1 /* 2 * Copyright (C) 2007 The Android Open Source Project 3 * 4 * Licensed under the Apache License, Version 2.0 (the "License"); 5 * you may not use this file except in compliance with the License. 6 * You may obtain a copy of the License at 7 * 8 * http://www.apache.org/licenses/LICENSE-2.0 9 * 10 * Unless required by applicable law or agreed to in writing, software 11 * distributed under the License is distributed on an "AS IS" BASIS, 12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 13 * See the License for the specific language governing permissions and 14 * limitations under the License. 15 */ 16 17 package android.util; 18 19 import java.util.regex.Matcher; 20 import java.util.regex.Pattern; 21 22 /** 23 * Commonly used regular expression patterns. 24 */ 25 public class Patterns { 26 /** 27 * Regular expression to match all IANA top-level domains. 28 * List accurate as of 2011/07/18. List taken from: 29 * http://data.iana.org/TLD/tlds-alpha-by-domain.txt 30 * This pattern is auto-generated by frameworks/ex/common/tools/make-iana-tld-pattern.py 31 * 32 * @deprecated Due to the recent profileration of gTLDs, this API is 33 * expected to become out-of-date very quickly. Therefore it is now 34 * deprecated. 35 */ 36 @Deprecated 37 public static final String TOP_LEVEL_DOMAIN_STR = 38 "((aero|arpa|asia|a[cdefgilmnoqrstuwxz])" 39 + "|(biz|b[abdefghijmnorstvwyz])" 40 + "|(cat|com|coop|c[acdfghiklmnoruvxyz])" 41 + "|d[ejkmoz]" 42 + "|(edu|e[cegrstu])" 43 + "|f[ijkmor]" 44 + "|(gov|g[abdefghilmnpqrstuwy])" 45 + "|h[kmnrtu]" 46 + "|(info|int|i[delmnoqrst])" 47 + "|(jobs|j[emop])" 48 + "|k[eghimnprwyz]" 49 + "|l[abcikrstuvy]" 50 + "|(mil|mobi|museum|m[acdeghklmnopqrstuvwxyz])" 51 + "|(name|net|n[acefgilopruz])" 52 + "|(org|om)" 53 + "|(pro|p[aefghklmnrstwy])" 54 + "|qa" 55 + "|r[eosuw]" 56 + "|s[abcdeghijklmnortuvyz]" 57 + "|(tel|travel|t[cdfghjklmnoprtvwz])" 58 + "|u[agksyz]" 59 + "|v[aceginu]" 60 + "|w[fs]" 61 + "|(\u03b4\u03bf\u03ba\u03b9\u03bc\u03ae|\u0438\u0441\u043f\u044b\u0442\u0430\u043d\u0438\u0435|\u0440\u0444|\u0441\u0440\u0431|\u05d8\u05e2\u05e1\u05d8|\u0622\u0632\u0645\u0627\u06cc\u0634\u06cc|\u0625\u062e\u062a\u0628\u0627\u0631|\u0627\u0644\u0627\u0631\u062f\u0646|\u0627\u0644\u062c\u0632\u0627\u0626\u0631|\u0627\u0644\u0633\u0639\u0648\u062f\u064a\u0629|\u0627\u0644\u0645\u063a\u0631\u0628|\u0627\u0645\u0