Synchronize the specifications of the POSIX character

classes alnum, graph, and print closer to the planned
Unicode proposal.

p4raw-id: //depot/perl@19231

diff --git a/lib/unicore/lib/Alnum.pl b/lib/unicore/lib/Alnum.pl
index 0ec13c5..5f2b1f0 100644 (file)
--- a/lib/unicore/lib/Alnum.pl
+++ b/lib/unicore/lib/Alnum.pl
@@ -13,10 +13,8 @@ return <<'END';
 0041   005A    
 0061   007A    
 00AA           
-00B2   00B3    
 00B5           
-00B9   00BA    
-00BC   00BE    
+00BA           
 00C0   00D6    
 00D8   00F6    
 00F8   0220    
@@ -85,7 +83,6 @@ return <<'END';
 09DC   09DD    
 09DF   09E3    
 09E6   09F1    
-09F4   09F9    
 0A02           
 0A05   0A0A    
 0A0F   0A10    
@@ -144,7 +141,7 @@ return <<'END';
 0BC6   0BC8    
 0BCA   0BCD    
 0BD7           
-0BE7   0BF2    
+0BE7   0BEF    
 0C01   0C03    
 0C05   0C0C    
 0C0E   0C10    
@@ -215,7 +212,7 @@ return <<'END';
 0EDC   0EDD    
 0F00           
 0F18   0F19    
-0F20   0F33    
+0F20   0F29    
 0F35           
 0F37           
 0F39           
@@ -263,13 +260,12 @@ return <<'END';
 1318   131E    
 1320   1346    
 1348   135A    
-1369   137C    
+1369   1371    
 13A0   13F4    
 1401   166C    
 166F   1676    
 1681   169A    
 16A0   16EA    
-16EE   16F0    
 1700   170C    
 170E   1714    
 1720   1734    
@@ -306,9 +302,8 @@ return <<'END';
 1FE0   1FEC    
 1FF2   1FF4    
 1FF6   1FFC    
-2070   2071    
-2074   2079    
-207F   2089    
+2071           
+207F           
 20D0   20EA    
 2102           
 2107           
@@ -323,14 +318,10 @@ return <<'END';
 2133   2139    
 213D   213F    
 2145   2149    
-2153   2183    
-2460   249B    
-24EA   24FE    
-2776   2793    
-3005   3007    
-3021   302F    
+3005   3006    
+302A   302F    
 3031   3035    
-3038   303C    
+303B   303C    
 3041   3096    
 3099   309A    
 309D   309F    
@@ -338,13 +329,8 @@ return <<'END';
 30FC   30FF    
 3105   312C    
 3131   318E    
-3192   3195    
 31A0   31B7    
 31F0   31FF    
-3220   3229    
-3251   325F    
-3280   3289    
-32B1   32BF    
 3400   4DB5    
 4E00   9FA5    
 A000   A48C    
@@ -377,8 +363,7 @@ FFCA        FFCF
 FFD2   FFD7    
 FFDA   FFDC    
 10300  1031E   
-10320  10323   
-10330  1034A   
+10330  10349   
 10400  10425   
 10428  1044D   
 1D165  1D169   

diff --git a/lib/unicore/lib/Graph.pl b/lib/unicore/lib/Graph.pl
index 212c4c6..1eff89d 100644 (file)
--- a/lib/unicore/lib/Graph.pl
+++ b/lib/unicore/lib/Graph.pl
@@ -277,7 +277,7 @@ return <<'END';
 1FDD   1FEF    
 1FF2   1FF4    
 1FF6   1FFE    
-2010   2027    
+2010   2029    
 2030   2052    
 2057           
 2070   2071    
@@ -331,7 +331,7 @@ return <<'END';
 A000   A48C    
 A490   A4C6    
 AC00   D7A3    
-E000   FA2D    
+F900   FA2D    
 FA30   FA6A    
 FB00   FB06    
 FB13   FB17    
@@ -394,6 +394,4 @@ FFFC        FFFD
 1D7CE  1D7FF   
 20000  2A6D6   
 2F800  2FA1D   
-F0000  FFFFD   
-100000 10FFFD  
 END

