2009-01-29 22:41:08 -05:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								! Copyright (C) 2005, 2009 Daniel Ehrenberg
							 | 
						
					
						
							
								
									
										
										
										
											2007-09-20 18:09:08 -04:00
										 
									 
								 
							 | 
							
								
							 | 
							
								
							 | 
							
							
								! See http://factorcode.org/license.txt for BSD license.
							 | 
						
					
						
							
								
									
										
										
										
											2009-03-21 02:11:45 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								USING: kernel sequences unicode.categories.syntax math math.order
							 | 
						
					
						
							
								
									
										
										
										
											2009-03-21 04:39:30 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								combinators hints combinators.short-circuit ;
							 | 
						
					
						
							
								
									
										
										
										
											2007-09-20 18:09:08 -04:00
										 
									 
								 
							 | 
							
								
							 | 
							
								
							 | 
							
							
								IN: xml.char-classes
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							
								
									
										
										
										
											2009-03-21 02:11:45 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								CATEGORY: 1.0name-start
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								    Ll Lu Lo Lt Nl | {
							 | 
						
					
						
							
								
									
										
										
										
											2011-11-23 21:49:33 -05:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								        [ 0x2BB 0x2C1 between? ]
							 | 
						
					
						
							
								
									
										
										
										
											2009-03-21 02:11:45 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								        [ "\u000559\u0006E5\u0006E6_:" member? ]
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								    } 1|| ;
							 | 
						
					
						
							
								
									
										
										
										
											2007-09-20 18:09:08 -04:00
										 
									 
								 
							 | 
							
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							
								
									
										
										
										
											2009-03-21 02:11:45 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								CATEGORY: 1.0name-char
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								    Ll Lu Lo Lt Nl Mc Me Mn Lm Nd |
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								    "_-.\u000387:" member? ;
							 | 
						
					
						
							
								
									
										
										
										
											2007-09-20 18:09:08 -04:00
										 
									 
								 
							 | 
							
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							
								
									
										
										
										
											2009-03-21 02:11:45 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								CATEGORY: 1.1name-start
							 | 
						
					
						
							
								
									
										
										
										
											2009-03-21 04:39:30 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								    Ll Lu Lo Lm Nl |
							 | 
						
					
						
							
								
									
										
										
										
											2009-03-21 02:11:45 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								    "_:" member? ;
							 | 
						
					
						
							
								
									
										
										
										
											2007-09-20 18:09:08 -04:00
										 
									 
								 
							 | 
							
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							
								
									
										
										
										
											2009-03-21 02:11:45 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								CATEGORY: 1.1name-char
							 | 
						
					
						
							
								
									
										
										
										
											2009-03-21 04:39:30 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								    Ll Lu Lo Lm Nl Mc Mn Nd Pc Cf |
							 | 
						
					
						
							
								
									
										
										
										
											2009-03-21 02:11:45 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								    "_-.\u0000b7:" member? ;
							 | 
						
					
						
							
								
									
										
										
										
											2007-09-20 18:09:08 -04:00
										 
									 
								 
							 | 
							
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								: name-start? ( 1.0? char -- ? )
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								    swap [ 1.0name-start? ] [ 1.1name-start? ] if ;
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								: name-char? ( 1.0? char -- ? )
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								    swap [ 1.0name-char? ] [ 1.1name-char? ] if ;
							 | 
						
					
						
							
								
									
										
										
										
											2009-01-19 23:25:15 -05:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							
								
									
										
										
										
											2011-09-30 15:47:38 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								HINTS: name-start? { object fixnum } ;
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								HINTS: name-char? { object fixnum } ;
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								<PRIVATE
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								: 1.0-text? ( char -- ? )
							 | 
						
					
						
							
								
									
										
										
										
											2009-01-19 23:25:15 -05:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								    ! 1.0:
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								    ! #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]
							 | 
						
					
						
							
								
									
										
										
										
											2011-09-30 15:47:38 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								    {
							 | 
						
					
						
							
								
									
										
										
										
											2011-11-23 21:49:33 -05:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								        [ 0x20 0xD7FF between? ]
							 | 
						
					
						
							
								
									
										
										
										
											2011-09-30 15:47:38 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								        [ "\t\r\n" member? ]
							 | 
						
					
						
							
								
									
										
										
										
											2011-11-23 21:49:33 -05:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								        [ 0xE000 0xFFFD between? ]
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								        [ 0x10000 0x10FFFF between? ]
							 | 
						
					
						
							
								
									
										
										
										
											2011-09-30 15:47:38 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								    } 1|| ; inline
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								: 1.1-text? ( char -- ? )
							 | 
						
					
						
							
								
									
										
										
										
											2009-01-19 23:25:15 -05:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								    ! 1.1:
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								    ! [#x1-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								    {
							 | 
						
					
						
							
								
									
										
										
										
											2011-11-23 21:49:33 -05:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								        [ 0x1 0xD7FF between? ]
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								        [ 0xE000 0xFFFD between? ]
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								        [ 0x10000 0x10FFFF between? ]
							 | 
						
					
						
							
								
									
										
										
										
											2011-09-30 15:47:38 -04:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								    } 1|| ; inline
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								PRIVATE>
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								: text? ( 1.0? char -- ? )
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								    swap [ 1.0-text? ] [ 1.1-text? ] if ;
							 | 
						
					
						
							
								
									
										
										
										
											2009-01-29 22:41:08 -05:00
										 
									 
								 
							 | 
							
								
									
										
									
								
							 | 
							
								
							 | 
							
							
								
							 | 
						
					
						
							| 
								
							 | 
							
								
							 | 
							
								
							 | 
							
							
								HINTS: text? { object fixnum } ;
							 |