diff options
| author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-19 01:47:29 +0000 |
|---|---|---|
| committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-19 01:47:29 +0000 |
| commit | 0ebf5bdf043a27fd3dfb7f92e0cb63d88954c44d (patch) | |
| tree | a31f07c9bcca9d56ce61e9a1ffd30ef350d513aa /intl/icu/source/test/testdata/regextst.txt | |
| parent | Initial commit. (diff) | |
| download | firefox-esr-0ebf5bdf043a27fd3dfb7f92e0cb63d88954c44d.tar.xz firefox-esr-0ebf5bdf043a27fd3dfb7f92e0cb63d88954c44d.zip | |
Adding upstream version 115.8.0esr.upstream/115.8.0esr
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'intl/icu/source/test/testdata/regextst.txt')
| -rw-r--r-- | intl/icu/source/test/testdata/regextst.txt | 2792 |
1 files changed, 2792 insertions, 0 deletions
diff --git a/intl/icu/source/test/testdata/regextst.txt b/intl/icu/source/test/testdata/regextst.txt new file mode 100644 index 0000000000..0909dc77e2 --- /dev/null +++ b/intl/icu/source/test/testdata/regextst.txt @@ -0,0 +1,2792 @@ +# Copyright (C) 2016 and later: Unicode, Inc. and others. +# License & terms of use: http://www.unicode.org/copyright.html +# 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>" +# the quotes on the pattern and match string can be " or ' or / +# <tagged string> = text, with the start and end of each +# capture group tagged with <n>...</n>. The overall match, +# if any, is group 0, as in <0>matched text</0> +# A region can be specified with <r>...</r> tags. +# Standard ICU unescape will be applied, allowing \u, \U, etc. to appear. +# +# <flags> = any combination of +# i case insensitive match +# x free spacing and comments +# s dot-matches-all mode +# m multi-line mode. +# ($ and ^ match at embedded new-lines) +# D Unix Lines mode (only recognize 0x0a as new-line) +# Q UREGEX_LITERAL flag. Entire pattern is literal string. +# v If icu configured without break iteration, this +# regex test pattern should not compile. +# e set the UREGEX_ERROR_ON_UNKNOWN_ESCAPES flag +# d dump the compiled pattern +# t trace operation of match engine. +# 2-9 a digit between 2 and 9, specifies the number of +# times to execute find(). The expected results are +# for the last find() in the sequence. +# G Only check match / no match. Do not check capture groups. +# E Pattern compilation error expected +# L Use LookingAt() rather than find() +# M Use matches() rather than find(). +# +# a Use non-Anchoring Bounds. +# b Use Transparent Bounds. +# The a and b options only make a difference if +# a <r>region</r> has been specified in the string. +# z|Z hitEnd was expected(z) or not expected (Z). +# With neither, hitEnd is not checked. +# y|Y Require End expected(y) or not expected (Y). +# +# White space must be present between the flags and the match string. +# + +# Look-ahead expressions +# +"(?!0{5})(\d{5})" "<0><1>00001</1></0>zzzz" +"(?!0{5})(\d{5})z" "<0><1>00001</1>z</0>zzz" +"(?!0{5})(\d{5})(?!y)" "<0><1>00001</1></0>zzzz" +"abc(?=def)" "<0>abc</0>def" +"(.*)(?=c)" "<0><1>ab</1></0>cdef" + +"(?:.*)(?=c)" "<r>ab</r>cdef" +"(?:.*)(?=c)" b "<r><0>ab</0></r>cdef" # transparent bounds +"(?:.*)(?=c)" bM "<r><0>ab</0></r>cdef" # transparent bounds + +"(?:.*)(?=(c))" b "<0>ab</0><1>c</1>def" # Capture in look-ahead +"(?=(.)\1\1)\1" "abcc<0><1>d</1></0>ddefg" # Backrefs to look-ahead capture + +".(?!\p{L})" "abc<0>d</0> " # Negated look-ahead +".(?!(\p{L}))" "abc<0>d</0> " # Negated look-ahead, no capture + # visible outside of look-ahead +"and(?=roid)" L "<0>and</0>roid" +"and(?=roid)" M "<r>and</r>roid" +"and(?=roid)" bM "<r><0>and</0></r>roid" + +"and(?!roid)" L "<0>and</0>roix" +"and(?!roid)" L "android" + +"and(?!roid)" M "<r><0>and</0></r>roid" # Opaque bounds +"and(?!roid)" bM "<r>and</r>roid" +"and(?!roid)" bM "<r><0>and</0></r>roix" + +# +# Negated Lookahead, various regions and region transparency +# +"abc(?!def)" "<0>abc</0>xyz" +"abc(?!def)" "abcdef" +"abc(?!def)" "<r><0>abc</0></r>def" +"abc(?!def)" b "<r>abc</r>def" +"abc(?!def)" b "<r><0>abc</0></r>xyz" + +# +# Nested Lookahead / Behind +# +"one(?=(?:(?!<out>).)*</out>)" "<out><0>one</0> stuff</out>" +"one(?=(?:(?!<out>).)*</out>)" "<out>one <out></out>" + +# More nesting lookaround: pattern matches "qq" when not preceded by 'a' and followed by 'z' +"(?<!a(?!...z))qq" "<0>qq</0>c" +"(?<!a(?!...z))qq" "f<0>qq</0>c" +"(?<!a(?!...z))qq" "aqqz" + +# More nested lookaround: match any two chars preceded and followed by an upper case letter. +# With gratuitous nesting of look-arounds and capture from the look-arounds. + +"(?=(?<=(\p{Lu})(?=..(\p{Lu})))).." "<1>A</1><0>jk</0><2>B</2>" +"(?=(?<=(\p{Lu})(?=..(\p{Lu})))).." "ajkB" +"(?=(?<=(\p{Lu})(?=..(\p{Lu})))).." "Ajkb" + +# Nested lookaround cases from bug ICU-20564 +"(?<=(?<=((?=)){0}+))" "<0></0>abc" +"(?<=c(?<=c((?=c)){1}+))" "c<0><1></1></0>cc" + +# +# Anchoring Bounds +# +"^def$" "abc<r><0>def</0></r>ghi" # anchoring (default) bounds +"^def$" a "abc<r>def</r>ghi" # non-anchoring bounds +"^def" a "<r><0>def</0></r>ghi" # non-anchoring bounds +"def$" a "abc<r><0>def</0></r>" # non-anchoring bounds + +"^.*$" m "<0>line 1</0>\n line 2" +"^.*$" m2 "line 1\n<0> line 2</0>" +"^.*$" m3 "line 1\n line 2" +"^.*$" m "li<r><0>ne </0></r>1\n line 2" # anchoring bounds +"^.*$" m2 "li<r>ne </r>1\n line 2" # anchoring bounds +"^.*$" am "li<r>ne </r>1\n line 2" # non-anchoring bounds +"^.*$" am "li\n<r><0>ne </0></r>\n1\n line 2" # non-anchoring bounds + +# +# HitEnd and RequireEnd for new-lines just before end-of-input +# +"xyz$" yz "<0>xyz</0>\n" +"xyz$" yz "<0>xyz</0>\x{d}\x{a}" + +"xyz$" myz "<0>xyz</0>" # multi-line mode +"xyz$" mYZ "<0>xyz</0>\n" +"xyz$" mYZ "<0>xyz</0>\r\n" +"xyz$" mYZ "<0>xyz</0>\x{85}abcd" + +"xyz$" Yz "xyz\nx" +"xyz$" Yz "xyza" +"xyz$" yz "<0>xyz</0>" + +# +# HitEnd +# +"abcd" Lz "a" +"abcd" Lz "ab" +"abcd" Lz "abc" +"abcd" LZ "<0>abcd</0>" +"abcd" LZ "<0>abcd</0>e" +"abcd" LZ "abcx" +"abcd" LZ "abx" +"abcd" Lzi "a" +"abcd" Lzi "ab" +"abcd" Lzi "abc" +"abcd" LZi "<0>abcd</0>" +"abcd" LZi "<0>abcd</0>e" +"abcd" LZi "abcx" +"abcd" LZi "abx" + +# +# All Unicode line endings recognized. +# 0a, 0b, 0c, 0d, 0x85, 0x2028, 0x2029 +# Multi-line and non-multiline mode take different paths, so repeated tests. +# +"^def$" mYZ "abc\x{a}<0>def</0>\x{a}ghi" +"^def$" mYZ "abc\x{b}<0>def</0>\x{b}ghi" +"^def$" mYZ "abc\x{c}<0>def</0>\x{c}ghi" +"^def$" mYZ "abc\x{d}<0>def</0>\x{d}ghi" +"^def$" mYZ "abc\x{85}<0>def</0>\x{85}ghi" +"^def$" mYZ "abc\x{2028}<0>def</0>\x{2028}ghi" +"^def$" mYZ "abc\x{2029}<0>def</0>\x{2029}ghi" +"^def$" mYZ "abc\r\n<0>def</0>\r\nghi" + +"^def$" yz "<0>def</0>\x{a}" +"^def$" yz "<0>def</0>\x{b}" +"^def$" yz "<0>def</0>\x{c}" +"^def$" yz "<0>def</0>\x{d}" +"^def$" yz "<0>def</0>\x{85}" +"^def$" yz "<0>def</0>\x{2028}" +"^def$" yz "<0>def</0>\x{2029}" +"^def$" yz "<0>def</0>\r\n" +"^def$" yz "<0>def</0>" + + +"^def$" "<0>def</0>\x{2028" #TODO: should be an error of some sort. + +# +# UNIX_LINES mode +# +"abc$" D "<0>abc</0>\n" +"abc$" D "abc\r" +"abc$" D "abc\u0085" +"a.b" D "<0>a\rb</0>" +"a.b" D "a\nb" +"(?d)abc$" "<0>abc</0>\n" +"(?d)abc$" "abc\r" +"abc$" mD "<0>abc</0>\ndef" +"abc$" mD "abc\rdef" + +".*def" L "abc\r def xyz" # Normal mode, LookingAt() stops at \r +".*def" DL "<0>abc\r def</0> xyz" # Unix Lines mode, \r not line end. +".*def" DL "abc\n def xyz" + +"(?d)a.b" "a\nb" +"(?d)a.b" "<0>a\rb</0>" + +"^abc" m "xyz\r<0>abc</0>" +"^abc" Dm "xyz\rabc" +"^abc" Dm "xyz\n<0>abc</0>" + + + +# Capturing parens +".(..)." "<0>a<1>bc</1>d</0>" + ".*\A( +hello)" "<0><1> hello</1></0>" +"(hello)|(goodbye)" "<0><1>hello</1></0>" +"(hello)|(goodbye)" "<0><2>goodbye</2></0>" +"abc( +( inner(X?) +) xyz)" "leading cruft <0>abc<1> <2> inner<3></3> </2> xyz</1></0> cruft" +"\s*([ixsmdt]*)([:letter:]*)" "<0> <1>d</1><2></2></0> " +"(a|b)c*d" "a<0><1>b</1>cd</0>" + +# Non-capturing parens (?: stuff). Groups, but does not capture. +"(?:abc)*(tail)" "<0>abcabcabc<1>tail</1></0>" + +# Non-greedy *? quantifier +".*?(abc)" "<0> abx <1>abc</1></0> abc abc abc" +".*(abc)" "<0> abx abc abc abc <1>abc</1></0>" + +"((?:abc |xyz )*?)abc " "<0><1>xyz </1>abc </0>abc abc " +"((?:abc |xyz )*)abc " "<0><1>xyz abc abc </1>abc </0>" + +# Non-greedy +? quantifier +"(a+?)(a*)" "<0><1>a</1><2>aaaaaaaaaaaa</2></0>" +"(a+)(a*)" "<0><1>aaaaaaaaaaaaa</1><2></2></0>" + +"((ab)+?)((ab)*)" "<0><1><2>ab</2></1><3>ababababab<4>ab</4></3></0>" +"((ab)+)((ab)*)" "<0><1>abababababab<2>ab</2></1><3></3></0>" + +# Non-greedy ?? quantifier +"(ab)(ab)??(ab)??(ab)??(ab)??c" "<0><1>ab</1><4>ab</4><5>ab</5>c</0>" + +# Unicode Properties as naked elements in a pattern +"\p{Lu}+" "here we go ... <0>ABC</0> and no more." +"(\p{L}+)(\P{L}*?) (\p{Zs}*)" "7999<0><1>letters</1><2>4949%^&*(</2> <3> </3></0>" + +# \w and \W +"\w+" " $%^&*( <0>hello123</0>%^&*(" +"\W+" "<0> $%^&*( </0>hello123%^&*(" + +# \A match at beginning of input only. + ".*\Ahello" "<0>hello</0> hello" + ".*hello" "<0>hello hello</0>" +".*\Ahello" "stuff\nhello" # don't match after embedded new-line. + +# \b \B +# +".*?\b(.).*" "<0> $%^&*( <1>h</1>ello123%^&*()gxx</0>" +"\ba\b" "-<0>a</0>" +"\by\b" "xy" +"[ \b]" "<0>b</0>" # in a set, \b is a literal b. + +# Finds first chars of up to 5 words +"(?:.*?\b(\w))?(?:.*?\b(\w))?(?:.*?\b(\w))?(?:.*?\b(\w))?(?:.*?\b(\w))?" "<0><1>T</1>the <2>q</2>ick <3>b</3>rown <4>f</4></0>ox" + +"H.*?((?:\B.)+)" "<0>H<1>ello</1></0> " +".*?((?:\B.)+).*?((?:\B.)+).*?((?:\B.)+)" "<0>H<1>ello</1> <2> </2>g<3>oodbye</3></0> " + +"(?:.*?\b(.))?(?:.*?\b(.))?(?:.*?\b(.))?(?:.*?\b(.))?(?:.*?\b(.))?.*" "<0> \u0301 \u0301<1>A</1>\u0302BC\u0303\u0304<2> </2>\u0305 \u0306<3>X</3>\u0307Y\u0308</0>" + + +# +# Unicode word boundary mode +# +"(?w).*?\b" v "<0></0>hello, world" +"(?w).*?(\b.+?\b).*" v "<0><1> </1>123.45 </0>" +"(?w).*?(\b\d.*?\b).*" v "<0> <1>123.45</1> </0>" +".*?(\b.+?\b).*" "<0> <1>123</1>.45 </0>" +"(?w:.*?(\b\d.*?\b).*)" v "<0> <1>123.45</1> </0>" +"(?w:.*?(\b.+?\b).*)" v "<0><1>don't</1> </0>" +"(?w:.+?(\b\S.+?\b).*)" v "<0> <1>don't</1> </0>" +"(?w:(\b.+?)(\b.+?)(\b.+?)(\b.+?)(\b.+?)(\b.+?)(\b.+?).*)" v "<0><1>.</1><2> </2><3>,</3><4>:</4><5>$</5><6>37,000.50</6><7> </7> </0>" + +# +# Unicode word boundaries with Regions +# +"(?w).*?\b" v "abc<r><0>def</0></r>ghi" +"(?w).*?\b" v2 "abc<r>def<0></0></r>ghi" +"(?w).*?\b" v3 "abc<r>def</r>ghi" +#"(?w).*?\b" vb "abc<r><0>def</0></r>ghi" # TODO: bug. Ticket 6073 +#"(?w).*?\b" vb2 "abc<r>def</r>ghi" + + + +# . does not match new-lines +"." "\u000a\u000d\u0085\u000c\u000b\u2028\u2029<0>X</0>\u000aY" +"A." "A\u000a "# no match + +# \d for decimal digits +"\d*" "<0>0123456789\u0660\u06F9\u0969\u0A66\u17E2\uFF10\U0001D7CE\U0001D7FF</0>non-digits" +"\D+" "<0>non digits</0>" +"\D*(\d*)(\D*)" "<0>non-digits<1>3456666</1><2>more non digits</2></0>" + +# \Q...\E quote mode +"hel\Qlo, worl\Ed" "<0>hello, world</0>" +"\Q$*^^(*)?\A\E(a*)" "<0>$*^^(*)?\\A<1>aaaaaaaaaaaaaaa</1></0>" +"[abc\Q]\r\E]+" "<0>aaaccc]]]\\\\\\</0>\r..." # \Q ... \E escape in a [set] + +# UREGEX_LITERAL - entire pattern is a literal string, no escapes recognized. +# Note that data strings in test cases still get escape processing. +"abc\an\r\E\\abcd\u0031bye" Q "lead<0>abc\\an\\r\\E\\\\abcd\\u0031bye</0>extra" +"case insensitive \\ (l)iteral" Qi "stuff!! <0>cAsE InSenSiTiVE \\\\ (L)ITeral</0>" + +# \S and \s space characters +"\s+" "not_space<0> \t \r \n \u3000 \u2004 \u2028 \u2029</0>xyz" +"(\S+).*?(\S+).*" "<0><1>Not-spaces</1> <2>more-non-spaces</2> </0>" + +# \X consume one Grapheme Cluster. +"(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?" v "<0><1>A</1><2>B</2><3> </3><4>\r\n</4></0>" +"(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?" v "<0><1>A\u0301</1><2>\n</2><3>\u0305</3><4>a\u0302\u0303\u0304</4></0>" +"(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?" v "<0><1>\u1100\u1161\u11a8</1><2>\u115f\u11a2\u11f9</2></0>" +"(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?" v "<0><1>\u1100\uac01</1><2>\uac02</2><3>\uac03\u11b0</3></0>" +"(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?" v "<0><1>\u1100\u1101\uac02\u0301</1><2>\u1100</2></0>" +# Regional indicator pairs are grapheme clusters +"(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?" v "<0><1>\U0001f1e6\U0001f1e8</1><2>\U0001f1ea\U0001f1ff</2></0>" +# Grapheme Break rule 9b: Prepend x +"(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?(\X)?" v "<0><1>\U000111C2x</1></0>" + +# Grapheme clusters that straddle a match region. Matching is pinned to the region limits, +# giving boundaries inside grapheme clusters +"(\X)?(\X)?(\X)?" v "a\u0301<r><0><1>\u0301\u0301</1><2>z\u0302</2></0></r>\u0302\u0302" +# Same as previous test case, but without the region limits. +"(\X)?(\X)?(\X)?" v "<0><1>a\u0301\u0301\u0301</1><2>z\u0302\u0302\u0302</2></0>" + +# ^ matches only at beginning of line +".*^(Hello)" "<0><1>Hello</1></0> Hello Hello Hello Goodbye" +".*(Hello)" "<0>Hello Hello Hello <1>Hello</1></0> Goodbye" +".*^(Hello)" " Hello Hello Hello Hello Goodbye"# No Match + +# $ matches only at end of line, or before a newline preceding the end of line +".*?(Goodbye)$" zy "<0>Hello Goodbye Goodbye <1>Goodbye</1></0>" +".*?(Goodbye)" ZY "<0>Hello <1>Goodbye</1></0> Goodbye Goodbye" +".*?(Goodbye)$" z "Hello Goodbye> Goodbye Goodbye "# No Match + +".*?(Goodbye)$" zy "<0>Hello Goodbye Goodbye <1>Goodbye</1></0>\n" +".*?(Goodbye)$" zy "<0>Hello Goodbye Goodbye <1>Goodbye</1></0>\n" +".*?(Goodbye)$" zy "<0>Hello Goodbye Goodbye <1>Goodbye</1></0>\r\n" +".*?(Goodbye)$" z "Hello Goodbye Goodbye Goodbye\n\n"# No Match + +# \Z matches at end of input, like $ with default flags. +".*?(Goodbye)\Z" zy "<0>Hello Goodbye Goodbye <1>Goodbye</1></0>" +".*?(Goodbye)" ZY "<0>Hello <1>Goodbye</1></0> Goodbye Goodbye" +".*?(Goodbye)\Z" z "Hello Goodbye> Goodbye Goodbye "# No Match +"here$" z "here\nthe end"# No Match + +".*?(Goodbye)\Z" "<0>Hello Goodbye Goodbye <1>Goodbye</1></0>\n" +".*?(Goodbye)\Z" "<0>Hello Goodbye Goodbye <1>Goodbye</1></0>\n" +".*?(Goodbye)\Z" "<0>Hello Goodbye Goodbye <1>Goodbye</1></0>\r\n" +".*?(Goodbye)\Z" "Hello Goodbye Goodbye Goodbye\n\n"# No Match + +# \z matches only at the end of string. +# no special treatment of new lines. +# no dependencies on flag settings. +".*?(Goodbye)\z" zy "<0>Hello Goodbye Goodbye <1>Goodbye</1></0>" +".*?(Goodbye)\z" z "Hello Goodbye Goodbye Goodbye "# No Match +"here$" z "here\nthe end"# No Match + +".*?(Goodbye)\z" z "Hello Goodbye Goodbye Goodbye\n"# No Match +".*?(Goodbye)\n\z" zy "<0>Hello Goodbye Goodbye <1>Goodbye</1>\n</0>" +"abc\z|def" ZY "abc<0>def</0>" + +# (?# comment) doesn't muck up pattern +"Hello (?# this is a comment) world" " <0>Hello world</0>..." + +# Check some implementation corner cases base on the way literal strings are compiled. +"A" "<0>A</0>" +"AB" "<0>AB</0>ABABAB" +"AB+" "<0>ABBB</0>A" +"AB+" "<0>AB</0>ABAB" +"ABC+" "<0>ABC</0>ABC" +"ABC+" "<0>ABCCCC</0>ABC" +"(?:ABC)+" "<0>ABCABCABC</0>D" +"(?:ABC)DEF+" "<0>ABCDEFFF</0>D" +"AB\.C\eD\u0666E" "<0>AB.C\u001BD\u0666E</0>F" +"ab\Bde" "<0>abde</0>" + +# loop breaking +"(a?)*" "<0><1></1></0>xyz" +"(a?)+" "<0><1></1></0>xyz" +"^(?:a?b?)*$" "a--" +"(x?)*xyz" "<0>xx<1></1>xyz</0>" # Sligthtly weird, but correct. The "last" time through (x?), + # it matches the empty string. + +# Set expressions, basic operators and escapes work +# +"[\d]+" "<0>0123</0>abc/.," +"[^\d]+" "0123<0>abc/.,</0>" +"[\D]+" "0123<0>abc/.,</0>" +"[^\D]+" "<0>0123</0>abc/.," + +"[\s]+" "<0> \t</0>abc/.," +"[^\s]+" " \t<0>abc/.,</0>" +"[\S]+" " \t<0>abc/.,</0>" +"[^\S]+" "<0> \t</0>abc/.," + +"[\w]+" "<0>abc123</0> .,;" +"[^\w]+" "abc123<0> .,;</0>" +"[\W]+" "abc123<0> .,;</0>" +"[^\W]+" "<0>abc123</0> .,;" + +"[\z]+" "abc<0>zzz</0>def" # \z has no special meaning +"[^\z]+" "<0>abc</0>zzzdef" +"[\^]+" "abc<0>^^</0>" +"[^\^]+" "<0>abc</0>^^" + +"[\u0041c]+" "<0>AcAc</0>def" +"[\U00010002]+" "<0>\ud800\udc02</0>\U00010003" +"[^\U00010002]+" "<0>Hello</0>\x{10002}" +"[\x61b]+" "<0>abab</0>cde" +#"[\x6z]+" "\x06" #TODO: single hex digits should fail +"[\x{9}\x{75}\x{6d6}\x{6ba6}\x{6146B}\x{10ffe3}]+" "<0>\u0009\u0075\u06d6\u6ba6\U0006146B\U0010ffe3</0>abc" + +"[\N{LATIN CAPITAL LETTER TONE SIX}ab\N{VARIATION SELECTOR-70} ]+" "x<0> \u0184\U000E0135 ab</0>c" +"[\N{LATIN SMALL LETTER C}-\N{LATIN SMALL LETTER F}]+" "ab<0>cdef</0>ghi" + + + +# +# [set expressions], check the precedence of '-', '&', '--', '&&' +# '-' and '&', for compatibility with ICU UnicodeSet, have the same +# precedence as the implicit Union between adjacent items. +# '--' and '&&', for compatibility with Java, have lower precedence than +# the implicit Union operations. '--' and '&&' themselves +# have the same precedence, and group left to right. +# +"[[a-m]-[f-w]p]+" "<0>dep</0>fgwxyz" +"[^[a-m]-[f-w]p]+" "dep<0>fgwxyz</0>" + +"[[a-m]--[f-w]p]+" "<0>de</0>pfgwxyz" +"[^[a-m]--[f-w]p]+" "de<0>pfgwxyz</0>" + +"[[a-m]&[e-s]w]+" "<0>efmw</0>adnst" +"[^[a-m]&[e-s]w]+" "efmw<0>adnst</0>" + +"[[a-m]&[e-s]]+" "<0>efm</0>adnst" + + + +# {min,max} iteration qualifier +"A{3}BC" "<0>AAABC</0>" + +"(ABC){2,3}AB" "no matchAB" +"(ABC){2,3}AB" "ABCAB" +"(ABC){2,3}AB" "<0>ABC<1>ABC</1>AB</0>" +"(ABC){2,3}AB" "<0>ABCABC<1>ABC</1>AB</0>" +"(ABC){2,3}AB" "<0>ABCABC<1>ABC</1>AB</0>CAB" + +"(ABC){2}AB" "ABCAB" +"(ABC){2}AB" "<0>ABC<1>ABC</1>AB</0>" +"(ABC){2}AB" "<0>ABC<1>ABC</1>AB</0>CAB" +"(ABC){2}AB" "<0>ABC<1>ABC</1>AB</0>CABCAB" + +"(ABC){2,}AB" "ABCAB" +"(ABC){2,}AB" "<0>ABC<1>ABC</1>AB</0>" +"(ABC){2,}AB" "<0>ABCABC<1>ABC</1>AB</0>" +"(ABC){2,}AB" "<0>ABCABCABC<1>ABC</1>AB</0>" + +"X{0,0}ABC" "<0>ABC</0>" +"X{0,1}ABC" "<0>ABC</0>" + +"(?:Hello(!{1,3}) there){1}" "Hello there" +"(?:Hello(!{1,3}) there){1}" "<0>Hello<1>!</1> there</0>" +"(?:Hello(!{1,3}) there){1}" "<0>Hello<1>!!</1> there</0>" +"(?:Hello(!{1,3}) there){1}" "<0>Hello<1>!!!</1> there</0>" +"(?:Hello(!{1,3}) there){1}" "Hello!!!! there" + +# Nongreedy {min,max}? intervals +"(ABC){2,3}?AB" "no matchAB" +"(ABC){2,3}?AB" "ABCAB" +"(ABC){2,3}?AB" "<0>ABC<1>ABC</1>AB</0>" +"(ABC){2,3}?AB" "<0>ABC<1>ABC</1>AB</0>CAB" +"(ABC){2,3}?AB" "<0>ABC<1>ABC</1>AB</0>CABCAB" +"(ABC){2,3}?AX" "<0>ABCABC<1>ABC</1>AX</0>" +"(ABC){2,3}?AX" "ABC<0>ABCABC<1>ABC</1>AX</0>" + +# Possessive {min,max}+ intervals +"(ABC){2,3}+ABC" "ABCABCABC" +"(ABC){1,2}+ABC" "<0>ABC<1>ABC</1>ABC</0>" +"(?:(.)\1){2,5}+." "<0>aabbcc<1>d</1>de</0>x" + + +# Atomic Grouping +"(?>.*)abc" "abcabcabc" # no match. .* consumed entire string. +"(?>(abc{2,4}?))(c*)" "<0><1>abcc</1><2>ccc</2></0>ddd" +"(\.\d\d(?>[1-9]?))\d+" "1.625" +"(\.\d\d(?>[1-9]?))\d+" "1<0><1>.625</1>0</0>" + +# Possessive *+ +"(abc)*+a" "abcabcabc" +"(abc)*+a" "<0>abc<1>abc</1>a</0>b" +"(a*b)*+a" "<0><1>aaaab</1>a</0>aaa" + +# Possessive ?+ +"c?+ddd" "<0>cddd</0>" +"c?+cddd" "cddd" +"c?cddd" "<0>cddd</0>" + +# Back Reference +"(?:ab(..)cd\1)*" "<0>ab23cd23ab<1>ww</1>cdww</0>abxxcdyy" +"ab(?:c|(d?))(\1)" "<0>ab<1><2></2></1></0>c" +"ab(?:c|(d?))(\1)" "<0>ab<1>d</1><2>d</2></0>" +"ab(?:c|(d?))(\1)" "<0>ab<1></1><2></2></0>e" +"ab(?:c|(d?))(\1)" "<0>ab<1></1><2></2></0>" + +# Back References that hit/don't hit end +"(abcd) \1" z "abcd abc" +"(abcd) \1" Z "<0><1>abcd</1> abcd</0>" +"(abcd) \1" Z "<0><1>abcd</1> abcd</0> " + +# Case Insensitive back references that hit/don't hit end. +"(abcd) \1" zi "abcd abc" +"(abcd) \1" Zi "<0><1>abcd</1> ABCD</0>" +"(abcd) \1" Zi "<0><1>abcd</1> ABCD</0> " + +# Back references that hit/don't hit boundary limits. + +"(abcd) \1" z "<r>abcd abc</r>d " +"(abcd) \1" Z "<r><0><1>abcd</1> abcd</0></r> " +"(abcd) \1" Z "<r><0><1>abcd</1> abcd</0> </r>" + +"(abcd) \1" zi "<r>abcd abc</r>d " +"(abcd) \1" Zi "<r><0><1>abcd</1> abcd</0></r> " +"(abcd) \1" Zi "<r><0><1>abcd</1> abcd</0> </r>" + +# Back reference that fails match near the end of input without actually hitting the end. +"(abcd) \1" ZL "abcd abd" +"(abcd) \1" ZLi "abcd abd" + +# 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" +"ssl" i "abc<0>ẞl</0>xyz" +"FIND" i "can <0>find</0> ?" # fi ligature, \ufb01 +"find" i "can <0>FIND</0> ?" +"ῧ" i "xxx<0>ῧ</0>xxx" # Composed char (match string) decomposes when case-folded (pattern) + +# White space handling +"a b" "ab" +"abc " "abc" +"abc " "<0>abc </0>" +"ab[cd e]z" "<0>ab z</0>" +"ab\ c" "<0>ab c</0> " +"ab c" "<0>ab c</0> " +"ab c" x "ab c " +"ab\ c" x "<0>ab c</0> " + +# +# Pattern Flags +# +"(?u)abc" "<0>abc</0>" +"(?