Remove obsolete macros/tables for \X

A previous commit changed how \X is implemented, and now we don't need
these anymore.

diff --git a/embedvar.h b/embedvar.h
index dde2340..7ba52fb 100644 (file)
--- a/embedvar.h
+++ b/embedvar.h
@@ -333,8 +333,6 @@
 #define PL_unitcheckav_save    (vTHX->Iunitcheckav_save)
 #define PL_unlockhook          (vTHX->Iunlockhook)
 #define PL_unsafe              (vTHX->Iunsafe)
-#define PL_utf8_X_extend       (vTHX->Iutf8_X_extend)
-#define PL_utf8_X_regular_begin        (vTHX->Iutf8_X_regular_begin)
 #define PL_utf8_charname_begin (vTHX->Iutf8_charname_begin)
 #define PL_utf8_charname_continue      (vTHX->Iutf8_charname_continue)
 #define PL_utf8_foldable       (vTHX->Iutf8_foldable)

diff --git a/intrpvar.h b/intrpvar.h
index dc44b31..5154094 100644 (file)
--- a/intrpvar.h
+++ b/intrpvar.h
@@ -598,8 +598,6 @@ PERLVAR(I, HasMultiCharFold,   SV *)
 
 /* utf8 character class swashes */
 PERLVAR(I, utf8_mark,  SV *)
-PERLVAR(I, utf8_X_regular_begin, SV *)
-PERLVAR(I, utf8_X_extend, SV *)
 PERLVAR(I, utf8_toupper, SV *)
 PERLVAR(I, utf8_totitle, SV *)
 PERLVAR(I, utf8_tolower, SV *)

diff --git a/lib/unicore/mktables b/lib/unicore/mktables
index 511ad02..8eb6dfc 100644 (file)
--- a/lib/unicore/mktables
+++ b/lib/unicore/mktables
@@ -14119,107 +14119,6 @@ sub compile_perl() {
                                     + utf8::unicode_to_native(0xA0) # NBSP
                         );
 
-    # These two tables are for matching \X, which is based on the 'extended'
-    # grapheme cluster, which came in 5.1; create empty ones if not already
-    # present.  The straight 'grapheme cluster' (non-extended) is used prior
-    # to 5.1, and differs from the extended (see
-    # http://www.unicode.org/reports/tr29/) only by these two tables, so we
-    # get the older definition automatically when they are empty.
-    my $gcb = property_ref('Grapheme_Cluster_Break');
-    my $perl_prepend = $perl->add_match_table('_X_GCB_Prepend',
-                                        Perl_Extension => 1,
-                                        Fate => $INTERNAL_ONLY);
-    if (defined (my $gcb_prepend = $gcb->table('Prepend'))) {
-        $perl_prepend->set_equivalent_to($gcb_prepend, Related => 1);
-    }
-    else {
-        push @tables_that_may_be_empty, $perl_prepend->complete_name;
-    }
-
-    # All the tables with _X_ in their names are used in defining \X handling,
-    # and are based on the Unicode GCB property.  Basically, \X matches:
-    #   CR LF
-    #   | Prepend* Begin Extend*
-    #   | .
-    # Begin is:           ( Special_Begin | ! Control )
-    # Begin is also:      ( Regular_Begin | Special_Begin )
-    #   where Regular_Begin is defined as ( ! Control - Special_Begin )
-    # Special_Begin is:   ( Regional-Indicator+ | Hangul-syllable )
-    # Extend is:          ( Grapheme_Extend | Spacing_Mark )
-    # Control is:         [ GCB_Control | CR | LF ]
-    # Hangul-syllable is: ( T+ | ( L* ( L | ( LVT | ( V | LV ) V* ) T* ) ))
-
-    foreach my $gcb_name (qw{ L V T LV LVT }) {
-
-        # The perl internal extension's name is the gcb table name prepended
-        # with an '_X_'
-        my $perl_table = $perl->add_match_table('_X_GCB_' . $gcb_name,
-                                        Perl_Extension => 1,
-                                        Fate => $INTERNAL_ONLY,
-                                        Initialize => $gcb->table($gcb_name),
-                                        );
-        # Version 1 had mostly different Hangul syllables that were removed
-        # from later versions, so some of the tables may not apply.
-        if ($v_version lt v2.0) {
-            push @tables_that_may_be_empty, $perl_table->complete_name;
-        }
-    }
-
-    # More GCB.  Populate a combined hangul syllables table
-    my $lv_lvt_v = $perl->add_match_table('_X_LV_LVT_V',
-                                          Perl_Extension => 1,
-                                          Fate => $INTERNAL_ONLY);
-    $lv_lvt_v += $gcb->table('LV') + $gcb->table('LVT') + $gcb->table('V');
-    $lv_lvt_v->add_comment('For use in \X; matches: gcb=LV | gcb=LVT | gcb=V');
-
-    my $ri = $perl->add_match_table('_X_RI', Perl_Extension => 1,
-                                    Fate => $INTERNAL_ONLY);
-    if ($v_version ge v6.2) {
-        $ri += $gcb->table('RI');
-    }
-    else {
-        push @tables_that_may_be_empty, $ri->full_name;
-    }
-
-    my $specials_begin = $perl->add_match_table('_X_Special_Begin_Start',
-                                       Perl_Extension => 1,
-                                       Fate => $INTERNAL_ONLY,
-                                       Initialize => $lv_lvt_v
-                                                   + $gcb->table('L')
-                                                   + $gcb->table('T')
-                                                   + $ri
-                                      );
-    $specials_begin->add_comment(join_lines( <<END
-For use in \\X; matches first (perhaps only) character of potential
-multi-character sequences that can begin an extended grapheme cluster.  They
-need special handling because of their complicated nature.
-END
-    ));
-    my $regular_begin = $perl->add_match_table('_X_Regular_Begin',
-                                       Perl_Extension => 1,
-                                       Fate => $INTERNAL_ONLY,
-                                       Initialize => ~ $gcb->table('Control')
-                                                   - $specials_begin
-                                                   - $gcb->table('CR')
-                                                   - $gcb->table('LF')
-                                      );
-    $regular_begin->add_comment(join_lines( <<END
-For use in \\X; matches first character of anything that can begin an extended
-grapheme cluster, except those that require special handling.
-END
-    ));
-
-    my $extend = $perl->add_match_table('_X_Extend', Perl_Extension => 1,
-                                        Fate => $INTERNAL_ONLY,
-                                        Initialize => $gcb->table('Extend')
-                                       );
-    if (defined (my $sm = $gcb->table('SpacingMark'))) {
-        $extend += $sm;
-    }
-    $extend->add_comment('For use in \X; matches: Extend | SpacingMark');
-
-    # End of GCB \X processing
-
     my @composition = ('Name', 'Unicode_1_Name', 'Name_Alias');
 
