1 # Copyright (C) 2016 and later: Unicode, Inc. and others. 2 # License & terms of use: http://www.unicode.org/copyright.html 3 # 4 # 5 # Copyright (C) 2002-2015, International Business Machines Corporation and others. 6 # All Rights Reserved. 7 # 8 # file: sent_el.txt 9 # 10 # ICU Sentence Break Rules 11 # See Unicode Standard Annex #29. 12 # These rules are based on UAX #29 Revision 26 for Unicode Version 8.0 13 # 14 15 !!quoted_literals_only; 16 17 # 18 # Character categories as defined in TR 29 19 # 20 $CR = [\p{Sentence_Break = CR}]; 21 $LF = [\p{Sentence_Break = LF}]; 22 $Extend = [\p{Sentence_Break = Extend}]; 23 $Sep = [\p{Sentence_Break = Sep}]; 24 $Format = [\p{Sentence_Break = Format}]; 25 $Sp = [\p{Sentence_Break = Sp}]; 26 $Lower = [\p{Sentence_Break = Lower}]; 27 $Upper = [\p{Sentence_Break = Upper}]; 28 $OLetter = [\p{Sentence_Break = OLetter}]; 29 $Numeric = [\p{Sentence_Break = Numeric}]; 30 $ATerm = [\p{Sentence_Break = ATerm}]; 31 $SContinue = [\p{Sentence_Break = SContinue}]; 32 $STerm = [\p{Sentence_Break = STerm} [\u003B \u037E]]; 33 $Close = [\p{Sentence_Break = Close}]; 34 35 # 36 # Define extended forms of the character classes, 37 # incorporate trailing Extend or Format chars. 38 # Rules 4 and 5. 39 40 $SpEx = $Sp ($Extend | $Format)*; 41 $LowerEx = $Lower ($Extend | $Format)*; 42 $UpperEx = $Upper ($Extend | $Format)*; 43 $OLetterEx = $OLetter ($Extend | $Format)*; 44 $NumericEx = $Numeric ($Extend | $Format)*; 45 $ATermEx = $ATerm ($Extend | $Format)*; 46 $SContinueEx= $SContinue ($Extend | $Format)*; 47 $STermEx = $STerm ($Extend | $Format)*; 48 $CloseEx = $Close ($Extend | $Format)*; 49 50 51 ## ------------------------------------------------- 52 53 !!chain; 54 55 # Rule 3 - break after separators. Keep CR/LF together. 56 # 57 $CR $LF; 58 59 60 # Rule 4 - Break after $Sep. 61 # Rule 5 - Ignore $Format and $Extend 62 # 63 [^$Sep $CR $LF]? ($Extend | $Format)*; 64 65 66 # Rule 6 67 $ATermEx $NumericEx; 68 69 # Rule 7 70 ($UpperEx | $LowerEx) $ATermEx $UpperEx; 71 72 #Rule 8 73 $NotLettersEx = [^$OLetter $Upper $Lower $Sep $CR $LF $ATerm $STerm] ($Extend | $Format)*; 74 $ATermEx $CloseEx* $SpEx* $NotLettersEx* $Lower; 75 76 # Rule 8a 77 ($STermEx | $ATermEx) $CloseEx* $SpEx* ($SContinueEx | $STermEx | $ATermEx); 78 79 #Rule 9, 10, 11 80 ($STermEx | $ATermEx) $CloseEx* $SpEx* ($Sep | $CR | $LF)?; 81 82 #Rule 12 83 [[^$STerm $ATerm $Close $Sp $Sep $LF $CR $Format $Extend]{bof}] ($Extend | $Format | $Close | $Sp)* .; 84 [[^$STerm $ATerm $Close $Sp $Sep $LF $CR $Format $Extend]{bof}] ($Extend | $Format | $Close | $Sp)* ([$Sep $LF $CR {eof}] | $CR $LF){100}; 85
