Add Yves Orton's script to regenerate regcharclass.h

p4raw-id: //depot/perl@31030

1 files changed, 117 insertions, 53 deletions

diff --git a/regcharclass.h b/regcharclass.h
index 3fc92d7bce..08b9eaeca3 100644
--- a/regcharclass.h
+++ b/regcharclass.h
@@ -1,13 +1,29 @@
-/****** WARNING WARNING WARNING ********/
-/*                                     */
-/* Autogenerated code, do not modify!  */
-/*                                     */
-/****** WARNING WARNING WARNING ********/
-
-/* LNBREAK Line Break: \j \J 
-  Codepoints:  0x0A, 0x0B, 0x0C, 0x0D, 0x0D.0x0A, 0x85, 0x2028, 0x2029
-  */
-#define is_LNBREAK(s,is_utf8) /*** Line Break: \j \J ***/                   \
+/*********************** WARNING WARNING WARNING ************************
+
+Do not modify this code directly: This file was autogenerated by 
+
+        Porting/regcharclass.pl 
+        
+from data contained within the script.  Change the script instead.
+
+Generated at: Mon Apr 23 11:33:52 2007 GMT
+
+************************ WARNING WARNING WARNING ************************/
+
+/*
+	LNBREAK: Line Break: \R
+
+	"\x0D\x0A"      # CRLF - Network (Windows) line ending 
+	0x0A            # LF  | LINE FEED
+	0x0B            # VT  | VERTICAL TAB
+	0x0C            # FF  | FORM FEED
+	0x0D            # CR  | CARRIAGE RETURN
+	0x85            # NEL | NEXT LINE
+	0x2028          # LINE SEPARATOR
+	0x2029          # PARAGRAPH SEPARATOR
+*/
+/*** GENERATED CODE ***/
+#define is_LNBREAK(s,is_utf8)                                               \
 ( ( ((U8*)s)[0]==13 ) ?                                                     \
     ( ( ((U8*)s)[1]==10 ) ? 2 : 1 ) :                                       \
   ( (10<=((U8*)s)[0] && ((U8*)s)[0]<=12) ? 1 :                              \
@@ -17,7 +33,8 @@
   (((( ((U8*)s)[0]==226 ) && ( ((U8*)s)[1]==128 )) && ( ((U8*)s)[2]==168 || ((U8*)s)[2]==169 )) ? 3 : 0) ) :\
   ( ((U8*)s)[0]==133 ) ) ) )                                                
 
-#define is_LNBREAK_safe(s,e,is_utf8) /*** Line Break: \j \J ***/            \
+/*** GENERATED CODE ***/
+#define is_LNBREAK_safe(s,e,is_utf8)                                        \
 ( ( (e) - (s) > 2 ) ?                                                       \
   ( ( ((U8*)s)[0]==13 ) ?                                                   \
     ( ( ((U8*)s)[1]==10 ) ? 2 : 1 ) :                                       \
@@ -39,7 +56,8 @@
 ( (!is_utf8) ?                                                              \
   ( ((U8*)s)[0]==133 ) : 0 ) ) : 0 ) ) )                                    
 
-#define is_LNBREAK_utf8(s) /*** Line Break: \j \J ***/                      \
+/*** GENERATED CODE ***/
+#define is_LNBREAK_utf8(s)                                                  \
 ( ( ((U8*)s)[0]==194 ) ?                                                    \
     ( ( ((U8*)s)[1]==133 ) ? 2 : 0 ) :                                      \
   ( ( ((U8*)s)[0]==226 ) ?                                                  \
@@ -48,7 +66,8 @@
     ( ( ((U8*)s)[1]==10 ) ? 2 : 1 ) :                                       \
   (10<=((U8*)s)[0] && ((U8*)s)[0]<=12) ) ) )                                
 
-#define is_LNBREAK_utf8_safe(s,e) /*** Line Break: \j \J ***/               \
+/*** GENERATED CODE ***/
+#define is_LNBREAK_utf8_safe(s,e)                                           \
 ( ( (e) - (s) > 2 ) ?                                                       \
   ( ( ((U8*)s)[0]==194 ) ?                                                  \
     ( ( ((U8*)s)[1]==133 ) ? 2 : 0 ) :                                      \
@@ -66,12 +85,14 @@
 ( ( (e) - (s) > 0 ) ?                                                       \
   (10<=((U8*)s)[0] && ((U8*)s)[0]<=13) : 0 ) ) )                            
 