     if (@named_sequences) {

diff --git a/regcharclass.h b/regcharclass.h
index 923be81..9f9c828 100644 (file)
--- a/regcharclass.h
+++ b/regcharclass.h
@@ -192,121 +192,6 @@
 ( ( ( 0xED == ((U8*)s)[0] ) && ( ((U8*)s)[1] >= 0xA0 ) ) ? 3 : 0 )
 
 /*
-       GCB_L: Grapheme_Cluster_Break=L
-
-       \p{_X_GCB_L}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_L_utf8(s)                                                    \
-( ( 0xE1 == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x84 == ((U8*)s)[1] ) ?                                             \
-       3                                                                   \
-    : ( ( 0x85 == ((U8*)s)[1] ) && ( ((U8*)s)[2] <= 0x9F ) ) ? 3 : 0 )      \
-: ( ( ( 0xEA == ((U8*)s)[0] ) && ( 0xA5 == ((U8*)s)[1] ) ) && ( 0xA0 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0xBC ) ) ? 3 : 0 )
-
-/*
-       GCB_LV_LVT_V: Grapheme_Cluster_Break=(LV or LVT or V)
-
-       \p{_X_LV_LVT_V}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_LV_LVT_V_utf8(s)                                             \
-( ( 0xE1 == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x85 == ((U8*)s)[1] ) ?                                             \
-       ( ( ((U8*)s)[2] >= 0xA0 ) ? 3 : 0 )                                 \
-    : ( ( 0x86 == ((U8*)s)[1] ) && ( ((U8*)s)[2] <= 0xA7 ) ) ? 3 : 0 )      \
-: ( 0xEA == ((U8*)s)[0] ) ?                                                 \
-    ( ( ((U8*)s)[1] >= 0xB0 ) ?                                             \
-       3                                                                   \
-    : 0 )                                                                   \
-: ( 0xEB == ((U8*)s)[0] || 0xEC == ((U8*)s)[0] ) ?                          \
-    3                                                                       \
-: ( 0xED == ((U8*)s)[0] ) ?                                                 \
-    ( ( ((U8*)s)[1] <= 0x9D ) ?                                             \
-       3                                                                   \
-    : ( 0x9E == ((U8*)s)[1] ) ?                                             \
-       ( ( ( ((U8*)s)[2] <= 0xA3 ) || ( ((U8*)s)[2] >= 0xB0 ) ) ? 3 : 0 )  \
-    : ( ( 0x9F == ((U8*)s)[1] ) && ( ((U8*)s)[2] <= 0x86 ) ) ? 3 : 0 )      \
-: 0 )
-
-/*
-       GCB_Prepend: Grapheme_Cluster_Break=Prepend
-
-       \p{_X_GCB_Prepend}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_Prepend_utf8(s)                                              \
-( 0 )
-
-/*
-       GCB_RI: Grapheme_Cluster_Break=RI
-
-       \p{_X_RI}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_RI_utf8(s)                                                   \
-( ( ( ( ( 0xF0 == ((U8*)s)[0] ) && ( 0x9F == ((U8*)s)[1] ) ) && ( 0x87 == ((U8*)s)[2] ) ) && ( ((U8*)s)[3] >= 0xA6 ) ) ? 4 : 0 )
-
-/*
-       GCB_SPECIAL_BEGIN_START: Grapheme_Cluster_Break=special_begin_starts
-
-       \p{_X_Special_Begin_Start}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_SPECIAL_BEGIN_START_utf8(s)                                  \
-( ( 0xE1 == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( ((U8*)s)[1] & 0xFC ) == 0x84 ) ?                                  \
-       3                                                                   \