diff --git a/lib/unicore/lib/Print.pl b/lib/unicore/lib/Print.pl
index d1eb1e2..ea9c553 100644 (file)
--- a/lib/unicore/lib/Print.pl
+++ b/lib/unicore/lib/Print.pl
@@ -278,7 +278,7 @@ return <<'END';
 1FF2   1FF4    
 1FF6   1FFE    
 2000   200B    
-2010   2027    
+2010   2029    
 202F   2052    
 2057           
 205F           
@@ -333,7 +333,7 @@ return <<'END';
 A000   A48C    
 A490   A4C6    
 AC00   D7A3    
-E000   FA2D    
+F900   FA2D    
 FA30   FA6A    
 FB00   FB06    
 FB13   FB17    
@@ -396,6 +396,4 @@ FFFC        FFFD
 1D7CE  1D7FF   
 20000  2A6D6   
 2F800  2FA1D   
-F0000  FFFFD   
-100000 10FFFD  
 END

diff --git a/lib/unicore/mktables b/lib/unicore/mktables
index 654301e..fa07346 100644 (file)
--- a/lib/unicore/mktables
+++ b/lib/unicore/mktables
@@ -2,6 +2,9 @@
 use strict;
 use Carp;
 
+die "$0: Please run me as ./mktables to avoid unnecessary differences\n"
+    unless $0 eq "./mktables";
+
 ##
 ## mktables -- create the runtime Perl Unicode files (lib/unicore/**/*.pl)
 ## from the Unicode database files (lib/unicore/*.txt).
@@ -625,7 +628,7 @@ sub UnicodeData_Txt()
 
         # 005F: SPACING UNDERSCORE
         $Cat{Word}->$op($code)  if $cat =~ /^[LMN]/ || $code == 0x005F;
-        $Cat{Alnum}->$op($code) if $cat =~ /^[LMN]/;
+        $Cat{Alnum}->$op($code) if $cat =~ /^[LM]|Nd/;
         $Cat{Alpha}->$op($code) if $cat =~ /^[LM]/;
 
 
@@ -647,7 +650,7 @@ sub UnicodeData_Txt()
                                     || $code == 0x2028 # 2028: LINE SEPARATOR
                                     || $code == 0x2029;# 2029: PARAGRAPH SEP.
 
-        $Cat{Blank}->$op($code) if $cat  =~ /^Z[^lp]$/
+        $Cat{Blank}->$op($code) if $cat eq "Zs"
                                 || $code == 0x0009  # 0009: HORIZONTAL TAB
                                 || $code == 0x0020; # 0020: SPACE
 
@@ -657,8 +660,9 @@ sub UnicodeData_Txt()
         $Cat{Title}->$op($code) if $cat eq "Lt";
         $Cat{ASCII}->$op($code) if $code <= 0x007F;
         $Cat{Cntrl}->$op($code) if $cat =~ /^C/;
-        $Cat{Graph}->$op($code) if $cat =~ /^([LMNPS]|Co)/;
-        $Cat{Print}->$op($code) if $cat =~ /^([LMNPS]|Co|Zs)/;
+        $Cat{Graph}->$op($code) if $cat =~ /^[^C]/
+                                && $cat ne "Zs";
+        $Cat{Print}->$op($code) if $cat =~ /^[^C]/;
         $Cat{Punct}->$op($code) if $cat =~ /^P/;
 
         $Cat{XDigit}->$op($code) if ($code >= 0x30 && $code <= 0x39)  ## 0..9

diff --git a/utf8.c b/utf8.c
index b112d4c..0dd9ad8 100644 (file)
--- a/utf8.c
+++ b/utf8.c
@@ -1235,7 +1235,7 @@ Perl_is_utf8_upper(pTHX_ U8 *p)
     if (!is_utf8_char(p))
        return FALSE;
     if (!PL_utf8_upper)
-       PL_utf8_upper = swash_init("utf8", "IsUpper", &PL_sv_undef, 0, 0);
+       PL_utf8_upper = swash_init("utf8", "IsUppercase", &PL_sv_undef, 0, 0);
     return swash_fetch(PL_utf8_upper, p, TRUE) != 0;
 }
 
@@ -1245,7 +1245,7 @@ Perl_is_utf8_lower(pTHX_ U8 *p)
     if (!is_utf8_char(p))
        return FALSE;
     if (!PL_utf8_lower)
-       PL_utf8_lower = swash_init("utf8", "IsLower", &PL_sv_undef, 0, 0);
+       PL_utf8_lower = swash_init("utf8", "IsLowercase", &PL_sv_undef, 0, 0);
     return swash_fetch(PL_utf8_lower, p, TRUE) != 0;
 }