/* -*- buffer-read-only: t -*- * * regcharclass.h * * Copyright (C) 2007, by Larry Wall and others * * You may distribute under the terms of either the GNU General Public * License or the Artistic License, as specified in the README file. * * !!!!!!! DO NOT EDIT THIS FILE !!!!!!! * This file is built by Porting/regcharclass.pl. * (Generated at: Tue Apr 24 12:19:13 2007 GMT) * Any changes made here will be lost! */ /* 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] == 0x0D ) ? \ ( ( ((U8*)s)[1] == 0x0A ) ? 2 : 1 ) : \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0C) ? 1 : \ ( (is_utf8) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0x85 ) ? 2 : 0 ) : \ (((( ((U8*)s)[0] == 0xE2 ) && ( ((U8*)s)[1] == 0x80 )) && ( ((U8*)s)[2] == 0xA8 || ((U8*)s)[2] == 0xA9 )) ? 3 : 0) ) :\ ( ((U8*)s)[0] == 0x85 ) ) ) ) /*** GENERATED CODE ***/ #define is_LNBREAK_safe(s,e,is_utf8) \ ( ( (e) - (s) > 2 ) ? \ ( ( ((U8*)s)[0] == 0x0D ) ? \ ( ( ((U8*)s)[1] == 0x0A ) ? 2 : 1 ) : \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0C) ? 1 : \ ( (is_utf8) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0x85 ) ? 2 : 0 ) : \ (((( ((U8*)s)[0] == 0xE2 ) && ( ((U8*)s)[1] == 0x80 )) && ( ((U8*)s)[2] == 0xA8 || ((U8*)s)[2] == 0xA9 )) ? 3 : 0) ) :\ ( ((U8*)s)[0] == 0x85 ) ) ) ) : \ ( ( (e) - (s) > 1 ) ? \ ( ( ((U8*)s)[0] == 0x0D ) ? \ ( ( ((U8*)s)[1] == 0x0A ) ? 2 : 1 ) : \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0C) ? 1 : \ ( (is_utf8) ? \ ((( ((U8*)s)[0] == 0xC2 ) && ( ((U8*)s)[1] == 0x85 )) ? 2 : 0) : \ ( ((U8*)s)[0] == 0x85 ) ) ) ) : \ ( ( (e) - (s) > 0 ) ? \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) ? 1 : \ ( (!is_utf8) ? \ ( ((U8*)s)[0] == 0x85 ) : 0 ) ) : 0 ) ) ) /*** GENERATED CODE ***/ #define is_LNBREAK_utf8(s) \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0x85 ) ? 2 : 0 ) : \ ( ( ((U8*)s)[0] == 0xE2 ) ? \ ((( ((U8*)s)[1] == 0x80 ) && ( ((U8*)s)[2] == 0xA8 || ((U8*)s)[2] == 0xA9 )) ? 3 : 0) :\ ( ( ((U8*)s)[0] == 0x0D ) ? \ ( ( ((U8*)s)[1] == 0x0A ) ? 2 : 1 ) : \ (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0C) ) ) ) /*** GENERATED CODE ***/ #define is_LNBREAK_utf8_safe(s,e) \ ( ( (e) - (s) > 2 ) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0x85 ) ? 2 : 0 ) : \ ( ( ((U8*)s)[0] == 0xE2 ) ? \ ((( ((U8*)s)[1] == 0x80 ) && ( ((U8*)s)[2] == 0xA8 || ((U8*)s)[2] == 0xA9 )) ? 