-u)abc" "<0>abc</0>" + +# +# \c escapes (Control-whatever) +# +"\cA" "<0>\u0001</0>" +"\ca" "<0>\u0001</0>" +"\c\x" "<0>\u001cx</0>" + + +#Multi-line mode +'b\s^' m "a\nb\n" +"(?m)^abc$" "abc \n abc\n<0>abc</0>\nabc" +"(?m)^abc$" 2 "abc \n abc\nabc\n<0>abc</0>" +"^abc$" 2 "abc \n abc\nabc\nabc" + +# Empty and full range +"[\u0000-\U0010ffff]+" "<0>abc\u0000\uffff\U00010000\U0010ffffzz</0>" +"[^\u0000-\U0010ffff]" "abc\u0000\uffff\U00010000\U0010ffffzz" +"[^a--a]+" "<0>abc\u0000\uffff\U00010000\U0010ffffzz</0>" + +# Free-spacing mode +"a b c # this is a comment" x "<0>abc</0> " +'^a (?#xxx) (?#yyy) {3}c' x "<0>aaac</0>" +"a b c [x y z]" x "abc " +"a b c [x y z]" x "a b c " +"a b c [x y z]" x "<0>abcx</0>yz" +"a b c [x y z]" x "<0>abcy</0>yz" + +# +# Look Behind +# +"(?<=a)b" "a<0>b</0>" +"(.*)(?<=[bc])" "<0><1>abc</1></0>d" +"(?<=(abc))def" "<1>abc</1><0>def</0>" # lookbehind precedes main match. +"(?<=ab|abc)xyz" "abwxyz" # ab matches, but not far enough. +"(?<=abc)cde" "abcde" +"(?<=abc|ab)cde" "ab<0>cde</0>" +"(?<=abc|ab)cde" "abc<0>cde</0>" + +"(?<=bc?c?c?)cd" "ab<0>cd</0>" +"(?<=bc?c?c?)cd" "abc<0>cd</0>" +"(?<=bc?c?c?)cd" "abcc<0>cd</0>" +"(?<=bc?c?c?)cd" "abccc<0>cd</0>" +"(?<=bc?c?c?)cd" "abcccccd" +"(?<=bc?c?c?)c+d" "ab<0>cccccd</0>" + +".*(?<=: ?)(\w*)" "<0>1:one 2: two 3:<1>three</1></0> " + +# +# Named Characters +# +"a\N{LATIN SMALL LETTER B}c" "<0>abc</0>" +"a\N{LATIN SMALL LETTER B}c" i "<0>abc</0>" +"a\N{LATIN SMALL LETTER B}c" i "<0>aBc</0>" +"a\N{LATIN SMALL LETTER B}c" "aBc" + +"\N{FULL STOP}*" "<0>...</0>abc" + +"$" "abc<0></0>" + +# +# Optimizations of .* at end of patterns +# +"abc.*" "<0>abcdef</0>" +"abc.*$" "<0>abcdef</0>" +"abc(.*)" "<0>abc<1>def</1></0>" +"abc(.*)" "<0>abc<1></1></0>" +"abc.*" "<0>abc</0>\ndef" +"abc.*" s "<0>abc\ndef</0>" +"abc.*$" s "<0>abc\ndef</0>" +"abc.*$" "abc\ndef" +"abc.*$" m "<0>abc</0>\ndef" +"abc.*\Z" m "abc\ndef" +"abc.*\Z" sm "<0>abc\ndef</0>" + +"abc*" "<0>abccc</0>d" +"abc*$" "<0>abccc</0>" +"ab(?:ab[xyz]\s)*" "<0>ababy abx </0>abc" + +"(?:(abc)|a)(?:bc)+" "<0>abc</0>" +"(?:(abc)|a)(?:bc)*" "<0><1>abc</1></0>" +"^[+\-]?[0-9]*\.?[0-9]*" "<0>123.456</0>" + +"ab.+yz" "<0>abc12345xyz</0>ttt" +"ab.+yz" s "<0>abc12345xyz</0>ttt" + +"ab.+yz" "abc123\n45xyzttt" +"ab.+yz" s "<0>abc12\n345xyz</0>ttt" + +"ab[0-9]+yz" "---abyz+++" +"ab[0-9]+yz" "---<0>ab1yz</0>+++" +"ab[0-9]+yz" "---<0>ab12yz</0>+++" +"ab[0-9]+yz" "---<0>ab123456yz</0>+++" + +"ab([0-9]+|[A-Z]+)yz" "---abyz+++" +"ab([0-9]+|[A-Z]+)yz" "---<0>ab<1>1</1>yz</0>+++" +"ab([0-9]+|[A-Z]+)yz" "---<0>ab<1>12</1>yz</0>+++" +"ab([0-9]+|[A-Z]+)yz" "---<0>ab<1>A</1>yz</0>+++" +"ab([0-9]+|[A-Z]+)yz" "---<0>ab<1>AB</1>yz</0>+++" +"ab([0-9]+|[A-Z]+)yz" "---<0>ab<1>ABCDE</1>yz</0>+++" + +# +# Hex format \x escaping +# +"ab\x63" "<0>abc</0>" +"ab\x09w" "<0>ab\u0009w</0>" +"ab\xabcdc" "<0>ab\u00abcdc</0>" +"ab\x{abcd}c" "<0>ab\uabcdc</0>" +"ab\x{101234}c" "<0>ab\U00101234c</0>" +"abα" "<0>abα</0>" + +# +# Octal Escaping. This conforms to Java conventions, not Perl. +"\0101\00\03\073\0154\01442" "<0>A\u0000\u0003\u003b\u006c\u0064\u0032</0>" +"\0776" "<0>\u003f\u0036</0>" # overflow, the 6 is literal. +"\0376xyz" "<0>\u00fexyz</0>" +"\08" E "<0>\u00008</0>" +"\0" E "x" + +# +# \u Surrogate Pairs +# +"\ud800\udc00" "<0>\U00010000</0>" +"\ud800\udc00*" "<0>\U00010000\U00010000\U00010000</0>\U00010001" +"\ud800\ud800\udc00" "<0>\ud800\U00010000</0>\U00010000\U00010000\U00010001" +"(\ud800)(\udc00)" "\U00010000" +"\U00010001+" "<0>\U00010001\U00010001</0>\udc01" + +# +# hitEnd with find() +# +"abc" Z "aa<0>abc</0> abcab" +"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" + + +# +# Bug 3225 + +"1|9" "<0>1</0>" +"1|9" "<0>9</0>" +"1*|9" "<0>1</0>" +"1*|9" "<0></0>9" + +"(?:a|ac)d" "<0>acd</0>" +"a|ac" "<0>a</0>c" + +# +# Bug 3320 +# +"(a([^ ]+)){0,} (c)" "<0><1>a<2>b</2></1> <3>c</3></0> " +"(a([^ ]+))* (c)" "<0><1>a<2>b</2></1> <3>c</3></0> " + +# +# Bug 3436 +# +"(.*?) *$" "<0><1>test</1> </0>" + +# +# Bug 4034 +# +"\D" "<0>A</0>BC\u00ffDEF" +"\d" "ABC\u00ffDEF" +"\D" "<0>\u00ff</0>DEF" +"\d" "\u00ffDEF" +"\D" "123<0>\u00ff</0>DEF" +"\D" "<0>\u0100</0>DEF" +"\D" "123<0>\u0100</0>DEF" + +# +#bug 4024, new line sequence handling +# +"(?m)^" "<0></0>AA\u000d\u000aBB\u000d\u000aCC\u000d\u000a" +"(?m)^" 2 "AA\u000d\u000a<0></0>BB\u000d\u000aCC\u000d\u000a" +"(?m)^" 3 "AA\u000d\u000aBB\u000d\u000a<0></0>CC\u000d\u000a" +"(?m)^" 4 "AA\u000d\u000aBB\u000d\u000aCC\u000d\u000a" + +"(?m)$" "AA<0></0>\u000d\u000aBB\u000d\u000aCC\u000d\u000a" +"(?m)$" 2 "AA\u000d\u000aBB<0></0>\u000d\u000aCC\u000d\u000a" +"(?m)$" 3 "AA\u000d\u000aBB\u000d\u000aCC<0></0>\u000d\u000a" +"(?m)$" 4 "AA\u000d\u000aBB\u000d\u000aCC\u000d\u000a<0></0>" +"(?m)$" 5 "AA\u000d\u000aBB\u000d\u000aCC\u000d\u000a" + +"$" "AA\u000d\u000aBB\u000d\u000aCC<0></0>\u000d\u000a" +"$" 2 "AA\u000d\u000aBB\u000d\u000aCC\u000d\u000a<0></0>" +"$" 3 "AA\u000d\u000aBB\u000d\u000aCC\u000d\u000a" + +"$" "\u000a\u0000a<0></0>\u000a" +"$" 2 "\u000a\u0000a\u000a<0></0>" +"$" 3 "\u000a\u0000a\u000a" + +"$" "<0></0>" +"$" 2 "" + +"$" "<0></0>\u000a" +"$" 2 "\u000a<0></0>" +"$" 3 "\u000a" + +"^" "<0></0>" +"^" 2 "" + +"\Z" "<0></0>" +"\Z" 2 "" +"\Z" 2 "\u000a<0></0>" +"\Z" "<0></0>\u000d\u000a" +"\Z" 2 "\u000d\u000a<0></0>" + + +# No matching ^ at interior new-lines if not in multi-line mode. +"^" "<0></0>AA\u000d\u000aBB\u000d\u000aCC\u000d\u000a" +"^" 2 "AA\u000d\u000aBB\u000d\u000aCC\u000d\u000a" + +# +# Dot-matches-any mode, and stopping at new-lines if off. +# +"." "<0>1</0>23\u000aXYZ" +"." 2 "1<0>2</0>3\u000aXYZ" +"." 3 "12<0>3</0>\u000aXYZ" +"." 4 "123\u000a<0>X</0>YZ" # . doesn't match newlines +"." 4 "123\u000b<0>X</0>YZ" +"." 4 "123\u000c<0>X</0>YZ" +"." 4 "123\u000d<0>X</0>YZ" +"." 4 "123\u000d\u000a<0>X</0>YZ" +"." 4 "123\u0085<0>X</0>YZ" +"." 4 "123\u2028<0>X</0>YZ" +"." 4 "123\u2029<0>X</0>YZ" +"." 4s "123<0>\u000a</0>XYZ" # . matches any +"." 4s "123<0>\u000b</0>XYZ" +"." 4s "123<0>\u000c</0>XYZ" +"." 4s "123<0>\u000d</0>XYZ" +"." 4s "123<0>\u000d\u000a</0>XYZ" +"." 4s "123<0>\u0085</0>XYZ" +"." 4s "123<0>\u2028</0>XYZ" +"." 4s "123<0>\u2029</0>XYZ" +".{6}" "123\u000a\u000dXYZ" +".{6}" s "<0>123\u000a\u000dX</0>Y" + + +# +# Ranges +# +".*" "abc<r><0>def</0></r>ghi" +"a" "aaa<r><0>a</0>aa</r>aaa" +"a" 2 "aaa<r>a<0>a</0>a</r>aaa" +"a" 3 "aaa<r>aa<0>a</0></r>aaa" +"a" 4 "aaa<r>aaa</r>aaa" +"a" "aaa<r><0>a</0>aa</r>aaa" + +# +# [set] parsing, systematically run through all of the parser states. +# +# +"[def]+" "abc<0>ddeeff</0>ghi" # set-open +"[^def]+" "<0>abc</0>defghi" +"[:digit:]+" "abc<0>123</0>def" +"[:^digit:]+" "<0>abc</0>123def" +"[\u005edef]+" "abc<0>de^f</0>ghi" + +"[]]+" "abc<0>]]]</0>[def" # set-open2 +"[^]]+" "<0>abc</0>]]][def" + +"[:Lu:]+" "abc<0>ABC</0>def" # set-posix +"[:Lu]+" "abc<0>uL::Lu</0>" +"[:^Lu]+" "abc<0>uL:^:Lu</0>" +"[:]+" "abc<0>:::</0>def" +"[:whats this:]" E " " +"[--]+" dE "-------" + +"[[nested]]+" "xyz[<0>nnetsteed</0>]abc" #set-start +"[\x{41}]+" "CB<0>AA</0>ZYX" +"[\[\]\\]+" "&*<0>[]\\</0>..." +"[*({<]+" "^&<0>{{(<<*</0>)))" + + +"[-def]+" "abc<0>def-ef-d</0>xyz" # set-start-dash +"[abc[--def]]" E " " + +"[x[&def]]+" "abc<0>def&</0>ghi" # set-start-amp +"[&& is bad at start]" E " " + +"[abc" E " " # set-after-lit +"[def]]" "abcdef" +"[def]]" "abcde<0>f]</0>]" + +"[[def][ghi]]+" "abc]<0>defghi</0>[xyz" # set-after-set +"[[def]ghi]+" "abc]<0>defghi</0>[xyz" +"[[[[[[[[[[[abc]" E " " +"[[abc]\p{Lu}]+" "def<0>abcABC</0>xyz" + +"[d-f]+" "abc<0>def</0>ghi" # set-after-range +"[d-f[x-z]]+" "abc<0>defxyzzz</0>gw" +"[\s\d]+" "abc<0> 123</0>def" +"[d-f\d]+" "abc<0>def123</0>ghi" +"[d-fr-t]+" "abc<0>defrst</0>uvw" + +"[abc--]" E " " # set-after-op +"[[def]&&]" E " " +"[-abcd---]+" "<0>abc</0>--" #[-abcd]--[-] +"[&abcd&&&ac]+" "b<0>ac&&ca</0>d" #[&abcd]&&[&ac] + +"[[abcd]&[ac]]+" "b<0>acac</0>d" # set-set-amp +"[[abcd]&&[ac]]+" "b<0>acac</0>d" +"[[abcd]&&ac]+" "b<0>acac</0>d" +"[[abcd]&ac]+" "<0>bacacd&&&</0>" + +"[abcd&[ac]]+" "<0>bacacd&&&</0>" #set-lit-amp +"[abcd&&[ac]]+" "b<0>acac</0>d" +"[abcd&&ac]+" "b<0>acac</0>d" + +"[[abcd]-[ac]]+" "a<0>bdbd</0>c" # set-set-dash +"[[abcd]--[ac]]+" "a<0>bdbd</0>c" +"[[abcd]--ac]+" "a<0>bdbd</0>c" +"[[abcd]-ac]+" "<0>bacacd---</0>" + +"[a-d--[b-c]]+" "b<0>adad</0>c" # set-range-dash +"[a-d--b-c]+" "b<0>adad</0>c" +"[a-d-[b-c]]+" "<0>bad-adc</0>" +"[a-d-b-c]+" "<0>bad-adc</0>" +"[\w--[b-c]]+" "b<0>adad</0>c" +"[\w--b-c]+" "b<0>adad</0>c" +"[\w-[b-c]]+" "<0>bad-adc</0>" +"[\w-b-c]+" "<0>bad-adc</0>" + +"[a-d&&[b-c]]+" "a<0>bcbc</0>d" # set-range-amp +"[a-d&&b-c]+" "a<0>bcbc</0>d" +"[a-d&[b-c]]+" "<0>abc&bcd</0>" +"[a-d&b-c]+" "<0>abc&bcd</0>" + +"[abcd--bc]+" "b<0>adda</0>c" # set-lit-dash +"[abcd--[bc]]+" "b<0>adda</0>c" +"[abcd-[bc]]+" "<0>bad--dac</0>xyz" +"[abcd-]+" "<0>bad--dac</0>xyz" + +"[abcd-\s]+" E "xyz<0>abcd --</0>xyz" # set-lit-dash-esc +"[abcd-\N{LATIN SMALL LETTER G}]+" "xyz-<0>abcdefg</0>hij-" +"[bcd-\{]+" "a<0>bcdefyz{</0>|}" + +"[\p{Ll}]+" "ABC<0>abc</0>^&*&" # set-escape +"[\P{Ll}]+" "abc<0>ABC^&*&</0>xyz" +"[\N{LATIN SMALL LETTER Q}]+" "mnop<0>qqq</0>rst" +"[\sa]+" "cb<0>a a </0>(*&" +"[\S]+" " <0>hello</0> " +"[\w]+" " <0>hello_world</0>! " +"[\W]+" "a<0> *$%#,</0>hello " +"[\d]+" "abc<0>123</0>def" +"[\D]+" "123<0>abc</0>567" +"[\$\#]+" "123<0>$#$#</0>\\" + +# +# Try each of the Java compatibility properties. +# These are checked here, while normal Unicode properties aren't, because +# these Java compatibility properties are implemented directly by regexp, while other +# properties are handled by ICU's Property and UnicodeSet APIs. +# +# These tests are only to verify that the names are recognized and the +# implementation isn't dead. They are not intended to verify that the +# function definitions are 100% correct. +# +"[:InBasic Latin:]+" "ΓΔΕΖΗΘ<0>hello, world.</0>ニヌネノハバパ" +"[:^InBasic Latin:]+" "<0>ΓΔΕΖΗΘ</0>hello, world.ニヌネノハバパ" +"\p{InBasicLatin}+" "ΓΔΕΖΗΘ<0>hello, world.</0>ニヌネノハバパ" +"\P{InBasicLatin}+" "<0>ΓΔΕΖΗΘ</0>hello, world.ニヌネノハバパ" +"\p{InGreek}+" "<0>ΓΔΕΖΗΘ</0>hello, world.ニヌネノハバパ" +"\p{InCombining Marks for Symbols}" "<0>\u20d0</0>" +"\p{Incombiningmarksforsymbols}" "<0>\u20d0</0>" + + +"\p{javaDefined}+" "\uffff<0>abcd</0>\U00045678" +"\p{javaDigit}+" "abc<0>1234</0>xyz" +"\p{javaIdentifierIgnorable}+" "abc<0>\u0000\u000e\u009f</0>xyz" +"\p{javaISOControl}+" "abc<0>\u0000\u000d\u0083</0>xyz" +"\p{javaJavaIdentifierPart}+" "#@!<0>abc123_$</0>;" +"\p{javaJavaIdentifierStart}+" "123\u0301<0>abc$_</0>%^&" +"\p{javaLetter}+" "123<0>abcDEF</0>&*()(" +"\p{javaLetterOrDigit}+" "$%^&*<0>123abcகஙசஜஞ</0>☺♘♚☔☎♬⚄⚡" +"\p{javaLowerCase}+" "ABC<0>def</0>&^%#:=" +"\p{javaMirrored}+" "ab$%<0>(){}[]</0>xyz" +"\p{javaSpaceChar}+" "abc<0> \u00ao\u2028</0>!@#" +"\p{javaSupplementaryCodePoint}+" "abc\uffff<0>\U00010000\U0010ffff</0>\u0000" +"\p{javaTitleCase}+" "abCE<0>Džῌᾨ</0>123" +"\p{javaUnicodeIdentifierStart}+" "123<0>abcⅣ</0>%^&&*" +"\p{javaUnicodeIdentifierPart}+" "%&&^<0>abc123\u0301\u0002</0>..." +"\p{javaUpperCase}+" "abc<0>ABC</0>123" +"\p{javaValidCodePoint}+" "<0>\u0000abc\ud800 unpaired \udfff |\U0010ffff</0>" +"\p{javaWhitespace}+" "abc\u00a0\u2007\u202f<0> \u0009\u001c\u001f\u2028</0>42" +"\p{all}+" "<0>123\u0000\U0010ffff</0>" +"\P{all}+" "123\u0000\U0010ffff" + +# [:word:] is implemented directly by regexp. Not a java compat property, but PCRE and others. + +"[:word:]+" ".??$<0>abc123ΓΔΕΖΗ_</0>%%%" +"\P{WORD}+" "<0>.??$</0>abc123ΓΔΕΖΗ_%%%" + +# +# Errors on unrecognized ASCII letter escape sequences. +# +"[abc\Y]+" "<0>abcY</0>" +"[abc\Y]+" eE "<0>abcY</0>" + +"(?:a|b|c|\Y)+" "<0>abcY</0>" +"(?:a|b|c|\Y)+" eE "<0>abcY</0>" + +"\Q\Y\E" e "<0>\\Y</0>" + +# +# Reported problem +# +"[a-\w]" E "x" + +# +# Bug 4045 +# +"A*" "<0>AAAA</0>" +"A*" 2 "AAAA<0></0>" +"A*" 3 "AAAA" +"A*" 4 "AAAA" +"A*" 5 "AAAA" +"A*" 6 "AAAA" +"A*" "<0></0>" +"A*" 2 "" +"A*" 3 "" +"A*" 4 "" +"A*" 5 "" + +# +# Bug 4046 +# +"(?m)^" "<0></0>AA\u000dBB\u000dCC\u000d" +"(?m)^" 2 "AA\u000d<0></0>BB\u000dCC\u000d" +"(?m)^" 3 "AA\u000dBB\u000d<0></0>CC\u000d" +"(?m)^" 4 "AA\u000dBB\u000dCC\u000d" +"(?m)^" 5 "AA\u000dBB\u000dCC\u000d" +"(?m)^" 6 "AA\u000dBB\u000dCC\u000d" + +"(?m)^" "<0></0>AA\u000d\u000aBB\u000d\u000aCC\u000d\u000a" +"(?m)^" 2 "AA\u000d\u000a<0></0>BB\u000d\u000aCC\u000d\u000a" +"(?m)^" 3 "AA\u000d\u000aBB\u000d\u000a<0></0>CC\u000d\u000a" +"(?m)^" 4 "AA\u000d\u000aBB\u000d\u000aCC\u000d\u000a" + +# +# Bug 4059 +# +"\w+" "<0>イチロー</0>" +"\b....\b." "<0>イチロー?</0>" + + +# +# Bug 4058 ICU Unicode Set patterns have an odd feature - +# A $ as the last character before the close bracket means match +# a \uffff, which means off the end of the string in transliterators. +# Didn't make sense for regular expressions, and is now fixed. +# +"[\$](P|C|D);" "<0>$<1>P</1>;</0>" +"[$](P|C|D);" "<0>$<1>P</1>;</0>" +"[$$](P|C|D);" "<0>$<1>P</1>;</0>" + +# +# bug 4888 Flag settings lost in some cases. +# +"((a){2})|(#)" is "no" +"((a){2})|(#)" is "<0><1>a<2>a</2></1></0>#" +"((a){2})|(#)" is "a<0><3>#</3></0>" + +"((a|b){2})|c" is "<0>c</0>" +"((a|b){2})|c" is "<0>C</0>" +"((a|b){2})|c" s "C" + +# +# bug 5617 ZWJ \u200d shouldn't cause word boundaries +# +".+?\b" "<0> </0>\u0935\u0915\u094D\u200D\u0924\u0947 " +".+?\b" 2 " <0>\u0935\u0915\u094D\u200D\u0924\u0947</0> " +".+?\b" 3 " \u0935\u0915\u094D\u200D\u0924\u0947 " + +# +# bug 5386 "^.*$" should match empty input +# +"^.*$" "<0></0>" +"^.*$" m "<0></0>" +"^.*$" "<0></0>\n" +"(?s)^.*$" "<0>\n</0>" + +# +# bug 5386 Empty pattern and empty input should match. +# +"" "<0></0>abc" +"" "<0></0>" + +# +# bug 5386 Range upper and lower bounds can be equal +# +"[a-a]" "<0>a</0>" + +# +# bug 5386 $* should not fail, should match empty string. +# +"$*" "<0></0>abc" + +# +# bug 5386 \Q ... \E escaping problem +# +"[a-z\Q-$\E]+" "QE<0>abc-def$</0>." + +# More reported 5386 Java comaptibility failures +# +"[^]*abb]*" "<0>kkkk</0>" +"\xa" "huh" # Java would like to be warned. +"^.*$" "<0></0>" + +# +# bug 5386 Empty left alternation should produce a zero length match. +# +"|a" "<0></0>a" +"$|ab" "<0>ab</0>" +"$|ba" "ab<0></0>" + +# +# bug 5386 Java compatibility for set expressions +# +"[a-z&&[cde]]+" "ab<0>cde</0>fg" + +# +# bug 6019 matches() needs to backtrack and check for a longer match if the +# first match(es) found don't match the entire input. +# +"a?|b" "<0></0>b" +"a?|b" M "<0>b</0>" +"a?|.*?u|stuff|d" M "<0>stuff</0>" +"a?|.*?(u)|stuff|d" M "<0>stuff<1>u</1></0>" +"a+?" "<0>a</0>aaaaaaaaaaaa" +"a+?" M "<0>aaaaaaaaaaaaa</0>" + +# +# Bug 7724. Expression to validate zip codes. +# +"(?!0{5})(\d{5})(?!-?0{4})(-?\d{4})?" "<0><1>94040</1><2>-3344</2></0>" +"(?!0{5})(\d{5})(?!-?0{4})(-?\d{4})?" "94040-0000" +"(?!0{5})(\d{5})(?!-?0{4})(-?\d{4})?" "00000-3344" + +# +# Bug 8666. Assertion failure on match, bad operand to JMP_SAV_X opcode. +# +"((.??)+|A)*" "<0><1><2></2></1></0>AAAAABBBBBCCCCCDDDDEEEEE" + +# +# Bug 8826. Incorrect results with case insensitive matches. +# +"AS(X)" i "aßx" +"AS.*" i "aßx" # Expansion of sharp s can't split between pattern terms. +"ASßS" i "<0>aßß</0>" # All one literal string, does match. +"ASß{1}S" i "aßß" # Pattern with terms, no match. +"aßx" i "<0>assx</0>" +"aßx" i "<0>ASSX</0>" +"aßx" i "<0>aßx</0>" +"ASS(.)" i "<0>aß<1>x</1></0>" + +# Case Insensitive, probe some corner cases. +"ass+" i "aß" # Second 's' in pattern is qualified, can't combine with first. +"as+" i "aß" +"aßs" i "as" # Can't match half of a ß +"aß+" i "<0>assssssss</0>s" +"aß+" i "<0>assßSssSSS</0>s" +"a(ß?)+" i "<0>assssssss<1></1></0>s" +"a(ß?)+" i "<0>a<1></1></0>zzzzzzzzs" + +"\U00010400" i "<0>\U00010428</0>" # case folded supplemental code point. + +"sstuff" i "<0>ßtuff</0>" # exercise optimizations on what chars can start a match. +"sstuff" i "s<0>ßtuff</0>" # exercise optimizations on what chars can start a match. +"ßtuff" i "s<0>sstuff</0>" +"ßtuff" i "s<0>Sstuff</0>" + +"a(..)\1" i "<0>A<1>bc</1>BC</0>def" +"(ß)\1" i "aa<0><1>ss</1>ß</0>zz" # Case insensitive back reference +"..(.)\1" i "<0>aa<1>ß</1>ss</0>" +"ab(..)\1" i "xx<0>ab<1>ss</1>ß</0>ss" + +" (ss) ((\1.*)|(.*))" i "<0> <1>ss</1> <2><4>sß</4></2></0>" # The back reference 'ss' must not match in 'sß' + +# Bug 9057 +# \u200c and \u200d should be word characters. +# +"\w+" " <0>abc\u200cdef\u200dghi</0> " +"\w+" i " <0>abc\u200cdef\u200dghi</0> " +"[\w]+" " <0>abc\u200cdef\u200dghi</0> " +"[\w]+" i " <0>abc\u200cdef\u200dghi</0> " + +# Bug 9283 +# uregex_open fails for look-behind assertion + case-insensitive + +"(ab)?(?<=ab)cd|ef" i "<0><1>ab</1>cd</0>" + +# Bug 9719 Loop breaking on (zero length match){3,} (unlimited upper bound). +# + +"(?:abc){1,}abc" "<0>abcabcabcabcabc</0>" +"(?:2*){2,}?a2\z" "<0>2a2</0>" +"(?:2*){2,}?a2\z" "2a3" +"(?:x?+){3,}+yz" "w<0>yz</0>" +"(2*){2,}?a2\\z" "2a3" +"(2*){2,}?a2\\z" "<0>2<1></1>a2\\z</0>" +"(2*){2,}?a2\z" "<0>2<1></1>a2</0>" + + +# Bug 10024 +# Incorrect (unbounded) longest match length with {1, 20} style quantifiers. +# Unbounded match is disallowed in look-behind expressions. +# Max match length is used to limit where to check for look-behind matches. + +"(?<=a{1,5})bc" "aaaa<0>bc</0>def" +"(?<=(?:aa){3,20})bc" "aaaaaa<0>bc</0>def" +"(?<!abc {1,100}|def {1,100}|ghi {1,100})jkl" "def jkl" +"(?<!abc {1,100}|def {1,100}|ghi {1,100})jkl" "rst <0>jkl</0>" +"(?<=a{11})bc" "aaaaaaaaaaa<0>bc</0>" +"(?<=a{11})bc" "aaaaaaaaaabc" +"(?<=a{1,})bc" E "aaaa<0>bc</0>def" # U_REGEX_LOOK_BEHIND_LIMIT error. +"(?<=(?:){11})bc" "<0>bc</0>" # Empty (?:) expression. + +# Bug 10835 +# Match Start Set not being correctly computed for case insensitive patterns. +# (Test here is to dump the compiled pattern & manually check the start set.) + +"(private|secret|confidential|classified|restricted)" i "hmm, <0><1>Classified</1></0> stuff" +"(private|secret|confidential|classified|restricted)" "hmm, Classified stuff" + +# Bug 10844 + +"^([\w\d:]+)$" "<0><1>DiesIst1Beispiel:text</1></0>" +"^([\w\d:]+)$" i "<0><1>DiesIst1Beispiel:text</1></0>" +"^(\w+\d\w+:\w+)$" "<0><1>DiesIst1Beispiel:text</1></0>" +"^(\w+\d\w+:\w+)$" i "<0><1>DiesIst1Beispiel:text</1></0>" + +# Bug 11049 +# Edge cases in find() when pattern match begins with set of code points +# and the match begins at the end of the string. + +"A|B|C" "hello <0>A</0>" +"A|B|C" "hello \U00011234" +"A|B|\U00012345" "hello <0>\U00012345</0>" +"A|B|\U00010000" "hello \ud800" + +# Bug 11369 +# Incorrect optimization of patterns with a zero length quantifier {0} + +"(.|b)(|b){0}\$(?#xxx){3}(?>\D*)" "AAAAABBBBBCCCCCDDDDEEEEE" +"(|b)ab(c)" "<0><1></1>ab<2>c</2></0>" +"(|b){0}a{3}(D*)" "<0>aaa<2></2></0>" +"(|b){0,1}a{3}(D*)" "<0><1></1>aaa<2></2></0>" +"((|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" + +# Bug 11507 Capture of an unpaired surrogate shouldn't allow a back reference to +# match half of a surrogate pair, but only another unpaired surrogate. +# +"pre(.)post\1" "pre\ud800post\ud800\udc00" +"pre(.)post\1" "<0>pre<1>\ud800</1>post\ud800</0> fin" +"pre(.)post\1" i "pre\ud800post\ud800\udc00" # case insensiteve backrefs take a different code path +"pre(.)post\1" i "<0>pre<1>\ud800</1>post\ud800</0> fin" + +# Bug 11554 +# +# Maximum match length computation was assuming UTF-16. +# Used in look-behind matches to constrain how far back to look. + +"(?<=a\x{100000})spam" "***a\x{100000}<0>spam</0>**" +"(?<=aą)spam" "**aą<0>spam</0>**" +"(?<=ąabc)spam" "**ąabc<0>spam</0>**" + +"(?<=a\x{100000})spam" "***a\x{100001}spam**" +"(?<=aą)spam" "**bąspam**" +"(?<=ąabc)spam" "**ąabxspam**" + +# with negative look-behind + +"(?<!a\x{100000})spam" "***a\x{100000}spam**" +"(?<!aą)spam" "**aąspam**" +"(?<!ąabc)spam" "**ąabcspam**" + +"(?<!a\x{100000})spam" "***a\x{100001}<0>spam</0>**" +"(?<!aą)spam" "**bą<0>spam</0>**" +"(?<!