627\u0631\u0627\u062a|\u0628\u06be\u0627\u0631\u062a|\u062a\u0648\u0646\u0633|\u0633\u0648\u0631\u064a\u0629|\u0641\u0644\u0633\u0637\u064a\u0646|\u0642\u0637\u0631|\u0645\u0635\u0631|\u092a\u0930\u0940\u0915\u094d\u0937\u093e|\u092d\u093e\u0930\u0924|\u09ad\u09be\u09b0\u09a4|\u0a2d\u0a3e\u0a30\u0a24|\u0aad\u0abe\u0ab0\u0aa4|\u0b87\u0ba8\u0bcd\u0ba4\u0bbf\u0baf\u0bbe|\u0b87\u0bb2\u0b99\u0bcd\u0b95\u0bc8|\u0b9a\u0bbf\u0b99\u0bcd\u0b95\u0baa\u0bcd\u0baa\u0bc2\u0bb0\u0bcd|\u0baa\u0bb0\u0bbf\u0b9f\u0bcd\u0b9a\u0bc8|\u0c2d\u0c3e\u0c30\u0c24\u0c4d|\u0dbd\u0d82\u0d9a\u0dcf|\u0e44\u0e17\u0e22|\u30c6\u30b9\u30c8|\u4e2d\u56fd|\u4e2d\u570b|\u53f0\u6e7e|\u53f0\u7063|\u65b0\u52a0\u5761|\u6d4b\u8bd5|\u6e2c\u8a66|\u9999\u6e2f|\ud14c\uc2a4\ud2b8|\ud55c\uad6d|xn\\-\\-0zwm56d|xn\\-\\-11b5bs3a9aj6g|xn\\-\\-3e0b707e|xn\\-\\-45brj9c|xn\\-\\-80akhbyknj4f|xn\\-\\-90a3ac|xn\\-\\-9t4b11yi5a|xn\\-\\-clchc0ea0b2g2a9gcd|xn\\-\\-deba0ad|xn\\-\\-fiqs8s|xn\\-\\-fiqz9s|xn\\-\\-fpcrj9c3d|xn\\-\\-fzc2c9e2c|xn\\-\\-g6w251d|xn\\-\\-gecrj9c|xn\\-\\-h2brj9c|xn\\-\\-hgbk6aj7f53bba|xn\\-\\-hlcj6aya9esc7a|xn\\-\\-j6w193g|xn\\-\\-jxalpdlp|xn\\-\\-kgbechtv|xn\\-\\-kprw13d|xn\\-\\-kpry57d|xn\\-\\-lgbbat1ad8j|xn\\-\\-mgbaam7a8h|xn\\-\\-mgbayh7gpa|xn\\-\\-mgbbh1a71e|xn\\-\\-mgbc0a9azcg|xn\\-\\-mgberp4a5d4ar|xn\\-\\-o3cw4h|xn\\-\\-ogbpf8fl|xn\\-\\-p1ai|xn\\-\\-pgbs0dh|xn\\-\\-s9brj9c|xn\\-\\-wgbh1c|xn\\-\\-wgbl6a|xn\\-\\-xkc2al3hye2a|xn\\-\\-xkc2dl3a5ee0h|xn\\-\\-yfro4i67o|xn\\-\\-ygbi2ammx|xn\\-\\-zckzah|xxx)" 62 + "|y[et]" 63 + "|z[amw])"; 64 65 /** 66 * Regular expression pattern to match all IANA top-level domains. 67 * @deprecated This API is deprecated. See {@link #TOP_LEVEL_DOMAIN_STR}. 68 */ 69 @Deprecated 70 public static final Pattern TOP_LEVEL_DOMAIN = 71 Pattern.compile(TOP_LEVEL_DOMAIN_STR); 72 73 /** 74 * Regular expression to match all IANA top-level domains for WEB_URL. 75 * List accurate as of 2011/07/18. List taken from: 76 * http://data.iana.org/TLD/tlds-alpha-by-domain.txt 77 * This pattern is auto-generated by frameworks/ex/common/tools/make-iana-tld-pattern.py 78 * 79 * @deprecated This API is deprecated. See {@link #TOP_LEVEL_DOMAIN_STR}. 