3 : 0) :\ ( ( ((U8*)s)[0] == 0x0D ) ? \ ( ( ((U8*)s)[1] == 0x0A ) ? 2 : 1 ) : \ (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0C) ) ) ) : \ ( ( (e) - (s) > 1 ) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0x85 ) ? 2 : 0 ) : \ ( ( ((U8*)s)[0] == 0x0D ) ? \ ( ( ((U8*)s)[1] == 0x0A ) ? 2 : 1 ) : \ (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0C) ) ) : \ ( ( (e) - (s) > 0 ) ? \ (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) : 0 ) ) ) /*** GENERATED CODE ***/ #define is_LNBREAK_latin1(s) \ ( ( ((U8*)s)[0] == 0x0D ) ? \ ( ( ((U8*)s)[1] == 0x0A ) ? 2 : 1 ) : \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0C) || ((U8*)s)[0] == 0x85 ) ) /*** GENERATED CODE ***/ #define is_LNBREAK_latin1_safe(s,e) \ ( ( (e) - (s) > 1 ) ? \ ( ( ((U8*)s)[0] == 0x0D ) ? \ ( ( ((U8*)s)[1] == 0x0A ) ? 2 : 1 ) : \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0C) || ((U8*)s)[0] == 0x85 ) ) :\ ( ( (e) - (s) > 0 ) ? \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) || ((U8*)s)[0] == 0x85 ) : 0 ) ) /*** GENERATED CODE ***/ #define is_LNBREAK_cp(cp) \ ( (0x0A <= cp && cp <= 0x0D) ||( cp > 0x0D && \ ( cp == 0x85 ||( cp > 0x85 && \ ( cp == 0x2028 ||( cp > 0x2028 && \ cp == 0x2029 ) ) ) ) ) ) /* 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] == 0x09 || ((U8*)s)[0] == 0x20 ) ? 1 : \ ( (is_utf8) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0xA0 ) ? 2 : 0 ) : \ ( ( ((U8*)s)[0] == 0xE1 ) ? \ ( ( ((U8*)s)[1] == 0xA0 ) ? \ ( ( ((U8*)s)[2] == 0x8E ) ? 3 : 0 ) : \ ((( ((U8*)s)[1] == 0x9A ) && ( ((U8*)s)[2] == 0x80 )) ? 3 : 0) ) : \ ( ( ((U8*)s)[0] == 0xE2 ) ? \ ( ( ((U8*)s)[1] == 0x81 ) ? \ ( ( ((U8*)s)[2] == 0x9F ) ? 3 : 0 ) : \ ((( ((U8*)s)[1] == 0x80 ) && ( (0x80 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x8A) || ((U8*)s)[2] == 0xAF )) ? 3 : 0) ) :\ (((( ((U8*)s)[0] == 0xE3 ) && ( ((U8*)s)[1] == 0x80 )) && ( ((U8*)s)[2] == 0x80 )) ? 3 : 0) ) ) ) :\ ( ((U8*)s)[0] == 0xA0 ) ) ) /*** GENERATED CODE ***/ #define is_HORIZWS_safe(s,e,is_utf8) \ ( ( (e) - (s) > 2 ) ? \ ( ( ((U8*)s)[0] == 0x09 || ((U8*)s)[0] == 0x20 ) ? 1 : \ ( (is_utf8) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0xA0 ) ? 2 : 0 ) : \ ( ( ((U8*)s)[0] == 0xE1 ) ? \ ( ( ((U8*)s)[1] == 0xA0 ) ? \ ( ( ((U8*)s)[2] == 0x8E ) ? 3 : 0 ) : \ ((( ((U8*)s)[1] == 0x9A ) && ( ((U8*)s)[2] == 0x80 )) ? 3 : 0) ) : \ ( ( ((U8*)s)[0] == 0xE2 ) ? \ ( ( ((U8*)s)[1] == 0x81 ) ? \ ( ( ((U8*)s)[2] == 0x9F ) ? 