-    : 0 )                                                                   \
-: ( 0xEA == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0xA5 == ((U8*)s)[1] ) ?                                             \
-       ( ( 0xA0 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0xBC ) ? 3 : 0 )          \
-    : ( ((U8*)s)[1] >= 0xB0 ) ?                                             \
-       3                                                                   \
-    : 0 )                                                                   \
-: ( 0xEB == ((U8*)s)[0] || 0xEC == ((U8*)s)[0] ) ?                          \
-    3                                                                       \
-: ( 0xED == ((U8*)s)[0] ) ?                                                 \
-    ( ( ((U8*)s)[1] <= 0x9D ) ?                                             \
-       3                                                                   \
-    : ( 0x9E == ((U8*)s)[1] ) ?                                             \
-       ( ( ( ((U8*)s)[2] <= 0xA3 ) || ( ((U8*)s)[2] >= 0xB0 ) ) ? 3 : 0 )  \
-    : ( ( 0x9F == ((U8*)s)[1] ) && ( ( ((U8*)s)[2] <= 0x86 ) || ( 0x8B <= ((U8*)s)[2] && ((U8*)s)[2] <= 0xBB ) ) ) ? 3 : 0 )\
-: ( ( ( ( 0xF0 == ((U8*)s)[0] ) && ( 0x9F == ((U8*)s)[1] ) ) && ( 0x87 == ((U8*)s)[2] ) ) && ( ((U8*)s)[3] >= 0xA6 ) ) ? 4 : 0 )
-
-/*
-       GCB_T: Grapheme_Cluster_Break=T
-
-       \p{_X_GCB_T}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_T_utf8(s)                                                    \
-( ( 0xE1 == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x86 == ((U8*)s)[1] ) ?                                             \
-       ( ( ((U8*)s)[2] >= 0xA8 ) ? 3 : 0 )                                 \
-    : ( 0x87 == ((U8*)s)[1] ) ?                                             \
-       3                                                                   \
-    : 0 )                                                                   \
-: ( ( ( 0xED == ((U8*)s)[0] ) && ( 0x9F == ((U8*)s)[1] ) ) && ( 0x8B <= ((U8*)s)[2] && ((U8*)s)[2] <= 0xBB ) ) ? 3 : 0 )
-
-/*
-       GCB_V: Grapheme_Cluster_Break=V
-
-       \p{_X_GCB_V}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_V_utf8(s)                                                    \
-( ( 0xE1 == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x85 == ((U8*)s)[1] ) ?                                             \
-       ( ( ((U8*)s)[2] >= 0xA0 ) ? 3 : 0 )                                 \
-    : ( ( 0x86 == ((U8*)s)[1] ) && ( ((U8*)s)[2] <= 0xA7 ) ) ? 3 : 0 )      \
-: ( 0xED == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x9E == ((U8*)s)[1] ) ?                                             \
-       ( ( ((U8*)s)[2] >= 0xB0 ) ? 3 : 0 )                                 \
-    : ( ( 0x9F == ((U8*)s)[1] ) && ( ((U8*)s)[2] <= 0x86 ) ) ? 3 : 0 )      \
-: 0 )
-
-/*
        QUOTEMETA: Meta-characters that \Q should quote
 
        \p{_Perl_Quotemeta}
@@ -904,113 +789,6 @@
 ( ( ( ( ( 0xDD == ((U8*)s)[0] ) && ( 0x65 == ((U8*)s)[1] || 0x66 == ((U8*)s)[1] ) ) && ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( ((U8*)s)[2] & 0xFC ) == 0x70 ) ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ) ? 4 : 0 )
 