80 */ 81 @Deprecated 82 public static final String TOP_LEVEL_DOMAIN_STR_FOR_WEB_URL = 83 "(?:" 84 + "(?:aero|arpa|asia|a[cdefgilmnoqrstuwxz])" 85 + "|(?:biz|b[abdefghijmnorstvwyz])" 86 + "|(?:cat|com|coop|c[acdfghiklmnoruvxyz])" 87 + "|d[ejkmoz]" 88 + "|(?:edu|e[cegrstu])" 89 + "|f[ijkmor]" 90 + "|(?:gov|g[abdefghilmnpqrstuwy])" 91 + "|h[kmnrtu]" 92 + "|(?:info|int|i[delmnoqrst])" 93 + "|(?:jobs|j[emop])" 94 + "|k[eghimnprwyz]" 95 + "|l[abcikrstuvy]" 96 + "|(?:mil|mobi|museum|m[acdeghklmnopqrstuvwxyz])" 97 + "|(?:name|net|n[acefgilopruz])" 98 + "|(?:org|om)" 99 + "|(?:pro|p[aefghklmnrstwy])" 100 + "|qa" 101 + "|r[eosuw]" 102 + "|s[abcdeghijklmnortuvyz]" 103 + "|(?:tel|travel|t[cdfghjklmnoprtvwz])" 104 + "|u[agksyz]" 105 + "|v[aceginu]" 106 + "|w[fs]" 107 + "|(?:\u03b4\u03bf\u03ba\u03b9\u03bc\u03ae|\u0438\u0441\u043f\u044b\u0442\u0430\u043d\u0438\u0435|\u0440\u0444|\u0441\u0440\u0431|\u05d8\u05e2\u05e1\u05d8|\u0622\u0632\u0645\u0627\u06cc\u0634\u06cc|\u0625\u062e\u062a\u0628\u0627\u0631|\u0627\u0644\u0627\u0631\u062f\u0646|\u0627\u0644\u062c\u0632\u0627\u0626\u0631|\u0627\u0644\u0633\u0639\u0648\u062f\u064a\u0629|\u0627\u0644\u0645\u063a\u0631\u0628|\u0627\u0645\u0627\u0631\u0627\u062a|\u0628\u06be\u0627\u0631\u062a|\u062a\u0648\u0646\u0633|\u0633\u0648\u0631\u064a\u0629|\u0641\u0644\u0633\u0637\u064a\u0646|\u0642\u0637\u0631|\u0645\u0635\u0631|\u092a\u0930\u0940\u0915\u094d\u0937\u093e|\u092d\u093e\u0930\u0924|\u09ad\u09be\u09b0\u09a4|\u0a2d\u0a3e\u0a30\u0a24|\u0aad\u0abe\u0ab0\u0aa4|\u0b87\u0ba8\u0bcd\u0ba4\u0bbf\u0baf\u0bbe|\u0b87\u0bb2\u0b99\u0bcd\u0b95\u0bc8|\u0b9a\u0bbf\u0b99\u0bcd\u0b95\u0baa\u0bcd\u0baa\u0bc2\u0bb0\u0bcd|\u0baa\u0bb0\u0bbf\u0b9f\u0bcd\u0b9a\u0bc8|\u0c2d\u0c3e\u0c30\u0c24\u0c4d|\u0dbd\u0d82\u0d9a\u0dcf|\u0e44\u0e17\u0e22|\u30c6\u30b9\u30c8|\u4e2d\u56fd|\u4e2d\u570b|\u53f0\u6e7e|\u53f0\u7063|\u65b0\u52a0\u5761|\u6d4b\u8bd5|\u6e2c\u8a66|\u9999\u6e2f|\ud14c\uc2a4\ud2b8|\ud55c\uad6d|xn\\-\\-0zwm56d|xn\\-\\-11b5bs3a9aj6g|xn\\-\\-3e0b707e|xn\\-\\-45brj9c|xn\\-\\-80akhbyknj4f|xn\\-\\-90a3ac|xn\\-\\-9t4b11yi5a|xn\\-\\-clchc0ea0b2g2a9gcd|xn\\-\\-deba0ad|xn\\-\\-fiqs8s|xn\\-\\-fiqz9s|xn\\-\\-fpcrj9c3d|xn\\-\\-fzc2c9e2c|xn\\-\\-g6w251d|xn\\-\\-gecrj9c|xn\\-\\-h2brj9c|xn\\-\\-hgbk6aj7f53bba|xn\\-\\-hlcj6aya9esc7a|xn\\-\\-j6w193g|xn\\-\\-jxalpdlp|xn\\-\\-kgbechtv|xn\\-\\-kprw13d|xn\\-\\-kpry57d|xn\\-\\-lgbbat1ad8j|xn\\-\\-mgbaam7a8h|xn\\-\\-mgbayh7gpa|xn\\-\\-mgbbh1a71e|xn\\-\\-mgbc0a9azcg|xn\\-\\-mgberp4a5d4ar|xn\\-\\-o3cw4h|xn\\-\\-ogbpf8fl|xn\\-\\-p1ai|xn\\-\\-pgbs0dh|xn\\-\\-s9brj9c|xn\\-\\-wgbh1c|xn\\-\\-wgbl6a|xn\\-\\-xkc2al3hye2a|xn\\-\\-xkc2dl3a5ee0h|xn\\-\\-yfro4i67o|xn\\-\\-ygbi2ammx|xn\\-\\-zckzah|xxx)" 108 + "|y[et]" 109 + "|z[amw]))"; 110 111 /** 112 * Regular expression to match all IANA top-level domains. 