3 : 0 ) : \ ((( ((U8*)s)[1] == 0x80 ) && ( (0x80 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x8A) || ((U8*)s)[2] == 0xAF )) ? 3 : 0) ) :\ (((( ((U8*)s)[0] == 0xE3 ) && ( ((U8*)s)[1] == 0x80 )) && ( ((U8*)s)[2] == 0x80 )) ? 3 : 0) ) ) ) :\ ( ((U8*)s)[0] == 0xA0 ) ) ) : \ ( ( (e) - (s) > 1 ) ? \ ( ( ((U8*)s)[0] == 0x09 || ((U8*)s)[0] == 0x20 ) ? 1 : \ ( (is_utf8) ? \ ((( ((U8*)s)[0] == 0xC2 ) && ( ((U8*)s)[1] == 0xA0 )) ? 2 : 0) : \ ( ((U8*)s)[0] == 0xA0 ) ) ) : \ ( ( (e) - (s) > 0 ) ? \ ( ( ((U8*)s)[0] == 0x09 || ((U8*)s)[0] == 0x20 ) ? 1 : \ ( (!is_utf8) ? \ ( ((U8*)s)[0] == 0xA0 ) : 0 ) ) : 0 ) ) ) /*** GENERATED CODE ***/ #define is_HORIZWS_utf8(s) \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0xA0 ) ? 2 : 0 ) : \ ( ( ((U8*)s)[0] == 0xE1 ) ? \ ( ( ((U8*)s)[1] == 0xA0 ) ? \ ( ( ((U8*)s)[2] == 0x8E ) ? 3 : 0 ) : \ ((( ((U8*)s)[1] == 0x9A ) && ( ((U8*)s)[2] == 0x80 )) ? 3 : 0) ) : \ ( ( ((U8*)s)[0] == 0xE2 ) ? \ ( ( ((U8*)s)[1] == 0x81 ) ? \ ( ( ((U8*)s)[2] == 0x9F ) ? 3 : 0 ) : \ ((( ((U8*)s)[1] == 0x80 ) && ( (0x80 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x8A) || ((U8*)s)[2] == 0xAF )) ? 3 : 0) ) :\ ( ( ((U8*)s)[0] == 0xE3 ) ? \ ((( ((U8*)s)[1] == 0x80 ) && ( ((U8*)s)[2] == 0x80 )) ? 3 : 0) : \ ( ((U8*)s)[0] == 0x09 || ((U8*)s)[0] == 0x20 ) ) ) ) ) /*** GENERATED CODE ***/ #define is_HORIZWS_utf8_safe(s,e) \ ( ( (e) - (s) > 2 ) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0xA0 ) ? 2 : 0 ) : \ ( ( ((U8*)s)[0] == 0xE1 ) ? \ ( ( ((U8*)s)[1] == 0xA0 ) ? \ ( ( ((U8*)s)[2] == 0x8E ) ? 3 : 0 ) : \ ((( ((U8*)s)[1] == 0x9A ) && ( ((U8*)s)[2] == 0x80 )) ? 3 : 0) ) : \ ( ( ((U8*)s)[0] == 0xE2 ) ? \ ( ( ((U8*)s)[1] == 0x81 ) ? \ ( ( ((U8*)s)[2] == 0x9F ) ? 3 : 0 ) : \ ((( ((U8*)s)[1] == 0x80 ) && ( (0x80 <= ((U8*)s)[2] && ((U8*)s)[2] <= 0x8A) || ((U8*)s)[2] == 0xAF )) ? 3 : 0) ) :\ ( ( ((U8*)s)[0] == 0xE3 ) ? \ ((( ((U8*)s)[1] == 0x80 ) && ( ((U8*)s)[2] == 0x80 )) ? 3 : 0) : \ ( ((U8*)s)[0] == 0x09 || ((U8*)s)[0] == 0x20 ) ) ) ) ) : \ ( ( (e) - (s) > 1 ) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0xA0 ) ? 