 /*
-       GCB_L: Grapheme_Cluster_Break=L
-
-       \p{_X_GCB_L}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_L_utf8(s)                                                    \
-( ( 0xBB == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( ( 0x49 == ((U8*)s)[1] || 0x4A == ((U8*)s)[1] ) || 0x51 == ((U8*)s)[1] ) && ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( ((U8*)s)[2] & 0xFC ) == 0x70 ) ) ? 3 : 0 )\
-: ( ( ( ( 0xDD == ((U8*)s)[0] ) && ( 0x51 == ((U8*)s)[1] ) ) && ( 0x52 == ((U8*)s)[2] ) ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || 0x70 == ((U8*)s)[3] ) ) ? 4 : 0 )
-
-/*
-       GCB_LV_LVT_V: Grapheme_Cluster_Break=(LV or LVT or V)
-
-       \p{_X_LV_LVT_V}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_LV_LVT_V_utf8(s)                                             \
-( ( 0xBB == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( ((U8*)s)[1] & 0xFE ) == 0x52 ) ?                                  \
-       ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( ((U8*)s)[2] & 0xFC ) == 0x70 ) ? 3 : 0 )\
-    : ( ( 0x54 == ((U8*)s)[1] ) && ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x48 ) ) ? 3 : 0 )\
-: ( 0xDD == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( 0x52 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0x59 ) || ( ((U8*)s)[1] & 0xFE ) == 0x62 ) ?\
-       ( ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( ((U8*)s)[2] & 0xFC ) == 0x70 ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ) ? 4 : 0 )\
-    : ( 0x64 == ((U8*)s)[1] ) ?                                             \
-       ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || 0x70 == ((U8*)s)[2] ) ?\
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ? 4 : 0 )\
-       : ( 0x71 == ((U8*)s)[2] ) ?                                         \
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x44 ) || ( 0x57 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ? 4 : 0 )\
-       : ( ( 0x72 == ((U8*)s)[2] ) && ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x47 ) ) ? 4 : 0 )\
-    : 0 )                                                                   \
-: 0 )
-
-/*
-       GCB_Prepend: Grapheme_Cluster_Break=Prepend
-
-       \p{_X_GCB_Prepend}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_Prepend_utf8(s)                                              \
-( 0 )
-
-/*
-       GCB_RI: Grapheme_Cluster_Break=RI
-
-       \p{_X_RI}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_RI_utf8(s)                                                   \
-( ( ( ( ( 0xDF == ((U8*)s)[0] ) && ( 0x70 == ((U8*)s)[1] ) ) && ( 0x56 == ((U8*)s)[2] ) ) && ( ( 0x47 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ) ? 4 : 0 )
-
-/*
-       GCB_SPECIAL_BEGIN_START: Grapheme_Cluster_Break=special_begin_starts
-
-       \p{_X_Special_Begin_Start}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_SPECIAL_BEGIN_START_utf8(s)                                  \
-( ( 0xBB == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( ( 0x49 == ((U8*)s)[1] || 0x4A == ((U8*)s)[1] ) || ( 0x51 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0x56 ) ) && ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( ((U8*)s)[2] & 0xFC ) == 0x70 ) ) ? 3 : 0 )\
-: ( 0xDD == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x51 == ((U8*)s)[1] ) ?                                             \
-       ( ( ( 0x52 == ((U8*)s)[2] ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || 0x70 == ((U8*)s)[3] ) ) ? 4 : 0 )\
-    : ( ( 0x52 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0x59 ) || ( ((U8*)s)[1] & 0xFE ) == 0x62 ) ?\
-       ( ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( ((U8*)s)[2] & 0xFC ) == 0x70 ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ) ? 4 : 0 )\
-    : ( 0x64 == ((U8*)s)[1] ) ?                                             \
-       ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || 0x70 == ((U8*)s)[2] ) ?\
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ? 4 : 0 )\
-       : ( 0x71 == ((U8*)s)[2] ) ?                                         \
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x44 ) || ( 0x57 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ? 4 : 0 )\
-       : ( 0x72 == ((U8*)s)[2] ) ?                                         \
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x47 ) || ( 0x52 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ? 4 : 0 )\
-       : ( ( 0x73 == ((U8*)s)[2] ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) ) ) ? 4 : 0 )\
-    : 0 )                                                                   \
-: ( ( ( ( 0xDF == ((U8*)s)[0] ) && ( 0x70 == ((U8*)s)[1] ) ) && ( 0x56 == ((U8*)s)[2] ) ) && ( ( 0x47 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ) ? 4 : 0 )
-
-/*
-       GCB_T: Grapheme_Cluster_Break=T
-
-       \p{_X_GCB_T}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_T_utf8(s)                                                    \
-( ( 0xBB == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x54 == ((U8*)s)[1] ) ?                                             \
-       ( ( ( 0x49 == ((U8*)s)[2] || 0x4A == ((U8*)s)[2] ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( ((U8*)s)[2] & 0xFC ) == 0x70 ) ? 3 : 0 )\
-    : ( ( 0x55 == ((U8*)s)[1] || 0x56 == ((U8*)s)[1] ) && ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( ((U8*)s)[2] & 0xFC ) == 0x70 ) ) ? 3 : 0 )\
-: ( ( 0xDD == ((U8*)s)[0] ) && ( 0x64 == ((U8*)s)[1] ) ) ? ( ( 0x72 == ((U8*)s)[2] ) ?\
-           ( ( ( 0x52 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ? 4 : 0 )\
-       : ( ( 0x73 == ((U8*)s)[2] ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) ) ) ? 4 : 0 ) : 0 )
-
-/*
-       GCB_V: Grapheme_Cluster_Break=V
-
-       \p{_X_GCB_V}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_V_utf8(s)                                                    \
-( ( 0xBB == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( ((U8*)s)[1] & 0xFE ) == 0x52 ) ?                                  \
-       ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( ((U8*)s)[2] & 0xFC ) == 0x70 ) ? 3 : 0 )\
-    : ( ( 0x54 == ((U8*)s)[1] ) && ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x48 ) ) ? 3 : 0 )\
-: ( ( 0xDD == ((U8*)s)[0] ) && ( 0x64 == ((U8*)s)[1] ) ) ? ( ( 0x71 == ((U8*)s)[2] ) ?\
-           ( ( ( 0x57 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( ((U8*)s)[3] & 0xFC ) == 0x70 ) ? 4 : 0 )\
-       : ( ( 0x72 == ((U8*)s)[2] ) && ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x47 ) ) ? 4 : 0 ) : 0 )
-
-/*
        UTF8_CHAR: Matches legal UTF-EBCDIC encoded characters from 2 through 3 bytes
 