-#define is_LNBREAK_latin1(s) /*** Line Break: \j \J ***/                    \
+/*** GENERATED CODE ***/
+#define is_LNBREAK_latin1(s)                                                \
 ( ( ((U8*)s)[0]==13 ) ?                                                     \
     ( ( ((U8*)s)[1]==10 ) ? 2 : 1 ) :                                       \
   ( (10<=((U8*)s)[0] && ((U8*)s)[0]<=12) || ((U8*)s)[0]==133 ) )            
 
-#define is_LNBREAK_latin1_safe(s,e) /*** Line Break: \j \J ***/             \
+/*** GENERATED CODE ***/
+#define is_LNBREAK_latin1_safe(s,e)                                         \
 ( ( (e) - (s) > 1 ) ?                                                       \
   ( ( ((U8*)s)[0]==13 ) ?                                                   \
     ( ( ((U8*)s)[1]==10 ) ? 2 : 1 ) :                                       \
@@ -79,21 +100,38 @@
 ( ( (e) - (s) > 0 ) ?                                                       \
   ( (10<=((U8*)s)[0] && ((U8*)s)[0]<=13) || ((U8*)s)[0]==133 ) : 0 ) )      
 
-#define is_LNBREAK_cp(cp) /*** Line Break: \j \J ***/                       \
-( (10<=cp && cp<=13) || ( cp > 13 && ( cp==133 || ( cp > 133 && ( cp==8232 || ( cp > 8232 && cp==8233 ) ) ) ) ) )
-
-/****** WARNING WARNING WARNING ********/
-/*                                     */
-/* Autogenerated code, do not modify!  */
-/*                                     */
-/****** WARNING WARNING WARNING ********/
-
-/* HORIZWS Horizontal Whitespace: \h \H 
-  Codepoints:  0x09, 0x20, 0xA0, 0x1680, 0x180E, 0x2000, 0x2001, 0x2002,
-  0x2003, 0x2004, 0x2005, 0x2006, 0x2007, 0x2008, 0x2009, 0x200A, 0x202F,
-  0x205F, 0x3000
-  */
-#define is_HORIZWS(s,is_utf8) /*** Horizontal Whitespace: \h \H ***/        \
+/*** GENERATED CODE ***/
+#define is_LNBREAK_cp(cp)                                                   \
+( (10<=cp && cp<=13) || ( cp > 13 &&                                        \
+( cp==133 || ( cp > 133 &&                                                  \
+( cp==8232 || ( cp > 8232 &&                                                \
+cp==8233 ) ) ) ) ) )                                                        
+
+/*
+	HORIZWS: Horizontal Whitespace: \h \H
+
+	0x09            # HT
+	0x20            # SPACE
+	0xa0            # NBSP
+	0x1680          # OGHAM SPACE MARK
+	0x180e          # MONGOLIAN VOWEL SEPARATOR
+	0x2000          # EN QUAD
+	0x2001          # EM QUAD
+	0x2002          # EN SPACE
+	0x2003          # EM SPACE
+	0x2004          # THREE-PER-EM SPACE
+	0x2005          # FOUR-PER-EM SPACE
+	0x2006          # SIX-PER-EM SPACE
+	0x2007          # FIGURE SPACE
+	0x2008          # PUNCTUATION SPACE
+	0x2009          # THIN SPACE
+	0x200A          # HAIR SPACE
+	0x202f          # NARROW NO-BREAK SPACE
+	0x205f          # MEDIUM MATHEMATICAL SPACE
+	0x3000          # IDEOGRAPHIC SPACE
+*/
+/*** GENERATED CODE ***/
+#define is_HORIZWS(s,is_utf8)                                               \
 ( ( ((U8*)s)[0]==9 || ((U8*)s)[0]==32 ) ? 1 :                               \
 ( (is_utf8) ?                                                               \
   ( ( ((U8*)s)[0]==194 ) ?                                                  \
@@ -109,7 +147,8 @@
   (((( ((U8*)s)[0]==227 ) && ( ((U8*)s)[1]==128 )) && ( ((U8*)s)[2]==128 )) ? 3 : 0) ) ) ) :\
   ( ((U8*)s)[0]==160 ) ) )                                                  
 
