ICU 56 update step 1

source/test/testdata/regextst.txt

-# Copyright (c) 2001-2014 International Business Machines
+# Copyright (c) 2001-2015 International Business Machines
 # Corporation and others. All Rights Reserved.
 #
 #  file:
 #
 #   ICU regular expression test cases.
 #
 #   format:   one test case per line,
 #               <test case>    =  <pattern>   <flags>  <match string>  [# comment]
 #               <pattern>      =  "<regular expression pattern>"
 #               <match string> =  "<tagged string>"
@@ skipping 494 matching lines @@
 # Back reference to a zero-length match.  They are always a successful match.
 "ab(x?)cd(\1)ef"               "<0>ab<1></1>cd<2></2>ef</0>"
 "ab(x?)cd(\1)ef"            i  "<0>ab<1></1>cd<2></2>ef</0>"
 
 # Back refs to capture groups that didn't participate in the match.
 "ab(?:(c)|(d))\1"              "abde"
 "ab(?:(c)|(d))\1"              "<0>ab<1>c</1>c</0>e"
 "ab(?:(c)|(d))\1"            i "abde"
 "ab(?:(c)|(d))\1"            i "<0>ab<1>c</1>c</0>e"
 
+# Named back references
+"(?<one>abcd)\k<one>"          "<0><1>abcd</1>abcd</0>"
+"(no)?(?<one>abcd)\k<one>"     "<0><2>abcd</2>abcd</0>"
+
+"(?<a_1>...)"               E  "  "   # backref names are ascii letters & numbers only"
+"(?<1a>...)"                E  "  "   # backref names must begin with a letter"
+"(?<a>.)(?<a>.)"            E  "  "   # Repeated names are illegal.
+
+
 # Case Insensitive
 "aBc"                    i      "<0>ABC</0>"
 "a[^bc]d"                i      "ABD"
 '((((((((((a))))))))))\10' i    "<0><1><2><3><4><5><6><7><8><9><10>A</10></9></8></7></6></5></4></3></2></1>A</0>"
 
 "(?:(?i)a)b"                    "<0>Ab</0>"
 "ab(?i)cd"                          "<0>abCd</0>"
 "ab$cd"                         "abcd"
 
 "ssl"                      i    "abc<0>ßl</0>xyz"
@@ skipping 152 matching lines @@
 "abc"                       2Z    "aaabc  <0>abc</0>ab"
 "abc"                       3z    "aa>abc  abcab"
 
 #
 # \ escaping
 #
 "abc\jkl"                         "<0>abcjkl</0>"    # escape of a non-special letter is just itself.
 "abc[ \j]kl"                      "<0>abcjkl</0>"
 
