ICU-57149.0.1.tar.gz

[apple/icu.git] / icuSources / i18n / regexcst.txt

diff --git a/icuSources/i18n/regexcst.txt b/icuSources/i18n/regexcst.txt
index 77ebd9606b47356a5535e38a0afc84c992ff5e8f..1b88e44654021cc4f0ab541550bd343be26e6ef4 100644 (file)
--- a/icuSources/i18n/regexcst.txt
+++ b/icuSources/i18n/regexcst.txt
@@ -1,7 +1,7 @@
 
 #*****************************************************************************
 #
 
 #*****************************************************************************
 #

-#   Copyright (C) 2002-2007, International Business Machines Corporation and others.
+#   Copyright (C) 2002-2015, International Business Machines Corporation and others.

 #   All Rights Reserved.
 #
 #*****************************************************************************
 #   All Rights Reserved.
 #
 #*****************************************************************************


@@ -147,6 +147,7 @@ open-paren-extended:
 open-paren-lookbehind:
     '='                  n  term            ^expr-cont              doOpenLookBehind       #  (?<=
     '!'                  n  term            ^expr-cont              doOpenLookBehindNeg    #  (?<!
 open-paren-lookbehind:
     '='                  n  term            ^expr-cont              doOpenLookBehind       #  (?<=
     '!'                  n  term            ^expr-cont              doOpenLookBehindNeg    #  (?<!

+    ascii_letter            named-capture                           doBeginNamedCapture    #  (?<name

     default                 errorDeath                              doBadOpenParenType
 
 
     default                 errorDeath                              doBadOpenParenType
 
 


@@ -174,6 +175,14 @@ paren-flag:
     ':'                  n  term              ^expr-quant           doMatchModeParen
     default                 errorDeath                              doBadModeFlag
 
     ':'                  n  term              ^expr-quant           doMatchModeParen
     default                 errorDeath                              doBadModeFlag
 

+#
+#  named-capture    (?<name> ... ), position currently on the name.
+#
+named-capture:
+    ascii_letter         n  named-capture                           doContinueNamedCapture
+    digit_char           n  named-capture                           doContinueNamedCapture
+    '>'                  n  term               ^expr-quant          doOpenCaptureParen      # common w non-named capture.
+    default                 errorDeath                              doBadNamedCapture

 
 #
 #  quant-star     Scanning a '*' quantifier.  Need to look ahead to decide
 
 #
 #  quant-star     Scanning a '*' quantifier.  Need to look ahead to decide


@@ -241,12 +250,18 @@ backslash:
    'd'                   n  expr-quant                              doBackslashd
    'D'                   n  expr-quant                              doBackslashD
    'G'                   n  term                                    doBackslashG
    'd'                   n  expr-quant                              doBackslashd
    'D'                   n  expr-quant                              doBackslashD
    'G'                   n  term                                    doBackslashG

+   'h'                   n  expr-quant                              doBackslashh
+   'H'                   n  expr-quant                              doBackslashH
+   'k'                   n  named-backref

    'N'                      expr-quant                              doNamedChar      #   \N{NAME}  named char
    'p'                      expr-quant                              doProperty       #   \p{Lu}  style property
    'P'                      expr-quant                              doProperty
    'N'                      expr-quant                              doNamedChar      #   \N{NAME}  named char
    'p'                      expr-quant                              doProperty       #   \p{Lu}  style property
    'P'                      expr-quant                              doProperty

+   'R'                   n  expr-quant                              doBackslashR

    'Q'                   n  term                                    doEnterQuoteMode
    'S'                   n  expr-quant                              doBackslashS
    's'                   n  expr-quant                              doBackslashs
    'Q'                   n  term                                    doEnterQuoteMode
    'S'                   n  expr-quant                              doBackslashS
    's'                   n  expr-quant                              doBackslashs

+   'v'                   n  expr-quant                              doBackslashv
+   'V'                   n  expr-quant                              doBackslashV

    'W'                   n  expr-quant                              doBackslashW
    'w'                   n  expr-quant                              doBackslashw
    'X'                   n  expr-quant                              doBackslashX
    'W'                   n  expr-quant                              doBackslashW
    'w'                   n  expr-quant                              doBackslashw
    'X'                   n  expr-quant                              doBackslashX


@@ -257,6 +272,24 @@ backslash:
    default               n  expr-quant                              doEscapedLiteralChar
 
 
    default               n  expr-quant                              doEscapedLiteralChar
 
 

+# named-backref   Scanned \k
+#                 Leading to \k<captureName>
+#                 Failure to get the full sequence is an error.
+#
+named-backref:
+    '<'                  n  named-backref-2                         doBeginNamedBackRef
+    default                 errorDeath                              doBadNamedCapture
+
+named-backref-2:
+    ascii_letter         n  named-backref-3                         doContinueNamedBackRef
+    default                 errorDeath                              doBadNamedCapture
+
+named-backref-3:
+    ascii_letter         n  named-backref-3                         doContinueNamedBackRef
+    digit_char           n  named-backref-3                         doContinueNamedBackRef
+    '>'                  n  expr-quant                              doCompleteNamedBackRef
+    default                 errorDeath                              doBadNamedCapture
+

 
 #
 # [set expression] parsing,
 
 #
 # [set expression] parsing,


@@ -444,6 +477,10 @@ set-escape:
    'W'                   n  set-after-range                         doSetBackslash_W
    'd'                   n  set-after-range                         doSetBackslash_d
    'D'                   n  set-after-range                         doSetBackslash_D
    'W'                   n  set-after-range                         doSetBackslash_W
    'd'                   n  set-after-range                         doSetBackslash_d
    'D'                   n  set-after-range                         doSetBackslash_D

+   'h'                   n  set-after-range                         doSetBackslash_h
+   'H'                   n  set-after-range                         doSetBackslash_H
+   'v'                   n  set-after-range                         doSetBackslash_v
+   'V'                   n  set-after-range                         doSetBackslash_V

    default               n  set-after-lit                           doSetLiteralEscaped 
 
 #
    default               n  set-after-lit                           doSetLiteralEscaped 
 
 #