5D#, 16#1EE5D#),   -- ARABIC MATHEMATICAL TAILED DOTLESS NOON
+     (16#1EE5F#, 16#1EE5F#),   -- ARABIC MATHEMATICAL TAILED DOTLESS QAF
+     (16#1EE61#, 16#1EE62#),   -- ARABIC MATHEMATICAL STRETCHED BEH..ARABIC MATHEMATICAL STRETCHED JEEM
+     (16#1EE64#, 16#1EE64#),   -- ARABIC MATHEMATICAL STRETCHED HEH
+     (16#1EE67#, 16#1EE6A#),   -- ARABIC MATHEMATICAL STRETCHED HAH..ARABIC MATHEMATICAL STRETCHED KAF
+     (16#1EE6C#, 16#1EE72#),   -- ARABIC MATHEMATICAL STRETCHED MEEM..ARABIC MATHEMATICAL STRETCHED QAF
+     (16#1EE74#, 16#1EE77#),   -- ARABIC MATHEMATICAL STRETCHED SHEEN..ARABIC MATHEMATICAL STRETCHED KHAH
+     (16#1EE79#, 16#1EE7C#),   -- ARABIC MATHEMATICAL STRETCHED DAD..ARABIC MATHEMATICAL STRETCHED DOTLESS BEH
+     (16#1EE7E#, 16#1EE7E#),   -- ARABIC MATHEMATICAL STRETCHED DOTLESS FEH
+     (16#1EE80#, 16#1EE89#),   -- ARABIC MATHEMATICAL LOOPED ALEF..ARABIC MATHEMATICAL LOOPED YEH
+     (16#1EE8B#, 16#1EE9B#),   -- ARABIC MATHEMATICAL LOOPED LAM..ARABIC MATHEMATICAL LOOPED GHAIN
+     (16#1EEA1#, 16#1EEA3#),   -- ARABIC MATHEMATICAL DOUBLE-STRUCK BEH..ARABIC MATHEMATICAL DOUBLE-STRUCK DAL
+     (16#1EEA5#, 16#1EEA9#),   -- ARABIC MATHEMATICAL DOUBLE-STRUCK WAW..ARABIC MATHEMATICAL DOUBLE-STRUCK YEH
+     (16#1EEAB#, 16#1EEBB#),   -- ARABIC MATHEMATICAL DOUBLE-STRUCK LAM..ARABIC MATHEMATICAL DOUBLE-STRUCK GHAIN
+     (16#1F100#, 16#1F10A#),   -- DIGIT ZERO FULL STOP..DIGIT NINE COMMA
+     (16#1F110#, 16#1F12E#),   -- PARENTHESIZED LATIN CAPITAL LETTER A..CIRCLED WZ
+     (16#1F130#, 16#1F14F#),   -- SQUARED LATIN CAPITAL LETTER A..SQUARED WC
+     (16#1F16A#, 16#1F16C#),   -- RAISED MC SIGN..RAISED MR SIGN
+     (16#1F190#, 16#1F190#),   -- SQUARE DJ
+     (16#1F200#, 16#1F202#),   -- SQUARE HIRAGANA HOKA..SQUARED KATAKANA SA
+     (16#1F210#, 16#1F23B#),   -- SQUARED CJK UNIFIED IDEOGRAPH-624B..SQUARED CJK UNIFIED IDEOGRAPH-914D
+     (16#1F240#, 16#1F248#),   -- TORTOISE SHELL BRACKETED CJK UNIFIED IDEOGRAPH-672C..TORTOISE SHELL BRACKETED CJK UNIFIED IDEOGRAPH-6557
+     (16#1F250#, 16#1F251#),   -- CIRCLED IDEOGRAPH ADVANTAGE..CIRCLED IDEOGRAPH ACCEPT
+     (16#1FBF0#, 16#1FBF9#),   -- SEGMENTED DIGIT ZERO..SEGMENTED DIGIT NINE
+     (16#2F800#, 16#2FA1D#));  -- CJK COMPATIBILITY IDEOGRAPH-2F800..CJK COMPATIBILITY IDEOGRAPH-2FA1D
+
+   pragma Warnings (On);
+   --  Temporary until pragma Warnings at start can be activated ???
+
    type Decomposition_Mapping is record
       Item               : UTF_32;
       First_Char_Mapping : UTF_32;
@@ -12001,6 +12395,15 @@ package body System.UTF_32 is
       return Non_Graphic (C);
    end Is_UTF_32_Non_Graphic;
 
+   --------------------
+   -- Is_UTF_32_NFKC --
+   --------------------
+
+   function Is_UTF_32_NFKC (U : UTF_32) return Boolean is
+   begin
+      return U < 160 or else Range_Search (U, UTF_32_NFKC_QC_No) = 0;
+   end Is_UTF_32_NFKC;
+
    ---------------------
    -- Is_UTF_32_Other --
    ---------------------

diff --git a/gcc/ada/libgnat/s-utf_32.ads b/gcc/ada/libgnat/s-utf_32.ads
index b8e4e3e..e3f0e00 100644 (file)
--- a/gcc/ada/libgnat/s-utf_32.ads
+++ b/gcc/ada/libgnat/s-utf_32.ads
@@ -189,6 +189,12 @@ package System.UTF_32 is
    --  letters to upper case using this routine. A corresponding routine to
    --  fold to lower case is also provided.
 
+   function Is_UTF_32_NFKC (U : UTF_32) return Boolean;
+   pragma Inline (Is_UTF_32_NFKC);
+   --  Return True if U could be present in a string normalized to
+   --  Normalization Form KC (as defined by Clause 21 of ISO/IEC 10646:2017),
+   --  otherwise returns False.
+
    function Is_UTF_32_Basic (U : UTF_32) return Boolean;
    pragma Inline (Is_UTF_32_Basic);
    --  Return True if U has no Decomposition Mapping in the code charts of

diff --git a/gcc/ada/scng.adb b/gcc/ada/scng.adb
index fd3dacc..2bac3a8 100644 (file)
--- a/gcc/ada/scng.adb
+++ b/gcc/ada/scng.adb
@@ -2485,10 +2485,17 @@ package body Scng is
                           ("wide character not allowed in identifier", Wptr);
                      end if;
 
+                     --  AI12-0004: An identifier shall only contain characters
+                     --  that may be present in Normalization Form KC.
+
+                     if not Is_UTF_32_NFKC (UTF_32 (Code)) then
+                        Error_Msg
+                          ("invalid wide character in identifier", Wptr);
+
                      --  If OK letter, store it folding to upper case. Note
                      --  that we include the folded letter in the checksum.
 
-                     if Is_UTF_32_Letter (Cat) then
+                     elsif Is_UTF_32_Letter (Cat) then
                         Code :=
                           Char_Code (UTF_32_To_Upper_Case (UTF_32 (Code)));
                         Accumulate_Checksum (Code);