 #
+# \R  all newline sequences.
+#
+"abc\Rxyz"                        "<0>abc\u000axyz</0>gh"
+"abc\Rxyz"                        "<0>abc\u000bxyz</0>gh"
+"abc\Rxyz"                        "<0>abc\u000cxyz</0>gh"
+"abc\Rxyz"                        "<0>abc\u000dxyz</0>gh"
+"abc\Rxyz"                        "<0>abc\u0085xyz</0>gh"
+"abc\Rxyz"                        "<0>abc\u2028xyz</0>gh"
+"abc\Rxyz"                        "<0>abc\u2029xyz</0>gh"
+"abc\Rxyz"                        "<0>abc\u000d\u000axyz</0>gh"
+
+"abc\R\nxyz"                      "abc\u000d\u000axyzgh"          # \R cannot match only the CR from a CR/LF sequence.
+"abc\r\nxyz"                      "<0>abc\u000d\u000axyz</0>gh"
+
+"abc\Rxyz"                        "abc\u0009xyz"                  # Assorted non-matches.
+"abc\Rxyz"                        "abc\u000exyz"
+"abc\Rxyz"                        "abc\u202axyz"
+
+# \v \V single character new line sequences.
+
+"abc\vxyz"                        "<0>abc\u000axyz</0>gh"
+"abc\vxyz"                        "<0>abc\u000bxyz</0>gh"
+"abc\vxyz"                        "<0>abc\u000cxyz</0>gh"
+"abc\vxyz"                        "<0>abc\u000dxyz</0>gh"
+"abc\vxyz"                        "<0>abc\u0085xyz</0>gh"
+"abc\vxyz"                        "<0>abc\u2028xyz</0>gh"
+"abc\vxyz"                        "<0>abc\u2029xyz</0>gh"
+"abc\vxyz"                        "abc\u000d\u000axyzgh"
+"abc\vxyz"                        "abc?xyzgh"
+
+"abc[\v]xyz"                      "<0>abc\u000axyz</0>gh"
+"abc[\v]xyz"                      "<0>abc\u000bxyz</0>gh"
+"abc[\v]xyz"                      "<0>abc\u000cxyz</0>gh"
+"abc[\v]xyz"                      "<0>abc\u000dxyz</0>gh"
+"abc[\v]xyz"                      "<0>abc\u0085xyz</0>gh"
+"abc[\v]xyz"                      "<0>abc\u2028xyz</0>gh"
+"abc[\v]xyz"                      "<0>abc\u2029xyz</0>gh"
+"abc[\v]xyz"                      "abc\u000d\u000axyzgh"
+"abc[\v]xyz"                      "abc?xyzgh"
+
+"abc\Vxyz"                        "abc\u000axyzgh"
+"abc\Vxyz"                        "abc\u000bxyzgh"
+"abc\Vxyz"                        "abc\u000cxyzgh"
+"abc\Vxyz"                        "abc\u000dxyzgh"
+"abc\Vxyz"                        "abc\u0085xyzgh"
+"abc\Vxyz"                        "abc\u2028xyzgh"
+"abc\Vxyz"                        "abc\u2029xyzgh"
+"abc\Vxyz"                        "abc\u000d\u000axyzgh"
+"abc\Vxyz"                        "<0>abc?xyz</0>gh"
+
+# \h \H horizontal white space. Defined as gc=space_separator plus ascii tab
+
+"abc\hxyz"                        "<0>abc xyz</0>gh"
+"abc\Hxyz"                        "abc xyzgh"
+"abc\hxyz"                        "<0>abc\u2003xyz</0>gh"
+"abc\Hxyz"                        "abc\u2003xyzgh"
+"abc\hxyz"                        "<0>abc\u0009xyz</0>gh"
+"abc\Hxyz"                        "abc\u0009xyzgh"
+"abc\hxyz"                        "abc?xyzgh"
+"abc\Hxyz"                        "<0>abc?xyz</0>gh"
+
+"abc[\h]xyz"                      "<0>abc xyz</0>gh"
+"abc[\H]xyz"                      "abc xyzgh"
+"abc[\h]xyz"                      "<0>abc\u2003xyz</0>gh"
+"abc[\H]xyz"                      "abc\u2003xyzgh"
+"abc[\h]xyz"                      "<0>abc\u0009xyz</0>gh"
+"abc[\H]xyz"                      "abc\u0009xyzgh"
+"abc[\h]xyz"                      "abc?xyzgh"
+"abc[\H]xyz"                      "<0>abc?xyz</0>gh"
+
+
+#
 # Bug xxxx
 #
 "(?:\-|(\-?\d+\d\d\d))?(?:\-|\-(\d\d))?(?:\-|\-(\d\d))?(T)?(?:(\d\d):(\d\d):(\d\d)(\.\d+)?)?(?:(?:((?:\+|\-)\d\d):(\d\d))|(Z))?"   MG  "<0>-1234-21-31T41:51:61.789+71:81</0>"
 