ąabc)spam" "**ąabx<0>spam</0>**" + +# Bug #12930 +# +# Minimum Match Length computation, int32_t overflow on an empty set in the pattern. +# The empty set, with no match possible, has a min match length of INT32_MAX. +# Was incremented subsequently. Caused assertion failure on pattern compile. + +"[^\u0000-\U0010ffff]bc?" "bc no match" +"[^\u0000-\U0010ffff]?bc?" "<0>bc</0> has a match" + +# Bug #12160 Hit End behavior after find fails to find. +# To match Java, should be true if find fails to find. +# +"abc" Z "<0>abc</0> abc abc xyz" +"abc" Z2 "abc <0>abc</0> abc xyz" +"abc" Z3 "abc abc <0>abc</0> xyz" +"abc" z4 "abc abc abc xyz" + +# Bug #13844 Verify that non-standard Java property names are recognized. +"[\p{IsAlphabetic}]" " <0>A</0>" +"[\P{IsAlphabetic}]" "A<0> </0>" +"[\p{IsIdeographic}]" "A<0>〆</0>" +"[\P{IsIdeographic}]" "〆<0>A</0>" +"[\p{IsLetter}]" " <0>A</0>" +"[\P{IsLetter}]" "A<0> </0>" +"[\p{Letter}]" " <0>A</0>" +"[\p{IsLowercase}]" "A<0>a</0>" +"[\P{IsLowercase}]" "a<0>A</0>" +"[\p{IsUppercase}]" "a<0>A</0>" +"[\P{IsUppercase}]" "A<0>a</0>" +"[\p{IsTitlecase}]" "D<0>Dz</0>" +"[\P{IsTitlecase}]" "Dz<0>D</0>" +"[\p{IsPunctuation}]" " <0>&</0>" +"[\P{IsPunctuation}]" "&<0> </0>" +"[\p{IsControl}]" " <0>\x{82}</0>" +"[\P{IsControl}]" "\x{82}<0> </0>" +"[\p{IsWhite_Space}]" "x<0> </0>" +"[\P{IsWhite_Space}]" " <0>x</0>" +"[\p{IsDigit}]" " <0>4</0>" +"[\P{IsDigit}]" "4<0> </0>" +"[\p{IsHex_Digit}]" " <0>F</0>" +"[\P{IsHex_Digit}]" "F<0> </0>" +"[\p{IsJoin_Control}]" " <0>\x{200d}</0>" +"[\P{IsJoin_Control}]" "\x{200d}<0> </0>" +"[\p{IsNoncharacter_Code_Point}]" "A<0>\x{5fffe}</0>" +"[\p{IsAssigned}]" "\x{10ffff}<0>a</0>" +"[\P{IsAssigned}]" "a<0>\x{10ffff}</0>" + +"[\p{InBasic Latin}]" "〆<0>A</0>" +"[\p{InBasicLatin}]" "〆<0>A</0>" +"[\p{InBasic-Latin}]" "〆<0>A</0>" # ICU accepts '-'; Java does not. +"[\p{InBasic_Latin}]" "〆<0>A</0>" +"[\p{Inbasiclatin}]" "〆<0>A</0>" +"[\p{inbasiclatin}]" E "〆<0>A</0>" # "In" must be cased as shown. Property name part is case insensitive. +"[\p{InCombining_Marks_for_Symbols}]" "a<0>\x{20DD}</0>" # COMBINING ENCLOSING CIRCLE + +"[\p{all}]*" "<0>\x{00}abc\x{10ffff}</0>" +"[\p{javaBadProperty}]" E "whatever" +"[\p{IsBadProperty}]" E "whatever" +"[\p{InBadBlock}]" E "whatever" +"[\p{In}]" E "whatever" +"[\p{Is}]" E "whatever" +"[\p{java}]" "x<0>ꦉ</0>" # Note: "java" is a valid script code. + +"[\p{javaLowerCase}]+" "A<0>a</0>" +"[\p{javaLowerCase}]+" i "<0>Aa</0>" +"[\P{javaLowerCase}]+" "<0>A</0>a" +"[\P{javaLowerCase}]+" i "Aa" # No Match because case fold of the set happens first, then negation. + # JDK is not case insensitive w named properties, even though + # the insensitive match flag is set. A JDK bug? + +"[a-z]+" i "<0>Aa</0>" # Matches JDK behavior. +"[^a-z]+" i "Aa" # (no match) which is JDK behavior. Case fold first, then negation. + +# Bug 20385. Assertion failure while compiling a negative look-behind expression consisting of a set with +# no contents. Meaning the [set] can never match. There is no syntax to directly express +# an empty set, so generate it by negating (^) a set of all code points. +# Also check empty sets in other contexts. + +"(?<![^[^a]a])" "<0></0>abc" + +"(?<![^\u0000-\U0010ffff])" "<0></0>abc" +"x(?<![^\u0000-\U0010ffff])" "<0>x</0>abc" +"x(?<![^\u0000-\U0010ffff]{1,5})" "<0>x</0>abc" +"x(?<![^\u0000-\U0010ffff]{0,5})" "xabc" + +"(?<=[^\u0000-\U0010ffff])" "abc" +"(x?<=[^\u0000-\U0010ffff])" "abc" +"x(?<=[^\u0000-\U0010ffff]{1,5})" "xabc" +"x(?<=[^\u0000-\U0010ffff]{0,5})" "<0>x</0>abc" + +"[^\u0000-\U0010ffff]" "a" +"[^[^\u0000-\U0010ffff]]" "<0>a</0>" + +"This is a string with (?:one |two |three )endings" "<0>This is a string with two endings</0>" + +# Bug ICU-20544. Similar to 20385, above. Assertion failure with a negative look-behind assertion containing +# a set with no contents. Look-behind pattern includes more than just the empty set. + +"(?<![ⰿ&&m]c)" "<0></0>abc" # note: first 'ⰿ' is \u2c3f, hence empty set. +"(?<![^\u0000-\U0010ffff]c)" "<0></0>abc" +"(?<=[^[^]]†)" "abc" # Problem also exists w positive look-behind + +# Bug ICU-20391. Crash in computation of minimum match length with nested look-around patterns. +# +"(?<=(?<=((?=)){0}+)" E "aaa" +"(?<=(?<=((?=)){0}+))" "<0></0>" +"(?<=c(?<=b((?=a)){1}+))" "aaa" +"abc(?=de(?=f))...g" "<0>abcdefg</0>" +"abc(?=de(?=f))...g" "abcdxfg" + +# Bug ICU-20618 Assertion failure with nested look-around expressions. +# +"(?<=(?<=b?(?=a)))" "hello, world." + +# Bug ICU-20939 +# Incorrect word \b boundaries w UTF-8 input and non-ASCII text +# +"(?w)\b" v2 "äää<0></0> äää" + +# Bug ICU-21492 Assertion failure with nested look-around expressions. +# +"(?<=(?:(?<=(?:(?<=(?:(?<=)){2})){3})){4}" E "<0></0>" # orig failure from bug report, w mismatched parens. +"(?:(?<=(?:(?<=)){2}))" "<0></0>" # Simplified case, with a valid pattern. + +# Random debugging, Temporary +# + +# +# Regexps from http://www.regexlib.com +# +"^[a-zA-Z]{1,2}[0-9][0-9A-Za-z]{0,1} {0,1}[0-9][A-Za-z]{2}$" G "<0>G1 1AA</0>" +"^[a-zA-Z]{1,2}[0-9][0-9A-Za-z]{0,1} {0,1}[0-9][A-Za-z]{2}$" G "<0>EH10 2QQ</0>" +"^[a-zA-Z]{1,2}[0-9][0-9A-Za-z]{0,1} {0,1}[0-9][A-Za-z]{2}$" G "<0>SW1 1ZZ</0>" +"^[a-zA-Z]{1,2}[0-9][0-9A-Za-z]{0,1} {0,1}[0-9][A-Za-z]{2}$" "G111 1AA" +"^[a-zA-Z]{1,2}[0-9][0-9A-Za-z]{0,1} {0,1}[0-9][A-Za-z]{2}$" "X10 WW" +"^[a-zA-Z]{1,2}[0-9][0-9A-Za-z]{0,1} {0,1}[0-9][A-Za-z]{2}$" "DDD 5WW" +#"^[\w\-]+(?:\.[\w\-]+)*@(?:[\w\-]+\.)+[a-zA-Z]{2,7}$" dG "<0>joe.tillis@unit.army.mil</0>" # TODO: \w in pattern +#"^[\w-]+(?:\.[\w-]+)*@(?:[\w-]+\.)+[a-zA-Z]{2,7}$" G "<0>jack_rabbit@slims.com</0>" # TODO: \w in pattern +#"^[\w-]+(?:\.[\w-]+)*@(?:[\w-]+\.)+[a-zA-Z]{2,7}$" G "<0>foo99@foo.co.uk</0>" # TODO: \w in pattern +#"^[\w-]+(?:\.[\w-]+)*@(?:[\w-]+\.)+[a-zA-Z]{2,7}$" "find_the_mistake.@foo.org" # TODO: \w in pattern +#"^[\w-]+(?:\.[\w-]+)*@(?:[\w-]+\.)+[a-zA-Z]{2,7}$" ".prefix.@some.net" +"^([a-zA-Z0-9_\-\.]+)@((\[[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.)|(([a-zA-Z0-9\-]+\.)+))([a-zA-Z]{2,4}|[0-9]{1,3})(\]?)$" G "<0>asmith@mactec.com</0>" +"^([a-zA-Z0-9_\-\.]+)@((\[[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.)|(([a-zA-Z0-9\-]+\.)+))([a-zA-Z]{2,4}|[0-9]{1,3})(\]?)$" G "<0>foo12@foo.edu</0>" +"^([a-zA-Z0-9_\-\.]+)@((\[[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.)|(([a-zA-Z0-9\-]+\.)+))([a-zA-Z]{2,4}|[0-9]{1,3})(\]?)$" G "<0>bob.smith@foo.tv</0>" +"^([a-zA-Z0-9_\-\.]+)@((\[[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.)|(([a-zA-Z0-9\-]+\.)+))([a-zA-Z]{2,4}|[0-9]{1,3})(\]?)$" "joe" +"^([a-zA-Z0-9_\-\.]+)@((\[[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.)|(([a-zA-Z0-9\-]+\.)+))([a-zA-Z]{2,4}|[0-9]{1,3})(\]?)$" "@foo.com" +"^([a-zA-Z0-9_\-\.]+)@((\[[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.)|(([a-zA-Z0-9\-]+\.)+))([a-zA-Z]{2,4}|[0-9]{1,3})(\]?)$" "a@a" +"^\d{1,2}\/\d{1,2}\/\d{4}$" G "<0>4/1/2001</0>" +"^\d{1,2}\/\d{1,2}\/\d{4}$" G "<0>12/12/2001</0>" +"^\d{1,2}\/\d{1,2}\/\d{4}$" G "<0>55/5/3434</0>" +"^\d{1,2}\/\d{1,2}\/\d{4}$" "1/1/01" +"^\d{1,2}\/\d{1,2}\/\d{4}$" "12 Jan 01" +"^\d{1,2}\/\d{1,2}\/\d{4}$" "1-1-2001" +"^(?:(?:(?:0?[13578]|1[02])(\/|-|\.)31)\1|(?:(?:0?[1,3-9]|1[0-2])(\/|-|\.)(?:29|30)\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:0?2(\/|-|\.)29\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:(?:0?[1-9])|(?:1[0-2]))(\/|-|\.)(?:0?[1-9]|1\d|2[0-8])\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" G "<0>01.1.02</0>" +"^(?:(?:(?:0?[13578]|1[02])(\/|-|\.)31)\1|(?:(?:0?[1,3-9]|1[0-2])(\/|-|\.)(?:29|30)\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:0?2(\/|-|\.)29\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:(?:0?[1-9])|(?:1[0-2]))(\/|-|\.)(?:0?[1-9]|1\d|2[0-8])\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" G "<0>11-30-2001</0>" +"^(?:(?:(?:0?[13578]|1[02])(\/|-|\.)31)\1|(?:(?:0?[1,3-9]|1[0-2])(\/|-|\.)(?:29|30)\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:0?2(\/|-|\.)29\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:(?:0?[1-9])|(?:1[0-2]))(\/|-|\.)(?:0?[1-9]|1\d|2[0-8])\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" G "<0>2/29/2000</0>" +"^(?:(?:(?:0?[13578]|1[02])(\/|-|\.)31)\1|(?:(?:0?[1,3-9]|1[0-2])(\/|-|\.)(?:29|30)\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:0?2(\/|-|\.)29\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:(?:0?[1-9])|(?:1[0-2]))(\/|-|\.)(?:0?[1-9]|1\d|2[0-8])\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" "02/29/01" +"^(?:(?:(?:0?[13578]|1[02])(\/|-|\.)31)\1|(?:(?:0?[1,3-9]|1[0-2])(\/|-|\.)(?:29|30)\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:0?2(\/|-|\.)29\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:(?:0?[1-9])|(?:1[0-2]))(\/|-|\.)(?:0?[1-9]|1\d|2[0-8])\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" "13/01/2002" +"^(?:(?:(?:0?[13578]|1[02])(\/|-|\.)31)\1|(?:(?:0?[1,3-9]|1[0-2])(\/|-|\.)(?:29|30)\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:0?2(\/|-|\.)29\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:(?:0?[1-9])|(?:1[0-2]))(\/|-|\.)(?:0?[1-9]|1\d|2[0-8])\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" "11/00/02" +"^(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9])\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[0-9])$" G "<0>127.0.0.1</0>" +"^(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9])\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[0-9])$" G "<0>255.255.255.0</0>" +"^(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9])\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[0-9])$" G "<0>192.168.0.1</0>" +"^(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9])\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[0-9])$" "1200.5.4.3" +"^(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9])\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[0-9])$" "abc.def.ghi.jkl" +"^(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9])\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0)\.(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[0-9])$" "255.foo.bar.1" +"(AUX|PRN|NUL|COM\d|LPT\d)+\s*$" G "<0>COM1</0>" +"(AUX|PRN|NUL|COM\d|LPT\d)+\s*$" G "<0>AUX</0>" +"(AUX|PRN|NUL|COM\d|LPT\d)+\s*$" G "<0>LPT1</0>" +"(AUX|PRN|NUL|COM\d|LPT\d)+\s*$" "image.jpg" +"(AUX|PRN|NUL|COM\d|LPT\d)+\s*$" "index.html" +"(AUX|PRN|NUL|COM\d|LPT\d)+\s*$" "readme.txt" +"^(?:(?:31(\/|-|\.)(?:0?[13578]|1[02]))\1|(?:(?:29|30)(\/|-|\.)(?:0?[1,3-9]|1[0-2])\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:29(\/|-|\.)0?2\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:0?[1-9]|1\d|2[0-8])(\/|-|\.)(?:(?:0?[1-9])|(?:1[0-2]))\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" G "<0>29/02/1972</0>" +"^(?:(?:31(\/|-|\.)(?:0?[13578]|1[02]))\1|(?:(?:29|30)(\/|-|\.)(?:0?[1,3-9]|1[0-2])\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:29(\/|-|\.)0?2\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:0?[1-9]|1\d|2[0-8])(\/|-|\.)(?:(?:0?[1-9])|(?:1[0-2]))\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" G "<0>5-9-98</0>" +"^(?:(?:31(\/|-|\.)(?:0?[13578]|1[02]))\1|(?:(?:29|30)(\/|-|\.)(?:0?[1,3-9]|1[0-2])\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:29(\/|-|\.)0?2\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:0?[1-9]|1\d|2[0-8])(\/|-|\.)(?:(?:0?[1-9])|(?:1[0-2]))\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" G "<0>10-11-2002</0>" +"^(?:(?:31(\/|-|\.)(?:0?[13578]|1[02]))\1|(?:(?:29|30)(\/|-|\.)(?:0?[1,3-9]|1[0-2])\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:29(\/|-|\.)0?2\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:0?[1-9]|1\d|2[0-8])(\/|-|\.)(?:(?:0?[1-9])|(?:1[0-2]))\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" "29/02/2003" +"^(?:(?:31(\/|-|\.)(?:0?[13578]|1[02]))\1|(?:(?:29|30)(\/|-|\.)(?:0?[1,3-9]|1[0-2])\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:29(\/|-|\.)0?2\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:0?[1-9]|1\d|2[0-8])(\/|-|\.)(?:(?:0?[1-9])|(?:1[0-2]))\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" "12/13/2002" +"^(?:(?:31(\/|-|\.)(?:0?[13578]|1[02]))\1|(?:(?:29|30)(\/|-|\.)(?:0?[1,3-9]|1[0-2])\2))(?:(?:1[6-9]|[2-9]\d)?\d{2})$|^(?:29(\/|-|\.)0?2\3(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00))))$|^(?:0?[1-9]|1\d|2[0-8])(\/|-|\.)(?:(?:0?[1-9])|(?:1[0-2]))\4(?:(?:1[6-9]|[2-9]\d)?\d{2})$" "1-1-1500" +"^(user=([a-z0-9]+,)*(([a-z0-9]+){1});)?(group=([a-z0-9]+,)*(([a-z0-9]+){1});)?(level=[0-9]+;)?$" G "<0>user=foo,bar,quux;group=manager,admin;level=100;</0>" +"^(user=([a-z0-9]+,)*(([a-z0-9]+){1});)?(group=([a-z0-9]+,)*(([a-z0-9]+){1});)?(level=[0-9]+;)?$" G "<0>group=nobody;level=24;</0>" +"^(user=([a-z0-9]+,)*(([a-z0-9]+){1});)?(group=([a-z0-9]+,)*(([a-z0-9]+){1});)?(level=[0-9]+;)?$" "user=foo" +"^(user=([a-z0-9]+,)*(([a-z0-9]+){1});)?(group=([a-z0-9]+,)*(([a-z0-9]+){1});)?(level=[0-9]+;)?$" "blahh" +"^(\(?\+?[0-9]*\)?)?[0-9_\- \(\)]*$" G "<0>(+44)(0)20-12341234</0>" +"^(\(?\+?[0-9]*\)?)?[0-9_\- \(\)]*$" G "<0>02012341234</0>" +"^(\(?\+?[0-9]*\)?)?[0-9_\- \(\)]*$" G "<0>+44 (0) 1234-1234</0>" +"^(\(?\+?[0-9]*\)?)?[0-9_\- \(\)]*$" "(44+)020-12341234" +"^(\(?\+?[0-9]*\)?)?[0-9_\- \(\)]*$" "12341234(+020)" +"\b(\w+)\s+\1\b" G "<0>Tell the the preacher</0>" +"\b(\w+)\s+\1\b" G "<0>some some</0>" +"\b(\w+)\s+\1\b" G "<0>hubba hubba</0>" +"\b(\w+)\s+\1\b" "once an annual report" +"\b(\w+)\s+\1\b" "mandate dated submissions" +"\b(\w+)\s+\1\b" "Hubba hubba" +"(^\+[0-9]{2}|^\+[0-9]{2}\(0\)|^\(\+[0-9]{2}\)\(0\)|^00[0-9]{2}|^0)([0-9]{9}$|[0-9\-\s]{10}$)" G "<0>+31235256677</0>" +"(^\+[0-9]{2}|^\+[0-9]{2}\(0\)|^\(\+[0-9]{2}\)\(0\)|^00[0-9]{2}|^0)([0-9]{9}$|[0-9\-\s]{10}$)" G "<0>+31(0)235256677</0>" +"(^\+[0-9]{2}|^\+[0-9]{2}\(0\)|^\(\+[0-9]{2}\)\(0\)|^00[0-9]{2}|^0)([0-9]{9}$|[0-9\-\s]{10}$)" G "<0>023-5256677</0>" +"(^\+[0-9]{2}|^\+[0-9]{2}\(0\)|^\(\+[0-9]{2}\)\(0\)|^00[0-9]{2}|^0)([0-9]{9}$|[0-9\-\s]{10}$)" "+3123525667788999" +"(^\+[0-9]{2}|^\+[0-9]{2}\(0\)|^\(\+[0-9]{2}\)\(0\)|^00[0-9]{2}|^0)([0-9]{9}$|[0-9\-\s]{10}$)" "3123525667788" +"(^\+[0-9]{2}|^\+[0-9]{2}\(0\)|^\(\+[0-9]{2}\)\(0\)|^00[0-9]{2}|^0)([0-9]{9}$|[0-9\-\s]{10}$)" "232-2566778" +"^[-+]?\d*\.?\d*$" G "<0>123</0>" +"^[-+]?\d*\.?\d*$" G "<0>+3.14159</0>" +"^[-+]?\d*\.?\d*$" G "<0>-3.14159</0>" +"^[-+]?\d*\.?\d*$" "abc" +"^[-+]?\d*\.?\d*$" "3.4.5" +"^[-+]?\d*\.?\d*$" "$99.95" +"^\$?([1-9]{1}[0-9]{0,2}(\,[0-9]{3})*(\.[0-9]{0,2})?|[1-9]{1}[0-9]{0,}(\.[0-9]{0,2})?|0(\.[0-9]{0,2})?|(\.[0-9]{1,2})?)$" G "<0>$1,234.50</0>" +"^\$?([1-9]{1}[0-9]{0,2}(\,[0-9]{3})*(\.[0-9]{0,2})?|[1-9]{1}[0-9]{0,}(\.[0-9]{0,2})?|0(\.[0-9]{0,2})?|(\.[0-9]{1,2})?)$" G "<0>$0.70</0>" +"^\$?([1-9]{1}[0-9]{0,2}(\,[0-9]{3})*(\.[0-9]{0,2})?|[1-9]{1}[0-9]{0,}(\.[0-9]{0,2})?|0(\.[0-9]{0,2})?|(\.[0-9]{1,2})?)$" G "<0>.7</0>" +"^\$?([1-9]{1}[0-9]{0,2}(\,[0-9]{3})*(\.[0-9]{0,2})?|[1-9]{1}[0-9]{0,}(\.[0-9]{0,2})?|0(\.[0-9]{0,2})?|(\.[0-9]{1,2})?)$" "$0,123.50" +"^\$?([1-9]{1}[0-9]{0,2}(\,[0-9]{3})*(\.[0-9]{0,2})?|[1-9]{1}[0-9]{0,}(\.[0-9]{0,2})?|0(\.[0-9]{0,2})?|(\.[0-9]{1,2})?)$" "$00.5" +"^[A-Z]{2}[0-9]{6}[A-DFM]{1}$" G "<0>AB123456D</0>" +"^[A-Z]{2}[0-9]{6}[A-DFM]{1}$" G "<0>AB123456F</0>" +"^[A-Z]{2}[0-9]{6}[A-DFM]{1}$" G "<0>AB123456M</0>" +"^[A-Z]{2}[0-9]{6}[A-DFM]{1}$" "AB123456E" +"^[A-Z]{2}[0-9]{6}[A-DFM]{1}$" "ab123456d" +#"(http|ftp|https):\/\/[\w]+(.[\w]+)([\w\-\.,@?^=%&:/~\+#]*[\w\-\@?^=%&/~\+#])?" G "<0>http://regxlib.com/Default.aspx</0>" # TODO: \w in pattern +#"(http|ftp|https):\/\/[\w]+(.[\w]+)([\w\-\.,@?^=%&:/~\+#]*[\w\-\@?^=%&/~\+#])?" G "<0>http://electronics.cnet.com/electronics/0-6342366-8-8994967-1.html</0>" # TODO: \w in pattern +#"(http|ftp|https):\/\/[\w]+(.[\w]+)([\w\-\.,@?^=%&:/~\+#]*[\w\-\@?^=%&/~\+#])?" "www.yahoo.com" # TODO: \w in pattern +"^[0-9]{4}\s{0,1}[a-zA-Z]{2}$" G "<0>2034AK</0>" +"^[0-9]{4}\s{0,1}[a-zA-Z]{2}$" G "<0>2034 AK</0>" +"^[0-9]{4}\s{0,1}[a-zA-Z]{2}$" G "<0>2034 ak</0>" +"^[0-9]{4}\s{0,1}[a-zA-Z]{2}$" "2034 AK" +"^[0-9]{4}\s{0,1}[a-zA-Z]{2}$" "321321 AKSSAA" +"((\d{2})|(\d))\/((\d{2})|(\d))\/((\d{4})|(\d{2}))" G "<0>4/5/91</0>" +"((\d{2})|(\d))\/((\d{2})|(\d))\/((\d{4})|(\d{2}))" G "<0>04/5/1991</0>" +"((\d{2})|(\d))\/((\d{2})|(\d))\/((\d{4})|(\d{2}))" G "<0>4/05/89</0>" +"((\d{2})|(\d))\/((\d{2})|(\d))\/((\d{4})|(\d{2}))" "4/5/1" +#"(^|\s|\()((([1-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-]((2[0-9]){1}|(3[01]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:])|(^|\s|\()((([0-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-](([11-31]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:|$|\>])){1}){1}){1}){1}" G "<0>01/01/2001 </0>" #TODO - \s in pattern. +"(^|\s|\()((([1-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-]((2[0-9]){1}|(3[01]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:])|(^|\s|\()((([0-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-](([11-31]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:|$|\>])){1}){1}){1}){1}" G "<0>01-01-2001:</0>" +"(^|\s|\()((([1-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-]((2[0-9]){1}|(3[01]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:])|(^|\s|\()((([0-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-](([11-31]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:|$|\>])){1}){1}){1}){1}" G "<0>(1-1-01)</0>" +"(^|\s|\()((([1-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-]((2[0-9]){1}|(3[01]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:])|(^|\s|\()((([0-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-](([11-31]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:|$|\>])){1}){1}){1}){1}" "13/1/2001" +"(^|\s|\()((([1-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-]((2[0-9]){1}|(3[01]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:])|(^|\s|\()((([0-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-](([11-31]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:|$|\>])){1}){1}){1}){1}" "1-32-2001" +"(^|\s|\()((([1-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-]((2[0-9]){1}|(3[01]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:])|(^|\s|\()((([0-9]){1}|([0][1-9]){1}|([1][012]){1}){1}[\/-](([11-31]){1}|([01][1-9]){1}|([1-9]){1}){1}[\/-](((19|20)([0-9][0-9]){1}|([0-9][0-9]){1})){1}(([\s|\)|:|$|\>])){1}){1}){1}){1}" "1-1-1801" +"^\d{3}\s?\d{3}$" G "<0>400 099</0>" +"^\d{3}\s?\d{3}$" G "<0>400099</0>" +"^\d{3}\s?\d{3}$" G "<0>400050</0>" +"^\d{3}\s?\d{3}$" "2345678" +"^\d{3}\s?\d{3}$" "12345" +"^\d{3}\s?\d{3}$" "asdf" +"^\D?(\d{3})\D?\D?(\d{3})\D?(\d{4})$" G "<0>(111) 222-3333</0>" +"^\D?(\d{3})\D?\D?(\d{3})\D?(\d{4})$" G "<0>1112223333</0>" +"^\D?(\d{3})\D?\D?(\d{3})\D?(\d{4})$" G "<0>111-222-3333</0>" +"^\D?(\d{3})\D?\D?(\d{3})\D?(\d{4})$" "11122223333" +"^\D?(\d{3})\D?\D?(\d{3})\D?(\d{4})$" "11112223333" +"^\D?(\d{3})\D?\D?(\d{3})\D?(\d{4})$" "11122233333" +"^#?([a-f]|[A-F]|[0-9]){3}(([a-f]|[A-F]|[0-9]){3})?$" G "<0>#00ccff</0>" +"^#?([a-f]|[A-F]|[0-9]){3}(([a-f]|[A-F]|[0-9]){3})?$" G "<0>#039</0>" +"^#?([a-f]|[A-F]|[0-9]){3}(([a-f]|[A-F]|[0-9]){3})?$" G "<0>ffffcc</0>" +"^#?([a-f]|[A-F]|[0-9]){3}(([a-f]|[A-F]|[0-9]){3})?$" "blue" +"^#?([a-f]|[A-F]|[0-9]){3}(([a-f]|[A-F]|[0-9]){3})?$" "0x000000" +"^#?([a-f]|[A-F]|[0-9]){3}(([a-f]|[A-F]|[0-9]){3})?$" "#ff000" +"^([0-9a-fA-F][0-9a-fA-F]:){5}([0-9a-fA-F][0-9a-fA-F])$" G "<0>01:23:45:67:89:ab</0>" +"^([0-9a-fA-F][0-9a-fA-F]:){5}([0-9a-fA-F][0-9a-fA-F])$" G "<0>01:23:45:67:89:AB</0>" +"^([0-9a-fA-F][0-9a-fA-F]:){5}([0-9a-fA-F][0-9a-fA-F])$" G "<0>fE:dC:bA:98:76:54</0>" +"^([0-9a-fA-F][0-9a-fA-F]:){5}([0-9a-fA-F][0-9a-fA-F])$" "01:23:45:67:89:ab:cd" +"^([0-9a-fA-F][0-9a-fA-F]:){5}([0-9a-fA-F][0-9a-fA-F])$" "01:23:45:67:89:Az" +"^([0-9a-fA-F][0-9a-fA-F]:){5}([0-9a-fA-F][0-9a-fA-F])$" "01:23:45:56:" +"^(http|https|ftp)\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(:[a-zA-Z0-9]*)?/?([a-zA-Z0-9\-\._\?\,\'/\\\+\&%\$#\=~])*$" G "<0>http://www.blah.com/~joe</0>" +"^(http|https|ftp)\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(:[a-zA-Z0-9]*)?/?([a-zA-Z0-9\-\._\?\,\'/\\\+\&%\$#\=~])*$" G "<0>ftp://ftp.blah.co.uk:2828/blah%20blah.gif</0>" +"^(http|https|ftp)\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(:[a-zA-Z0-9]*)?/?([a-zA-Z0-9\-\._\?\,\'/\\\+\&%\$#\=~])*$" G "<0>https://blah.gov/blah-blah.as</0>" +"^(http|https|ftp)\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(:[a-zA-Z0-9]*)?/?([a-zA-Z0-9\-\._\?\,\'/\\\+\&%\$#\=~])*$" "www.blah.com" +"^(http|https|ftp)\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(:[a-zA-Z0-9]*)?/?([a-zA-Z0-9\-\._\?\,\'/\\\+\&%\$#\=~])*$" "http://www.blah.com/I have spaces!" +"^(http|https|ftp)\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(:[a-zA-Z0-9]*)?/?([a-zA-Z0-9\-\._\?