113 * 114 * List accurate as of 2015/11/24. List taken from: 115 * http://data.iana.org/TLD/tlds-alpha-by-domain.txt 116 * This pattern is auto-generated by frameworks/ex/common/tools/make-iana-tld-pattern.py 117 * 118 * @hide 119 */ 120 static final String IANA_TOP_LEVEL_DOMAINS = 121 "(?:" 122 + "(?:aaa|aarp|abb|abbott|abogado|academy|accenture|accountant|accountants|aco|active" 123 + "|actor|ads|adult|aeg|aero|afl|agency|aig|airforce|airtel|allfinanz|alsace|amica|amsterdam" 124 + "|android|apartments|app|apple|aquarelle|aramco|archi|army|arpa|arte|asia|associates" 125 + "|attorney|auction|audio|auto|autos|axa|azure|a[cdefgilmoqrstuwxz])" 126 + "|(?:band|bank|bar|barcelona|barclaycard|barclays|bargains|bauhaus|bayern|bbc|bbva" 127 + "|bcn|beats|beer|bentley|berlin|best|bet|bharti|bible|bid|bike|bing|bingo|bio|biz|black" 128 + "|blackfriday|bloomberg|blue|bms|bmw|bnl|bnpparibas|boats|bom|bond|boo|boots|boutique" 129 + "|bradesco|bridgestone|broadway|broker|brother|brussels|budapest|build|builders|business" 130 + "|buzz|bzh|b[abdefghijmnorstvwyz])" 131 + "|(?:cab|cafe|cal|camera|camp|cancerresearch|canon|capetown|capital|car|caravan|cards" 132 + "|care|career|careers|cars|cartier|casa|cash|casino|cat|catering|cba|cbn|ceb|center|ceo" 133 + "|cern|cfa|cfd|chanel|channel|chat|cheap|chloe|christmas|chrome|church|cipriani|cisco" 134 + "|citic|city|cityeats|claims|cleaning|click|clinic|clothing|cloud|club|clubmed|coach" 135 + "|codes|coffee|college|cologne|com|commbank|community|company|computer|comsec|condos" 136 + "|construction|consulting|contractors|cooking|cool|coop|corsica|country|coupons|courses" 137 + "|credit|creditcard|creditunion|cricket|crown|crs|cruises|csc|cuisinella|cymru|cyou|c[acdfghiklmnoruvwxyz])" 138 + "|(?:dabur|dad|dance|date|dating|datsun|day|dclk|deals|degree|delivery|dell|delta" 139 + "|democrat|dental|dentist|desi|design|dev|diamonds|diet|digital|direct|directory|discount" 140 + "|dnp|docs|dog|doha|domains|doosan|download|drive|durban|dvag|d[ejkmoz])" 141 + "|(?:earth|eat|edu|education|email|emerck|energy|engineer|engineering|enterprises" 142 + "|epson|equipment|erni|esq|estate|eurovision|eus|events|everbank|exchange|expert|exposed" 