2 : 0 ) : \ ( ((U8*)s)[0] == 0x09 || ((U8*)s)[0] == 0x20 ) ) : \ ( ( (e) - (s) > 0 ) ? \ ( ((U8*)s)[0] == 0x09 || ((U8*)s)[0] == 0x20 ) : 0 ) ) ) /*** GENERATED CODE ***/ #define is_HORIZWS_latin1(s) \ ( ((U8*)s)[0] == 0x09 || ((U8*)s)[0] == 0x20 || ((U8*)s)[0] == 0xA0 ) /*** GENERATED CODE ***/ #define is_HORIZWS_latin1_safe(s,e) \ ( ( (e) - (s) > 0 ) ? \ ( ((U8*)s)[0] == 0x09 || ((U8*)s)[0] == 0x20 || ((U8*)s)[0] == 0xA0 ) : 0 ) /*** GENERATED CODE ***/ #define is_HORIZWS_cp(cp) \ ( cp == 0x09 ||( cp > 0x09 && \ ( cp == 0x20 ||( cp > 0x20 && \ ( cp == 0xA0 ||( cp > 0xA0 && \ ( cp == 0x1680 ||( cp > 0x1680 && \ ( cp == 0x180E ||( cp > 0x180E && \ ( (0x2000 <= cp && cp <= 0x200A) ||( cp > 0x200A && \ ( cp == 0x202F ||( cp > 0x202F && \ ( cp == 0x205F ||( cp > 0x205F && \ cp == 0x3000 ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) /* VERTWS: 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) \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) ? 1 : \ ( (is_utf8) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0x85 ) ? 2 : 0 ) : \ (((( ((U8*)s)[0] == 0xE2 ) && ( ((U8*)s)[1] == 0x80 )) && ( ((U8*)s)[2] == 0xA8 || ((U8*)s)[2] == 0xA9 )) ? 3 : 0) ) :\ ( ((U8*)s)[0] == 0x85 ) ) ) /*** GENERATED CODE ***/ #define is_VERTWS_safe(s,e,is_utf8) \ ( ( (e) - (s) > 2 ) ? \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) ? 1 : \ ( (is_utf8) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0x85 ) ? 2 : 0 ) : \ (((( ((U8*)s)[0] == 0xE2 ) && ( ((U8*)s)[1] == 0x80 )) && ( ((U8*)s)[2] == 0xA8 || ((U8*)s)[2] == 0xA9 )) ? 3 : 0) ) :\ ( ((U8*)s)[0] == 0x85 ) ) ) : \ ( ( (e) - (s) > 1 ) ? \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) ? 1 : \ ( (is_utf8) ? \ ((( ((U8*)s)[0] == 0xC2 ) && ( ((U8*)s)[1] == 0x85 )) ? 2 : 0) : \ ( ((U8*)s)[0] == 0x85 ) ) ) : \ ( ( (e) - (s) > 0 ) ? \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) ? 1 : \ ( (!is_utf8) ? \ ( ((U8*)s)[0] == 0x85 ) : 0 ) ) : 0 ) ) ) /*** GENERATED CODE ***/ #define is_VERTWS_utf8(s) \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0x85 ) ? 2 : 0 ) : \ ( ( ((U8*)s)[0] == 0xE2 ) ? \ ((( ((U8*)s)[1] == 0x80 ) && ( ((U8*)s)[2] == 0xA8 || ((U8*)s)[2] == 0xA9 )) ? 3 : 0) :\ (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) ) ) /*** GENERATED CODE ***/ #define is_VERTWS_utf8_safe(s,e) \ ( ( (e) - (s) > 2 ) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0x85 ) ? 2 : 0 ) : \ ( ( ((U8*)s)[0] == 0xE2 ) ? \ ((( ((U8*)s)[1] == 0x80 ) && ( ((U8*)s)[2] == 0xA8 || ((U8*)s)[2] == 0xA9 )) ? 3 : 0) :\ (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) ) ) : \ ( ( (e) - (s) > 1 ) ? \ ( ( ((U8*)s)[0] == 0xC2 ) ? \ ( ( ((U8*)s)[1] == 0x85 ) ? 