        0xA0 - 0x3FFF
@@ -1623,113 +1401,6 @@
 ( ( ( ( ( 0xDD == ((U8*)s)[0] ) && ( ( ((U8*)s)[1] & 0xFE ) == 0x64 ) ) && ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || 0x5F == ((U8*)s)[2] || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x72 ) ) ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ) ? 4 : 0 )
 
 /*
-       GCB_L: Grapheme_Cluster_Break=L
-
-       \p{_X_GCB_L}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_L_utf8(s)                                                    \
-( ( 0xBC == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( ( 0x49 == ((U8*)s)[1] || 0x4A == ((U8*)s)[1] ) || 0x51 == ((U8*)s)[1] ) && ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || 0x5F == ((U8*)s)[2] || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x72 ) ) ) ? 3 : 0 )\
-: ( ( ( ( 0xDD == ((U8*)s)[0] ) && ( 0x51 == ((U8*)s)[1] ) ) && ( 0x52 == ((U8*)s)[2] ) ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) ) ) ? 4 : 0 )
-
-/*
-       GCB_LV_LVT_V: Grapheme_Cluster_Break=(LV or LVT or V)
-
-       \p{_X_LV_LVT_V}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_LV_LVT_V_utf8(s)                                             \
-( ( 0xBC == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( ((U8*)s)[1] & 0xFE ) == 0x52 ) ?                                  \
-       ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || 0x5F == ((U8*)s)[2] || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x72 ) ) ? 3 : 0 )\
-    : ( ( 0x54 == ((U8*)s)[1] ) && ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x48 ) ) ? 3 : 0 )\
-: ( 0xDD == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( 0x52 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0x59 ) || 0x5F == ((U8*)s)[1] || 0x62 == ((U8*)s)[1] ) ?\
-       ( ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || 0x5F == ((U8*)s)[2] || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x72 ) ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ) ? 4 : 0 )\
-    : ( 0x63 == ((U8*)s)[1] ) ?                                             \
-       ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || 0x5F == ((U8*)s)[2] || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) ) ?\
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ? 4 : 0 )\
-       : ( 0x70 == ((U8*)s)[2] ) ?                                         \
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x44 ) || ( 0x57 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ? 4 : 0 )\
-       : ( ( 0x71 == ((U8*)s)[2] ) && ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x47 ) ) ? 4 : 0 )\
-    : 0 )                                                                   \
-: 0 )
-
-/*
-       GCB_Prepend: Grapheme_Cluster_Break=Prepend
-
-       \p{_X_GCB_Prepend}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_Prepend_utf8(s)                                              \
-( 0 )
-
-/*
-       GCB_RI: Grapheme_Cluster_Break=RI
-
-       \p{_X_RI}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_RI_utf8(s)                                                   \
-( ( ( ( ( 0xDF == ((U8*)s)[0] ) && ( 0x6A == ((U8*)s)[1] ) ) && ( 0x56 == ((U8*)s)[2] ) ) && ( ( 0x47 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ) ? 4 : 0 )
-
-/*
-       GCB_SPECIAL_BEGIN_START: Grapheme_Cluster_Break=special_begin_starts
-
-       \p{_X_Special_Begin_Start}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_SPECIAL_BEGIN_START_utf8(s)                                  \
-( ( 0xBC == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( ( 0x49 == ((U8*)s)[1] || 0x4A == ((U8*)s)[1] ) || ( 0x51 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0x56 ) ) && ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || 0x5F == ((U8*)s)[2] || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x72 ) ) ) ? 3 : 0 )\
-: ( 0xDD == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x51 == ((U8*)s)[1] ) ?                                             \
-       ( ( ( 0x52 == ((U8*)s)[2] ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) ) ) ? 4 : 0 )\
-    : ( ( 0x52 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0x59 ) || 0x5F == ((U8*)s)[1] || 0x62 == ((U8*)s)[1] ) ?\
-       ( ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || 0x5F == ((U8*)s)[2] || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x72 ) ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ) ? 4 : 0 )\
-    : ( 0x63 == ((U8*)s)[1] ) ?                                             \
-       ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || 0x5F == ((U8*)s)[2] || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) ) ?\
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ? 4 : 0 )\
-       : ( 0x70 == ((U8*)s)[2] ) ?                                         \
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x44 ) || ( 0x57 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ? 4 : 0 )\
-       : ( 0x71 == ((U8*)s)[2] ) ?                                         \
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x47 ) || ( 0x52 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ? 4 : 0 )\
-       : ( ( 0x72 == ((U8*)s)[2] ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) ) ) ? 4 : 0 )\
-    : 0 )                                                                   \
-: ( ( ( ( 0xDF == ((U8*)s)[0] ) && ( 0x6A == ((U8*)s)[1] ) ) && ( 0x56 == ((U8*)s)[2] ) ) && ( ( 0x47 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ) ? 4 : 0 )
-
-/*
-       GCB_T: Grapheme_Cluster_Break=T
-
-       \p{_X_GCB_T}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_T_utf8(s)                                                    \
-( ( 0xBC == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x54 == ((U8*)s)[1] ) ?                                             \
-       ( ( ( 0x49 == ((U8*)s)[2] || 0x4A == ((U8*)s)[2] ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || 0x5F == ((U8*)s)[2] || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x72 ) ) ? 3 : 0 )\
-    : ( ( 0x55 == ((U8*)s)[1] || 0x56 == ((U8*)s)[1] ) && ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || 0x5F == ((U8*)s)[2] || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x72 ) ) ) ? 3 : 0 )\
-: ( ( 0xDD == ((U8*)s)[0] ) && ( 0x63 == ((U8*)s)[1] ) ) ? ( ( 0x71 == ((U8*)s)[2] ) ?\
-           ( ( ( 0x52 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ? 4 : 0 )\
-       : ( ( 0x72 == ((U8*)s)[2] ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x4A ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) ) ) ? 4 : 0 ) : 0 )
-
-/*
-       GCB_V: Grapheme_Cluster_Break=V
-
-       \p{_X_GCB_V}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_V_utf8(s)                                                    \
-( ( 0xBC == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( ((U8*)s)[1] & 0xFE ) == 0x52 ) ?                                  \
-       ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x4A ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || 0x5F == ((U8*)s)[2] || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x6A ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x72 ) ) ? 3 : 0 )\
-    : ( ( 0x54 == ((U8*)s)[1] ) && ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x48 ) ) ? 3 : 0 )\
-: ( ( 0xDD == ((U8*)s)[0] ) && ( 0x63 == ((U8*)s)[1] ) ) ? ( ( 0x70 == ((U8*)s)[2] ) ?\
-           ( ( ( 0x57 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || 0x5F == ((U8*)s)[3] || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x6A ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ? 4 : 0 )\
-       : ( ( 0x71 == ((U8*)s)[2] ) && ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x47 ) ) ? 4 : 0 ) : 0 )
-
-/*
        UTF8_CHAR: Matches legal UTF-EBCDIC encoded characters from 2 through 3 bytes
 