\,\'/\\\+\&%\$#\=~])*$" "ftp://blah_underscore/[nope]" +"^(([0-2]\d|[3][0-1])\/([0]\d|[1][0-2])\/[2][0]\d{2})$|^(([0-2]\d|[3][0-1])\/([0]\d|[1][0-2])\/[2][0]\d{2}\s([0-1]\d|[2][0-3])\:[0-5]\d\:[0-5]\d)$" G "<0>12/01/2002</0>" +"^(([0-2]\d|[3][0-1])\/([0]\d|[1][0-2])\/[2][0]\d{2})$|^(([0-2]\d|[3][0-1])\/([0]\d|[1][0-2])\/[2][0]\d{2}\s([0-1]\d|[2][0-3])\:[0-5]\d\:[0-5]\d)$" G "<0>12/01/2002 12:32:10</0>" +"^(([0-2]\d|[3][0-1])\/([0]\d|[1][0-2])\/[2][0]\d{2})$|^(([0-2]\d|[3][0-1])\/([0]\d|[1][0-2])\/[2][0]\d{2}\s([0-1]\d|[2][0-3])\:[0-5]\d\:[0-5]\d)$" "32/12/2002" +"^(([0-2]\d|[3][0-1])\/([0]\d|[1][0-2])\/[2][0]\d{2})$|^(([0-2]\d|[3][0-1])\/([0]\d|[1][0-2])\/[2][0]\d{2}\s([0-1]\d|[2][0-3])\:[0-5]\d\:[0-5]\d)$" "12/13/2001" +"^(([0-2]\d|[3][0-1])\/([0]\d|[1][0-2])\/[2][0]\d{2})$|^(([0-2]\d|[3][0-1])\/([0]\d|[1][0-2])\/[2][0]\d{2}\s([0-1]\d|[2][0-3])\:[0-5]\d\:[0-5]\d)$" "12/02/06" +"^[0-9](\.[0-9]+)?$" G "<0>1.2345</0>" +"^[0-9](\.[0-9]+)?$" G "<0>0.00001</0>" +"^[0-9](\.[0-9]+)?$" G "<0>7</0>" +"^[0-9](\.[0-9]+)?$" "12.2" +"^[0-9](\.[0-9]+)?$" "1.10.1" +"^[0-9](\.[0-9]+)?$" "15.98" +"^(?:[mM]{1,3})?(?:(?:[cC][dDmM])|(?:[dD]?(?:[cC]{1,3})?))?[lL]?(([xX])(?:\2{1,2}|[lL]|[cC])?)?((([iI])((\5{1,2})|[vV]|[xX]|[lL])?)|([vV]?([iI]{1,3})?))?$" G "<0>III</0>" +"^(?:[mM]{1,3})?(?:(?:[cC][dDmM])|(?:[dD]?(?:[cC]{1,3})?))?[lL]?(([xX])(?:\2{1,2}|[lL]|[cC])?)?((([iI])((\5{1,2})|[vV]|[xX]|[lL])?)|([vV]?([iI]{1,3})?))?$" G "<0>xiv</0>" +"^(?:[mM]{1,3})?(?:(?:[cC][dDmM])|(?:[dD]?(?:[cC]{1,3})?))?[lL]?(([xX])(?:\2{1,2}|[lL]|[cC])?)?((([iI])((\5{1,2})|[vV]|[xX]|[lL])?)|([vV]?([iI]{1,3})?))?$" G "<0>MCMLXLIX</0>" +"^(?:[mM]{1,3})?(?:(?:[cC][dDmM])|(?:[dD]?(?:[cC]{1,3})?))?[lL]?(([xX])(?:\2{1,2}|[lL]|[cC])?)?((([iI])((\5{1,2})|[vV]|[xX]|[lL])?)|([vV]?([iI]{1,3})?))?$" "iiV" +"^(?:[mM]{1,3})?(?:(?:[cC][dDmM])|(?:[dD]?(?:[cC]{1,3})?))?[lL]?(([xX])(?:\2{1,2}|[lL]|[cC])?)?((([iI])((\5{1,2})|[vV]|[xX]|[lL])?)|([vV]?([iI]{1,3})?))?$" "MCCM" +"^(?:[mM]{1,3})?(?:(?:[cC][dDmM])|(?:[dD]?(?:[cC]{1,3})?))?[lL]?(([xX])(?:\2{1,2}|[lL]|[cC])?)?((([iI])((\5{1,2})|[vV]|[xX]|[lL])?)|([vV]?([iI]{1,3})?))?$" "XXXX" +"^[-+]?[0-9]+[.]?[0-9]*([eE][-+]?[0-9]+)?$" G "<0>123</0>" +"^[-+]?[0-9]+[.]?[0-9]*([eE][-+]?[0-9]+)?$" G "<0>-123.35</0>" +"^[-+]?[0-9]+[.]?[0-9]*([eE][-+]?[0-9]+)?$" G "<0>-123.35e-2</0>" +"^[-+]?[0-9]+[.]?[0-9]*([eE][-+]?[0-9]+)?$" "abc" +"^[-+]?[0-9]+[.]?[0-9]*([eE][-+]?[0-9]+)?$" "123.32e" +"^[-+]?[0-9]+[.]?[0-9]*([eE][-+]?[0-9]+)?$" "123.32.3" +"^[a-zA-Z]+(([\'\,\.\- ][a-zA-Z ])?[a-zA-Z]*)*$" G "<0>T.F. Johnson</0>" +"^[a-zA-Z]+(([\'\,\.\- ][a-zA-Z ])?[a-zA-Z]*)*$" G "<0>John O'Neil</0>" +"^[a-zA-Z]+(([\'\,\.\- ][a-zA-Z ])?[a-zA-Z]*)*$" G "<0>Mary-Kate Johnson</0>" +"^[a-zA-Z]+(([\'\,\.\- ][a-zA-Z ])?[a-zA-Z]*)*$" "sam_johnson" +"^[a-zA-Z]+(([\'\,\.\- ][a-zA-Z ])?[a-zA-Z]*)*$" "Joe--Bob Jones" +"^[a-zA-Z]+(([\'\,\.\- ][a-zA-Z ])?[a-zA-Z]*)*$" "dfjsd0rd" +"^(20|21|22|23|[0-1]\d)[0-5]\d$" G "<0>1200</0>" +"^(20|21|22|23|[0-1]\d)[0-5]\d$" G "<0>1645</0>" +"^(20|21|22|23|[0-1]\d)[0-5]\d$" G "<0>2359</0>" +"^(20|21|22|23|[0-1]\d)[0-5]\d$" "2400" +"^(20|21|22|23|[0-1]\d)[0-5]\d$" "asbc" +"^(20|21|22|23|[0-1]\d)[0-5]\d$" "12:45" +/<[^>]*\n?.*=("|')?(.*\.jpg)("|')?.*\n?[^<]*>/ G '<0><td background="../img/img.jpg" ></0>' +/<[^>]*\n?.*=("|')?(.*\.jpg)("|')?.*\n?[^<]*>/ G "<0><img src=img.jpg ></0>" +/<[^>]*\n?.*=("|')?(.*\.jpg)("|')?.*\n?[^<]*>/ G "<0><img src='img.jpg'></0>" +/<[^>]*\n?.*=("|')?(.*\.jpg)("|')?.*\n?[^<]*>/ "= img.jpg" +/<[^>]*\n?.*=("|')?(.*\.jpg)("|')?.*\n?[^<]*>/ "img.jpg" +"^(\d{5}-\d{4}|\d{5})$|^([a-zA-Z]\d[a-zA-Z] \d[a-zA-Z]\d)$" G "<0>78754</0>" +"^(\d{5}-\d{4}|\d{5})$|^([a-zA-Z]\d[a-zA-Z] \d[a-zA-Z]\d)$" G "<0>78754-1234</0>" +"^(\d{5}-\d{4}|\d{5})$|^([a-zA-Z]\d[a-zA-Z] \d[a-zA-Z]\d)$" G "<0>G3H 6A3</0>" +"^(\d{5}-\d{4}|\d{5})$|^([a-zA-Z]\d[a-zA-Z] \d[a-zA-Z]\d)$" "78754-12aA" +"^(\d{5}-\d{4}|\d{5})$|^([a-zA-Z]\d[a-zA-Z] \d[a-zA-Z]\d)$" "7875A" +"^(\d{5}-\d{4}|\d{5})$|^([a-zA-Z]\d[a-zA-Z] \d[a-zA-Z]\d)$" "g3h6a3" +#"^([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))$" G "<0>bob@somewhere.com</0>" # TODO: \w in pattern +#"^([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))$" G "<0>bob.jones@[1.1.1.1]</0 # TODO: \w in pattern>" +#"^([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))$" G "<0>bob@a.b.c.d.info</0>" # TODO: \w in pattern +#"^([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))$" "bob@com" # TODO: \w in pattern +#"^([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))$" "bob.jones@some.where" # TODO: \w in pattern +#"^([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))$" "bob@1.1.1.123" # TODO: \w in pattern +#"^(([-\w \.]+)|(""[-\w \.]+"") )?<([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))>$" G "<0><ab@cd.ef></0>" # TODO: \w in pattern +#"^(([-\w \.]+)|(""[-\w \.]+"") )?<([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))>$" G "<0>bob A. jones <ab@cd.ef></0>" # TODO: \w in pattern +#"^(([-\w \.]+)|(""[-\w \.]+"") )?<([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))>$" G "<0>bob A. jones <ab@[1.1.1.111]></0>" # TODO: \w in pattern +#"^(([-\w \.]+)|(""[-\w \.]+"") )?<([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))>$" "ab@cd.ef" # TODO: \w in pattern +#"^(([-\w \.]+)|(""[-\w \.]+"") )?<([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))>$" ""bob A. jones <ab@cd.ef>" # TODO: \w in pattern +#"^(([-\w \.]+)|(""[-\w \.]+"") )?<([\w\-\.]+)@((\[([0-9]{1,3}\.){3}[0-9]{1,3}\])|(([\w\-]+\.)+)([a-zA-Z]{2,4}))>$" "bob A. jones <ab@1.1.1.111>" # TODO: \w in pattern +"^[A-Za-z]{1,2}[0-9A-Za-z]{1,2}[ ]?[0-9]{0,1}[A-Za-z]{2}$" G "<0>SW112LE</0>" +"^[A-Za-z]{1,2}[0-9A-Za-z]{1,2}[ ]?[0-9]{0,1}[A-Za-z]{2}$" G "<0>SW11 2LE</0>" +"^[A-Za-z]{1,2}[0-9A-Za-z]{1,2}[ ]?[0-9]{0,1}[A-Za-z]{2}$" G "<0>CR05LE</0>" +"^[A-Za-z]{1,2}[0-9A-Za-z]{1,2}[ ]?[0-9]{0,1}[A-Za-z]{2}$" "12CR0LE" +"^[A-Za-z]{1,2}[0-9A-Za-z]{1,2}[ ]?[0-9]{0,1}[A-Za-z]{2}$" "12CR 0LE" +"^[A-Za-z]{1,2}[0-9A-Za-z]{1,2}[ ]?[0-9]{0,1}[A-Za-z]{2}$" "SWLE05" +"20\d{2}(-|\/)((0[1-9])|(1[0-2]))(-|\/)((0[1-9])|([1-2][0-9])|(3[0-1]))(T|\s)(([0-1][0-9])|(2[0-3])):([0-5][0-9]):([0-5][0-9])" G "<0>2099-12-31T23:59:59</0>" +"20\d{2}(-|\/)((0[1-9])|(1[0-2]))(-|\/)((0[1-9])|([1-2][0-9])|(3[0-1]))(T|\s)(([0-1][0-9])|(2[0-3])):([0-5][0-9]):([0-5][0-9])" G "<0>2002/02/09 16:30:00</0>" +"20\d{2}(-|\/)((0[1-9])|(1[0-2]))(-|\/)((0[1-9])|([1-2][0-9])|(3[0-1]))(T|\s)(([0-1][0-9])|(2[0-3])):([0-5][0-9]):([0-5][0-9])" G "<0>2000-01-01T00:00:00</0>" +"20\d{2}(-|\/)((0[1-9])|(1[0-2]))(-|\/)((0[1-9])|([1-2][0-9])|(3[0-1]))(T|\s)(([0-1][0-9])|(2[0-3])):([0-5][0-9]):([0-5][0-9])" "2000-13-31T00:00:00" +"20\d{2}(-|\/)((0[1-9])|(1[0-2]))(-|\/)((0[1-9])|([1-2][0-9])|(3[0-1]))(T|\s)(([0-1][0-9])|(2[0-3])):([0-5][0-9]):([0-5][0-9])" "2002/02/33 24:00:00" +"20\d{2}(-|\/)((0[1-9])|(1[0-2]))(-|\/)((0[1-9])|([1-2][0-9])|(3[0-1]))(T|\s)(([0-1][0-9])|(2[0-3])):([0-5][0-9]):([0-5][0-9])" "2000-01-01 60:00:00" +"^((?:4\d{3})|(?:5[1-5]\d{2})|(?:6011)|(?:3[68]\d{2})|(?:30[012345]\d))[ -]?(\d{4})[ -]?(\d{4})[ -]?(\d{4}|3[4,7]\d{13})$" G "<0>6011567812345678</0>" +"^((?:4\d{3})|(?:5[1-5]\d{2})|(?:6011)|(?:3[68]\d{2})|(?:30[012345]\d))[ -]?(\d{4})[ -]?(\d{4})[ -]?(\d{4}|3[4,7]\d{13})$" G "<0>6011 5678 1234 5678</0>" +"^((?:4\d{3})|(?:5[1-5]\d{2})|(?:6011)|(?:3[68]\d{2})|(?:30[012345]\d))[ -]?(\d{4})[ -]?(\d{4})[ -]?(\d{4}|3[4,7]\d{13})$" G "<0>6011-5678-1234-5678</0>" +"^((?:4\d{3})|(?:5[1-5]\d{2})|(?:6011)|(?:3[68]\d{2})|(?:30[012345]\d))[ -]?(\d{4})[ -]?(\d{4})[ -]?(\d{4}|3[4,7]\d{13})$" "1234567890123456" +"^((((0[13578])|(1[02]))[\/]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\/]?(([0-2][0-9])|(30)))|(02[\/]?[0-2][0-9]))[\/]?\d{4}$" G "<0>01/01/2001</0>" +"^((((0[13578])|(1[02]))[\/]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\/]?(([0-2][0-9])|(30)))|(02[\/]?[0-2][0-9]))[\/]?\d{4}$" G "<0>02/29/2002</0>" +"^((((0[13578])|(1[02]))[\/]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\/]?(([0-2][0-9])|(30)))|(02[\/]?[0-2][0-9]))[\/]?\d{4}$" G "<0>12/31/2002</0>" +"^((((0[13578])|(1[02]))[\/]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\/]?(([0-2][0-9])|(30)))|(02[\/]?[0-2][0-9]))[\/]?\d{4}$" "1/1/02" +"^((((0[13578])|(1[02]))[\/]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\/]?(([0-2][0-9])|(30)))|(02[\/]?[0-2][0-9]))[\/]?\d{4}$" "02/30/2002" +"^((((0[13578])|(1[02]))[\/]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\/]?(([0-2][0-9])|(30)))|(02[\/]?[0-2][0-9]))[\/]?\d{4}$" "1/25/2002" +#"^(?=[^\&])(?:(?<scheme>[^:/?#]+):)?(?://(?<authority>[^/?#]*))?(?<path>[^?#]*)(?:\?(?<query>[^#]*))?(?:#(?<fragment>.*))?" G "<0>http://regexlib.com/REDetails.aspx?regexp_id=x#Details</0>" # out of context, can't work stand-alone +#"^(?=[^\&])(?:(?<scheme>[^:/?#]+):)?(?://(?<authority>[^/?#]*))?(?<path>[^?#]*)(?:\?(?<query>[^#]*))?(?:#(?<fragment>.*))?" "&" # out of context, can't work stand-alone +"^[-+]?\d+(\.\d+)?$" G "<0>123</0>" +"^[-+]?\d+(\.\d+)?$" G "<0>-123.45</0>" +"^[-+]?\d+(\.\d+)?$" G "<0>+123.56</0>" +"^[-+]?\d+(\.\d+)?$" "123x" +"^[-+]?\d+(\.\d+)?$" ".123" +"^[-+]?\d+(\.\d+)?$" "-123." +"^(\d{4}[- ]){3}\d{4}|\d{16}$" G "<0>1234-1234-1234-1234</0>" +"^(\d{4}[- ]){3}\d{4}|\d{16}$" G "<0>1234 1234 1234 1234</0>" +"^(\d{4}[- ]){3}\d{4}|\d{16}$" G "<0>1234123412341234</0>" +"^(\d{4}[- ]){3}\d{4}|\d{16}$" "Visa" +"^(\d{4}[- ]){3}\d{4}|\d{16}$" "1234" +"^(\d{4}[- ]){3}\d{4}|\d{16}$" "123-1234-12345" +"^((4\d{3})|(5[1-5]\d{2})|(6011))-?\d{4}-?\d{4}-?\d{4}|3[4,7]\d{13}$" G "<0>6011-1111-1111-1111</0>" +"^((4\d{3})|(5[1-5]\d{2})|(6011))-?\d{4}-?\d{4}-?\d{4}|3[4,7]\d{13}$" G "<0>5423-1111-1111-1111</0>" +"^((4\d{3})|(5[1-5]\d{2})|(6011))-?\d{4}-?\d{4}-?\d{4}|3[4,7]\d{13}$" G "<0>341111111111111</0>" +"^((4\d{3})|(5[1-5]\d{2})|(6011))-?\d{4}-?\d{4}-?\d{4}|3[4,7]\d{13}$" "4111-111-111-111" +"^((4\d{3})|(5[1-5]\d{2})|(6011))-?\d{4}-?\d{4}-?\d{4}|3[4,7]\d{13}$" "3411-1111-1111-111" +"^((4\d{3})|(5[1-5]\d{2})|(6011))-?\d{4}-?\d{4}-?\d{4}|3[4,7]\d{13}$" "Visa" +"^[A-Z0-9]{8}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{12}$" G "<0>4D28C5AD-6482-41CD-B84E-4573F384BB5C</0>" +"^[A-Z0-9]{8}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{12}$" G "<0>B1E1282C-A35C-4D5A-BF8B-7A3A51D9E388</0>" +"^[A-Z0-9]{8}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{12}$" G "91036A4A-A0F4-43F0-8CD" +"^[A-Z0-9]{8}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{12}$" "{B1E1282C-A35C-4D3A-BF8B-7A3A51D9E388}" +"^[A-Z0-9]{8}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{12}$" "AAAAAAAAAAAAAAAAA" +"^[A-Z0-9]{8}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{4}-[A-Z0-9]{12}$" "B;E1282C-A35C-4D3A-BF8B-7A3A51D9E38" +"(^(4|5)\d{3}-?\d{4}-?\d{4}-?\d{4}|(4|5)\d{15})|(^(6011)-?\d{4}-?\d{4}-?\d{4}|(6011)-?\d{12})|(^((3\d{3}))-\d{6}-\d{5}|^((3\d{14})))" G "<0>4111-1234-1234-1234</0>" +"(^(4|5)\d{3}-?\d{4}-?\d{4}-?\d{4}|(4|5)\d{15})|(^(6011)-?\d{4}-?\d{4}-?\d{4}|(6011)-?\d{12})|(^((3\d{3}))-\d{6}-\d{5}|^((3\d{14})))" G "<0>6011123412341234</0>" +"(^(4|5)\d{3}-?\d{4}-?\d{4}-?\d{4}|(4|5)\d{15})|(^(6011)-?\d{4}-?\d{4}-?\d{4}|(6011)-?\d{12})|(^((3\d{3}))-\d{6}-\d{5}|^((3\d{14})))" G "<0>3711-123456-12345</0>" +"(^(4|5)\d{3}-?\d{4}-?\d{4}-?\d{4}|(4|5)\d{15})|(^(6011)-?\d{4}-?\d{4}-?\d{4}|(6011)-?\d{12})|(^((3\d{3}))-\d{6}-\d{5}|^((3\d{14})))" "1234567890123456" +"(^(4|5)\d{3}-?\d{4}-?\d{4}-?\d{4}|(4|5)\d{15})|(^(6011)-?\d{4}-?\d{4}-?\d{4}|(6011)-?\d{12})|(^((3\d{3}))-\d{6}-\d{5}|^((3\d{14})))" "4111-123-1234-1234" +"(^(4|5)\d{3}-?\d{4}-?\d{4}-?\d{4}|(4|5)\d{15})|(^(6011)-?\d{4}-?\d{4}-?\d{4}|(6011)-?\d{12})|(^((3\d{3}))-\d{6}-\d{5}|^((3\d{14})))" "412-1234-1234-1234" +#'\[link="(?<link>((.|\n)*?))"\](?<text>((.|\n)*?))\[\/link\]' G '<0>[link="http://www.yahoo.com"]Yahoo[/link]</0>' #named capture +#'\[link="(?<link>((.|\n)*?))"\](?<text>((.|\n)*?))\[\/link\]' "[link]http://www.yahoo.com[/link]" #named capture +#'\[link="(?<link>((.|\n)*?))"\](?<text>((.|\n)*?))\[\/link\]' "[link=http://www.yahoo.com]Yahoo[/link]" #named capture +"^[a-zA-Z0-9]+$" G "<0>10a</0>" +"^[a-zA-Z0-9]+$" G "<0>ABC</0>" +"^[a-zA-Z0-9]+$" G "<0>A3fg</0>" +"^[a-zA-Z0-9]+$" "45.3" +"^[a-zA-Z0-9]+$" "this or that" +"^[a-zA-Z0-9]+$" "$23" +"((\(\d{3}\) ?)|(\d{3}-))?\d{3}-\d{4}" G "<0>(123) 456-7890</0>" +"((\(\d{3}\) ?)|(\d{3}-))?\d{3}-\d{4}" G "<0>123-456-7890</0>" +"((\(\d{3}\) ?)|(\d{3}-))?\d{3}-\d{4}" "1234567890" +"^[a-zA-Z]\w{3,14}$" G "<0>abcd</0>" +"^[a-zA-Z]\w{3,14}$" G "<0>aBc45DSD_sdf</0>" +"^[a-zA-Z]\w{3,14}$" G "<0>password</0>" +"^[a-zA-Z]\w{3,14}$" "afv" +"^[a-zA-Z]\w{3,14}$" "1234" +"^[a-zA-Z]\w{3,14}$" "reallylongpassword" +"^[A-Z]{1,2}[1-9][0-9]?[A-Z]? [0-9][A-Z]{2,}|GIR 0AA$" G "<0>G1 1AA </0>" +"^[A-Z]{1,2}[1-9][0-9]?[A-Z]? [0-9][A-Z]{2,}|GIR 0AA$" G "<0>GIR 0AA</0>" +"^[A-Z]{1,2}[1-9][0-9]?[A-Z]? [0-9][A-Z]{2,}|GIR 0AA$" G "<0>SW1 1ZZ</0>" +"^[A-Z]{1,2}[1-9][0-9]?[A-Z]? [0-9][A-Z]{2,}|GIR 0AA$" "BT01 3RT" +"^[A-Z]{1,2}[1-9][0-9]?[A-Z]? [0-9][A-Z]{2,}|GIR 0AA$" "G111 1AA" +"^0[23489]{1}(\-)?[^0\D]{1}\d{6}$" G "<0>03-6106666</0>" +"^0[23489]{1}(\-)?[^0\D]{1}\d{6}$" G "<0>036106666</0>" +"^0[23489]{1}(\-)?[^0\D]{1}\d{6}$" G "<0>02-5523344</0>" +"^0[23489]{1}(\-)?[^0\D]{1}\d{6}$" "00-6106666" +"^0[23489]{1}(\-)?[^0\D]{1}\d{6}$" "03-0106666" +"^0[23489]{1}(\-)?[^0\D]{1}\d{6}$" "02-55812346" +"^0(5[012345678]|6[47]){1}(\-)?[^0\D]{1}\d{5}$" G "<0>050-346634</0>" +"^0(5[012345678]|6[47]){1}(\-)?[^0\D]{1}\d{5}$" G "<0>058633633</0>" +"^0(5[012345678]|6[47]){1}(\-)?[^0\D]{1}\d{5}$" G "<0>064-228226</0>" +"^0(5[012345678]|6[47]){1}(\-)?[^0\D]{1}\d{5}$" "059-336622" +"^0(5[012345678]|6[47]){1}(\-)?[^0\D]{1}\d{5}$" "064-022663" +"^0(5[012345678]|6[47]){1}(\-)?[^0\D]{1}\d{5}$" "0545454545" +"^([A-Z]{1,2}[0-9]{1,2}|[A-Z]{3}|[A-Z]{1,2}[0-9][A-Z])( |-)[0-9][A-Z]{2}" G "<0>AA11 1AA</0>" +"^([A-Z]{1,2}[0-9]{1,2}|[A-Z]{3}|[A-Z]{1,2}[0-9][A-Z])( |-)[0-9][A-Z]{2}" G "<0>AA1A 1AA</0>" +"^([A-Z]{1,2}[0-9]{1,2}|[A-Z]{3}|[A-Z]{1,2}[0-9][A-Z])( |-)[0-9][A-Z]{2}" G "<0>A11-1AA</0>" +"^([A-Z]{1,2}[0-9]{1,2}|[A-Z]{3}|[A-Z]{1,2}[0-9][A-Z])( |-)[0-9][A-Z]{2}" "111 AAA" +"^([A-Z]{1,2}[0-9]{1,2}|[A-Z]{3}|[A-Z]{1,2}[0-9][A-Z])( |-)[0-9][A-Z]{2}" "1AAA 1AA" +"^([A-Z]{1,2}[0-9]{1,2}|[A-Z]{3}|[A-Z]{1,2}[0-9][A-Z])( |-)[0-9][A-Z]{2}" "A1AA 1AA" +"@{2}((\S)+)@{2}" G "<0>@@test@@</0>" +"@{2}((\S)+)@{2}" G "<0>@@name@@</0>" +"@{2}((\S)+)@{2}" G "<0>@@2342@@</0>" +"@{2}((\S)+)@{2}" "@test@" +"@{2}((\S)+)@{2}" "@@na me@@" +"@{2}((\S)+)@{2}" "@@ name@@" +"([0-1][0-9]|2[0-3]):[0-5][0-9]" G "<0>00:00</0>" +"([0-1][0-9]|2[0-3]):[0-5][0-9]" G "<0>13:59</0>" +"([0-1][0-9]|2[0-3]):[0-5][0-9]" G "<0>23:59</0>" +"([0-1][0-9]|2[0-3]):[0-5][0-9]" "24:00" +"([0-1][0-9]|2[0-3]):[0-5][0-9]" "23:60" +"^[+-]?([0-9]*\.?[0-9]+|[0-9]+\.?[0-9]*)([eE][+-]?[0-9]+)?$" G "<0>23</0>" +"^[+-]?([0-9]*\.?[0-9]+|[0-9]+\.?[0-9]*)([eE][+-]?[0-9]+)?$" G "<0>-17.e23</0>" +"^[+-]?([0-9]*\.?[0-9]+|[0-9]+\.?[0-9]*)([eE][+-]?[0-9]+)?$" G "<0>+.23e+2</0>" +"^[+-]?([0-9]*\.?[0-9]+|[0-9]+\.?[0-9]*)([eE][+-]?[0-9]+)?$" "+.e2" +"^[+-]?([0-9]*\.?[0-9]+|[0-9]+\.?[0-9]*)([eE][+-]?[0-9]+)?$" "23.17.5" +"^[+-]?([0-9]*\.?[0-9]+|[0-9]+\.?[0-9]*)([eE][+-]?[0-9]+)?$" "10e2.0" +"^([1-zA-Z0-1@.\s ]{1,255})$" G "<0>email@email.com</0>" +"^([1-zA-Z0-1@.\s ]{1,255})$" G "<0>My Name</0>" +"^([1-zA-Z0-1@.\s ]{1,255})$" G "<0>asdf12df</0>" +"^([1-zA-Z0-1@.\s ]{1,255})$" "‘,\*&$<>" +"^([1-zA-Z0-1@.\s ]{1,255})$" "1001' string" +"^((0[1-9])|(1[0-2]))\/(\d{4})$" G "<0>12/2002</0>" +"^((0[1-9])|(1[0-2]))\/(\d{4})$" G "<0>11/1900</0>" +"^((0[1-9])|(1[0-2]))\/(\d{4})$" G "<0>02/1977</0>" +"^((0[1-9])|(1[0-2]))\/(\d{4})$" "1/1977" +"^((0[1-9])|(1[0-2]))\/(\d{4})$" "00/000" +"^((0[1-9])|(1[0-2]))\/(\d{4})$" "15/2002" +"^\(\d{1,2}(\s\d{1,2}){1,2}\)\s(\d{1,2}(\s\d{1,2}){1,2})((-(\d{1,4})){0,1})$" G "<0>(0 34 56) 34 56 67</0>" +"^\(\d{1,2}(\s\d{1,2}){1,2}\)\s(\d{1,2}(\s\d{1,2}){1,2})((-(\d{1,4})){0,1})$" G "<0>(03 45) 5 67 67</0>" +"^\(\d{1,2}(\s\d{1,2}){1,2}\)\s(\d{1,2}(\s\d{1,2}){1,2})((-(\d{1,4})){0,1})$" G "<0>(0 45) 2 33 45-45</0>" +"^\(\d{1,2}(\s\d{1,2}){1,2}\)\s(\d{1,2}(\s\d{1,2}){1,2})((-(\d{1,4})){0,1})$" "(2345) 34 34" +"^\(\d{1,2}(\s\d{1,2}){1,2}\)\s(\d{1,2}(\s\d{1,2}){1,2})((-(\d{1,4})){0,1})$" "(0 56) 456 456" +"^\(\d{1,2}(\s\d{1,2}){1,2}\)\s(\d{1,2}(\s\d{1,2}){1,2})((-(\d{1,4})){0,1})$" "(3 45) 2 34-45678" +"(?:\d|I{1,3})?\s?\w{2,}\.?\s*\d{1,}\:\d{1,}-?,?\d{0,2}(?:,\d{0,2}){0,2}" G "<0>Genesis 3:3-4,6</0>" +"(?:\d|I{1,3})?\s?\w{2,}\.?\s*\d{1,}\:\d{1,}-?,?\d{0,2}(?:,\d{0,2}){0,2}" G "<0>II Sam 2:11,2</0>" +"(?:\d|I{1,3})?\s?\w{2,}\.?\s*\d{1,}\:\d{1,}-?,?\d{0,2}(?:,\d{0,2}){0,2}" G "<0>2 Tim 3:16</0>" +"(?:\d|I{1,3})?\s?\w{2,}\.?\s*\d{1,}\:\d{1,}-?,?\d{0,2}(?:,\d{0,2}){0,2}" "Genesis chap 3, verse 3" +"(?:\d|I{1,3})?\s?\w{2,}\.?\s*\d{1,}\:\d{1,}-?,?\d{0,2}(?:,\d{0,2}){0,2}" "2nd Samuel 2" +"(\[[Ii][Mm][Gg]\])(\S+?)(\[\/[Ii][Mm][Gg]\])" G "<0>[IMG]http://bleh.jpg[/IMG]</0>" +"(\[[Ii][Mm][Gg]\])(\S+?)(\[\/[Ii][Mm][Gg]\])" G "<0>[ImG]bleh[/imG]</0>" +"(\[[Ii][Mm][Gg]\])(\S+?)(\[\/[Ii][Mm][Gg]\])" G "<0>[img]ftp://login:pass@bleh.gif[/img]</0>" +"(\[[Ii][Mm][Gg]\])(\S+?)(\[\/[Ii][Mm][Gg]\])" '<img src="bleh.jpg">' +"^([0-9]{1,2})[./-]+([0-9]{1,2})[./-]+([0-9]{2}|[0-9]{4})$" G "<0>10/03/1979</0>" +"^([0-9]{1,2})[./-]+([0-9]{1,2})[./-]+([0-9]{2}|[0-9]{4})$" G "<0>1-1-02</0>" +"^([0-9]{1,2})[./-]+([0-9]{1,2})[./-]+([0-9]{2}|[0-9]{4})$" G "<0>01.1.2003</0>" +"^([0-9]{1,2})[./-]+([0-9]{1,2})[./-]+([0-9]{2}|[0-9]{4})$" "10/03/197" +"^([0-9]{1,2})[./-]+([0-9]{1,2})[./-]+([0-9]{2}|[0-9]{4})$" "01-02-003" +"^([0-9]{1,2})[./-]+([0-9]{1,2})[./-]+([0-9]{2}|[0-9]{4})$" "01 02 03" +#"^(?(^00000(|-0000))|(\d{5}(|-\d{4})))$" G "<0>12345</0>" # No Conditionals? +#"^(?(^00000(|-0000))|(\d{5}(|-\d{4})))$" G "<0>12345-6789</0>" # No Conditionals? +#"^(?(^00000(|-0000))|(\d{5}(|-\d{4})))$" "00000" # No Conditionals? +#"^(?(^00000(|-0000))|(\d{5}(|-\d{4})))$" "00000-0000" # No Conditionals? +#"^(?(^00000(|-0000))|(\d{5}(|-\d{4})))$" "a4650-465s" # No Conditionals? +"^((0?[1-9])|((1|2)[0-9])|30|31)$" G "<0>01</0>" +"^((0?[1-9])|((1|2)[0-9])|30|31)$" G "<0>12</0>" +"^((0?[1-9])|((1|2)[0-9])|30|31)$" G "<0>31</0>" +"^((0?[1-9])|((1|2)[0-9])|30|31)$" "123" +"^((0?[1-9])|((1|2)[0-9])|30|31)$" "32" +"^((0?[1-9])|((1|2)[0-9])|30|31)$" "abc" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?(\d{3}([\s\-./\\])?\d{4}|[a-zA-Z0-9]{7})$" G "<0>1.222.333.1234</0>" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?(\d{3}([\s\-./\\])?\d{4}|[a-zA-Z0-9]{7})$" G "<0>1-223-123-1232</0>" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?(\d{3}([\s\-./\\])?\d{4}|[a-zA-Z0-9]{7})$" G "<0>12223334444</0>" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?(\d{3}([\s\-./\\])?\d{4}|[a-zA-Z0-9]{7})$" "1.1.123123.123" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?(\d{3}([\s\-./\\])?\d{4}|[a-zA-Z0-9]{7})$" "12-1322-112-31" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?(\d{3}([\s\-./\\])?\d{4}|[a-zA-Z0-9]{7})$" "11231321131" +"^([A-PR-UWYZ0-9][A-HK-Y0-9][AEHMNPRTVXY0-9]?[ABEHMNPRVWXY0-9]? {1,2}[0-9][ABD-HJLN-UW-Z]{2}|GIR 0AA)$" G "<0>DN3 6GB</0>" +"^([A-PR-UWYZ0-9][A-HK-Y0-9][AEHMNPRTVXY0-9]?[ABEHMNPRVWXY0-9]? {1,2}[0-9][ABD-HJLN-UW-Z]{2}|GIR 0AA)$" G "<0>SW42 4RG</0>" +"^([A-PR-UWYZ0-9][A-HK-Y0-9][AEHMNPRTVXY0-9]?[ABEHMNPRVWXY0-9]? {1,2}[0-9][ABD-HJLN-UW-Z]{2}|GIR 0AA)$" G "<0>GIR 0AA</0>" +"^([A-PR-UWYZ0-9][A-HK-Y0-9][AEHMNPRTVXY0-9]?[ABEHMNPRVWXY0-9]? {1,2}[0-9][ABD-HJLN-UW-Z]{2}|GIR 0AA)$" "SEW4 5TY" +"^([A-PR-UWYZ0-9][A-HK-Y0-9][AEHMNPRTVXY0-9]?[ABEHMNPRVWXY0-9]? {1,2}[0-9][ABD-HJLN-UW-Z]{2}|GIR 0AA)$" "AA2C 4FG" +"^([A-PR-UWYZ0-9][A-HK-Y0-9][AEHMNPRTVXY0-9]?[ABEHMNPRVWXY0-9]? {1,2}[0-9][ABD-HJLN-UW-Z]{2}|GIR 0AA)$" "AA2 4CV" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z]).{4,8}$" G "<0>asD1</0>" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z]).{4,8}$" G "<0>asDF1234</0>" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z]).{4,8}$" G "<0>ASPgo123</0>" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z]).{4,8}$" "asdf" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z]).{4,8}$" "1234" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z]).{4,8}$" "ASDF12345" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?([0-9]{3}([\s\-./\\])?[0-9]{4}|[a-zA-Z0-9]{7}|([0-9]{3}[-][a-zA-Z0-9]{4}))" G "<0>1.222.333.1234</0>" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?([0-9]{3}([\s\-./\\])?[0-9]{4}|[a-zA-Z0-9]{7}|([0-9]{3}[-][a-zA-Z0-9]{4}))" G "<0>1-223-123-1232</0>" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?([0-9]{3}([\s\-./\\])?[0-9]{4}|[a-zA-Z0-9]{7}|([0-9]{3}[-][a-zA-Z0-9]{4}))" G "<0>1-888-425-DELL</0>" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?