-#define is_HORIZWS_safe(s,e,is_utf8) /*** Horizontal Whitespace: \h \H ***/ \
+/*** GENERATED CODE ***/
+#define is_HORIZWS_safe(s,e,is_utf8)                                        \
 ( ( (e) - (s) > 2 ) ?                                                       \
   ( ( ((U8*)s)[0]==9 || ((U8*)s)[0]==32 ) ? 1 :                             \
 ( (is_utf8) ?                                                               \
@@ -135,7 +174,8 @@
 ( (!is_utf8) ?                                                              \
   ( ((U8*)s)[0]==160 ) : 0 ) ) : 0 ) ) )                                    
 
-#define is_HORIZWS_utf8(s) /*** Horizontal Whitespace: \h \H ***/           \
+/*** GENERATED CODE ***/
+#define is_HORIZWS_utf8(s)                                                  \
 ( ( ((U8*)s)[0]==194 ) ?                                                    \
     ( ( ((U8*)s)[1]==160 ) ? 2 : 0 ) :                                      \
   ( ( ((U8*)s)[0]==225 ) ?                                                  \
@@ -150,7 +190,8 @@
     ((( ((U8*)s)[1]==128 ) && ( ((U8*)s)[2]==128 )) ? 3 : 0) :              \
   ( ((U8*)s)[0]==9 || ((U8*)s)[0]==32 ) ) ) ) )                             
 
-#define is_HORIZWS_utf8_safe(s,e) /*** Horizontal Whitespace: \h \H ***/    \
+/*** GENERATED CODE ***/
+#define is_HORIZWS_utf8_safe(s,e)                                           \
 ( ( (e) - (s) > 2 ) ?                                                       \
   ( ( ((U8*)s)[0]==194 ) ?                                                  \
     ( ( ((U8*)s)[1]==160 ) ? 2 : 0 ) :                                      \
@@ -172,26 +213,40 @@
 ( ( (e) - (s) > 0 ) ?                                                       \
   ( ((U8*)s)[0]==9 || ((U8*)s)[0]==32 ) : 0 ) ) )                           
 
-#define is_HORIZWS_latin1(s) /*** Horizontal Whitespace: \h \H ***/         \
+/*** GENERATED CODE ***/
+#define is_HORIZWS_latin1(s)                                                \
 ( ((U8*)s)[0]==9 || ((U8*)s)[0]==32 || ((U8*)s)[0]==160 )                   
 
-#define is_HORIZWS_latin1_safe(s,e) /*** Horizontal Whitespace: \h \H ***/  \
+/*** GENERATED CODE ***/
+#define is_HORIZWS_latin1_safe(s,e)                                         \
 ( ( (e) - (s) > 0 ) ?                                                       \
   ( ((U8*)s)[0]==9 || ((U8*)s)[0]==32 || ((U8*)s)[0]==160 ) : 0 )           
 
-#define is_HORIZWS_cp(cp) /*** Horizontal Whitespace: \h \H ***/            \
-( cp==9 || ( cp > 9 && ( cp==32 || ( cp > 32 && ( cp==160 || ( cp > 160 && ( cp==5760 || ( cp > 5760 && ( cp==6158 || ( cp > 6158 && ( (8192<=cp && cp<=8202) || ( cp > 8202 && ( cp==8239 || ( cp > 8239 && ( cp==8287 || ( cp > 8287 && cp==12288 ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) )
+/*** GENERATED CODE ***/
+#define is_HORIZWS_cp(cp)                                                   \
+( cp==9 || ( cp > 9 &&                                                      \
+( cp==32 || ( cp > 32 &&                                                    \
+( cp==160 || ( cp > 160 &&                                                  \
+( cp==5760 || ( cp > 5760 &&                                                \
+( cp==6158 || ( cp > 6158 &&                                                \
+( (8192<=cp && cp<=8202) || ( cp > 8202 &&                                  \
+( cp==8239 || ( cp > 8239 &&                                                \
+( cp==8287 || ( cp > 8287 &&                                                \
+cp==12288 ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) )                                   
 
-/****** WARNING WARNING WARNING ********/
-/*                                     */
-/* Autogenerated code, do not modify!  */
-/*                                     */
-/****** WARNING WARNING WARNING ********/
+/*
+	VERTWS: Vertical Whitespace: \v \V
 