        0xA0 - 0x3FFF
@@ -2350,113 +2021,6 @@
 ( ( ( ( ( 0xDC == ((U8*)s)[0] ) && ( ( ((U8*)s)[1] & 0xFE ) == 0x66 ) ) && ( ( ( ((U8*)s)[2] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[2] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[2] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[2] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[2] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[2] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[2] & 0xFA ) == 0x70 ) ) ) && ( ( ( ((U8*)s)[3] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[3] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[3] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[3] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[3] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[3] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[3] & 0xFA ) == 0x70 ) ) ) ? 4 : 0 )
 
 /*
-       GCB_L: Grapheme_Cluster_Break=L
-
-       \p{_X_GCB_L}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_L_utf8(s)                                                    \
-( ( 0xBA == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( 0x49 == ((U8*)s)[1] || ( 0x51 == ((U8*)s)[1] || 0x52 == ((U8*)s)[1] ) ) && ( ( ( ((U8*)s)[2] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[2] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[2] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[2] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[2] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[2] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[2] & 0xFA ) == 0x70 ) ) ) ? 3 : 0 )\
-: ( ( ( ( 0xDC == ((U8*)s)[0] ) && ( 0x52 == ((U8*)s)[1] ) ) && ( 0x53 == ((U8*)s)[2] ) ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x49 ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ) ? 4 : 0 )
-
-/*
-       GCB_LV_LVT_V: Grapheme_Cluster_Break=(LV or LVT or V)
-
-       \p{_X_LV_LVT_V}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_LV_LVT_V_utf8(s)                                             \
-( ( 0xBA == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x53 == ((U8*)s)[1] || 0x54 == ((U8*)s)[1] ) ?                      \
-       ( ( ( ( ((U8*)s)[2] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[2] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[2] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[2] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[2] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[2] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[2] & 0xFA ) == 0x70 ) ) ? 3 : 0 )\
-    : ( ( 0x55 == ((U8*)s)[1] ) && ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x48 ) ) ? 3 : 0 )\
-: ( 0xDC == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( 0x53 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0x59 ) || ( 0x62 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0x64 ) ) ?\
-       ( ( ( ( ( ((U8*)s)[2] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[2] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[2] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[2] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[2] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[2] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[2] & 0xFA ) == 0x70 ) ) && ( ( ( ((U8*)s)[3] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[3] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[3] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[3] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[3] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[3] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[3] & 0xFA ) == 0x70 ) ) ) ? 4 : 0 )\
-    : ( 0x65 == ((U8*)s)[1] ) ?                                             \
-       ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x49 ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x69 ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x72 ) ) ?\
-           ( ( ( ( ((U8*)s)[3] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[3] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[3] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[3] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[3] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[3] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[3] & 0xFA ) == 0x70 ) ) ? 4 : 0 )\
-       : ( 0x73 == ((U8*)s)[2] ) ?                                         \
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x44 ) || ( ((U8*)s)[3] & 0xFE ) == 0x58 || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x75 ) ) ? 4 : 0 )\
-       : ( ( 0x74 == ((U8*)s)[2] ) && ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x47 ) ) ? 4 : 0 )\
-    : 0 )                                                                   \
-: 0 )
-
-/*
-       GCB_Prepend: Grapheme_Cluster_Break=Prepend
-
-       \p{_X_GCB_Prepend}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_Prepend_utf8(s)                                              \
-( 0 )
-
-/*
-       GCB_RI: Grapheme_Cluster_Break=RI
-
-       \p{_X_RI}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_RI_utf8(s)                                                   \
-( ( ( ( ( 0xDE == ((U8*)s)[0] ) && ( 0x72 == ((U8*)s)[1] ) ) && ( 0x57 == ((U8*)s)[2] ) ) && ( ( 0x47 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x49 ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x75 ) ) ) ? 4 : 0 )
-
-/*
-       GCB_SPECIAL_BEGIN_START: Grapheme_Cluster_Break=special_begin_starts
-
-       \p{_X_Special_Begin_Start}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_SPECIAL_BEGIN_START_utf8(s)                                  \
-( ( 0xBA == ((U8*)s)[0] ) ?                                                 \