([0-9]{3}([\s\-./\\])?[0-9]{4}|[a-zA-Z0-9]{7}|([0-9]{3}[-][a-zA-Z0-9]{4}))" "1.1.123123.123" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?([0-9]{3}([\s\-./\\])?[0-9]{4}|[a-zA-Z0-9]{7}|([0-9]{3}[-][a-zA-Z0-9]{4}))" "12-1322-112-31" +"^([0-1]([\s\-./\\])?)?(\(?[2-9]\d{2}\)?|[2-9]\d{3})([\s\-./\\])?([0-9]{3}([\s\-./\\])?[0-9]{4}|[a-zA-Z0-9]{7}|([0-9]{3}[-][a-zA-Z0-9]{4}))" "1-800-CALL-DEL" +"^(([0]?[1-9]|1[0-2])(:)([0-5][0-9]))$" G "<0>09:00</0>" +"^(([0]?[1-9]|1[0-2])(:)([0-5][0-9]))$" G "<0>9:00</0>" +"^(([0]?[1-9]|1[0-2])(:)([0-5][0-9]))$" G "<0>11:35</0>" +"^(([0]?[1-9]|1[0-2])(:)([0-5][0-9]))$" "13:00" +"^(([0]?[1-9]|1[0-2])(:)([0-5][0-9]))$" "9.00" +"^(([0]?[1-9]|1[0-2])(:)([0-5][0-9]))$" "6:60" +"^([1-9]|[1-9]\d|1\d{2}|2[0-4]\d|25[0-5])$" G "<0>1</0>" +"^([1-9]|[1-9]\d|1\d{2}|2[0-4]\d|25[0-5])$" G "<0>108</0>" +"^([1-9]|[1-9]\d|1\d{2}|2[0-4]\d|25[0-5])$" G "<0>255</0>" +"^([1-9]|[1-9]\d|1\d{2}|2[0-4]\d|25[0-5])$" "01" +"^([1-9]|[1-9]\d|1\d{2}|2[0-4]\d|25[0-5])$" "256" +"^((((0[13578])|([13578])|(1[02]))[\/](([1-9])|([0-2][0-9])|(3[01])))|(((0[469])|([469])|(11))[\/](([1-9])|([0-2][0-9])|(30)))|((2|02)[\/](([1-9])|([0-2][0-9]))))[\/]\d{4}$|^\d{4}$" G "<0>01/01/2001</0>" +"^((((0[13578])|([13578])|(1[02]))[\/](([1-9])|([0-2][0-9])|(3[01])))|(((0[469])|([469])|(11))[\/](([1-9])|([0-2][0-9])|(30)))|((2|02)[\/](([1-9])|([0-2][0-9]))))[\/]\d{4}$|^\d{4}$" G "<0>1/01/2001</0>" +"^((((0[13578])|([13578])|(1[02]))[\/](([1-9])|([0-2][0-9])|(3[01])))|(((0[469])|([469])|(11))[\/](([1-9])|([0-2][0-9])|(30)))|((2|02)[\/](([1-9])|([0-2][0-9]))))[\/]\d{4}$|^\d{4}$" G "<0>2002</0>" +"^((((0[13578])|([13578])|(1[02]))[\/](([1-9])|([0-2][0-9])|(3[01])))|(((0[469])|([469])|(11))[\/](([1-9])|([0-2][0-9])|(30)))|((2|02)[\/](([1-9])|([0-2][0-9]))))[\/]\d{4}$|^\d{4}$" "2/30/2002" +"^((((0[13578])|([13578])|(1[02]))[\/](([1-9])|([0-2][0-9])|(3[01])))|(((0[469])|([469])|(11))[\/](([1-9])|([0-2][0-9])|(30)))|((2|02)[\/](([1-9])|([0-2][0-9]))))[\/]\d{4}$|^\d{4}$" "13/23/2002" +"^((((0[13578])|([13578])|(1[02]))[\/](([1-9])|([0-2][0-9])|(3[01])))|(((0[469])|([469])|(11))[\/](([1-9])|([0-2][0-9])|(30)))|((2|02)[\/](([1-9])|([0-2][0-9]))))[\/]\d{4}$|^\d{4}$" "12345" +"^[A-Za-z]{2}[0-9]{6}[A-Za-z]{1}$" G "<0>SP939393H</0>" +"^[A-Za-z]{2}[0-9]{6}[A-Za-z]{1}$" G "<0>PX123456D</0>" +"^[A-Za-z]{2}[0-9]{6}[A-Za-z]{1}$" G "<0>SW355667G</0>" +"^[A-Za-z]{2}[0-9]{6}[A-Za-z]{1}$" "12SP9393H" +"^[A-Za-z]{2}[0-9]{6}[A-Za-z]{1}$" "S3P93930D" +"^[A-Za-z]{2}[0-9]{6}[A-Za-z]{1}$" "11223344SP00ddSS" +"(^0[78][2347][0-9]{7})" G "<0>0834128458</0>" +"(^0[78][2347][0-9]{7})" G "<0>0749526308</0>" +"(^0[78][2347][0-9]{7})" "0861212308" +"(^0[78][2347][0-9]{7})" "0892549851" +"^([A-HJ-TP-Z]{1}\d{4}[A-Z]{3}|[a-z]{1}\d{4}[a-hj-tp-z]{3})$" G "<0>C1406HHA</0>" +"^([A-HJ-TP-Z]{1}\d{4}[A-Z]{3}|[a-z]{1}\d{4}[a-hj-tp-z]{3})$" G "<0>A4126AAB</0>" +"^([A-HJ-TP-Z]{1}\d{4}[A-Z]{3}|[a-z]{1}\d{4}[a-hj-tp-z]{3})$" G "<0>c1406hha</0>" +"^([A-HJ-TP-Z]{1}\d{4}[A-Z]{3}|[a-z]{1}\d{4}[a-hj-tp-z]{3})$" "c1406HHA" +"^([A-HJ-TP-Z]{1}\d{4}[A-Z]{3}|[a-z]{1}\d{4}[a-hj-tp-z]{3})$" "4126" +"^([A-HJ-TP-Z]{1}\d{4}[A-Z]{3}|[a-z]{1}\d{4}[a-hj-tp-z]{3})$" "C1406hha" +"^(((25[0-5]|2[0-4][0-9]|19[0-1]|19[3-9]|18[0-9]|17[0-1]|17[3-9]|1[0-6][0-9]|1[1-9]|[2-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9]))|(192\.(25[0-5]|2[0-4][0-9]|16[0-7]|169|1[0-5][0-9]|1[7-9][0-9]|[1-9][0-9]|[0-9]))|(172\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|1[0-5]|3[2-9]|[4-9][0-9]|[0-9])))\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])$" G "<0>66.129.71.120</0>" +"^(((25[0-5]|2[0-4][0-9]|19[0-1]|19[3-9]|18[0-9]|17[0-1]|17[3-9]|1[0-6][0-9]|1[1-9]|[2-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9]))|(192\.(25[0-5]|2[0-4][0-9]|16[0-7]|169|1[0-5][0-9]|1[7-9][0-9]|[1-9][0-9]|[0-9]))|(172\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|1[0-5]|3[2-9]|[4-9][0-9]|[0-9])))\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])$" G "<0>207.46.230.218</0>" +"^(((25[0-5]|2[0-4][0-9]|19[0-1]|19[3-9]|18[0-9]|17[0-1]|17[3-9]|1[0-6][0-9]|1[1-9]|[2-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9]))|(192\.(25[0-5]|2[0-4][0-9]|16[0-7]|169|1[0-5][0-9]|1[7-9][0-9]|[1-9][0-9]|[0-9]))|(172\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|1[0-5]|3[2-9]|[4-9][0-9]|[0-9])))\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])$" G "<0>64.58.76.225</0>" +"^(((25[0-5]|2[0-4][0-9]|19[0-1]|19[3-9]|18[0-9]|17[0-1]|17[3-9]|1[0-6][0-9]|1[1-9]|[2-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9]))|(192\.(25[0-5]|2[0-4][0-9]|16[0-7]|169|1[0-5][0-9]|1[7-9][0-9]|[1-9][0-9]|[0-9]))|(172\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|1[0-5]|3[2-9]|[4-9][0-9]|[0-9])))\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])$" "10.0.5.4" +"^(((25[0-5]|2[0-4][0-9]|19[0-1]|19[3-9]|18[0-9]|17[0-1]|17[3-9]|1[0-6][0-9]|1[1-9]|[2-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9]))|(192\.(25[0-5]|2[0-4][0-9]|16[0-7]|169|1[0-5][0-9]|1[7-9][0-9]|[1-9][0-9]|[0-9]))|(172\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|1[0-5]|3[2-9]|[4-9][0-9]|[0-9])))\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])$" "192.168.0.1" +"^(((25[0-5]|2[0-4][0-9]|19[0-1]|19[3-9]|18[0-9]|17[0-1]|17[3-9]|1[0-6][0-9]|1[1-9]|[2-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9]))|(192\.(25[0-5]|2[0-4][0-9]|16[0-7]|169|1[0-5][0-9]|1[7-9][0-9]|[1-9][0-9]|[0-9]))|(172\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|1[0-5]|3[2-9]|[4-9][0-9]|[0-9])))\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])\.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9][0-9]|[0-9])$" "my ip address" +"^([\w\d\-\.]+)@{1}(([\w\d\-]{1,67})|([\w\d\-]+\.[\w\d\-]{1,67}))\.(([a-zA-Z\d]{2,4})(\.[a-zA-Z\d]{2})?)$" G "<0>foo@foo.com</0>" +"^([\w\d\-\.]+)@{1}(([\w\d\-]{1,67})|([\w\d\-]+\.[\w\d\-]{1,67}))\.(([a-zA-Z\d]{2,4})(\.[a-zA-Z\d]{2})?)$" G "<0>foo@foo-foo.com.au</0>" +"^([\w\d\-\.]+)@{1}(([\w\d\-]{1,67})|([\w\d\-]+\.[\w\d\-]{1,67}))\.(([a-zA-Z\d]{2,4})(\.[a-zA-Z\d]{2})?)$" G "<0>foo@foo.foo.info</0>" +"^([\w\d\-\.]+)@{1}(([\w\d\-]{1,67})|([\w\d\-]+\.[\w\d\-]{1,67}))\.(([a-zA-Z\d]{2,4})(\.[a-zA-Z\d]{2})?)$" "foo@.com" +"^([\w\d\-\.]+)@{1}(([\w\d\-]{1,67})|([\w\d\-]+\.[\w\d\-]{1,67}))\.(([a-zA-Z\d]{2,4})(\.[a-zA-Z\d]{2})?)$" "foo@foo..com" +"^([\w\d\-\.]+)@{1}(([\w\d\-]{1,67})|([\w\d\-]+\.[\w\d\-]{1,67}))\.(([a-zA-Z\d]{2,4})(\.[a-zA-Z\d]{2})?)$" "foo@me@.com" +"/\*[\d\D]*?\*/" G "<0>/* my comment */</0>" +"/\*[\d\D]*?\*/" G "<0>/* my multiline comment */</0>" +"/\*[\d\D]*?\*/" G "<0>/* my nested comment */</0>" +"/\*[\d\D]*?\*/" "*/ anything here /*" +"/\*[\d\D]*?\*/" "anything between 2 separate comments" +"/\*[\d\D]*?\*/" "\* *\" +"/\*[\p{N}\P{N}]*?\*/" G "<0>/* my comment */</0>" +"/\*[\p{N}\P{N}]*?\*/" G "<0>/* my multiline comment */</0>" +"/\*[\p{N}\P{N}]*?\*/" G "<0>/* my nested comment */</0>" +"/\*[\p{N}\P{N}]*?\*/" "*/ anything here /*" +"/\*[\p{N}\P{N}]*?\*/" "anything between 2 separate comments" +"/\*[\p{N}\P{N}]*?\*/" "\* *\" +"((0?[13578]|10|12)(-|\/)((0[0-9])|([12])([0-9]?)|(3[01]?))(-|\/)((\d{4})|(\d{2}))|(0?[2469]|11)(-|\/)((0[0-9])|([12])([0-9]?)|(3[0]?))(-|\/)((\d{4}|\d{2})))" G "<0>1/31/2002</0>" +"((0?[13578]|10|12)(-|\/)((0[0-9])|([12])([0-9]?)|(3[01]?))(-|\/)((\d{4})|(\d{2}))|(0?[2469]|11)(-|\/)((0[0-9])|([12])([0-9]?)|(3[0]?))(-|\/)((\d{4}|\d{2})))" G "<0>04-30-02</0>" +"((0?[13578]|10|12)(-|\/)((0[0-9])|([12])([0-9]?)|(3[01]?))(-|\/)((\d{4})|(\d{2}))|(0?[2469]|11)(-|\/)((0[0-9])|([12])([0-9]?)|(3[0]?))(-|\/)((\d{4}|\d{2})))" G "<0>12-01/2002</0>" +"((0?[13578]|10|12)(-|\/)((0[0-9])|([12])([0-9]?)|(3[01]?))(-|\/)((\d{4})|(\d{2}))|(0?[2469]|11)(-|\/)((0[0-9])|([12])([0-9]?)|(3[0]?))(-|\/)((\d{4}|\d{2})))" "2/31/2002" +"((0?[13578]|10|12)(-|\/)((0[0-9])|([12])([0-9]?)|(3[01]?))(-|\/)((\d{4})|(\d{2}))|(0?[2469]|11)(-|\/)((0[0-9])|([12])([0-9]?)|(3[0]?))(-|\/)((\d{4}|\d{2})))" "13/0/02" +"((0?[13578]|10|12)(-|\/)((0[0-9])|([12])([0-9]?)|(3[01]?))(-|\/)((\d{4})|(\d{2}))|(0?[2469]|11)(-|\/)((0[0-9])|([12])([0-9]?)|(3[0]?))(-|\/)((\d{4}|\d{2})))" "Jan 1, 2001" +'^(([^<>;()\[\]\\.,;:@"]+(\.[^<>()\[\]\\.,;:@"]+)*)|(".+"))@((([a-z]([-a-z0-9]*[a-z0-9])?)|(#[0-9]+)|(\[((([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\.){3}(([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\]))\.)*(([a-z]([-a-z0-9]*[a-z0-9])?)|(#[0-9]+)|(\[((([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\.){3}(([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\]))$' G "<0>blah@[10.0.0.1]</0>" +'^(([^<>;()\[\]\\.,;:@"]+(\.[^<>()\[\]\\.,;:@"]+)*)|(".+"))@((([a-z]([-a-z0-9]*[a-z0-9])?)|(#[0-9]+)|(\[((([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\.){3}(([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\]))\.)*(([a-z]([-a-z0-9]*[a-z0-9])?)|(#[0-9]+)|(\[((([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\.){3}(([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\]))$' G "<0>a@b.c</0>" +'^(([^<>;()\[\]\\.,;:@"]+(\.[^<>()\[\]\\.,;:@"]+)*)|(".+"))@((([a-z]([-a-z0-9]*[a-z0-9])?)|(#[0-9]+)|(\[((([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\.){3}(([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\]))\.)*(([a-z]([-a-z0-9]*[a-z0-9])?)|(#[0-9]+)|(\[((([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\.){3}(([01]?[0-9]{0,2})|(2(([0-4][0-9])|(5[0-5]))))\]))$' "non@match@." +"^\d{9}[\d|X]$" G "<0>1234123412</0>" +"^\d{9}[\d|X]$" G "<0>123412341X</0>" +"^\d{9}[\d|X]$" "not an isbn" +"^\d{9}(\d|X)$" G "<0>1234123412</0>" +"^\d{9}(\d|X)$" G "<0>123412341X</0>" +"^\d{9}(\d|X)$" "not an isbn" +"^(([1-9])|(0[1-9])|(1[0-2]))\/(([0-9])|([0-2][0-9])|(3[0-1]))\/(([0-9][0-9])|([1-2][0,9][0-9][0-9]))$" G "<0>01/01/2001</0>" +"^(([1-9])|(0[1-9])|(1[0-2]))\/(([0-9])|([0-2][0-9])|(3[0-1]))\/(([0-9][0-9])|([1-2][0,9][0-9][0-9]))$" G "<0>1/1/1999</0>" +"^(([1-9])|(0[1-9])|(1[0-2]))\/(([0-9])|([0-2][0-9])|(3[0-1]))\/(([0-9][0-9])|([1-2][0,9][0-9][0-9]))$" G "<0>10/20/2080</0>" +"^(([1-9])|(0[1-9])|(1[0-2]))\/(([0-9])|([0-2][0-9])|(3[0-1]))\/(([0-9][0-9])|([1-2][0,9][0-9][0-9]))$" "13/01/2001" +"^(([1-9])|(0[1-9])|(1[0-2]))\/(([0-9])|([0-2][0-9])|(3[0-1]))\/(([0-9][0-9])|([1-2][0,9][0-9][0-9]))$" "1/1/1800" +"^(([1-9])|(0[1-9])|(1[0-2]))\/(([0-9])|([0-2][0-9])|(3[0-1]))\/(([0-9][0-9])|([1-2][0,9][0-9][0-9]))$" "10/32/2080" +"^\d*\.?((25)|(50)|(5)|(75)|(0)|(00))?$" G "<0>0.25</0>" +"^\d*\.?((25)|(50)|(5)|(75)|(0)|(00))?$" G "<0>.75</0>" +"^\d*\.?((25)|(50)|(5)|(75)|(0)|(00))?$" G "<0>123.50</0>" +"^\d*\.?((25)|(50)|(5)|(75)|(0)|(00))?$" ".77" +"^\d*\.?((25)|(50)|(5)|(75)|(0)|(00))?$" "1.435" +"^(s-|S-){0,1}[0-9]{3}\s?[0-9]{2}$" G "<0>12345</0>" +"^(s-|S-){0,1}[0-9]{3}\s?[0-9]{2}$" G "<0>932 68</0>" +"^(s-|S-){0,1}[0-9]{3}\s?[0-9]{2}$" G "<0>S-621 46</0>" +"^(s-|S-){0,1}[0-9]{3}\s?[0-9]{2}$" "5367" +"^(s-|S-){0,1}[0-9]{3}\s?[0-9]{2}$" "425611" +"^(s-|S-){0,1}[0-9]{3}\s?[0-9]{2}$" "31 545" +"^\d{5}(-\d{4})?$" G "<0>48222</0>" +"^\d{5}(-\d{4})?$" G "<0>48222-1746</0>" +"^\d{5}(-\d{4})?$" "4632" +"^\d{5}(-\d{4})?$" "Blake" +"^\d{5}(-\d{4})?$" "37333-32" +'^(?!^(PRN|AUX|CLOCK\$|NUL|CON|COM\d|LPT\d|\..*)(\..+)?$)[^\x00-\x1f\\?*:\";|/]+$' G "<0>test.txt</0>" +'^(?!^(PRN|AUX|CLOCK\$|NUL|CON|COM\d|LPT\d|\..*)(\..+)?$)[^\x00-\x1f\\?*:\";|/]+$' G "<0>test.jpg.txt</0>" +'^(?!^(PRN|AUX|CLOCK\$|NUL|CON|COM\d|LPT\d|\..*)(\..+)?$)[^\x00-\x1f\\?*:\";|/]+$' G "<0>a&b c.bmp</0>" +'^(?!^(PRN|AUX|CLOCK\$|NUL|CON|COM\d|LPT\d|\..*)(\..+)?$)[^\x00-\x1f\\?*:\";|/]+$' "CON" +'^(?!^(PRN|AUX|CLOCK\$|NUL|CON|COM\d|LPT\d|\..*)(\..+)?$)[^\x00-\x1f\\?*:\";|/]+$' ".pdf" +'^(?!^(PRN|AUX|CLOCK\$|NUL|CON|COM\d|LPT\d|\..*)(\..+)?$)[^\x00-\x1f\\?*:\";|/]+$' "test:2.pdf" +"^(\d{1,3}'(\d{3}')*\d{3}(\.\d{1,3})?|\d{1,3}(\.\d{3})?)$" G "<0>1'235.140</0>" +"^(\d{1,3}'(\d{3}')*\d{3}(\.\d{1,3})?|\d{1,3}(\.\d{3})?)$" G "<0>1'222'333.120</0>" +"^(\d{1,3}'(\d{3}')*\d{3}(\.\d{1,3})?|\d{1,3}(\.\d{3})?)$" G "<0>456</0>" +"^(\d{1,3}'(\d{3}')*\d{3}(\.\d{1,3})?|\d{1,3}(\.\d{3})?)$" "1234.500" +"^(\d{1,3}'(\d{3}')*\d{3}(\.\d{1,3})?|\d{1,3}(\.\d{3})?)$" "78'45.123" +"^(\d{1,3}'(\d{3}')*\d{3}(\.\d{1,3})?|\d{1,3}(\.\d{3})?)$" "123,0012" +"^[a-zA-Z][0-9][a-zA-Z]\s?[0-9][a-zA-Z][0-9]$" G "<0>T2p 3c7</0>" +"^[a-zA-Z][0-9][a-zA-Z]\s?[0-9][a-zA-Z][0-9]$" G "<0>T3P3c7</0>" +"^[a-zA-Z][0-9][a-zA-Z]\s?[0-9][a-zA-Z][0-9]$" G "<0>T2P 3C7</0>" +"^[a-zA-Z][0-9][a-zA-Z]\s?[0-9][a-zA-Z][0-9]$" "123456" +"^[a-zA-Z][0-9][a-zA-Z]\s?[0-9][a-zA-Z][0-9]$" "3C7T2P" +"^[a-zA-Z][0-9][a-zA-Z]\s?[0-9][a-zA-Z][0-9]$" "11T21RWW" +"^\$[0-9]+(\.[0-9][0-9])?$" G "<0>$1.50</0>" +"^\$[0-9]+(\.[0-9][0-9])?$" G "<0>$49</0>" +"^\$[0-9]+(\.[0-9][0-9])?$" G "<0>$0.50</0>" +"^\$[0-9]+(\.[0-9][0-9])?$" "1.5" +"^\$[0-9]+(\.[0-9][0-9])?$" "$1.333" +"^\$[0-9]+(\.[0-9][0-9])?$" "this $5.12 fails" +"\b((25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\.){3}(25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\b" G "<0>217.6.9.89</0>" +"\b((25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\.){3}(25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\b" G "<0>0.0.0.0</0>" +"\b((25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\.){3}(25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\b" G "<0>255.255.255.255</0>" +"\b((25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\.){3}(25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\b" "256.0.0.0" +"\b((25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\.){3}(25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\b" "0978.3.3.3" +"\b((25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\.){3}(25[0-5]|2[0-4]\d|[01]\d\d|\d?\d)\b" "65.4t.54.3" +"((mailto\:|(news|(ht|f)tp(s?))\://){1}\S+)" G "<0>http://www.aspemporium.com</0>" +"((mailto\:|(news|(ht|f)tp(s?))\://){1}\S+)" G "<0>mailto:dominionx@hotmail.com</0>" +"((mailto\:|(news|(ht|f)tp(s?))\://){1}\S+)" G "<0>ftp://ftp.test.com</0>" +"((mailto\:|(news|(ht|f)tp(s?))\://){1}\S+)" "www.aspemporium.com" +"((mailto\:|(news|(ht|f)tp(s?))\://){1}\S+)" "dominionx@hotmail.com" +"((mailto\:|(news|(ht|f)tp(s?))\://){1}\S+)" "bloggs" +"\(([0-9]{2}|0{1}((x|[0-9]){2}[0-9]{2}))\)\s*[0-9]{3,4}[- ]*[0-9]{4}" G "<0>(12) 123 1234</0>" +"\(([0-9]{2}|0{1}((x|[0-9]){2}[0-9]{2}))\)\s*[0-9]{3,4}[- ]*[0-9]{4}" G "<0>(01512) 123 1234</0>" +"\(([0-9]{2}|0{1}((x|[0-9]){2}[0-9]{2}))\)\s*[0-9]{3,4}[- ]*[0-9]{4}" G "<0>(0xx12) 1234 1234</0>" +"\(([0-9]{2}|0{1}((x|[0-9]){2}[0-9]{2}))\)\s*[0-9]{3,4}[- ]*[0-9]{4}" "12 123 1234" +"\(([0-9]{2}|0{1}((x|[0-9]){2}[0-9]{2}))\)\s*[0-9]{3,4}[- ]*[0-9]{4}" "(012) 123/1234" +"\(([0-9]{2}|0{1}((x|[0-9]){2}[0-9]{2}))\)\s*[0-9]{3,4}[- ]*[0-9]{4}" "(012) 123 12345" +"^\w+[\w-\.]*\@\w+((-\w+)|(\w*))\.[a-z]{2,3}$" G "<0>bob-smith@foo.com</0>" +"^\w+[\w-\.]*\@\w+((-\w+)|(\w*))\.[a-z]{2,3}$" G "<0>bob.smith@foo.com</0>" +"^\w+[\w-\.]*\@\w+((-\w+)|(\w*))\.[a-z]{2,3}$" G "<0>bob_smith@foo.com</0>" +"^\w+[\w-\.]*\@\w+((-\w+)|(\w*))\.[a-z]{2,3}$" "-smith@foo.com" +"^\w+[\w-\.]*\@\w+((-\w+)|(\w*))\.[a-z]{2,3}$" ".smith@foo.com" +"^\w+[\w-\.]*\@\w+((-\w+)|(\w*))\.[a-z]{2,3}$" "smith@foo_com" +"^(?=.*\d).{4,8}$" G "<0>1234</0>" +"^(?=.*\d).{4,8}$" G "<0>asdf1234</0>" +"^(?=.*\d).{4,8}$" G "<0>asp123</0>" +"^(?=.*\d).{4,8}$" "asdf" +"^(?=.*\d).{4,8}$" "asdf12345" +"^(?=.*\d).{4,8}$" "password" +"[^A-Za-z0-9_@\.]|@{2,}|\.{5,}" G "<0>user name</0>" +"[^A-Za-z0-9_@\.]|@{2,}|\.{5,}" G "<0>user#name</0>" +"[^A-Za-z0-9_@\.]|@{2,}|\.{5,}" G "<0>.....</0>" +"[^A-Za-z0-9_@\.]|@{2,}|\.{5,}" "User_Name1" +"[^A-Za-z0-9_@\.]|@{2,}|\.{5,}" "username@foo.com" +"[^A-Za-z0-9_@\.]|@{2,}|\.{5,}" "user.name@mail.foo.com" +"^100$|^[0-9]{1,2}$|^[0-9]{1,2}\,[0-9]{1,3}$" G "<0>12,654</0>" +"^100$|^[0-9]{1,2}$|^[0-9]{1,2}\,[0-9]{1,3}$" G "<0>1,987</0>" +"^100$|^[0-9]{1,2}$|^[0-9]{1,2}\,[0-9]{1,3}$" "128,2" +"^100$|^[0-9]{1,2}$|^[0-9]{1,2}\,[0-9]{1,3}$" "12," +"^(http|https|ftp)\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(:[a-zA-Z0-9]*)?/?([a-zA-Z0-9\-\._\?\,\'/\\\+\&%\$#\=~])*[^\.\,\)\(\s]$" G "<0>https://www.restrictd.com/~myhome/</0>" +"^(http|https|ftp)\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(:[a-zA-Z0-9]*)?/?([a-zA-Z0-9\-\._\?\,\'/\\\+\&%\$#\=~])*[^\.\,\)\(\s]$" "http://www.krumedia.com." +"^(http|https|ftp)\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(:[a-zA-Z0-9]*)?/?([a-zA-Z0-9\-\._\?\,\'/\\\+\&%\$#\=~])*[^\.\,\)\(\s]$" "(http://www.krumedia.com)" +"^(http|https|ftp)\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(:[a-zA-Z0-9]*)?/?([a-zA-Z0-9\-\._\?\,\'/\\\+\&%\$#\=~])*[^\.\,\)\(\s]$" "http://www.krumedia.com," +"(\d{1,3},(\d{3},)*\d{3}(\.\d{1,3})?|\d{1,3}(\.\d{3})?)$" G "<0>2&651.50</0>" +"(\d{1,3},(\d{3},)*\d{3}(\.\d{1,3})?|\d{1,3}(\.\d{3})?)$" G "<0>987.895</0>" +"(\d{1,3},(\d{3},)*\d{3}(\.\d{1,3})?|\d{1,3}(\.\d{3})?)$" "25$%787*" +"\$[0-9]?[0-9]?[0-9]?((\,[0-9][0-9][0-9])*)?(\.[0-9][0-9]?)?$" G "<0>$1,456,983.00</0>" +"\$[0-9]?[0-9]?[0-9]?((\,[0-9][0-9][0-9])*)?(\.[0-9][0-9]?)?$" G "<0>$1,700.07</0>" +"\$[0-9]?[0-9]?[0-9]?((\,[0-9][0-9][0-9])*)?(\.[0-9][0-9]?)?$" G "<0>$68,944.23</0>" +"\$[0-9]?[0-9]?[0-9]?((\,[0-9][0-9][0-9])*)?(\.[0-9][0-9]?)?$" "$20,86.93" +"\$[0-9]?[0-9]?[0-9]?((\,[0-9][0-9][0-9])*)?(\.[0-9][0-9]?)?$" "$1098.84" +"\$[0-9]?[0-9]?[0-9]?((\,[0-9][0-9][0-9])*)?(\.[0-9][0-9]?)?$" "$150." +"\$[0-9]?[0-9]?[0-9]?((\,[0-9][0-9][0-9])*)?(\.[0-9][0-9])?$" G "<0>$28,009,987.88</0>" +"\$[0-9]?[0-9]?[0-9]?((\,[0-9][0-9][0-9])*)?(\.[0-9][0-9])?$" G "<0>$23,099.05</0>" +"\$[0-9]?[0-9]?[0-9]?((\,[0-9][0-9][0-9])*)?(\.[0-9][0-9])?$" G "<0>$.88</0>" +"\$[0-9]?[0-9]?[0-9]?((\,[0-9][0-9][0-9])*)?(\.[0-9][0-9])?$" "$234,5.99" +"^((((31\/(0?[13578]|1[02]))|((29|30)\/(0?[1,3-9]|1[0-2])))\/(1[6-9]|[2-9]\d)?\d{2})|(29\/0?2\/(((1[6-9]|[2-9]\d)?(0[48]|[2468][048]|[13579][26])|((16|[2468][048]|[3579][26])00))))|(0?[1-9]|1\d|2[0-8])\/((0?[1-9])|(1[0-2]))\/((1[6-9]|[2-9]\d)?\d{2})) (20|21|22|23|[0-1]?\d):[0-5]?\d:[0-5]?\d$" G "<0>29/02/2004 20:15:27</0>" +"^((((31\/(0?[13578]|1[02]))|((29|30)\/(0?[1,3-9]|1[0-2])))\/(1[6-9]|[2-9]\d)?\d{2})|(29\/0?2\/(((1[6-9]|[2-9]\d)?(0[48]|[2468][048]|[13579][26])|((16|[2468][048]|[3579][26])00))))|(0?[1-9]|1\d|2[0-8])\/((0?[1-9])|(1[0-2]))\/((1[6-9]|[2-9]\d)?\d{2})) (20|21|22|23|[0-1]?\d):[0-5]?\d:[0-5]?\d$" G "<0>29/2/04 8:9:5</0>" +"^((((31\/(0?[13578]|1[02]))|((29|30)\/(0?[1,3-9]|1[0-2])))\/(1[6-9]|[2-9]\d)?\d{2})|(29\/0?2\/(((1[6-9]|[2-9]\d)?(0[48]|[2468][048]|[13579][26])|((16|[2468][048]|[3579][26])00))))|(0?[1-9]|1\d|2[0-8])\/((0?[1-9])|(1[0-2]))\/((1[6-9]|[2-9]\d)?\d{2})) (20|21|22|23|[0-1]?\d):[0-5]?\d:[0-5]?\d$" G "<0>31/3/2004 9:20:17</0>" +"^((((31\/(0?[13578]|1[02]))|((29|30)\/(0?[1,3-9]|1[0-2])))\/(1[6-9]|[2-9]\d)?\d{2})|(29\/0?2\/(((1[6-9]|[2-9]\d)?(0[48]|[2468][048]|[13579][26])|((16|[2468][048]|[3579][26])00))))|(0?[1-9]|1\d|2[0-8])\/((0?[1-9])|(1[0-2]))\/((1[6-9]|[2-9]\d)?\d{2})) (20|21|22|23|[0-1]?\d):[0-5]?\d:[0-5]?\d$" "29/02/2003 20:15:15" +"^((((31\/(0?[13578]|1[02]))|((29|30)\/(0?[1,3-9]|1[0-2])))\/(1[6-9]|[2-9]\d)?\d{2})|(29\/0?2\/(((1[6-9]|[2-9]\d)?(0[48]|[2468][048]|[13579][26])|((16|[2468][048]|[3579][26])00))))|(0?[1-9]|1\d|2[0-8])\/((0?[1-9])|(1[0-2]))\/((1[6-9]|[2-9]\d)?\d{2})) (20|21|22|23|[0-1]?\d):[0-5]?\d:[0-5]?\d$" "2/29/04 20:15:15" +"^((((31\/(0?[13578]|1[02]))|((29|30)\/(0?[1,3-9]|1[0-2])))\/(1[6-9]|[2-9]\d)?\d{2})|(29\/0?2\/(((1[6-9]|[2-9]\d)?(0[48]|[2468][048]|[13579][26])|((16|[2468][048]|[3579][26])00))))|(0?[1-9]|1\d|2[0-8])\/((0?[1-9])|(1[0-2]))\/((1[6-9]|[2-9]\d)?\d{2})) (20|21|22|23|[0-1]?\d):[0-5]?\d:[0-5]?\d$" "31/3/4 9:20:17" +"^([a-zA-Z0-9_\-\.]+)@([a-zA-Z0-9_\-\.]+)\.([a-zA-Z]{2,5})$" G "<0>something@someserver.com</0>" +"^([a-zA-Z0-9_\-\.]+)@([a-zA-Z0-9_\-\.]+)\.([a-zA-Z]{2,5})$" G "<0>firstname.lastname@mailserver.domain.com</0>" +"^([a-zA-Z0-9_\-\.]+)@([a-zA-Z0-9_\-\.]+)\.([a-zA-Z]{2,5})$" G "<0>username-something@some-server.nl</0>" +"^([a-zA-Z0-9_\-\.]+)@([a-zA-Z0-9_\-\.]+)\.([a-zA-Z]{2,5})$" "username@someserver.domain.c" +"^([a-zA-Z0-9_\-\.]+)@([a-zA-Z0-9_\-\.]+)\.([a-zA-Z]{2,5})$" "somename@server.domain-com" +"^([a-zA-Z0-9_\-\.]+)@([a-zA-Z0-9_\-\.]+)\.