-/* VERTWS Vertical Whitespace: \v \V 
-  Codepoints:  0x0A, 0x0B, 0x0C, 0x0D, 0x85, 0x2028, 0x2029
-  */
-#define is_VERTWS(s,is_utf8) /*** Vertical Whitespace: \v \V ***/           \
+	0x0A            # LF
+	0x0B            # VT
+	0x0C            # FF  
+	0x0D            # CR
+	0x85            # NEL
+	0x2028          # LINE SEPARATOR
+	0x2029          # PARAGRAPH SEPARATOR
+*/
+/*** GENERATED CODE ***/
+#define is_VERTWS(s,is_utf8)                                                \
 ( (10<=((U8*)s)[0] && ((U8*)s)[0]<=13) ? 1 :                                \
 ( (is_utf8) ?                                                               \
   ( ( ((U8*)s)[0]==194 ) ?                                                  \
@@ -199,7 +254,8 @@
   (((( ((U8*)s)[0]==226 ) && ( ((U8*)s)[1]==128 )) && ( ((U8*)s)[2]==168 || ((U8*)s)[2]==169 )) ? 3 : 0) ) :\
   ( ((U8*)s)[0]==133 ) ) )                                                  
 
-#define is_VERTWS_safe(s,e,is_utf8) /*** Vertical Whitespace: \v \V ***/    \
+/*** GENERATED CODE ***/
+#define is_VERTWS_safe(s,e,is_utf8)                                         \
 ( ( (e) - (s) > 2 ) ?                                                       \
   ( (10<=((U8*)s)[0] && ((U8*)s)[0]<=13) ? 1 :                              \
 ( (is_utf8) ?                                                               \
@@ -217,14 +273,16 @@
 ( (!is_utf8) ?                                                              \
   ( ((U8*)s)[0]==133 ) : 0 ) ) : 0 ) ) )                                    
 
-#define is_VERTWS_utf8(s) /*** Vertical Whitespace: \v \V ***/              \
+/*** GENERATED CODE ***/
+#define is_VERTWS_utf8(s)                                                   \
 ( ( ((U8*)s)[0]==194 ) ?                                                    \
     ( ( ((U8*)s)[1]==133 ) ? 2 : 0 ) :                                      \
   ( ( ((U8*)s)[0]==226 ) ?                                                  \
     ((( ((U8*)s)[1]==128 ) && ( ((U8*)s)[2]==168 || ((U8*)s)[2]==169 )) ? 3 : 0) :\
   (10<=((U8*)s)[0] && ((U8*)s)[0]<=13) ) )                                  
 
-#define is_VERTWS_utf8_safe(s,e) /*** Vertical Whitespace: \v \V ***/       \
+/*** GENERATED CODE ***/
+#define is_VERTWS_utf8_safe(s,e)                                            \
 ( ( (e) - (s) > 2 ) ?                                                       \
   ( ( ((U8*)s)[0]==194 ) ?                                                  \
     ( ( ((U8*)s)[1]==133 ) ? 2 : 0 ) :                                      \
@@ -238,13 +296,19 @@
 ( ( (e) - (s) > 0 ) ?                                                       \
   (10<=((U8*)s)[0] && ((U8*)s)[0]<=13) : 0 ) ) )                            
 
-#define is_VERTWS_latin1(s) /*** Vertical Whitespace: \v \V ***/            \
+/*** GENERATED CODE ***/
+#define is_VERTWS_latin1(s)                                                 \
 ( (10<=((U8*)s)[0] && ((U8*)s)[0]<=13) || ((U8*)s)[0]==133 )                
 
-#define is_VERTWS_latin1_safe(s,e) /*** Vertical Whitespace: \v \V ***/     \
+/*** GENERATED CODE ***/
+#define is_VERTWS_latin1_safe(s,e)                                          \
 ( ( (e) - (s) > 0 ) ?                                                       \
   ( (10<=((U8*)s)[0] && ((U8*)s)[0]<=13) || ((U8*)s)[0]==133 ) : 0 )        
 
-#define is_VERTWS_cp(cp) /*** Vertical Whitespace: \v \V ***/               \
-( (10<=cp && cp<=13) || ( cp > 13 && ( cp==133 || ( cp > 133 && ( cp==8232 || ( cp > 8232 && cp==8233 ) ) ) ) ) )
+/*** GENERATED CODE ***/
+#define is_VERTWS_cp(cp)                                                    \
+( (10<=cp && cp<=13) || ( cp > 13 &&                                        \
+( cp==133 || ( cp > 133 &&                                                  \
+( cp==8232 || ( cp > 8232 &&                                                \
+cp==8233 ) ) ) ) ) )