2 : 0 ) : \ (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) ) : \ ( ( (e) - (s) > 0 ) ? \ (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) : 0 ) ) ) /*** GENERATED CODE ***/ #define is_VERTWS_latin1(s) \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) || ((U8*)s)[0] == 0x85 ) /*** GENERATED CODE ***/ #define is_VERTWS_latin1_safe(s,e) \ ( ( (e) - (s) > 0 ) ? \ ( (0x0A <= ((U8*)s)[0] && ((U8*)s)[0] <= 0x0D) || ((U8*)s)[0] == 0x85 ) : 0 ) /*** GENERATED CODE ***/ #define is_VERTWS_cp(cp) \ ( (0x0A <= cp && cp <= 0x0D) ||( cp > 0x0D && \ ( cp == 0x85 ||( cp > 0x85 && \ ( cp == 0x2028 ||( cp > 0x2028 && \ cp == 0x2029 ) ) ) ) ) ) /* TRICKYFOLD: Problematic fold case letters. 0x00DF # LATIN SMALL LETTER SHARP S 0x0390 # GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS 0x03B0 # GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS */ /*** GENERATED CODE ***/ #define is_TRICKYFOLD(s,is_utf8) \ ( (is_utf8) ? \ ( ( ((U8*)s)[0] == 0xC3 ) ? \ ( ( ((U8*)s)[1] == 0x9F ) ? 2 : 0 ) : \ ((( ((U8*)s)[0] == 0xCE ) && ( ((U8*)s)[1] == 0x90 || ((U8*)s)[1] == 0xB0 )) ? 2 : 0) ) :\ ( ((U8*)s)[0] == 0xDF ) ) /*** GENERATED CODE ***/ #define is_TRICKYFOLD_safe(s,e,is_utf8) \ ( ( (e) - (s) > 1 ) ? \ ( (is_utf8) ? \ ( ( ((U8*)s)[0] == 0xC3 ) ? \ ( ( ((U8*)s)[1] == 0x9F ) ? 2 : 0 ) : \ ((( ((U8*)s)[0] == 0xCE ) && ( ((U8*)s)[1] == 0x90 || ((U8*)s)[1] == 0xB0 )) ? 2 : 0) ) :\ ( ((U8*)s)[0] == 0xDF ) ) : \ ((( (e) - (s) > 0 ) && (!is_utf8)) ? ( ((U8*)s)[0] == 0xDF ) : 0) ) /*** GENERATED CODE ***/ #define is_TRICKYFOLD_utf8(s) \ ( ( ((U8*)s)[0] == 0xC3 ) ? \ ( ( ((U8*)s)[1] == 0x9F ) ? 2 : 0 ) : \ ((( ((U8*)s)[0] == 0xCE ) && ( ((U8*)s)[1] == 0x90 || ((U8*)s)[1] == 0xB0 )) ? 2 : 0) ) /*** GENERATED CODE ***/ #define is_TRICKYFOLD_utf8_safe(s,e) \ ( ( (e) - (s) > 1 ) ? \ ( ( ((U8*)s)[0] == 0xC3 ) ? \ ( ( ((U8*)s)[1] == 0x9F ) ? 2 : 0 ) : \ ((( ((U8*)s)[0] == 0xCE ) && ( ((U8*)s)[1] == 0x90 || ((U8*)s)[1] == 0xB0 )) ? 2 : 0) ) : 0 ) /*** GENERATED CODE ***/ #define is_TRICKYFOLD_latin1(s) \ ( ((U8*)s)[0] == 0xDF ) /*** GENERATED CODE ***/ #define is_TRICKYFOLD_latin1_safe(s,e) \ ( ( (e) - (s) > 0 ) ? \ ( ((U8*)s)[0] == 0xDF ) : 0 ) /*** GENERATED CODE ***/ #define is_TRICKYFOLD_cp(cp) \ ( cp == 0xDF ||( cp > 0xDF && \ ( cp == 0x390 ||( cp > 0x390 && \ cp == 0x3B0 ) ) ) ) /* ex: set ro: */