([a-zA-Z]{2,5})$" "someone@something.se_eo" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)|(^([0-9]|[1][0-9]|[2][0-3])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)" G "<0>8am</0>" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)|(^([0-9]|[1][0-9]|[2][0-3])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)" G "<0>8 am</0>" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)|(^([0-9]|[1][0-9]|[2][0-3])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)" G "<0>8:00 am</0>" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)|(^([0-9]|[1][0-9]|[2][0-3])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)" "8a" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)|(^([0-9]|[1][0-9]|[2][0-3])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)" "8 a" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)|(^([0-9]|[1][0-9]|[2][0-3])(\s{0,1})(AM|PM|am|pm|aM|Am|pM|Pm{2,2})$)" "8:00 a" +"^([0-9]{2})?(\([0-9]{2})\)([0-9]{3}|[0-9]{4})-[0-9]{4}$" G "<0>55(21)123-4567</0>" +"^([0-9]{2})?(\([0-9]{2})\)([0-9]{3}|[0-9]{4})-[0-9]{4}$" G "<0>(11)1234-5678</0>" +"^([0-9]{2})?(\([0-9]{2})\)([0-9]{3}|[0-9]{4})-[0-9]{4}$" G "<0>55(71)4562-2234</0>" +"^([0-9]{2})?(\([0-9]{2})\)([0-9]{3}|[0-9]{4})-[0-9]{4}$" "3434-3432" +"^([0-9]{2})?(\([0-9]{2})\)([0-9]{3}|[0-9]{4})-[0-9]{4}$" "4(23)232-3232" +"^([0-9]{2})?(\([0-9]{2})\)([0-9]{3}|[0-9]{4})-[0-9]{4}$" "55(2)232-232" +"^((([0]?[1-9]|1[0-2])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?))$" G "<0>1:01 AM</0>" +"^((([0]?[1-9]|1[0-2])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?))$" G "<0>23:52:01</0>" +"^((([0]?[1-9]|1[0-2])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?))$" G "<0>03.24.36 AM</0>" +"^((([0]?[1-9]|1[0-2])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?))$" "19:31 AM" +"^((([0]?[1-9]|1[0-2])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?))$" "9:9 PM" +"^((([0]?[1-9]|1[0-2])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)[0-5][0-9]((:|\.)[0-5][0-9])?))$" "25:60:61" +"^\d{0,2}(\.\d{1,2})?$" G "<0>99.99</0>" +"^\d{0,2}(\.\d{1,2})?$" G "<0>99</0>" +"^\d{0,2}(\.\d{1,2})?$" G "<0>.99</0>" +"^\d{0,2}(\.\d{1,2})?$" "999.999" +"^\d{0,2}(\.\d{1,2})?$" "999" +"^\d{0,2}(\.\d{1,2})?$" ".999" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z])(?!.*\s).{4,8}$" G "<0>1agdA*$#</0>" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z])(?!.*\s).{4,8}$" G "<0>1agdA*$#</0>" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z])(?!.*\s).{4,8}$" G "<0>1agdA*$#</0>" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z])(?!.*\s).{4,8}$" "wyrn%@*&$# f" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z])(?!.*\s).{4,8}$" "mbndkfh782" +"^(?=.*\d)(?=.*[a-z])(?=.*[A-Z])(?!.*\s).{4,8}$" "BNfhjdhfjd&*)%#$)" +"^([a-zA-Z0-9][-a-zA-Z0-9]*[a-zA-Z0-9]\.)+([a-zA-Z0-9]{3,5})$" G "<0>freshmeat.net</0>" +"^([a-zA-Z0-9][-a-zA-Z0-9]*[a-zA-Z0-9]\.)+([a-zA-Z0-9]{3,5})$" G "<0>123.com</0>" +"^([a-zA-Z0-9][-a-zA-Z0-9]*[a-zA-Z0-9]\.)+([a-zA-Z0-9]{3,5})$" G "<0>TempLate-toolkKt.orG</0>" +"^([a-zA-Z0-9][-a-zA-Z0-9]*[a-zA-Z0-9]\.)+([a-zA-Z0-9]{3,5})$" "-dog.com" +"^([a-zA-Z0-9][-a-zA-Z0-9]*[a-zA-Z0-9]\.)+([a-zA-Z0-9]{3,5})$" "?boy.net" +"^([a-zA-Z0-9][-a-zA-Z0-9]*[a-zA-Z0-9]\.)+([a-zA-Z0-9]{3,5})$" "this.domain" +"^[^']*$" G "<0>asljas</0>" +"^[^']*$" G "<0>%/&89uhuhadjkh</0>" +"^[^']*$" G '<0>"hi there!"</0>' +"^[^']*$" "'hi there!'" +"^[^']*$" "It's 9 o'clock" +"^[^']*$" "'''''" +"(^\(\)$|^\(((\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\),)*(\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\)){1}\)))$" G "<0>((24,((1,2,3),(3,4,5))))</0>" +"(^\(\)$|^\(((\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\),)*(\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\)){1}\)))$" G "<0>((1,((2,3,4),(4,5,6),(96,34,26))),(12,((1,3,4),(4,5,6),(7,8,9))))</0>" +"(^\(\)$|^\(((\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\),)*(\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\)){1}\)))$" G "<0>()</0>" +"(^\(\)$|^\(((\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\),)*(\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\)){1}\)))$" "(24,((1,2,3),(3,4,5)))" +"(^\(\)$|^\(((\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\),)*(\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\)){1}\)))$" "( )" +"(^\(\)$|^\(((\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\),)*(\([0-9]+,(\((\([0-9]+,[0-9]+,[0-9]+\),)*(\([0-9]+,[0-9]+,[0-9]+\)){1}\))+\)){1}\)))$" "((23,(12,3,4),(4,5,6)))" +"^[a-zA-Z0-9\s .\-_']+$" G "<0>dony d'gsa</0>" +"^[a-zA-Z0-9\s .\-_']+$" "^[a-zA-Z0-9\s.\-_']+$" +"^[_a-zA-Z0-9-]+(\.[_a-zA-Z0-9-]+)*@[a-zA-Z0-9-]+(\.[a-zA-Z0-9-]+)*\.(([0-9]{1,3})|([a-zA-Z]{2,3})|(aero|coop|info|museum|name))$" G "<0>example@example.com</0>" +"^[_a-zA-Z0-9-]+(\.[_a-zA-Z0-9-]+)*@[a-zA-Z0-9-]+(\.[a-zA-Z0-9-]+)*\.(([0-9]{1,3})|([a-zA-Z]{2,3})|(aero|coop|info|museum|name))$" G "<0>foo@bar.info</0>" +"^[_a-zA-Z0-9-]+(\.[_a-zA-Z0-9-]+)*@[a-zA-Z0-9-]+(\.[a-zA-Z0-9-]+)*\.(([0-9]{1,3})|([a-zA-Z]{2,3})|(aero|coop|info|museum|name))$" G "<0>blah@127.0.0.1</0>" +"^[_a-zA-Z0-9-]+(\.[_a-zA-Z0-9-]+)*@[a-zA-Z0-9-]+(\.[a-zA-Z0-9-]+)*\.(([0-9]{1,3})|([a-zA-Z]{2,3})|(aero|coop|info|museum|name))$" "broken@@example.com" +"^[_a-zA-Z0-9-]+(\.[_a-zA-Z0-9-]+)*@[a-zA-Z0-9-]+(\.[a-zA-Z0-9-]+)*\.(([0-9]{1,3})|([a-zA-Z]{2,3})|(aero|coop|info|museum|name))$" "foo@bar.infp" +"^[_a-zA-Z0-9-]+(\.[_a-zA-Z0-9-]+)*@[a-zA-Z0-9-]+(\.[a-zA-Z0-9-]+)*\.(([0-9]{1,3})|([a-zA-Z]{2,3})|(aero|coop|info|museum|name))$" "blah@.nospam.biz" +"^\d{5}(-\d{3})?$" G "<0>13165-000</0>" +"^\d{5}(-\d{3})?$" G "<0>38175-000</0>" +"^\d{5}(-\d{3})?$" G "<0>81470-276</0>" +"^\d{5}(-\d{3})?$" "13165-00" +"^\d{5}(-\d{3})?$" "38175-abc" +"^\d{5}(-\d{3})?$" "81470-2763" +"^\$(\d{1,3}(\,\d{3})*|(\d+))(\.\d{2})?$" G "<0>$0.84</0>" +"^\$(\d{1,3}(\,\d{3})*|(\d+))(\.\d{2})?$" G "<0>$123458</0>" +"^\$(\d{1,3}(\,\d{3})*|(\d+))(\.\d{2})?$" G "<0>$1,234,567.89</0>" +"^\$(\d{1,3}(\,\d{3})*|(\d+))(\.\d{2})?$" "$12,3456.01" +"^\$(\d{1,3}(\,\d{3})*|(\d+))(\.\d{2})?$" "12345" +"^\$(\d{1,3}(\,\d{3})*|(\d+))(\.\d{2})?$" "$1.234" +"([A-Z]:\\[^/:\*\?<>\|]+\.\w{2,6})|(\\{2}[^/:\*\?<>\|]+\.\w{2,6})" G "<0>C:\\temp\\this allows spaces\\web.config</0>" +"([A-Z]:\\[^/:\*\?<>\|]+\.\w{2,6})|(\\{2}[^/:\*\?<>\|]+\.\w{2,6})" G "<0>\\\\Andromeda\\share\\file name.123</0>" +"([A-Z]:\\[^/:\*\?<>\|]+\.\w{2,6})|(\\{2}[^/:\*\?<>\|]+\.\w{2,6})" "tz:\temp\ fi*le?na:m<e>.doc" +"([A-Z]:\\[^/:\*\?<>\|]+\.\w{2,6})|(\\{2}[^/:\*\?<>\|]+\.\w{2,6})" "\\Andromeda\share\filename.a" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])$)|(^([0-9]|[1][0-9]|[2][0-3])$)" G "<0>10:35</0>" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])$)|(^([0-9]|[1][0-9]|[2][0-3])$)" G "<0>9:20</0>" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])$)|(^([0-9]|[1][0-9]|[2][0-3])$)" G "<0>23</0>" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])$)|(^([0-9]|[1][0-9]|[2][0-3])$)" "24:00" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])$)|(^([0-9]|[1][0-9]|[2][0-3])$)" "20 PM" +"(^([0-9]|[0-1][0-9]|[2][0-3]):([0-5][0-9])$)|(^([0-9]|[1][0-9]|[2][0-3])$)" "20:15 PM" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(\.[0-9][0-9])?$" G "<0>$3,023,123.34</0>" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(\.[0-9][0-9])?$" G "<0>9,876,453</0>" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(\.[0-9][0-9])?$" G "<0>123456.78</0>" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(\.[0-9][0-9])?$" "4,33,234.34" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(\.[0-9][0-9])?$" "$1.234" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(\.[0-9][0-9])?$" "abc" +"^\$?\d+(\.(\d{2}))?$" G "<0>$2.43</0>" +"^\$?\d+(\.(\d{2}))?$" G "<0>2.02</0>" +"^\$?\d+(\.(\d{2}))?$" G "<0>$2112</0>" +"^\$?\d+(\.(\d{2}))?$" "2.1" +"^\$?\d+(\.(\d{2}))?$" "$.14" +"^\$?\d+(\.(\d{2}))?$" "$2,222.12" +/("[^"]*")|('[^\r]*)(\r\n)?/ G '<0>"my string"</0>' +/("[^"]*")|('[^\r]*)(\r\n)?/ G '<0>"a string with \u0027 in it"</0>' +/("[^"]*")|('[^\r]*)(\r\n)?/ G "<0>' comment</0>" +/("[^"]*")|('[^\r]*)(\r\n)?/ /asd "/ +"^[A-Za-z0-9]{8}-[A-Za-z0-9]{4}-[A-Za-z0-9]{4}-[A-Za-z0-9]{4}-[A-Za-z0-9]{12}$" G "<0>BFDB4D31-3E35-4DAB-AFCA-5E6E5C8F61EA</0>" +"^[A-Za-z0-9]{8}-[A-Za-z0-9]{4}-[A-Za-z0-9]{4}-[A-Za-z0-9]{4}-[A-Za-z0-9]{12}$" G "<0>BFDB4d31-3e35-4dab-afca-5e6e5c8f61ea</0>" +"^[A-Za-z0-9]{8}-[A-Za-z0-9]{4}-[A-Za-z0-9]{4}-[A-Za-z0-9]{4}-[A-Za-z0-9]{12}$" "qqqBFDB4D31-3E35-4DAB-AFCA-5E6E5C8F61EA" +"^[A-Za-z0-9]{8}-[A-Za-z0-9]{4}-[A-Za-z0-9]{4}-[A-Za-z0-9]{4}-[A-Za-z0-9]{12}$" "BFDB4D31-3E-4DAB-AFCA-5E6E5C8F61EA" +"^[A-Za-z0-9]{8}-[A-Za-z0-9]{4}-[A-Za-z0-9]{4}-[A-Za-z0-9]{4}-[A-Za-z0-9]{12}$" "BFDB4D31-3E35-4DAB-AF" +"^\d{2}(\x2e)(\d{3})(-\d{3})?$" G "<0>12.345-678</0>" +"^\d{2}(\x2e)(\d{3})(-\d{3})?$" G "<0>23.345-123</0>" +"^\d{2}(\x2e)(\d{3})(-\d{3})?$" G "<0>99.999</0>" +"^\d{2}(\x2e)(\d{3})(-\d{3})?$" "41222-222" +"^\d{2}(\x2e)(\d{3})(-\d{3})?$" "3.444-233" +"^\d{2}(\x2e)(\d{3})(-\d{3})?$" "43.324444" +"^\d{2}(\u002e)(\d{3})(-\d{3})?$" G "<0>12.345-678</0>" +"^\d{2}(\u002e)(\d{3})(-\d{3})?$" G "<0>23.345-123</0>" +"^\d{2}(\u002e)(\d{3})(-\d{3})?$" G "<0>99.999</0>" +"^\d{2}(\u002e)(\d{3})(-\d{3})?$" "41222-222" +"^\d{2}(\u002e)(\d{3})(-\d{3})?$" "3.444-233" +"^\d{2}(\u002e)(\d{3})(-\d{3})?$" "43.324444" +#"^(([a-zA-Z]:)|(\\{2}\w+)\$?)(\\(\w[\w ]*))+\.(txt|TXT)$" G "<0>c:\file.txt</0>" # TODO: debug +#"^(([a-zA-Z]:)|(\\{2}\w+)\$?)(\\(\w[\w ]*))+\.(txt|TXT)$" G "<0>c:\folder\sub folder\file.txt</0>" # TODO: debug +#"^(([a-zA-Z]:)|(\\{2}\w+)\$?)(\\(\w[\w ]*))+\.(txt|TXT)$" G "<0>\\network\folder\file.txt</0>" # TODO: debug +"^(([a-zA-Z]:)|(\\{2}\w+)\$?)(\\(\w[\w ]*))+\.(txt|TXT)$" "C:" +"^(([a-zA-Z]:)|(\\{2}\w+)\$?)(\\(\w[\w ]*))+\.(txt|TXT)$" "C:\file.xls" +"^(([a-zA-Z]:)|(\\{2}\w+)\$?)(\\(\w[\w ]*))+\.(txt|TXT)$" "folder.txt" +"^[a-zA-Z0-9]+([a-zA-Z0-9\-\.]+)?\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$" G "<0>my.domain.com</0>" +"^[a-zA-Z0-9]+([a-zA-Z0-9\-\.]+)?\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$" G "<0>regexlib.com</0>" +"^[a-zA-Z0-9]+([a-zA-Z0-9\-\.]+)?\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$" G "<0>big-reg.com</0>" +"^[a-zA-Z0-9]+([a-zA-Z0-9\-\.]+)?\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$" ".mydomain.com" +"^[a-zA-Z0-9]+([a-zA-Z0-9\-\.]+)?\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$" "regexlib.comm" +"^[a-zA-Z0-9]+([a-zA-Z0-9\-\.]+)?\.(com|org|net|mil|edu|COM|ORG|NET|MIL|EDU)$" "-bigreg.com" +"^\d{4}[\-\/\s]?((((0[13578])|(1[02]))[\-\/\s]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\-\/\s]?(([0-2][0-9])|(30)))|(02[\-\/\s]?[0-2][0-9]))$" G "<0>0001-12-31</0>" +"^\d{4}[\-\/\s ]?((((0[13578])|(1[02]))[\-\/\s ]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\-\/\s ]?(([0-2][0-9])|(30)))|(02[\-\/\s ]?[0-2][0-9]))$" G "<0>9999 09 30</0>" +"^\d{4}[\-\/\s]?((((0[13578])|(1[02]))[\-\/\s]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\-\/\s]?(([0-2][0-9])|(30)))|(02[\-\/\s]?[0-2][0-9]))$" G "<0>2002/03/03</0>" +"^\d{4}[\-\/\s]?((((0[13578])|(1[02]))[\-\/\s]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\-\/\s]?(([0-2][0-9])|(30)))|(02[\-\/\s]?[0-2][0-9]))$" "0001\02\30" +"^\d{4}[\-\/\s]?((((0[13578])|(1[02]))[\-\/\s]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\-\/\s]?(([0-2][0-9])|(30)))|(02[\-\/\s]?[0-2][0-9]))$" "9999.15.01" +"^\d{4}[\-\/\s]?((((0[13578])|(1[02]))[\-\/\s]?(([0-2][0-9])|(3[01])))|(((0[469])|(11))[\-\/\s]?(([0-2][0-9])|(30)))|(02[\-\/\s]?[0-2][0-9]))$" "2002/3/3" +"^http\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(/\S*)?$" G "<0>http://psychopop.org</0>" +"^http\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(/\S*)?$" G "<0>http://www.edsroom.com/newUser.asp</0>" +"^http\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(/\S*)?$" G "<0>http://unpleasant.jarrin.net/markov/inde</0>" +"^http\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(/\S*)?$" "ftp://psychopop.org" +"^http\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(/\S*)?$" "http://www.edsroom/" +"^http\://[a-zA-Z0-9\-\.]+\.[a-zA-Z]{2,3}(/\S*)?$" "http://un/pleasant.jarrin.net/markov/index.asp" +"^( [1-9]|[1-9]|0[1-9]|10|11|12)[0-5]\d$" G "<0>1145</0>" +"^( [1-9]|[1-9]|0[1-9]|10|11|12)[0-5]\d$" G "<0>933</0>" +"^( [1-9]|[1-9]|0[1-9]|10|11|12)[0-5]\d$" G "<0> 801</0>" +"^( [1-9]|[1-9]|0[1-9]|10|11|12)[0-5]\d$" "0000" +"^( [1-9]|[1-9]|0[1-9]|10|11|12)[0-5]\d$" "1330" +"^( [1-9]|[1-9]|0[1-9]|10|11|12)[0-5]\d$" "8:30" +"^\d{1,2}\/\d{2,4}$" G "<0>9/02</0>" +"^\d{1,2}\/\d{2,4}$" G "<0>09/2002</0>" +"^\d{1,2}\/\d{2,4}$" G "<0>09/02</0>" +"^\d{1,2}\/\d{2,4}$" "Fall 2002" +"^\d{1,2}\/\d{2,4}$" "Sept 2002" +"^(|(0[1-9])|(1[0-2]))\/((0[1-9])|(1\d)|(2\d)|(3[0-1]))\/((\d{4}))$" G "<0>01/01/2001</0>" +"^(|(0[1-9])|(1[0-2]))\/((0[1-9])|(1\d)|(2\d)|(3[0-1]))\/((\d{4}))$" G "<0>02/30/2001</0>" +"^(|(0[1-9])|(1[0-2]))\/((0[1-9])|(1\d)|(2\d)|(3[0-1]))\/((\d{4}))$" G "<0>12/31/2002</0>" +"^(|(0[1-9])|(1[0-2]))\/((0[1-9])|(1\d)|(2\d)|(3[0-1]))\/((\d{4}))$" "1/1/02" +"^(|(0[1-9])|(1[0-2]))\/((0[1-9])|(1\d)|(2\d)|(3[0-1]))\/((\d{4}))$" "1/1/2002" +"^(|(0[1-9])|(1[0-2]))\/((0[1-9])|(1\d)|(2\d)|(3[0-1]))\/((\d{4}))$" "1/25/2002" +"^(1?(-?\d{3})-?)?(\d{3})(-?\d{4})$" G "<0>15615552323</0>" +"^(1?(-?\d{3})-?)?(\d{3})(-?\d{4})$" G "<0>1-561-555-1212</0>" +"^(1?(-?\d{3})-?)?(\d{3})(-?\d{4})$" G "<0>5613333</0>" +"^(1?(-?\d{3})-?)?(\d{3})(-?\d{4})$" "1-555-5555" +"^(1?(-?\d{3})-?)?(\d{3})(-?\d{4})$" "15553333" +"^(1?(-?\d{3})-?)?(\d{3})(-?\d{4})$" "0-561-555-1212" +'<[^>]*name[\s]*=[\s]*"?[^\w_]*"?[^>]*>' G '<0><input type = text name = "bob"></0>' +'<[^>]*name[\s]*=[\s]*"?[^\w_]*"?[^>]*>' G '<0><select name = "fred"></0>' +#'<[^>]*name[\s]*=[\s]*"?[^\w_]*"?[^>]*>' G '<0><form></0>' #TODO: Debug +'<[^>]*name[\s]*=[\s]*"?[^\w_]*"?[^>]*>' "<input type = submit>" # TODO: \w in pattern +'<[^>]*name[\s]*=[\s]*"?[^\w_]*"?[^>]*>' '<font face = "arial">' # TODO: \w in pattern +'<[^>]*name[\s]*=[\s]*"?[^\w_]*"?[^>]*>' "The dirty brown fox stank like" +"^(1|01|2|02|3|03|4|04|5|05|6|06|7|07|8|08|9|09|10|11|12{1,2}):(([0-5]{1}[0-9]{1}\s{0,1})([AM|PM|am|pm]{2,2}))\W{0}$" G "<0>1:00 AM</0>" +"^(1|01|2|02|3|03|4|04|5|05|6|06|7|07|8|08|9|09|10|11|12{1,2}):(([0-5]{1}[0-9]{1}\s{0,1})([AM|PM|am|pm]{2,2}))\W{0}$" G "<0>12:00 PM</0>" +"^(1|01|2|02|3|03|4|04|5|05|6|06|7|07|8|08|9|09|10|11|12{1,2}):(([0-5]{1}[0-9]{1}\s{0,1})([AM|PM|am|pm]{2,2}))\W{0}$" G "<0>1:00am</0>" +"^(1|01|2|02|3|03|4|04|5|05|6|06|7|07|8|08|9|09|10|11|12{1,2}):(([0-5]{1}[0-9]{1}\s{0,1})([AM|PM|am|pm]{2,2}))\W{0}$" "24:00" +"^\d*$" G "<0>123</0>" +"^\d*$" G "<0>000</0>" +"^\d*$" G "<0>43</0>" +"^\d*$" "asbc" +"^\d*$" "-34" +"^\d*$" "3.1415" +"^[-+]?\d*$" G "<0>123</0>" +"^[-+]?\d*$" G "<0>-123</0>" +"^[-+]?\d*$" G "<0>+123</0>" +"^[-+]?\d*$" "abc" +"^[-+]?\d*$" "3.14159" +"^[-+]?\d*$" "-3.14159" +"^\d*\.?\d*$" G "<0>123</0>" +"^\d*\.?\d*$" G "<0>3.14159</0>" +"^\d*\.?\d*$" G "<0>.234</0>" +"^\d*\.?\d*$" "abc" +"^\d*\.?\d*$" "-3.14159" +"^\d*\.?\d*$" "3.4.2" +"^((\d{5}-\d{4})|(\d{5})|([A-Z]\d[A-Z]\s\d[A-Z]\d))$" G "<0>44240</0>" +"^((\d{5}-\d{4})|(\d{5})|([A-Z]\d[A-Z]\s\d[A-Z]\d))$" G "<0>44240-5555</0>" +"^((\d{5}-\d{4})|(\d{5})|([A-Z]\d[A-Z]\s\d[A-Z]\d))$" G "<0>T2P 3C7</0>" +"^((\d{5}-\d{4})|(\d{5})|([A-Z]\d[A-Z]\s\d[A-Z]\d))$" "44240ddd" +"^((\d{5}-\d{4})|(\d{5})|([A-Z]\d[A-Z]\s\d[A-Z]\d))$" "t44240-55" +"^((\d{5}-\d{4})|(\d{5})|([A-Z]\d[A-Z]\s\d[A-Z]\d))$" "t2p3c7" +"^[\\(]{0,1}([0-9]){3}[\\)]{0,1}[ ]?([^0-1]){1}([0-9]){2}[ ]?[-]?[ ]?([0-9]){4}[ ]*((x){0,1}([0-9]){1,5}){0,1}$" G "<0>(910)456-7890</0>" +"^[\\(]{0,1}([0-9]){3}[\\)]{0,1}[ ]?([^0-1]){1}([0-9]){2}[ ]?[-]?[ ]?([0-9]){4}[ ]*((x){0,1}([0-9]){1,5}){0,1}$" G "<0>(910)456-8970 x12</0>" +"^[\\(]{0,1}([0-9]){3}[\\)]{0,1}[ ]?([^0-1]){1}([0-9]){2}[ ]?[-]?[ ]?([0-9]){4}[ ]*((x){0,1}([0-9]){1,5}){0,1}$" G "<0>(910)456-8970 1211</0>" +"^[\\(]{0,1}([0-9]){3}[\\)]{0,1}[ ]?([^0-1]){1}([0-9]){2}[ ]?[-]?[ ]?([0-9]){4}[ ]*((x){0,1}([0-9]){1,5}){0,1}$" "(910) 156-7890" +"^[\\(]{0,1}([0-9]){3}[\\)]{0,1}[ ]?([^0-1]){1}([0-9]){2}[ ]?[-]?[ ]?([0-9]){4}[ ]*((x){0,1}([0-9]){1,5}){0,1}$" "(910) 056-7890" +"^[\\(]{0,1}([0-9]){3}[\\)]{0,1}[ ]?([^0-1]){1}([0-9]){2}[ ]?[-]?[ ]?([0-9]){4}[ ]*((x){0,1}([0-9]){1,5}){0,1}$" "(910) 556-7890 x" +"^((0?[1-9]|[12][1-9]|3[01])\.(0?[13578]|1[02])\.20[0-9]{2}|(0?[1-9]|[12][1-9]|30)\.(0?[13456789]|1[012])\.20[0-9]{2}|(0?[1-9]|1[1-9]|2[0-8])\.(0?[123456789]|1[012])\.20[0-9]{2}|(0?[1-9]|[12][1-9])\.(0?[123456789]|1[012])\.20(00|04|08|12|16|20|24|28|32|36|40|44|48|52|56|60|64|68|72|76|80|84|88|92|96))$" G "<0>31.01.2002</0>" +"^((0?[1-9]|[12][1-9]|3[01])\.(0?[13578]|1[02])\.20[0-9]{2}|(0?[1-9]|[12][1-9]|30)\.(0?[13456789]|1[012])\.20[0-9]{2}|(0?[1-9]|1[1-9]|2[0-8])\.(0?[123456789]|1[012])\.20[0-9]{2}|(0?[1-9]|[12][1-9])\.(0?[123456789]|1[012])\.20(00|04|08|12|16|20|24|28|32|36|40|44|48|52|56|60|64|68|72|76|80|84|88|92|96))$" G "<0>29.2.2004</0>" +"^((0?[1-9]|[12][1-9]|3[01])\.(0?[13578]|1[02])\.20[0-9]{2}|(0?[1-9]|[12][1-9]|30)\.(0?[13456789]|1[012])\.20[0-9]{2}|(0?[1-9]|1[1-9]|2[0-8])\.(0?[123456789]|1[012])\.20[0-9]{2}|(0?[1-9]|[12][1-9])\.(0?[123456789]|1[012])\.20(00|04|08|12|16|20|24|28|32|36|40|44|48|52|56|60|64|68|72|76|80|84|88|92|96))$" G "<0>09.02.2005</0>" +"^((0?[1-9]|[12][1-9]|3[01])\.(0?[13578]|1[02])\.20[0-9]{2}|(0?[1-9]|[12][1-9]|30)\.(0?[13456789]|1[012])\.20[0-9]{2}|(0?[1-9]|1[1-9]|2[0-8])\.(0?[123456789]|1[012])\.20[0-9]{2}|(0?[1-9]|[12][1-9])\.(0?[123456789]|1[012])\.20(00|04|08|12|16|20|24|28|32|36|40|44|48|52|56|60|64|68|72|76|80|84|88|92|96))$" "31.11.2002" +"^((0?[1-9]|[12][1-9]|3[01])\.(0?[13578]|1[02])\.20[0-9]{2}|(0?[1-9]|[12][1-9]|30)\.(0?[13456789]|1[012])\.20[0-9]{2}|(0?[1-9]|1[1-9]|2[0-8])\.(0?[123456789]|1[012])\.20[0-9]{2}|(0?[1-9]|[12][1-9])\.(0?[123456789]|1[012])\.20(00|04|08|12|16|20|24|28|32|36|40|44|48|52|56|60|64|68|72|76|80|84|88|92|96))$" "29.2.2002" +"^((0?[1-9]|[12][1-9]|3[01])\.(0?[13578]|1[02])\.20[0-9]{2}|(0?[1-9]|[12][1-9]|30)\.(0?[13456789]|1[012])\.20[0-9]{2}|(0?[1-9]|1[1-9]|2[0-8])\.(0?[123456789]|1[012])\.20[0-9]{2}|(0?[1-9]|[12][1-9])\.(0?[123456789]|1[012])\.20(00|04|08|12|16|20|24|28|32|36|40|44|48|52|56|60|64|68|72|76|80|84|88|92|96))$" "33.06.2000" +"^(0[1-9]|1[0-2])\/((0[1-9]|2\d)|3[0-1])\/(19\d\d|200[0-3])$" G "<0>12/31/2003</0>" +"^(0[1-9]|1[0-2])\/((0[1-9]|2\d)|3[0-1])\/(19\d\d|200[0-3])$" G "<0>01/01/1900</0>" +"^(0[1-9]|1[0-2])\/((0[1-9]|2\d)|3[0-1])\/(19\d\d|200[0-3])$" G "<0>11/31/2002</0>" +"^(0[1-9]|1[0-2])\/((0[1-9]|2\d)|3[0-1])\/(19\d\d|200[0-3])$" "1/1/2002" +"^(0[1-9]|1[0-2])\/((0[1-9]|2\d)|3[0-1])\/(19\d\d|200[0-3])$" "01/01/02" +"^(0[1-9]|1[0-2])\/((0[1-9]|2\d)|3[0-1])\/(19\d\d|200[0-3])$" "01/01/2004" +"^((((([13578])|(1[0-2]))[\-\/\s]?(([1-9])|([1-2][0-9])|(3[01])))|((([469])|(11))[\-\/\s]?(([1-9])|([1-2][0-9])|(30)))|(2[\-\/\s]?(([1-9])|([1-2][0-9]))))[\-\/\s]?\d{4})(\s((([1-9])|(1[02]))\:([0-5][0-9])((\s)|(\:([0-5][0-9])\s))([AM|PM|am|pm]{2,2})))?$" G "<0>3/3/2003</0>" +"^((((([13578])|(1[0-2]))[\-\/\s]?(([1-9])|([1-2][0-9])|(3[01])))|((([469])|(11))[\-\/\s]?(([1-9])|([1-2][0-9])|(30)))|(2[\-\/\s]?(([1-9])|([1-2][0-9]))))[\-\/\s]?\d{4})(\s((([1-9])|(1[02]))\:([0-5][0-9])((\s)|(\:([0-5][0-9])\s))([AM|PM|am|pm]{2,2})))?$" G "<0>3/3/2002 3:33 pm</0>" +"^((((([13578])|(1[0-2]))[\-\/\s]?(([1-9])|([1-2][0-9])|(3[01])))|((([469])|(11))[\-\/\s]?(([1-9])|([1-2][0-9])|(30)))|(2[\-\/\s]?(([1-9])|([1-2][0-9]))))[\-\/\s]?\d{4})(\s((([1-9])|(1[02]))\:([0-5][0-9])((\s)|(\:([0-5][0-9])\s))([AM|PM|am|pm]{2,2})))?$" G "<0>3/3/2003 3:33:33 am</0>" +"^((((([13578])|(1[0-2]))[\-\/\s]?(([1-9])|([1-2][0-9])|(3[01])))|((([469])|(11))[\-\/\s]?(([1-9])|([1-2][0-9])|(30)))|(2[\-\/\s]?(([1-9])|([1-2][0-9]))))[\-\/\s]?\d{4})(\s((([1-9])|(1[02]))\:([0-5][0-9])((\s)|(\:([0-5][0-9])\s))([AM|PM|am|pm]{2,2})))?$" "13/1/2002" +"^((((([13578])|(1[0-2]))[\-\/\s]?(([1-9])|([1-2][0-9])|(3[01])))|((([469])|(11))[\-\/\s]?(([1-9])|([1-2][0-9])|(30)))|(2[\-\/\s]?(([1-9])|([1-2][0-9]))))[\-\/\s]?\d{4})(\s((([1-9])|(1[02]))\:([0-5][0-9])((\s)|(\:([0-5][0-9])\s))([AM|PM|am|pm]{2,2})))?$" "3/3/2002 3:33" +"^((((([13578])|(1[0-2]))[\-\/\s]?(([1-9])|([1-2][0-9])|(3[01])))|((([469])|(11))[\-\/\s]?(([1-9])|([1-2][0-9])|(30)))|(2[\-\/\s]?(([1-9])|([1-2][0-9]))))[\-\/\s]?\d{4})(\s((([1-9])|(1[02]))\:([0-5][0-9])((\s)|(\:([0-5][0-9])\s))([AM|PM|am|pm]{2,2})))?$" "31/3/2002" +"([a-zA-Z]:(\\w+)*\\[a-zA-Z0_9]+)?.xls" G "<0>E:\DyAGT\SD01A_specV2.xls</0>" +"([a-zA-Z]:(\\w+)*\\[a-zA-Z0_9]+)?.xls" "E:\DyAGT\SD01A_specV2.txt" +"(((0[13578]|10|12)([-./])(0[1-9]|[12][0-9]|3[01])([-./])(\d{4}))|((0[469]|11)([-./])([0][1-9]|[12][0-9]|30)([-./])(\d{4}))|((2)([-./])(0[1-9]|1[0-9]|2[0-8])([-./])(\d{4}))|((2)(\.