 
 #
 # A random, complex, meaningless pattern that should at least compile
 #
 "(?![^\<C\f\0146\0270\}&&[|\02-\x3E\}|X-\|]]{7,}+)[|\\\x98\<\?\u4FCFr\,\0025\}\004|\0025-\0521]|(?<![|\01-\u829E])|(?<!\p{Alpha})|^|(?-s:[^\x15\\\x24F\a\,\a\u97D8[\x38\a[\0224-\0306[^\0020-\u6A57]]]]??)(?xix:[^|\{\[\0367\t\e\x8C\{\[\074c\]V[|b\fu\r\0175\<\07f\066s[^D-\x5D]]])(?xx:^{5,}+)(?d)(?=^\D)|(?!\G)(?>\G)(?![^|\]\070\ne\{\t\[\053\?\\\x51\a\075\0023-\[&&[|\022-\xEA\00-\u41C2&&[^|a-\xCC&&[^\037\uECB3\u3D9A\x31\|\<b\0206\uF2EC\01m\,\ak\a\03&&\p{Punct}]]]])(?-dxs:[|\06-\07|\e-\x63&&[|Tp\u18A3\00\|\xE4\05\061\015\0116C|\r\{\}\006\xEA\0367\xC4\01\0042\0267\xBB\01T\}\0100\?[|\[-\u459B|\x23\x91\rF\0376[|\?-\x94\0113-\\\s]]]]{6}?)(?<=[^\t-\x42H\04\f\03\0172\?i\u97B6\e\f\uDAC2])(?=\B)(?>[^\016\r\{\,\uA29D\034\02[\02-\[|\t\056\uF599\x62\e\<\032\uF0AC\0026\0205Q\|\\\06\0164[|\057-\u7A98&&[\061-g|\|\0276\n\042\011\e\xE8\x64B\04\u6D0EDW^\p{Lower}]]]]?)(?<=[^\n\\\t\u8E13\,\0114\u656E\xA5\]&&[\03-\026|\uF39D\01\{i\u3BC2\u14FE]])(?<=[^|\uAE62\054H\|\}&&^\p{Space}])(?sxx)(?<=[\f\006\a\r\xB4]{1,5})|(?x-xd:^{5}+)()"  "<0></0>abc"
 
@@ skipping 513 matching lines @@
 "((|b){0})a{3}(D*)"             "<0><1></1>aaa<3></3></0>"
 
 # Bug 11370
 #   Max match length computation of look-behind expression gives result that is too big to fit in the
 #   in the 24 bit operand portion of the compiled code. Expressions should fail to compile
 #   (Look-behind match length must be bounded. This case is treated as unbounded, an error.)
 
 "(?<!(0123456789a){10000000})x"         E  "no match"
 "(?<!\\ubeaf(\\ubeaf{11000}){11000})"   E  "no match"
 
+# Bug 11374 Bad integer overflow check in number conversion.
+#           4294967300 converts to 4 with 32 bit overflow.
+
+"x{4294967300}"                         E  "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
+"x{0,4294967300}"                       E  "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
+
+# Bug 11373
+#
+#    Overflow checking in max match length computation for loops.
+#    Value here is 10 * 100000 * 3000 = 3E9, overflowing a 32 bit signed value.
+#    Before fixing, this case gave an assertion failure.
+
+"(?<=((0123456789){100000}){3000})abc"  E  "abc"
+
+
 
 #  Random debugging, Temporary
 #
 
 "This is a string with (?:one |two |three )endings"   "<0>This is a string with two endings</0>"
 
 
 #
 #  Regexps from http://www.regexlib.com
 #
@@ skipping 1272 matching lines @@
 "(\w+)\s+\1"     "may day"
 "(\w+)\s+\1"     "gogo"
 "(\w+)\s+\1"     "1212"
 "^[a-zA-Z0-9\-\.]+\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$"   G "<0>3SquareBand.com</0>"
 "^[a-zA-Z0-9\-\.]+\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$"   G "<0>asp.net</0>"
 "^[a-zA-Z0-9\-\.]+\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$"   G "<0>army.mil</0>"
 "^[a-zA-Z0-9\-\.]+\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$"     "$SquareBand.com"
 "^[a-zA-Z0-9\-\.]+\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$"     "asp/dot.net"
 "^[a-zA-Z0-9\-\.]+\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$"     "army.military"