-    ( ( ( 0x49 == ((U8*)s)[1] || ( 0x51 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0x57 ) ) && ( ( ( ((U8*)s)[2] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[2] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[2] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[2] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[2] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[2] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[2] & 0xFA ) == 0x70 ) ) ) ? 3 : 0 )\
-: ( 0xDC == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x52 == ((U8*)s)[1] ) ?                                             \
-       ( ( ( 0x53 == ((U8*)s)[2] ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x49 ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x72 ) ) ) ? 4 : 0 )\
-    : ( ( 0x53 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0x59 ) || ( 0x62 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0x64 ) ) ?\
-       ( ( ( ( ( ((U8*)s)[2] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[2] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[2] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[2] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[2] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[2] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[2] & 0xFA ) == 0x70 ) ) && ( ( ( ((U8*)s)[3] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[3] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[3] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[3] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[3] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[3] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[3] & 0xFA ) == 0x70 ) ) ) ? 4 : 0 )\
-    : ( 0x65 == ((U8*)s)[1] ) ?                                             \
-       ( ( ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x49 ) || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x69 ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x72 ) ) ?\
-           ( ( ( ( ((U8*)s)[3] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[3] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[3] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[3] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[3] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[3] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[3] & 0xFA ) == 0x70 ) ) ? 4 : 0 )\
-       : ( 0x73 == ((U8*)s)[2] ) ?                                         \
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x44 ) || ( ((U8*)s)[3] & 0xFE ) == 0x58 || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x75 ) ) ? 4 : 0 )\
-       : ( 0x74 == ((U8*)s)[2] ) ?                                         \
-           ( ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x47 ) || ( 0x53 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x75 ) ) ? 4 : 0 )\
-       : ( ( 0x75 == ((U8*)s)[2] ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x49 ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) || ( ((U8*)s)[3] & 0xFE ) == 0x70 ) ) ? 4 : 0 )\
-    : 0 )                                                                   \
-: ( ( ( ( 0xDE == ((U8*)s)[0] ) && ( 0x72 == ((U8*)s)[1] ) ) && ( 0x57 == ((U8*)s)[2] ) ) && ( ( 0x47 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x49 ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x75 ) ) ) ? 4 : 0 )
-
-/*
-       GCB_T: Grapheme_Cluster_Break=T
-
-       \p{_X_GCB_T}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_T_utf8(s)                                                    \
-( ( 0xBA == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x55 == ((U8*)s)[1] ) ?                                             \
-       ( ( 0x49 == ((U8*)s)[2] || ( 0x51 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x59 ) || ( 0x62 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x69 ) || ( 0x70 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x75 ) ) ? 3 : 0 )\
-    : ( ( ( ((U8*)s)[1] & 0xFE ) == 0x56 ) && ( ( ( ((U8*)s)[2] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[2] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[2] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[2] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[2] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[2] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[2] & 0xFA ) == 0x70 ) ) ) ? 3 : 0 )\
-: ( ( 0xDC == ((U8*)s)[0] ) && ( 0x65 == ((U8*)s)[1] ) ) ? ( ( 0x74 == ((U8*)s)[2] ) ?\
-           ( ( ( 0x53 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x75 ) ) ? 4 : 0 )\
-       : ( ( 0x75 == ((U8*)s)[2] ) && ( ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x49 ) || ( 0x51 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x59 ) || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) || ( ((U8*)s)[3] & 0xFE ) == 0x70 ) ) ? 4 : 0 ) : 0 )
-
-/*
-       GCB_V: Grapheme_Cluster_Break=V
-
-       \p{_X_GCB_V}
-*/
-/*** GENERATED CODE ***/
-#define is_GCB_V_utf8(s)                                                    \
-( ( 0xBA == ((U8*)s)[0] ) ?                                                 \
-    ( ( 0x53 == ((U8*)s)[1] || 0x54 == ((U8*)s)[1] ) ?                      \
-       ( ( ( ( ((U8*)s)[2] & 0xEF ) == 0x41 ) || ( ( ((U8*)s)[2] & 0xCE ) == 0x42 ) || ( ( ((U8*)s)[2] & 0xEC ) == 0x44 ) || ( ( ((U8*)s)[2] & 0xEE ) == 0x48 ) || ( ( ((U8*)s)[2] & 0xFC ) == 0x64 ) || ( ( ((U8*)s)[2] & 0xFE ) == 0x68 ) || ( ( ((U8*)s)[2] & 0xFA ) == 0x70 ) ) ? 3 : 0 )\
-    : ( ( 0x55 == ((U8*)s)[1] ) && ( 0x41 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x48 ) ) ? 3 : 0 )\
-: ( ( 0xDC == ((U8*)s)[0] ) && ( 0x65 == ((U8*)s)[1] ) ) ? ( ( 0x73 == ((U8*)s)[2] ) ?\
-           ( ( ( ((U8*)s)[3] & 0xFE ) == 0x58 || ( 0x62 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x69 ) || ( 0x70 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x75 ) ) ? 4 : 0 )\
-       : ( ( 0x74 == ((U8*)s)[2] ) && ( 0x41 <= ((U8*)s)[3] && ((U8*)s)[3] <= 0x47 ) ) ? 4 : 0 ) : 0 )
-
-/*
        UTF8_CHAR: Matches legal UTF-EBCDIC encoded characters from 2 through 3 bytes
 