|-|\/)(29)([-./])([02468][048]00))|((2)([-./])(29)([-./])([13579][26]00))|((2)([-./])(29)([-./])([0-9][0-9][0][48]))|((2)([-./])(29)([-./])([0-9][0-9][2468][048]))|((2)([-./])(29)([-./])([0-9][0-9][13579][26])))" G "<0>02/29/2084</0>" +"(((0[13578]|10|12)([-./])(0[1-9]|[12][0-9]|3[01])([-./])(\d{4}))|((0[469]|11)([-./])([0][1-9]|[12][0-9]|30)([-./])(\d{4}))|((2)([-./])(0[1-9]|1[0-9]|2[0-8])([-./])(\d{4}))|((2)(\.|-|\/)(29)([-./])([02468][048]00))|((2)([-./])(29)([-./])([13579][26]00))|((2)([-./])(29)([-./])([0-9][0-9][0][48]))|((2)([-./])(29)([-./])([0-9][0-9][2468][048]))|((2)([-./])(29)([-./])([0-9][0-9][13579][26])))" G "<0>01/31/2000</0>" +"(((0[13578]|10|12)([-./])(0[1-9]|[12][0-9]|3[01])([-./])(\d{4}))|((0[469]|11)([-./])([0][1-9]|[12][0-9]|30)([-./])(\d{4}))|((2)([-./])(0[1-9]|1[0-9]|2[0-8])([-./])(\d{4}))|((2)(\.|-|\/)(29)([-./])([02468][048]00))|((2)([-./])(29)([-./])([13579][26]00))|((2)([-./])(29)([-./])([0-9][0-9][0][48]))|((2)([-./])(29)([-./])([0-9][0-9][2468][048]))|((2)([-./])(29)([-./])([0-9][0-9][13579][26])))" G "<0>11/30/2000</0>" +"(((0[13578]|10|12)([-./])(0[1-9]|[12][0-9]|3[01])([-./])(\d{4}))|((0[469]|11)([-./])([0][1-9]|[12][0-9]|30)([-./])(\d{4}))|((2)([-./])(0[1-9]|1[0-9]|2[0-8])([-./])(\d{4}))|((2)(\.|-|\/)(29)([-./])([02468][048]00))|((2)([-./])(29)([-./])([13579][26]00))|((2)([-./])(29)([-./])([0-9][0-9][0][48]))|((2)([-./])(29)([-./])([0-9][0-9][2468][048]))|((2)([-./])(29)([-./])([0-9][0-9][13579][26])))" "02/29/2083" +"(((0[13578]|10|12)([-./])(0[1-9]|[12][0-9]|3[01])([-./])(\d{4}))|((0[469]|11)([-./])([0][1-9]|[12][0-9]|30)([-./])(\d{4}))|((2)([-./])(0[1-9]|1[0-9]|2[0-8])([-./])(\d{4}))|((2)(\.|-|\/)(29)([-./])([02468][048]00))|((2)([-./])(29)([-./])([13579][26]00))|((2)([-./])(29)([-./])([0-9][0-9][0][48]))|((2)([-./])(29)([-./])([0-9][0-9][2468][048]))|((2)([-./])(29)([-./])([0-9][0-9][13579][26])))" "11/31/2000" +"(((0[13578]|10|12)([-./])(0[1-9]|[12][0-9]|3[01])([-./])(\d{4}))|((0[469]|11)([-./])([0][1-9]|[12][0-9]|30)([-./])(\d{4}))|((2)([-./])(0[1-9]|1[0-9]|2[0-8])([-./])(\d{4}))|((2)(\.|-|\/)(29)([-./])([02468][048]00))|((2)([-./])(29)([-./])([13579][26]00))|((2)([-./])(29)([-./])([0-9][0-9][0][48]))|((2)([-./])(29)([-./])([0-9][0-9][2468][048]))|((2)([-./])(29)([-./])([0-9][0-9][13579][26])))" "01/32/2000" +"^[a-zA-Z0-9\s .\-]+$" G "<0>2222 Mock St.</0>" # TODO: \s in patterns not implemented +"^[a-zA-Z0-9\s .\-]+$" G "<0>1 A St.</0>" +"^[a-zA-Z0-9\s .\-]+$" G "<0>555-1212</0>" +"^[a-zA-Z0-9\s.\-]+$" "[A Street]" +"^[a-zA-Z0-9\s.\-]+$" "(3 A St.)" +"^[a-zA-Z0-9\s.\-]+$" "{34 C Ave.}" +"^[a-zA-Z0-9\s.\-]+$" "Last.*?(\d+.?\d*)" +"^[a-zA-Z0-9\s .\-]+$" G "<TR><TD ALIGN=RIGHT> </TD><TD>Last</TD><TD ALIGN=RIGHT NOW" +"^[a-zA-Z0-9\s.\-]+$" "[AADDSS]" +"^([0-9]( |-)?)?(\(?[0-9]{3}\)?|[0-9]{3})( |-)?([0-9]{3}( |-)?[0-9]{4}|[a-zA-Z0-9]{7})$" G "<0>1-(123)-123-1234</0>" +"^([0-9]( |-)?)?(\(?[0-9]{3}\)?|[0-9]{3})( |-)?([0-9]{3}( |-)?[0-9]{4}|[a-zA-Z0-9]{7})$" G "<0>123 123 1234</0>" +"^([0-9]( |-)?)?(\(?[0-9]{3}\)?|[0-9]{3})( |-)?([0-9]{3}( |-)?[0-9]{4}|[a-zA-Z0-9]{7})$" G "<0>1-800-ALPHNUM</0>" +"^([0-9]( |-)?)?(\(?[0-9]{3}\)?|[0-9]{3})( |-)?([0-9]{3}( |-)?[0-9]{4}|[a-zA-Z0-9]{7})$" "1.123.123.1234" +"^([0-9]( |-)?)?(\(?[0-9]{3}\)?|[0-9]{3})( |-)?([0-9]{3}( |-)?[0-9]{4}|[a-zA-Z0-9]{7})$" "(123)-1234-123" +"^([0-9]( |-)?)?(\(?[0-9]{3}\)?|[0-9]{3})( |-)?([0-9]{3}( |-)?[0-9]{4}|[a-zA-Z0-9]{7})$" "123-1234" +"^([0-1][0-9]|[2][0-3]):([0-5][0-9])$" G "<0>02:04</0>" +"^([0-1][0-9]|[2][0-3]):([0-5][0-9])$" G "<0>16:56</0>" +"^([0-1][0-9]|[2][0-3]):([0-5][0-9])$" G "<0>23:59</0>" +"^([0-1][0-9]|[2][0-3]):([0-5][0-9])$" "02:00 PM" +"^([0-1][0-9]|[2][0-3]):([0-5][0-9])$" "PM2:00" +"^([0-1][0-9]|[2][0-3]):([0-5][0-9])$" "24:00" +"^[0,1]?\d{1}\/(([0-2]?\d{1})|([3][0,1]{1}))\/(([1]{1}[9]{1}[9]{1}\d{1})|([2-9]{1}\d{3}))$" G "<0>01/01/1990</0>" +"^[0,1]?\d{1}\/(([0-2]?\d{1})|([3][0,1]{1}))\/(([1]{1}[9]{1}[9]{1}\d{1})|([2-9]{1}\d{3}))$" G "<0>12/12/9999</0>" +"^[0,1]?\d{1}\/(([0-2]?\d{1})|([3][0,1]{1}))\/(([1]{1}[9]{1}[9]{1}\d{1})|([2-9]{1}\d{3}))$" G "<0>3/28/2001</0>" +"^[0,1]?\d{1}\/(([0-2]?\d{1})|([3][0,1]{1}))\/(([1]{1}[9]{1}[9]{1}\d{1})|([2-9]{1}\d{3}))$" "3-8-01" +"^[0,1]?\d{1}\/(([0-2]?\d{1})|([3][0,1]{1}))\/(([1]{1}[9]{1}[9]{1}\d{1})|([2-9]{1}\d{3}))$" "13/32/1001" +"^[0,1]?\d{1}\/(([0-2]?\d{1})|([3][0,1]{1}))\/(([1]{1}[9]{1}[9]{1}\d{1})|([2-9]{1}\d{3}))$" "03/32/1989" +"((\(\d{3}\)?)|(\d{3}))([\s \-./]?)(\d{3})([\s \-./]?)(\d{4})" G "<0>1.2123644567</0>" +"((\(\d{3}\)?)|(\d{3}))([\s \-./]?)(\d{3})([\s \-./]?)(\d{4})" G "<0>0-234.567/8912</0>" +"((\(\d{3}\)?)|(\d{3}))([\s \-./]?)(\d{3})([\s \-./]?)(\d{4})" G "<0>1-(212)-123 4567</0>" +"((\(\d{3}\)?)|(\d{3}))([\s \-./]?)(\d{3})([\s \-./]?)(\d{4})" "0-212364345" +"((\(\d{3}\)?)|(\d{3}))([\s \-./]?)(\d{3})([\s \-./]?)(\d{4})" "1212-364,4321" +"((\(\d{3}\)?)|(\d{3}))([\s \-./]?)(\d{3})([\s \-./]?)(\d{4})" "0212\345/6789" +"^([0-9]{6}[\s \-]{1}[0-9]{12}|[0-9]{18})$" G "<0>000000 000000000000</0>" +"^([0-9]{6}[\s \-]{1}[0-9]{12}|[0-9]{18})$" G "<0>000000-000000000000</0>" +"^([0-9]{6}[\s \-]{1}[0-9]{12}|[0-9]{18})$" G "<0>000000000000000000</0>" +"^([0-9]{6}[\s \-]{1}[0-9]{12}|[0-9]{18})$" "000000_000000000000" +"^(([1-9])|(0[1-9])|(1[0-2]))\/((0[1-9])|([1-31]))\/((\d{2})|(\d{4}))$" G "<0>01/01/2001</0>" +"^(([1-9])|(0[1-9])|(1[0-2]))\/((0[1-9])|([1-31]))\/((\d{2})|(\d{4}))$" G "<0>1/1/2001</0>" +"^(([1-9])|(0[1-9])|(1[0-2]))\/((0[1-9])|([1-31]))\/((\d{2})|(\d{4}))$" G "<0>01/1/01</0>" +"^(([1-9])|(0[1-9])|(1[0-2]))\/((0[1-9])|([1-31]))\/((\d{2})|(\d{4}))$" "13/01/2001" +"^(([1-9])|(0[1-9])|(1[0-2]))\/((0[1-9])|([1-31]))\/((\d{2})|(\d{4}))$" "1/2/100" +"^(([1-9])|(0[1-9])|(1[0-2]))\/((0[1-9])|([1-31]))\/((\d{2})|(\d{4}))$" "09/32/2001" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(.[0-9][0-9])?$" G "<0>$3,023,123.34</0>" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(.[0-9][0-9])?$" G "<0>9,876,453</0>" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(.[0-9][0-9])?$" G "<0>123456.78</0>" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(.[0-9][0-9])?$" "4,33,234.34" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(.[0-9][0-9])?$" "$1.234" +"^\$?([0-9]{1,3},([0-9]{3},)*[0-9]{3}|[0-9]+)(.[0-9][0-9])?$" "abc" +"^\d{5}$|^\d{5}-\d{4}$" G "<0>55555-5555</0>" +"^\d{5}$|^\d{5}-\d{4}$" G "<0>34564-3342</0>" +"^\d{5}$|^\d{5}-\d{4}$" G "<0>90210</0>" +"^\d{5}$|^\d{5}-\d{4}$" "434454444" +"^\d{5}$|^\d{5}-\d{4}$" "645-32-2345" +"^\d{5}$|^\d{5}-\d{4}$" "abc" +"^\d{3}-\d{2}-\d{4}$" G "<0>333-22-4444</0>" +"^\d{3}-\d{2}-\d{4}$" G "<0>123-45-6789</0>" +"^\d{3}-\d{2}-\d{4}$" "123456789" +"^\d{3}-\d{2}-\d{4}$" "SSN" +"^[2-9]\d{2}-\d{3}-\d{4}$" G "<0>800-555-5555</0>" +"^[2-9]\d{2}-\d{3}-\d{4}$" G "<0>333-444-5555</0>" +"^[2-9]\d{2}-\d{3}-\d{4}$" G "<0>212-666-1234</0>" +"^[2-9]\d{2}-\d{3}-\d{4}$" "000-000-0000" +"^[2-9]\d{2}-\d{3}-\d{4}$" "123-456-7890" +"^[2-9]\d{2}-\d{3}-\d{4}$" "2126661234" +"^\d{5}-\d{4}|\d{5}|[A-Z]\d[A-Z] \d[A-Z]\d$" G "<0>44240</0>" +"^\d{5}-\d{4}|\d{5}|[A-Z]\d[A-Z] \d[A-Z]\d$" G "<0>44240-5555</0>" +"^\d{5}-\d{4}|\d{5}|[A-Z]\d[A-Z] \d[A-Z]\d$" G "<0>G3H 6A3</0>" +"^\d{5}-\d{4}|\d{5}|[A-Z]\d[A-Z] \d[A-Z]\d$" "Ohio" +"^\d{5}-\d{4}|\d{5}|[A-Z]\d[A-Z] \d[A-Z]\d$" "abc" +"^\d{5}-\d{4}|\d{5}|[A-Z]\d[A-Z] \d[A-Z]\d$" "g3h6a3" +"[0-9]{4}\s*[a-zA-Z]{2}" G "<0>1054 WD</0>" +"[0-9]{4}\s*[a-zA-Z]{2}" G "<0>1054WD</0>" +"[0-9]{4}\s*[a-zA-Z]{2}" G "<0>1054 wd</0>" +"[0-9]{4}\s*[a-zA-Z]{2}" "10543" +"(^1300\d{6}$)|(^1800|1900|1902\d{6}$)|(^0[2|3|7|8]{1}[0-9]{8}$)|(^13\d{4}$)|(^04\d{2,3}\d{6}$)" G "<0>0732105432</0>" +"(^1300\d{6}$)|(^1800|1900|1902\d{6}$)|(^0[2|3|7|8]{1}[0-9]{8}$)|(^13\d{4}$)|(^04\d{2,3}\d{6}$)" G "<0>1300333444</0>" +"(^1300\d{6}$)|(^1800|1900|1902\d{6}$)|(^0[2|3|7|8]{1}[0-9]{8}$)|(^13\d{4}$)|(^04\d{2,3}\d{6}$)" G "<0>131313</0>" +"(^1300\d{6}$)|(^1800|1900|1902\d{6}$)|(^0[2|3|7|8]{1}[0-9]{8}$)|(^13\d{4}$)|(^04\d{2,3}\d{6}$)" "32105432" +"(^1300\d{6}$)|(^1800|1900|1902\d{6}$)|(^0[2|3|7|8]{1}[0-9]{8}$)|(^13\d{4}$)|(^04\d{2,3}\d{6}$)" "13000456" +"^((https?|ftp)\://((\[?(\d{1,3}\.){3}\d{1,3}\]?)|(([\-a-zA-Z0-9]+\.)+[a-zA-Z]{2,4}))(\:\d+)?(/[\-a-zA-Z0-9._?,'+\&%$#=~\\]+)*/?)$" G "<0>http://207.68.172.254/home.ashx</0>" +"^((https?|ftp)\://((\[?(\d{1,3}\.){3}\d{1,3}\]?)|(([\-a-zA-Z0-9]+\.)+[a-zA-Z]{2,4}))(\:\d+)?(/[\-a-zA-Z0-9._?,'+\&%$#=~\\]+)*/?)$" G "<0>ftp://ftp.netscape.com/</0>" +"^((https?|ftp)\://((\[?(\d{1,3}\.){3}\d{1,3}\]?)|(([\-a-zA-Z0-9]+\.)+[a-zA-Z]{2,4}))(\:\d+)?(/[\-a-zA-Z0-9._?,'+\&%$#=~\\]+)*/?)$" G "<0>https://www.brinkster.com/login.asp</0>" +"^((https?|ftp)\://((\[?(\d{1,3}\.){3}\d{1,3}\]?)|(([\-a-zA-Z0-9]+\.)+[a-zA-Z]{2,4}))(\:\d+)?(/[\-a-zA-Z0-9._?,'+\&%$#=~\\]+)*/?)$" "htp://mistake.com/" +"^((https?|ftp)\://((\[?(\d{1,3}\.){3}\d{1,3}\]?)|(([\-a-zA-Z0-9]+\.)+[a-zA-Z]{2,4}))(\:\d+)?(/[\-a-zA-Z0-9._?,'+\&%$#=~\\]+)*/?)$" "http://www_address.com/" +"^((https?|ftp)\://((\[?(\d{1,3}\.){3}\d{1,3}\]?)|(([\-a-zA-Z0-9]+\.)+[a-zA-Z]{2,4}))(\:\d+)?(/[\-a-zA-Z0-9._?,'+\&%$#=~\\]+)*/?)$" "ftp://www.files.com/file with spaces.txt" +"([0-9]{4})-([0-9]{1,2})-([0-9]{1,2})" G "<0>2002-11-03</0>" +"([0-9]{4})-([0-9]{1,2})-([0-9]{1,2})" G "<0>2007-17-08</0>" +"([0-9]{4})-([0-9]{1,2})-([0-9]{1,2})" G "<0>9999-99-99</0>" +"([0-9]{4})-([0-9]{1,2})-([0-9]{1,2})" "2002/17/18" +"([0-9]{4})-([0-9]{1,2})-([0-9]{1,2})" "2002.18.45" +"([0-9]{4})-([0-9]{1,2})-([0-9]{1,2})" "18.45.2002" +"^\$?(\d{1,3}(\,\d{3})*|(\d+))(\.\d{0,2})?$" G "<0>$0,234.50</0>" +"^\$?(\d{1,3}(\,\d{3})*|(\d+))(\.\d{0,2})?$" G "<0>0234.5</0>" +"^\$?(\d{1,3}(\,\d{3})*|(\d+))(\.\d{0,2})?$" G "<0>0,234.</0>" +"^\$?(\d{1,3}(\,\d{3})*|(\d+))(\.\d{0,2})?$" "$1,23,50" +"^\$?(\d{1,3}(\,\d{3})*|(\d+))(\.\d{0,2})?$" "$123.123" +"(^\d{5}-\d{3}|^\d{2}.\d{3}-\d{3}|\d{8})" G "<0>12.345-678</0>" +"(^\d{5}-\d{3}|^\d{2}.\d{3}-\d{3}|\d{8})" G "<0>12345-678</0>" +"(^\d{5}-\d{3}|^\d{2}.\d{3}-\d{3}|\d{8})" G "<0>12345678</0>" +"(^\d{5}-\d{3}|^\d{2}.\d{3}-\d{3}|\d{8})" "12.345678" +"(^\d{5}-\d{3}|^\d{2}.\d{3}-\d{3}|\d{8})" "12345-1" +"(^\d{5}-\d{3}|^\d{2}.\d{3}-\d{3}|\d{8})" "123" +'^([a-zA-Z]\:|\\)\\([^\\]+\\)*[^\/:*?"<>|]+\.htm(l)?$' G "<0>x:\\test\\testing.htm</0>" +'^([a-zA-Z]\:|\\)\\([^\\]+\\)*[^\/:*?"<>|]+\.htm(l)?$' G "<0>x:\\test\\test#$ ing.html</0>" +'^([a-zA-Z]\:|\\)\\([^\\]+\\)*[^\/:*?"<>|]+\.htm(l)?$' G "<0>\\\\test\testing.html</0>" +'^([a-zA-Z]\:|\\)\\([^\\]+\\)*[^\/:*?"<>|]+\.htm(l)?$' "x:\test\test/ing.htm" +'^([a-zA-Z]\:|\\)\\([^\\]+\\)*[^\/:*?"<>|]+\.htm(l)?$' "x:\test\test*.htm" +'^([a-zA-Z]\:|\\)\\([^\\]+\\)*[^\/:*?"<>|]+\.htm(l)?$' "\\test?<.htm" +"^[1-9]{1}[0-9]{3}$" G "<0>1234</0>" +"^[1-9]{1}[0-9]{3}$" "123" +"^[1-9]{1}[0-9]{3}$" "123A" +"^[A-Z]{1}( |-)?[1-9]{1}[0-9]{3}$" G "<0>A-1234</0>" +"^[A-Z]{1}( |-)?[1-9]{1}[0-9]{3}$" G "<0>A 1234</0>" +"^[A-Z]{1}( |-)?[1-9]{1}[0-9]{3}$" G "<0>A1234</0>" +"^[A-Z]{1}( |-)?[1-9]{1}[0-9]{3}$" "AA-1234" +"^[A-Z]{1}( |-)?[1-9]{1}[0-9]{3}$" "A12345" +"^(F-)?[0-9]{5}$" G "<0>12345</0>" +"^(F-)?[0-9]{5}$" G "<0>F-12345</0>" +"^(F-)?[0-9]{5}$" "F12345" +"^(F-)?[0-9]{5}$" "F-123456" +"^(F-)?[0-9]{5}$" "123456" +"^(V-|I-)?[0-9]{4}$" G "<0>1234</0>" +"^(V-|I-)?[0-9]{4}$" G "<0>V-1234</0>" +"^(V-|I-)?[0-9]{4}$" "12345" +"^[1-9]{1}[0-9]{3} ?[A-Z]{2}$" G "<0>1234 AB</0>" +"^[1-9]{1}[0-9]{3} ?[A-Z]{2}$" G "<0>1234AB</0>" +"^[1-9]{1}[0-9]{3} ?[A-Z]{2}$" "123AB" +"^[1-9]{1}[0-9]{3} ?[A-Z]{2}$" "1234AAA" +"^([1-9]{2}|[0-9][1-9]|[1-9][0-9])[0-9]{3}$" G "<0>12345</0>" +"^([1-9]{2}|[0-9][1-9]|[1-9][0-9])[0-9]{3}$" G "<0>10234</0>" +"^([1-9]{2}|[0-9][1-9]|[1-9][0-9])[0-9]{3}$" G "<0>01234</0>" +"^([1-9]{2}|[0-9][1-9]|[1-9][0-9])[0-9]{3}$" "00123" +"^(/w|/W|[^<>+?$%\{}\&])+$" G "<0>John Doe Sr.</0>" +"^(/w|/W|[^<>+?$%\{}\&])+$" G "<0>100 Elm St., Suite 25</0>" +"^(/w|/W|[^<>+?$%\{}\&])+$" G "<0>Valerie's Gift Shop</0>" +"^(/w|/W|[^<>+?$%\{}\&])+$" "<h1>Hey</h1>" +/<[a-zA-Z][^>]*\son\w+=(\w+|'[^']*'|"[^"]*")[^>]*>/ G '<0><IMG onmouseover="window.close()"></0>' +/<[a-zA-Z][^>]*\son\w+=(\w+|'[^']*'|"[^"]*")[^>]*>/ '<IMG src="star.gif">' +"(?!^0*$)(?!^0*\.0*$)^\d{1,5}(\.\d{1,3})?$" G "<0>1</0>" +"(?!^0*$)(?!^0*\.0*$)^\d{1,5}(\.\d{1,3})?$" G "<0>12345.123</0>" +"(?!^0*$)(?!^0*\.0*$)^\d{1,5}(\.\d{1,3})?$" G "<0>0.5</0>" +"(?!^0*$)(?!^0*\.0*$)^\d{1,5}(\.\d{1,3})?$" "0" +"(?!^0*$)(?!^0*\.0*$)^\d{1,5}(\.\d{1,3})?$" "0.0" +"(?!^0*$)(?!^0*\.0*$)^\d{1,5}(\.\d{1,3})?$" "123456.1234" +"^.+@[^\.].*\.[a-z]{2,}$" G "<0>whatever@somewhere.museum</0>" +"^.+@[^\.].*\.[a-z]{2,}$" G "<0>foreignchars@myforeigncharsdomain.nu</0>" +"^.+@[^\.].*\.[a-z]{2,}$" G "<0>me+mysomething@mydomain.com</0>" +"^.+@[^\.].*\.[a-z]{2,}$" "a@b.c" +"^.+@[^\.].*\.[a-z]{2,}$" "me@.my.com" +"^.+@[^\.].*\.[a-z]{2,}$" "a@b.comFOREIGNCHAR" +"^(\d{5}-\d{4}|\d{5})$" G "<0>12345</0>" +"^(\d{5}-\d{4}|\d{5})$" G "<0>12345-1234</0>" +"^(\d{5}-\d{4}|\d{5})$" "12345-12345" +"^(\d{5}-\d{4}|\d{5})$" "123" +"^(\d{5}-\d{4}|\d{5})$" "12345-abcd" +"^(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])$" G "<0>0.0.0.0</0>" +"^(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])$" G "<0>255.255.255.02</0>" +"^(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])$" G "<0>192.168.0.136</0>" +"^(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])$" "256.1.3.4" +"^(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])$" "023.44.33.22" +"^(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])\.(\d{1,2}|1\d\d|2[0-4]\d|25[0-5])$" "10.57.98.23." +"<img([^>]*[^/])>" G '<0><img src="bob"></0>' +"<img([^>]*[^/])>" '<img src="bob" />' +"<!--[\s\S]*?-->" G "<0><!-- comments --></0>" +"<!--[\s\S]*?-->" G "<0><!-- x = a > b - 3 --></0>" +"<!--[\s\S]*?-->" "<COMMENTS>this is a comment</COMMENTS>" +"<!--[\p{Zs}\P{Zs}]*?-->" G "<0><!-- comments --></0>" +"<!--[\p{Zs}\P{Zs}]*?-->" G "<0><!-- x = a > b - 3 --></0>" +"<!--[\p{Zs}\P{Zs}]*?-->" "<COMMENTS>this is a comment</COMMENTS>" +/<\u002f?(\w+)(\s+\w+=(\w+|"[^"]*"|'[^']*'))*>/ G "<0><TD></0>" +/<\u002f?(\w+)(\s+\w+=(\w+|"[^"]*"|'[^']*'))*>/ G '<0><TD bgColor="FFFFFF"></0>' +/<\u002f?(\w+)(\s+\w+=(\w+|"[^"]*"|'[^']*'))*>/ G "<0></TD></0>" +/<\u002f?(\w+)(\s+\w+=(\w+|"[^"]*"|'[^']*'))*>/ "No Tag Here ..." +"(\{\\f\d*)\\([^;]+;)" G "<0>{\\f0\\Some Font names here;</0>" +"(\{\\f\d*)\\([^;]+;)" G "<0>{\\f1\\fswiss\\fcharset0\\fprq2{\\*\\panose 020b0604020202020204}Arial;</0>" +"(\{\\f\d*)\\([^;]+;)" G "{\\f" +"(\{\\f\d*)\\([^;]+;)" "{f0fs20 some text}" +#"</?([a-zA-Z][-A-Za-z\d\.]{0,71})(\s+(\S+)(\s*=\s*([-\w\.]{1,1024}|"[^"]{0,1024}"|'[^']{0,1024}'))?)*\s*>" G '<0><IMG src='stars.gif' alt="space" height=1></0>' # TODO: Can't quote this pattern with the test syntax! +#"</?([a-zA-Z][-A-Za-z\d\.]{0,71})(\s+(\S+)(\s*=\s*([-\w\.]{1,1024}|"[^"]{0,1024}"|'[^']{0,1024}'))?)*\s*>" "this is not a tag" +"^([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\/([2][01]|[1][6-9])\d{2}(\s([0-1]\d|[2][0-3])(\:[0-5]\d){1,2})?$" G "<0>12/30/2002</0>" +"^([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\/([2][01]|[1][6-9])\d{2}(\s([0-1]\d|[2][0-3])(\:[0-5]\d){1,2})?$" G "<0>01/12/1998 13:30</0>" +"^([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\/([2][01]|[1][6-9])\d{2}(\s([0-1]\d|[2][0-3])(\:[0-5]\d){1,2})?$" G "<0>01/28/2002 22:35:00</0>" +"^([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\/([2][01]|[1][6-9])\d{2}(\s([0-1]\d|[2][0-3])(\:[0-5]\d){1,2})?$" "13/30/2002" +"^([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\/([2][01]|[1][6-9])\d{2}(\s([0-1]\d|[2][0-3])(\:[0-5]\d){1,2})?$" "01/12/1998 24:30" +"^([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\/([2][01]|[1][6-9])\d{2}(\s([0-1]\d|[2][0-3])(\:[0-5]\d){1,2})?$" "01/28/2002 22:35:64" +#"((?<strElement>(^[A-Z0-9-;=]*:))(?<strValue>(.*)))" G "<0>BEGIN:</0>" #named capture +#"((?<strElement>(^[A-Z0-9-;=]*:))(?<strValue>(.*)))" G "<0>TEL;WORK;VOICE:</0>" #named capture +#"((?<strElement>(^[A-Z0-9-;=]*:))(?<strValue>(.*)))" G "<0>TEL:</0>" #named capture +#"((?<strElement>(^[A-Z0-9-;=]*:))(?<strValue>(.*)))" "begin:" #named capture +#"((?<strElement>(^[A-Z0-9-;=]*:))(?<strValue>(.*)))" "TEL;PREF;" #named capture +'^<a\s+href\s*=\s*"http:\/\/([^"]*)"([^>]*)>(.*?(?=<\/a>))<\/a>$' G '<0><a href="http://www.mysite.com">my external link</a></0>' +'^<a\s+href\s*=\s*"http:\/\/([^"]*)"([^>]*)>(.*?(?=<\/a>))<\/a>$' G '<a href="http:/' +'^<a\s+href\s*=\s*"http:\/\/([^"]*)"([^>]*)>(.*?(?=<\/a>))<\/a>$' '<a href="myinternalpage.html">my internal link</a>' +"^([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\/([2][01]|[1][6-9])\d{2}(\s([0]\d|[1][0-2])(\:[0-5]\d){1,2})*\s*([aApP][mM]{0,2})?$" G "<0>12/31/2002</0>" +"^([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\/([2][01]|[1][6-9])\d{2}(\s([0]\d|[1][0-2])(\:[0-5]\d){1,2})*\s*([aApP][mM]{0,2})?$" G "<0>12/31/2002 08:00</0>" +"^([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\/([2][01]|[1][6-9])\d{2}(\s([0]\d|[1][0-2])(\:[0-5]\d){1,2})*\s*([aApP][mM]{0,2})?$" G "<0>12/31/2002 08:00 AM</0>" +"^([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\/([2][01]|[1][6-9])\d{2}(\s([0]\d|[1][0-2])(\:[0-5]\d){1,2})*\s*([aApP][mM]{0,2})?$" "12/31/02" +"^([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\/([2][01]|[1][6-9])\d{2}(\s([0]\d|[1][0-2])(\:[0-5]\d){1,2})*\s*([aApP][mM]{0,2})?$" "12/31/2002 14:00" +"<blockquote>(?:\s*([^<]+)<br>\s*)+</blockquote>" G "<0><blockquote>string1<br>string2<br>string3<br></blockquote></0>" +"<blockquote>(?:\s*([^<]+)<br>\s*)+</blockquote>" ".." +"^((0?[13578]|10|12)(-|\/)((0[0-9])|([12])([0-9]?)|(3[01]?))(-|\/)((19)([2-9])(\d{1})|(20)([01])(\d{1})|([8901])(\d{1}))|(0?[2469]|11)(-|\/)((0[0-9])|([12])([0-9]?)|(3[0]?))(-|\/)((19)([2-9])(\d{1})|(20)([01])(\d{1})|([8901])(\d{1})))$" G "<0>1/2/03</0>" +"^((0?[13578]|10|12)(-|\/)((0[0-9])|([12])([0-9]?)|(3[01]?))(-|\/)((19)([2-9])(\d{1})|(20)([01])(\d{1})|([8901])(\d{1}))|(0?[2469]|11)(-|\/)((0[0-9])|([12])([0-9]?)|(3[0]?))(-|\/)((19)([2-9])(\d{1})|(20)([01])(\d{1})|([8901])(\d{1})))$" G "<0>2/30/1999</0>" +"^((0?[13578]|10|12)(-|\/)((0[0-9])|([12])([0-9]?)|(3[01]?))(-|\/)((19)([2-9])(\d{1})|(20)([01])(\d{1})|([8901])(\d{1}))|(0?[2469]|11)(-|\/)((0[0-9])|([12])([0-9]?)|(3[0]?))(-|\/)((19)([2-9])(\d{1})|(20)([01])(\d{1})|([8901])(\d{1})))$" G "<0>03/04/19</0>" +"^((0?[13578]|10|12)(-|\/)((0[0-9])|([12])([0-9]?)|(3[01]?))(-|\/)((19)([2-9])(\d{1})|(20)([01])(\d{1})|([8901])(\d{1}))|(0?[2469]|11)(-|\/)((0[0-9])|([12])([0-9]?)|(3[0]?))(-|\/)((19)([2-9])(\d{1})|(20)([01])(\d{1})|([8901])(\d{1})))$" "3/4/2020" +"^((0?[13578]|10|12)(-|\/)((0[0-9])|([12])([0-9]?)|(3[01]?))(-|\/)((19)([2-9])(\d{1})|(20)([01])(\d{1})|([8901])(\d{1}))|(0?[2469]|11)(-|\/)((0[0-9])|([12])([0-9]?)|(3[0]?))(-|\/)((19)([2-9])(\d{1})|(20)([01])(\d{1})|([8901])(\d{1})))$" "3/4/1919" +'</?(\w+)(\s*\w*\s*=\s*("[^"]*"|\u0027[^\u0027]\u0027|[^>]*))*|/?>' G '<0><font color="blue"></0>' +'</?(\w+)(\s*\w*\s*=\s*("[^"]*"|\u0027[^\u0027]\u0027|[^>]*))*|/?>' G "<0></font></0>" +'</?(\w+)(\s*\w*\s*=\s*("[^"]*"|\u0027[^\u0027]\u0027|[^>]*))*|/?>' G "<0><br /></0>" +'</?(\w+)(\s*\w*\s*=\s*("[^"]*"|\u0027[^\u0027]\u0027|[^>]*))*|/?>' "this is a test..." +"^ *(1[0-2]|[1-9]):[0-5][0-9] *(a|p|A|P)(m|M) *$" G "<0>12:00am</0>" +"^ *(1[0-2]|[1-9]):[0-5][0-9] *(a|p|A|P)(m|M) *$" G "<0>1:00 PM</0>" +"^ *(1[0-2]|[1-9]):[0-5][0-9] *(a|p|A|P)(m|M) *$" G "<0> 12:59 pm</0>" +"^ *(1[0-2]|[1-9]):[0-5][0-9] *(a|p|A|P)(m|M) *$" "0:00" +"^ *(1[0-2]|[1-9]):[0-5][0-9] *(a|p|A|P)(m|M) *$" "0:01 am" +"^ *(1[0-2]|[1-9]):[0-5][0-9] *(a|p|A|P)(m|M) *$" "13:00 pm" +"\({1}[0-9]{3}\){1}\-{1}[0-9]{3}\-{1}[0-9]{4}" G "<0>(111)-111-1111</0>" +"\({1}[0-9]{3}\){1}\-{1}[0-9]{3}\-{1}[0-9]{4}" "11111111111" +"[^abc]" G "<0>def</0>" +"[^abc]" "abc" +"^(([0]?[1-9]|[1][0-2])[\/|\-|\.]([0-2]\d|[3][0-1]|[1-9])[\/|\-|\.]([2][0])?\d{2}\s+((([0][0-9]|[1][0-2]|[0-9])[\:|\-|\.]([0-5]\d)\s*([aApP][mM])?)|(([0-1][0-9]|[2][0-3]|[0-9])[\:|\-|\.]([0-5]\d))))$" G "<0>01/01/2002 04:42</0>" +"^(([0]?[1-9]|[1][0-2])[\/|\-|\.]([0-2]\d|[3][0-1]|[1-9])[\/|\-|\.]([2][0])?\d{2}\s+((([0][0-9]|[1][0-2]|[0-9])[\:|\-|\.]([0-5]\d)\s*([aApP][mM])?)|(([0-1][0-9]|[2][0-3]|[0-9])[\:|\-|\.]([0-5]\d))))$" G "<0>5-12-02 04:42 AM</0>" +"^(([0]?[1-9]|[1][0-2])[\/|\-|\.]([0-2]\d|[3][0-1]|[1-9])[\/|\-|\.]([2][0])?\d{2}\s+((([0][0-9]|[1][0-2]|[0-9])[\:|\-|\.]([0-5]\d)\s*([aApP][mM])?)|(([0-1][0-9]|[2][0-3]|[0-9])[\:|\-|\.]([0-5]\d))))$" G "<0>01.01/02 04-42aM</0>" +"^(([0]?[1-9]|[1][0-2])[\/|\-|\.]([0-2]\d|[3][0-1]|[1-9])[\/|\-|\.]([2][0])?\d{2}\s+((([0][0-9]|[1][0-2]|[0-9])[\:|\-|\.]([0-5]\d)\s*([aApP][mM])?)