        0xA0 - 0x3FFF
@@ -2908,5 +2472,5 @@
 #endif /* H_REGCHARCLASS */
 
 /* Generated from:
- * d3ce45436d251560bef015ee209e8a0661a0e564216585e9e84f25ec8ac715ae regen/regcharclass.pl
+ * 76d3d07dadbba1440342f5ab3701db8c9f6df5a851b1cfd08f7b09014d4efc52 regen/regcharclass.pl
  * ex: set ro: */

diff --git a/regen/regcharclass.pl b/regen/regcharclass.pl
index cd19563..9a83e54 100755 (executable)
--- a/regen/regcharclass.pl
+++ b/regen/regcharclass.pl
@@ -1613,34 +1613,6 @@ SURROGATE: Surrogate characters
 => UTF8 :fast
 \p{Gc=Cs}
 
-GCB_L: Grapheme_Cluster_Break=L
-=> UTF8 :fast
-\p{_X_GCB_L}
-
-GCB_LV_LVT_V: Grapheme_Cluster_Break=(LV or LVT or V)
-=> UTF8 :fast
-\p{_X_LV_LVT_V}
-
-GCB_Prepend: Grapheme_Cluster_Break=Prepend
-=> UTF8 :fast
-\p{_X_GCB_Prepend}
-
-GCB_RI: Grapheme_Cluster_Break=RI
-=> UTF8 :fast
-\p{_X_RI}
-
-GCB_SPECIAL_BEGIN_START: Grapheme_Cluster_Break=special_begin_starts
-=> UTF8 :fast
-\p{_X_Special_Begin_Start}
-
-GCB_T: Grapheme_Cluster_Break=T
-=> UTF8 :fast
-\p{_X_GCB_T}
-
-GCB_V: Grapheme_Cluster_Break=V
-=> UTF8 :fast
-\p{_X_GCB_V}
-
 # This program was run with this enabled, and the results copied to utf8.h;
 # then this was commented out because it takes so long to figure out these 2
 # million code points.  The results would not change unless utf8.h decides it

diff --git a/sv.c b/sv.c
index 3b65510..76cc801 100644 (file)
--- a/sv.c
+++ b/sv.c
@@ -14946,8 +14946,6 @@ perl_clone_using(PerlInterpreter *proto_perl, UV flags,
     }
     PL_GCB_invlist = sv_dup_inc(proto_perl->IGCB_invlist, param);
     PL_utf8_mark       = sv_dup_inc(proto_perl->Iutf8_mark, param);
-    PL_utf8_X_regular_begin    = sv_dup_inc(proto_perl->Iutf8_X_regular_begin, param);
-    PL_utf8_X_extend   = sv_dup_inc(proto_perl->Iutf8_X_extend, param);
     PL_utf8_toupper    = sv_dup_inc(proto_perl->Iutf8_toupper, param);
     PL_utf8_totitle    = sv_dup_inc(proto_perl->Iutf8_totitle, param);
     PL_utf8_tolower    = sv_dup_inc(proto_perl->Iutf8_tolower, param);