|(([0-1][0-9]|[2][0-3]|[0-9])[\:|\-|\.]([0-5]\d))))$" "01-12-1999 4:50PM" +"^(([0]?[1-9]|[1][0-2])[\/|\-|\.]([0-2]\d|[3][0-1]|[1-9])[\/|\-|\.]([2][0])?\d{2}\s+((([0][0-9]|[1][0-2]|[0-9])[\:|\-|\.]([0-5]\d)\s*([aApP][mM])?)|(([0-1][0-9]|[2][0-3]|[0-9])[\:|\-|\.]([0-5]\d))))$" "01-12-2002 15:10PM" +"^(([0]?[1-9]|[1][0-2])[\/|\-|\.]([0-2]\d|[3][0-1]|[1-9])[\/|\-|\.]([2][0])?\d{2}\s+((([0][0-9]|[1][0-2]|[0-9])[\:|\-|\.]([0-5]\d)\s*([aApP][mM])?)|(([0-1][0-9]|[2][0-3]|[0-9])[\:|\-|\.]([0-5]\d))))$" "01-12-002 8:20PM" +"^([1][12]|[0]?[1-9])[\/-]([3][01]|[12]\d|[0]?[1-9])[\/-](\d{4}|\d{2})$" G "<0>11-02-02</0>" +"^([1][12]|[0]?[1-9])[\/-]([3][01]|[12]\d|[0]?[1-9])[\/-](\d{4}|\d{2})$" G "<0>1-25-2002</0>" +"^([1][12]|[0]?[1-9])[\/-]([3][01]|[12]\d|[0]?[1-9])[\/-](\d{4}|\d{2})$" G "<0>01/25/2002</0>" +"^([1][12]|[0]?[1-9])[\/-]([3][01]|[12]\d|[0]?[1-9])[\/-](\d{4}|\d{2})$" "13-02-02" +"^([1][12]|[0]?[1-9])[\/-]([3][01]|[12]\d|[0]?[1-9])[\/-](\d{4}|\d{2})$" "11.02.02" +"^([1][12]|[0]?[1-9])[\/-]([3][01]|[12]\d|[0]?[1-9])[\/-](\d{4}|\d{2})$" "11/32/2002" +"(([0-1][0-9])|([2][0-3])):([0-5][0-9]):([0-5][0-9])" G "<0>09:30:00</0>" +"(([0-1][0-9])|([2][0-3])):([0-5][0-9]):([0-5][0-9])" G "<0>17:45:20</0>" +"(([0-1][0-9])|([2][0-3])):([0-5][0-9]):([0-5][0-9])" G "<0>23:59:59</0>" +"(([0-1][0-9])|([2][0-3])):([0-5][0-9]):([0-5][0-9])" "24:00:00" +"(((0[1-9]|[12][0-9]|3[01])([-./])(0[13578]|10|12)([-./])(\d{4}))|(([0][1-9]|[12][0-9]|30)([-./])(0[469]|11)([-./])(\d{4}))|((0[1-9]|1[0-9]|2[0-8])([-./])(02)([-./])(\d{4}))|((29)(\.|-|\/)(02)([-./])([02468][048]00))|((29)([-./])(02)([-./])([13579][26]00))|((29)([-./])(02)([-./])([0-9][0-9][0][48]))|((29)([-./])(02)([-./])([0-9][0-9][2468][048]))|((29)([-./])(02)([-./])([0-9][0-9][13579][26])))" G "<0>29/02/2000</0>" +"(((0[1-9]|[12][0-9]|3[01])([-./])(0[13578]|10|12)([-./])(\d{4}))|(([0][1-9]|[12][0-9]|30)([-./])(0[469]|11)([-./])(\d{4}))|((0[1-9]|1[0-9]|2[0-8])([-./])(02)([-./])(\d{4}))|((29)(\.|-|\/)(02)([-./])([02468][048]00))|((29)([-./])(02)([-./])([13579][26]00))|((29)([-./])(02)([-./])([0-9][0-9][0][48]))|((29)([-./])(02)([-./])([0-9][0-9][2468][048]))|((29)([-./])(02)([-./])([0-9][0-9][13579][26])))" G "<0>31/01/2000</0>" +"(((0[1-9]|[12][0-9]|3[01])([-./])(0[13578]|10|12)([-./])(\d{4}))|(([0][1-9]|[12][0-9]|30)([-./])(0[469]|11)([-./])(\d{4}))|((0[1-9]|1[0-9]|2[0-8])([-./])(02)([-./])(\d{4}))|((29)(\.|-|\/)(02)([-./])([02468][048]00))|((29)([-./])(02)([-./])([13579][26]00))|((29)([-./])(02)([-./])([0-9][0-9][0][48]))|((29)([-./])(02)([-./])([0-9][0-9][2468][048]))|((29)([-./])(02)([-./])([0-9][0-9][13579][26])))" G "<0>30-01-2000</0>" +"(((0[1-9]|[12][0-9]|3[01])([-./])(0[13578]|10|12)([-./])(\d{4}))|(([0][1-9]|[12][0-9]|30)([-./])(0[469]|11)([-./])(\d{4}))|((0[1-9]|1[0-9]|2[0-8])([-./])(02)([-./])(\d{4}))|((29)(\.|-|\/)(02)([-./])([02468][048]00))|((29)([-./])(02)([-./])([13579][26]00))|((29)([-./])(02)([-./])([0-9][0-9][0][48]))|((29)([-./])(02)([-./])([0-9][0-9][2468][048]))|((29)([-./])(02)([-./])([0-9][0-9][13579][26])))" "29/02/2002" +"(((0[1-9]|[12][0-9]|3[01])([-./])(0[13578]|10|12)([-./])(\d{4}))|(([0][1-9]|[12][0-9]|30)([-./])(0[469]|11)([-./])(\d{4}))|((0[1-9]|1[0-9]|2[0-8])([-./])(02)([-./])(\d{4}))|((29)(\.|-|\/)(02)([-./])([02468][048]00))|((29)([-./])(02)([-./])([13579][26]00))|((29)([-./])(02)([-./])([0-9][0-9][0][48]))|((29)([-./])(02)([-./])([0-9][0-9][2468][048]))|((29)([-./])(02)([-./])([0-9][0-9][13579][26])))" "32/01/2002" +"(((0[1-9]|[12][0-9]|3[01])([-./])(0[13578]|10|12)([-./])(\d{4}))|(([0][1-9]|[12][0-9]|30)([-./])(0[469]|11)([-./])(\d{4}))|((0[1-9]|1[0-9]|2[0-8])([-./])(02)([-./])(\d{4}))|((29)(\.|-|\/)(02)([-./])([02468][048]00))|((29)([-./])(02)([-./])([13579][26]00))|((29)([-./])(02)([-./])([0-9][0-9][0][48]))|((29)([-./])(02)([-./])([0-9][0-9][2468][048]))|((29)([-./])(02)([-./])([0-9][0-9][13579][26])))" "10/2/2002" +"^0[1-6]{1}(([0-9]{2}){4})|((\s[0-9]{2}){4})|((-[0-9]{2}){4})$" G "<0>01 46 70 89 12</0>" +"^0[1-6]{1}(([0-9]{2}){4})|((\s[0-9]{2}){4})|((-[0-9]{2}){4})$" G "<0>01-46-70-89-12</0>" +"^0[1-6]{1}(([0-9]{2}){4})|((\s[0-9]{2}){4})|((-[0-9]{2}){4})$" G "<0>0146708912</0>" +"^0[1-6]{1}(([0-9]{2}){4})|((\s[0-9]{2}){4})|((-[0-9]{2}){4})$" "01-46708912" +"^0[1-6]{1}(([0-9]{2}){4})|((\s[0-9]{2}){4})|((-[0-9]{2}){4})$" "01 46708912" +"^0[1-6]{1}(([0-9]{2}){4})|((\s[0-9]{2}){4})|((-[0-9]{2}){4})$" "+33235256677" +"^[0-9A-Za-z_ ]+(.[jJ][pP][gG]|.[gG][iI][fF])$" G "<0>good.gif</0>" +"^[0-9A-Za-z_ ]+(.[jJ][pP][gG]|.[gG][iI][fF])$" G "<0>go d.GIf</0>" +"^[0-9A-Za-z_ ]+(.[jJ][pP][gG]|.[gG][iI][fF])$" G "<0>goo_d.jPg</0>" +"^[0-9A-Za-z_ ]+(.[jJ][pP][gG]|.[gG][iI][fF])$" "junk" +"^[0-9A-Za-z_ ]+(.[jJ][pP][gG]|.[gG][iI][fF])$" "bad.bad.gif" +"^[0-9A-Za-z_ ]+(.[jJ][pP][gG]|.[gG][iI][fF])$" "slash\gif." +"<[^>\s]*\bauthor\b[^>]*>" G '<0><author name="Daniel"></0>' +"<[^>\s]*\bauthor\b[^>]*>" G "<0></sch:author></0>" +# "<[^>\s]*\bauthor\b[^>]*>" G '<0><pp:author name="Daniel"</0>' #Debug should work +"<[^> ]*\bauthor\b[^>]*>" G "<0></sch:author></0>" +"<[^> ]*\bauthor\b[^>]*>" G '<0><pp:author name="Daniel"></0>' +"<[^>\s]*\bauthor\b[^>]*>" "<other>" +"<[^>\s]*\bauthor\b[^>]*>" "</authors>" +"<[^>\s]*\bauthor\b[^>]*>" "<work>author</work>" +"^(?:(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00)))(\/|-|\.)(?:0?2\1(?:29))$)|(?:(?:1[6-9]|[2-9]\d)?\d{2})(\/|-|\.)(?:(?:(?:0?[13578]|1[02])\2(?:31))|(?:(?:0?[1,3-9]|1[0-2])\2(29|30))|(?:(?:0?[1-9])|(?:1[0-2]))\2(?:0?[1-9]|1\d|2[0-8]))$" G "<0>04/2/29</0>" +"^(?:(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00)))(\/|-|\.)(?:0?2\1(?:29))$)|(?:(?:1[6-9]|[2-9]\d)?\d{2})(\/|-|\.)(?:(?:(?:0?[13578]|1[02])\2(?:31))|(?:(?:0?[1,3-9]|1[0-2])\2(29|30))|(?:(?:0?[1-9])|(?:1[0-2]))\2(?:0?[1-9]|1\d|2[0-8]))$" G "<0>2002-4-30</0>" +"^(?:(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00)))(\/|-|\.)(?:0?2\1(?:29))$)|(?:(?:1[6-9]|[2-9]\d)?\d{2})(\/|-|\.)(?:(?:(?:0?[13578]|1[02])\2(?:31))|(?:(?:0?[1,3-9]|1[0-2])\2(29|30))|(?:(?:0?[1-9])|(?:1[0-2]))\2(?:0?[1-9]|1\d|2[0-8]))$" G "<0>02.10.31</0>" +"^(?:(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00)))(\/|-|\.)(?:0?2\1(?:29))$)|(?:(?:1[6-9]|[2-9]\d)?\d{2})(\/|-|\.)(?:(?:(?:0?[13578]|1[02])\2(?:31))|(?:(?:0?[1,3-9]|1[0-2])\2(29|30))|(?:(?:0?[1-9])|(?:1[0-2]))\2(?:0?[1-9]|1\d|2[0-8]))$" "2003/2/29" +"^(?:(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00)))(\/|-|\.)(?:0?2\1(?:29))$)|(?:(?:1[6-9]|[2-9]\d)?\d{2})(\/|-|\.)(?:(?:(?:0?[13578]|1[02])\2(?:31))|(?:(?:0?[1,3-9]|1[0-2])\2(29|30))|(?:(?:0?[1-9])|(?:1[0-2]))\2(?:0?[1-9]|1\d|2[0-8]))$" "02.4.31" +"^(?:(?:(?:(?:1[6-9]|[2-9]\d)?(?:0[48]|[2468][048]|[13579][26])|(?:(?:16|[2468][048]|[3579][26])00)))(\/|-|\.)(?:0?2\1(?:29))$)|(?:(?:1[6-9]|[2-9]\d)?\d{2})(\/|-|\.)(?:(?:(?:0?[13578]|1[02])\2(?:31))|(?:(?:0?[1,3-9]|1[0-2])\2(29|30))|(?:(?:0?[1-9])|(?:1[0-2]))\2(?:0?[1-9]|1\d|2[0-8]))$" "00/00/00" +'(\d*)\u0027*-*(\d*)/*(\d*)"' G '<0>5\u0027-3/16"</0>' +'(\d*)\u0027*-*(\d*)/*(\d*)"' G '<0>1\u0027-2"</0>' +'(\d*)\u0027*-*(\d*)/*(\d*)"' G '<0>5/16"</0>' +'(\d*)\u0027*-*(\d*)/*(\d*)"' '1 3/16' +"^[1-9]{1}$|^[1-4]{1}[0-9]{1}$|^50$" G "<0>1</0>" +"^[1-9]{1}$|^[1-4]{1}[0-9]{1}$|^50$" G "<0>23</0>" +"^[1-9]{1}$|^[1-4]{1}[0-9]{1}$|^50$" G "<0>50</0>" +"^[1-9]{1}$|^[1-4]{1}[0-9]{1}$|^50$" "0" +"^[1-9]{1}$|^[1-4]{1}[0-9]{1}$|^50$" "111" +"^[1-9]{1}$|^[1-4]{1}[0-9]{1}$|^50$" "xyz" +"^([ \u00c0-\u01ffa-zA-Z'])+$" G "<0>Jon Doe</0>" +"^([ \u00c0-\u01ffa-zA-Z'])+$" G "<0>J\u00f8rn</0>" +"^([ \u00c0-\u01ffa-zA-Z'])+$" G "<0>Mc'Neelan</0>" +"^([ \u00c0-\u01ffa-zA-Z'])+$" "Henry); hacking attempt" +"^((([0]?[1-9]|1[0-2])(:|\.)(00|15|30|45)?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)(00|15|30|45)?))$" G "<0>1:00 PM</0>" +"^((([0]?[1-9]|1[0-2])(:|\.)(00|15|30|45)?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)(00|15|30|45)?))$" G "<0>6:45 am</0>" +"^((([0]?[1-9]|1[0-2])(:|\.)(00|15|30|45)?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)(00|15|30|45)?))$" G "<0>17:30</0>" +"^((([0]?[1-9]|1[0-2])(:|\.)(00|15|30|45)?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)(00|15|30|45)?))$" "4:32 am" +"^((([0]?[1-9]|1[0-2])(:|\.)(00|15|30|45)?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)(00|15|30|45)?))$" "5:30:00 am" +"^((([0]?[1-9]|1[0-2])(:|\.)(00|15|30|45)?( )?(AM|am|aM|Am|PM|pm|pM|Pm))|(([0]?[0-9]|1[0-9]|2[0-3])(:|\.)(00|15|30|45)?))$" "17:01" +"(^\d*\.?\d*[1-9]+\d*$)|(^[1-9]+\d*\.\d*$)" G "<0>0.050</0>" +"(^\d*\.?\d*[1-9]+\d*$)|(^[1-9]+\d*\.\d*$)" G "<0>5.0000</0>" +"(^\d*\.?\d*[1-9]+\d*$)|(^[1-9]+\d*\.\d*$)" G "<0>5000</0>" +"(^\d*\.?\d*[1-9]+\d*$)|(^[1-9]+\d*\.\d*$)" "0" +"(^\d*\.?\d*[1-9]+\d*$)|(^[1-9]+\d*\.\d*$)" "0.0" +"(^\d*\.?\d*[1-9]+\d*$)|(^[1-9]+\d*\.\d*$)" ".0" +"^([A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^$" G "<0>Sacramento</0>" +"^([A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^$" "<0><2>San Francisco</2></0>" +"^([A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^$" "<0><3>San Luis Obispo</3></0>" +"^([A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^$" "SanFrancisco" +"^([A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^$" "SanLuisObispo" +"^([A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^([A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,}\040[A-Z]{1}[a-z]{1,})$|^$" "San francisco" +"^\{?[a-fA-F\d]{8}-([a-fA-F\d]{4}-){3}[a-fA-F\d]{12}\}?$" G "<0>{e02ff0e4-00ad-090A-c030-0d00a0008ba0}</0>" +"^\{?[a-fA-F\d]{8}-([a-fA-F\d]{4}-){3}[a-fA-F\d]{12}\}?$" G "<0>e02ff0e4-00ad-090A-c030-0d00a0008ba0</0>" +"^\{?[a-fA-F\d]{8}-([a-fA-F\d]{4}-){3}[a-fA-F\d]{12}\}?$" "0xe02ff0e400ad090Ac0300d00a0008ba0" +"^\{?[a-fA-F0-9]{8}-([a-fA-F0-9]{4}-){3}[a-fA-F0-9]{12}\}?$" G "<0>{e02ff0e4-00ad-090A-c030-0d00a0008ba0}</0>" +"^\{?[a-fA-F0-9]{8}-([a-fA-F0-9]{4}-){3}[a-fA-F0-9]{12}\}?$" G "<0>e02ff0e4-00ad-090A-c030-0d00a0008ba0</0>" +"^\{?[a-fA-F0-9]{8}-([a-fA-F0-9]{4}-){3}[a-fA-F0-9]{12}\}?$" "0xe02ff0e400ad090Ac0300d00a0008ba0" +"^([a-zA-Z0-9@*#]{8,15})$" G "<0>@12X*567</0>" +"^([a-zA-Z0-9@*#]{8,15})$" G "<0>1#Zv96g@*Yfasd4</0>" +"^([a-zA-Z0-9@*#]{8,15})$" G "<0>#67jhgt@erd</0>" +"^([a-zA-Z0-9@*#]{8,15})$" "$12X*567" +"^([a-zA-Z0-9@*#]{8,15})$" "1#Zv_96" +"^([a-zA-Z0-9@*#]{8,15})$" "+678jhgt@erd" +'(("|\u0027)[a-z0-9\/\.\?\=\&]*(\.htm|\.asp|\.php|\.jsp)[a-z0-9\/\.\?\=\&]*("|\u0027))|(href=*?[a-z0-9\/\.\?\=\&"\u0027]*)' G '<0>href="produktsida.asp?kategori2=218"</0>' +'(("|\u0027)[a-z0-9\/\.\?\=\&]*(\.htm|\.asp|\.php|\.jsp)[a-z0-9\/\.\?\=\&]*("|\u0027))|(href=*?[a-z0-9\/\.\?\=\&"\u0027]*)' G '<0>href="NuclearTesting.htm"</0>' +'(("|\u0027)[a-z0-9\/\.\?\=\&]*(\.htm|\.asp|\.php|\.jsp)[a-z0-9\/\.\?\=\&]*("|\u0027))|(href=*?[a-z0-9\/\.\?\=\&"\u0027]*)' 'U Suck' +"^(((((0[1-9])|(1\d)|(2[0-8]))-((0[1-9])|(1[0-2])))|((31-((0[13578])|(1[02])))|((29|30)-((0[1,3-9])|(1[0-2])))))-((20[0-9][0-9]))|(29-02-20(([02468][048])|([13579][26]))))$" G "<0>05-01-2002</0>" +"^(((((0[1-9])|(1\d)|(2[0-8]))-((0[1-9])|(1[0-2])))|((31-((0[13578])|(1[02])))|((29|30)-((0[1,3-9])|(1[0-2])))))-((20[0-9][0-9]))|(29-02-20(([02468][048])|([13579][26]))))$" G "<0>29-02-2004</0>" +"^(((((0[1-9])|(1\d)|(2[0-8]))-((0[1-9])|(1[0-2])))|((31-((0[13578])|(1[02])))|((29|30)-((0[1,3-9])|(1[0-2])))))-((20[0-9][0-9]))|(29-02-20(([02468][048])|([13579][26]))))$" G "<0>31-12-2002</0>" +"^(((((0[1-9])|(1\d)|(2[0-8]))-((0[1-9])|(1[0-2])))|((31-((0[13578])|(1[02])))|((29|30)-((0[1,3-9])|(1[0-2])))))-((20[0-9][0-9]))|(29-02-20(([02468][048])|([13579][26]))))$" "1-1-02" +"^(((((0[1-9])|(1\d)|(2[0-8]))-((0[1-9])|(1[0-2])))|((31-((0[13578])|(1[02])))|((29|30)-((0[1,3-9])|(1[0-2])))))-((20[0-9][0-9]))|(29-02-20(([02468][048])|([13579][26]))))$" "29-02-2002" +"^(((((0[1-9])|(1\d)|(2[0-8]))-((0[1-9])|(1[0-2])))|((31-((0[13578])|(1[02])))|((29|30)-((0[1,3-9])|(1[0-2])))))-((20[0-9][0-9]))|(29-02-20(([02468][048])|([13579][26]))))$" "31-11-2002" +"^\d*[0-9](|.\d*[0-9]|,\d*[0-9])?$" G "<0>123456.123456</0>" +"^\d*[0-9](|.\d*[0-9]|,\d*[0-9])?$" G "<0>123456,123456</0>" +"^\d*[0-9](|.\d*[0-9]|,\d*[0-9])?$" G "<0>123456</0>" +"^\d*[0-9](|.\d*[0-9]|,\d*[0-9])?$" "123a.123" +"^\d*[0-9](|.\d*[0-9]|,\d*[0-9])?$" "123a,123" +"^\d*[0-9](|.\d*[0-9]|,\d*[0-9])?$" "a" +"^(ac|AC|al|AL|am|AM|ap|AP|ba|BA|ce|CE|df|DF|es|ES|go|GO|ma|MA|mg|MG|ms|MS|mt|MT|pa|PA|pb|PB|pe|PE|pi|PI|pr|PR|rj|RJ|rn|RN|ro|RO|rr|RR|rs|RS|sc|SC|se|SE|sp|SP|to|TO)$" G "<0>AC</0>" +"^(ac|AC|al|AL|am|AM|ap|AP|ba|BA|ce|CE|df|DF|es|ES|go|GO|ma|MA|mg|MG|ms|MS|mt|MT|pa|PA|pb|PB|pe|PE|pi|PI|pr|PR|rj|RJ|rn|RN|ro|RO|rr|RR|rs|RS|sc|SC|se|SE|sp|SP|to|TO)$" G "<0>RJ</0>" +"^(ac|AC|al|AL|am|AM|ap|AP|ba|BA|ce|CE|df|DF|es|ES|go|GO|ma|MA|mg|MG|ms|MS|mt|MT|pa|PA|pb|PB|pe|PE|pi|PI|pr|PR|rj|RJ|rn|RN|ro|RO|rr|RR|rs|RS|sc|SC|se|SE|sp|SP|to|TO)$" G "<0>SP</0>" +"^(ac|AC|al|AL|am|AM|ap|AP|ba|BA|ce|CE|df|DF|es|ES|go|GO|ma|MA|mg|MG|ms|MS|mt|MT|pa|PA|pb|PB|pe|PE|pi|PI|pr|PR|rj|RJ|rn|RN|ro|RO|rr|RR|rs|RS|sc|SC|se|SE|sp|SP|to|TO)$" "XX" +"^(ac|AC|al|AL|am|AM|ap|AP|ba|BA|ce|CE|df|DF|es|ES|go|GO|ma|MA|mg|MG|ms|MS|mt|MT|pa|PA|pb|PB|pe|PE|pi|PI|pr|PR|rj|RJ|rn|RN|ro|RO|rr|RR|rs|RS|sc|SC|se|SE|sp|SP|to|TO)$" "AB" +"^(ac|AC|al|AL|am|AM|ap|AP|ba|BA|ce|CE|df|DF|es|ES|go|GO|ma|MA|mg|MG|ms|MS|mt|MT|pa|PA|pb|PB|pe|PE|pi|PI|pr|PR|rj|RJ|rn|RN|ro|RO|rr|RR|rs|RS|sc|SC|se|SE|sp|SP|to|TO)$" "HJ" +"^[+]?\d*$" G "<0>0123456789</0>" +"^[+]?\d*$" G "<0>1234</0>" +"^[+]?\d*$" G "<0>1</0>" +"^[+]?\d*$" "1.0?&" +"^[+]?\d*$" "a1" +"^[+]?\d*$" "2a-" +#/<[aA][ ]{0,}([a-zA-Z0-9"'_,.:;!?@$\&()%=\u002f ]|[\-]|[ \f]){0,}>((<(([a-zA-Z0-9"'_,.:;!?@$\&()%=\u002f ]|[\-]|[ \f]){0,})>([a-zA-Z0-9"'_,.:;!?@$\&()%=\u002f ]|[\-]|[ \f]){0,})|(([a-zA-Z0-9"'_,.:;!?@$\&()%=\u002f ]|[\-]|[ \f]){0,})){1,}/ G "<0><a href='javascript:functionA();'><i>this text is italicized</i></a></0>" #TODO: Need infinite loop breaking +#/<[aA][ ]{0,}([a-zA-Z0-9"'_,.:;!?@$\&()%=\u002f ]|[\-]|[ \f]){0,}>((<(([a-zA-Z0-9"'_,.:;!?@$\&()%=\u002f ]|[\-]|[ \f]){0,})>([a-zA-Z0-9"'_,.:;!?@$\&()%=\u002f ]|[\-]|[ \f]){0,})|(([a-zA-Z0-9"'_,.:;!?@$\&()%=\u002f ]|[\-]|[ \f]){0,})){1,}/ "<A href='#'><P</A></P>" #TODO: need infinite loop breaking. +"^([0-1]?[0-9]|[2][0-3]):([0-5][0-9])$" G "<0>0:00</0>" +"^([0-1]?[0-9]|[2][0-3]):([0-5][0-9])$" G "<0>23:00</0>" +"^([0-1]?[0-9]|[2][0-3]):([0-5][0-9])$" G "<0>00:59</0>" +"^([0-1]?[0-9]|[2][0-3]):([0-5][0-9])$" "0:0" +"^([0-1]?[0-9]|[2][0-3]):([0-5][0-9])$" "24:00" +"^([0-1]?[0-9]|[2][0-3]):([0-5][0-9])$" "00:60" +"^((0[1-9])|(1[0-2]))\/(\d{2})$" G "<0>11/03</0>" +"^((0[1-9])|(1[0-2]))\/(\d{2})$" G "<0>01/04</0>" +"^((0[1-9])|(1[0-2]))\/(\d{2})$" "13/03" +"^((0[1-9])|(1[0-2]))\/(\d{2})$" "10/2003" +"<script[^>]*>[\w|\t|\r|\W]*</script>" G '<0><script language=javascript>document.write("one");</script></0>' +"<script[^>]*>[\w|\t|\r|\W]*</script>" "--" +"<script[^>]*>[\w|\t|\r|\W]*</script>" "A-Z][a-z]+" +#"<script[^>]*>[\w|\t|\r|\W]*</script>" G "<0>strFirstName</0>" # Test Case damaged? +#"<script[^>]*>[\w|\t|\r|\W]*</script>" G "<0>intAgeInYears</0>" # Test Case damaged? +#"<script[^>]*>[\w|\t|\r|\W]*</script>" G "<0>Where the Wild Things Are</0>" # Test Case damaged? +"<script[^>]*>[\w|\t|\r|\W]*</script>" "123" +"<script[^>]*>[\w|\t|\r|\W]*</script>" "abc" +"<script[^>]*>[\w|\t|\r|\W]*</script>" "this has no caps in it" +"(^-\d*\.?\d*[1-9]+\d*$)|(^-[1-9]+\d*\.\d*$)" G "<0>-0.050</0>" +"(^-\d*\.?\d*[1-9]+\d*$)|(^-[1-9]+\d*\.\d*$)" G "<0>-5.000</0>" +"(^-\d*\.?\d*[1-9]+\d*$)|(^-[1-9]+\d*\.\d*$)" G "<0>-5</0>" +"(^-\d*\.?\d*[1-9]+\d*$)|(^-[1-9]+\d*\.\d*$)" "0" +"(^-\d*\.?\d*[1-9]+\d*$)|(^-[1-9]+\d*\.\d*$)" "0.0" +"(^-\d*\.?\d*[1-9]+\d*$)|(^-[1-9]+\d*\.\d*$)" ".0" +"^([2][0]\d{2}\/([0]\d|[1][0-2])\/([0-2]\d|[3][0-1]))$|^([2][0]\d{2}\/([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\s([0-1]\d|[2][0-3])\:[0-5]\d\:[0-5]\d)$" G "<0>2002/02/03</0>" +"^([2][0]\d{2}\/([0]\d|[1][0-2])\/([0-2]\d|[3][0-1]))$|^([2][0]\d{2}\/([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\s([0-1]\d|[2][0-3])\:[0-5]\d\:[0-5]\d)$" G "<0>2002/02/03 12:12:18</0>" +"^([2][0]\d{2}\/([0]\d|[1][0-2])\/([0-2]\d|[3][0-1]))$|^([2][0]\d{2}\/([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\s([0-1]\d|[2][0-3])\:[0-5]\d\:[0-5]\d)$" "2002/02/36" +"^([2][0]\d{2}\/([0]\d|[1][0-2])\/([0-2]\d|[3][0-1]))$|^([2][0]\d{2}\/([0]\d|[1][0-2])\/([0-2]\d|[3][0-1])\s([0-1]\d|[2][0-3])\:[0-5]\d\:[0-5]\d)$" "02/03/2002" +"^(\d|,)*\.?\d*$" G "<0>1,000</0>" +"^(\d|,)*\.?\d*$" G "<0>3,000.05</0>" +"^(\d|,)*\.?\d*$" G "<0>5,000,000</0>" +"^(\d|,)*\.?\d*$" "abc" +"^(\d|,)*\.?\d*$" "$100,000" +"^(\d|,)*\.?\d*$" "Forty" +"^\d$" G "<0>1</0>" +"^\d$" G "<0>2</0>" +"^\d$" G "<0>3</0>" +"^\d$" "a" +"^\d$" "324" +"^\d$" "num" +"^[0-9]+$" G "<0>1234567890</0>" +"^[0-9]+$" G "<0>1234567890</0>" +"^[0-9]+$" G "<0>1234567890</0>" +"^[0-9]+$" "http://none" +"^[0-9]+$" "http://none" +"^[0-9]+$" "http://none" +"^.{4,8}$" G "<0>asdf</0>" +"^.{4,8}$" G "<0>1234</0>" +"^.{4,8}$" G "<0>asdf1234</0>" +"^.{4,8}$" "asd" +"^.{4,8}$" "123" +"^.{4,8}$" "asdfe12345" +"^[\w\.=-]+@[\w\.-]+\.[\w]{2,3}$" G "<0>a@a.com</0>" +"^[\w\.=-]+@[\w\.-]+\.[\w]{2,3}$" G "<0>a@a.com.au</0>" +"^[\w\.=-]+@[\w\.-]+\.[\w]{2,3}$" G "<0>a@a.au</0>" +"^[\w\.=-]+@[\w\.-]+\.[\w]{2,3}$" "word" +"^[\w\.=-]+@[\w\.-]+\.[\w]{2,3}$" "word@" +"^[\w\.=-]+@[\w\.-]+\.[\w]{2,3}$" "@word" +"^\d{5}-\d{4}$" G "<0>22222-3333</0>" +"^\d{5}-\d{4}$" G "<0>34545-2367</0>" +"^\d{5}-\d{4}$" G "<0>56334-2343</0>" +"^\d{5}-\d{4}$" "123456789" +"^\d{5}-\d{4}$" "A3B 4C5" +"^\d{5}-\d{4}$" "55335" +"(a|b|c).(a.b)*.b+.c" G "<0>autbfc</0>" +"(a|b|c).(a.b)*.b+.c" "attc" +'"((\\")|[^"(\\")])+"' G '<0>"test"</0>' +'"((\\")|[^"(\\")])+"' G '<0>"escape\"quote"</0>' +'"((\\")|[^"(\\")])+"' G '<0>"\\""</0>' +'"((\\")|[^"(\\")])+"' "test" +'"((\\")|[^"(\\")])+"' '"test' +'"((\\")|[^"(\\")])+"' '""test\\"' +"((0[1-9])|(1[02]))/\d{2}" G "<0>01/00</0>" +"((0[1-9])|(1[02]))/\d{2}" G "<0>12/99</0>" +"((0[1-9])|(1[02]))/\d{2}" "13/00" +"((0[1-9])|(1[02]))/\d{2}" "12/AS" +"^[a-zA-Z]$" G "<0>a</0>" +"^[a-zA-Z]$" G "<0>B</0>" +"^[a-zA-Z]$" G "<0>c</0>" +"^[a-zA-Z]$" "0" +"^[a-zA-Z]$" "&" +"^[a-zA-Z]$" "AbC" +"^[a-zA-Z]+$" G "<0>abc</0>" +"^[a-zA-Z]+$" G "<0>ABC</0>" +"^[a-zA-Z]+$" G "<0>aBcDeF</0>" +"^[a-zA-Z]+$" "abc123" +"^[a-zA-Z]+$" "mr." +"^[a-zA-Z]+$" "a word" +"^\s*[a-zA-Z,\p{Zs}]+\s*$" G "<0>Smith, Ed</0>" +"^\s*[a-zA-Z,\p{Zs}]+\s*$" G "<0>Ed Smith</0>" +"^\s*[a-zA-Z,\p{Zs}]+\s*$" G "<0>aBcDeFgH</0>" +"^\s*[a-zA-Z,\p{Zs}]+\s*$" "a123" +"^\s*[a-zA-Z,\p{Zs}]+\s*$" "AB5" +"^\s*[a-zA-Z,\p{Zs}]+\s*$" "Mr. Ed" +"(\w+?@\w+?\u002E.+)" G "<0>bob@vsnl.com</0>" +"(\w+?@\w+?\u002E.+)" "[AABB]" +"^\d+$" G "<0>123</0>" +"^\d+$" G "<0>10</0>" +"^\d+$" G "<0>54</0>" +"^\d+$" "-54" +"^\d+$" "54.234" +"^\d+$" "abc" +"^(\+|-)?\d+$" G "<0>-34</0>" +"^(\+|-)?\d+$" G "<0>34</0>" +"^(\+|-)?\d+$" G "<0>+5</0>" +"^(\+|-)?\d+$" "abc" +"^(\+|-)?\d+$" "3.1415" +"^(\+|-)?\d+$" "-5.3" +"foo" G "<0>foo</0>" +"foo" "bar" +"^[1-5]$" G "<0>1</0>" +"^[1-5]$" G "<0>3</0>" +"^[1-5]$" G "<0>4</0>" +"^[1-5]$" "6" +"^[1-5]$" "23" +"^[1-5]$" "a" +"^[12345]$" G "<0>1</0>" +"^[12345]$" G "<0>2</0>" +"^[12345]$" G "<0>4</0>" +"^[12345]$" "6" +"^[12345]$" "-1" +"^[12345]$" "abc" +"^[\w-\.]+@([\w-]+\.)+[\w-]{2,4}$" G "<0>joe@aol.com</0>" +"^[\w-\.]+@([\w-]+\.)+[\w-]{2,4}$" G "<0>joe@wrox.co.uk</0>" +"^[\w-\.]+@([\w-]+\.)+[\w-]{2,4}$" G "<0>joe@domain.info</0>" +"^[\w-\.]+@([\w-]+\.)+[\w-]{2,4}$" "a@b" +"^[\w-\.]+@([\w-]+\.)+[\w-]{2,4}$" "notanemail" +"^[\w-\.]+@([\w-]+\.)+[\w-]{2,4}$" "joe@@." +"^\w+@[a-zA-Z_]+?\.[a-zA-Z]{2,3}$" G "<0>joe@aol.com</0>" +"^\w+@[a-zA-Z_]+?\.[a-zA-Z]{2,3}$" G "<0>ssmith@aspalliance.com</0>" +"^\w+@[a-zA-Z_]+?\.[a-zA-Z]{2,3}$" G "<0>a@b.cc</0>" +"^\w+@[a-zA-Z_]+?\.[a-zA-Z]{2,3}$" "joe@123aspx.com" +"^\w+@[a-zA-Z_]+?\.[a-zA-Z]{2,3}$" "joe@web.info" +"^\w+@[a-zA-Z_]+?\.[a-zA-Z]{2,3}$" "joe@company.co.uk" +"[\w-]+@([\w-]+\.)+[\w-]+" G "<0>joe@aol.com</0>" +"[\w-]+@([\w-]+\.)+[\w-]+" G "<0>a@b.c</0>" +"[\w-]+@([\w-]+\.)+[\w-]+" "asdf" +"[\w-]+@([\w-]+\.)+[\w-]+" "1234" +"\d{4}-?\d{4}-?\d{4}-?\d{4}" G "<0>1234-1234-1234-1234</0>" +"\d{4}-?\d{4}-?\d{4}-?\d{4}" G "<0>1234123412341234</0>" +"\d{4}-?\d{4}-?\d{4}-?\d{4}" "1234123412345" +"^\d{5}$" G "<0>33333</0>" +"^\d{5}$" G "<0>55555</0>" +"^\d{5}$" G "<0>23445</0>" +"^\d{5}$" "abcd" +"^\d{5}$" "1324" +"^\d{5}$" "as;lkjdf" +"(\w+)\s+\1" G "<0>hubba hubba</0>" +"(\w+)\s+\1" G "<0>mandate dated</0>" +"(\w+)\s+\1" G "<0>an annual</0>" +"(\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" + |