143 + "|express|e[cegrstu])" 144 + "|(?:fage|fail|fairwinds|faith|family|fan|fans|farm|fashion|feedback|ferrero|film" 145 + "|final|finance|financial|firmdale|fish|fishing|fit|fitness|flights|florist|flowers|flsmidth" 146 + "|fly|foo|football|forex|forsale|forum|foundation|frl|frogans|fund|furniture|futbol|fyi" 147 + "|f[ijkmor])" 148 + "|(?:gal|gallery|game|garden|gbiz|gdn|gea|gent|genting|ggee|gift|gifts|gives|giving" 149 + "|glass|gle|global|globo|gmail|gmo|gmx|gold|goldpoint|golf|goo|goog|google|gop|gov|grainger" 150 + "|graphics|gratis|green|gripe|group|gucci|guge|guide|guitars|guru|g[abdefghilmnpqrstuwy])" 151 + "|(?:hamburg|hangout|haus|healthcare|help|here|hermes|hiphop|hitachi|hiv|hockey|holdings" 152 + "|holiday|homedepot|homes|honda|horse|host|hosting|hoteles|hotmail|house|how|hsbc|hyundai" 153 + "|h[kmnrtu])" 154 + "|(?:ibm|icbc|ice|icu|ifm|iinet|immo|immobilien|industries|infiniti|info|ing|ink|institute" 155 + "|insure|int|international|investments|ipiranga|irish|ist|istanbul|itau|iwc|i[delmnoqrst])" 156 + "|(?:jaguar|java|jcb|jetzt|jewelry|jlc|jll|jobs|joburg|jprs|juegos|j[emop])" 157 + "|(?:kaufen|kddi|kia|kim|kinder|kitchen|kiwi|koeln|komatsu|krd|kred|kyoto|k[eghimnprwyz])" 158 + "|(?:lacaixa|lancaster|land|landrover|lasalle|lat|latrobe|law|lawyer|lds|lease|leclerc" 159 + "|legal|lexus|lgbt|liaison|lidl|life|lifestyle|lighting|limited|limo|linde|link|live" 160 + "|lixil|loan|loans|lol|london|lotte|lotto|love|ltd|ltda|lupin|luxe|luxury|l[abcikrstuvy])" 161 + "|(?:madrid|maif|maison|man|management|mango|market|marketing|markets|marriott|mba" 162 + "|media|meet|melbourne|meme|memorial|men|menu|meo|miami|microsoft|mil|mini|mma|mobi|moda" 163 + "|moe|moi|mom|monash|money|montblanc|mormon|mortgage|moscow|motorcycles|mov|movie|movistar" 164 + "|mtn|mtpc|mtr|museum|mutuelle|m[acdeghklmnopqrstuvwxyz])" 165 + "|(?:nadex|nagoya|name|navy|nec|net|netbank|network|neustar|new|news|nexus|ngo|nhk" 166 + "|nico|ninja|nissan|nokia|nra|nrw|ntt|nyc|n[acefgilopruz])" 167 + "|(?:obi|office|okinawa|omega|one|ong|onl|online|ooo|oracle|orange|org|organic|osaka" 168 + "|otsuka|ovh|om)" 169 + "|(?:page|panerai|paris|partners|parts|party|pet|pharmacy|philips|photo|photography" 170 + "|photos|physio|piaget|pics|pictet|pictures|ping|pink|pizza|place|play|playstation|plumbing" 171 + "|plus|pohl|poker|porn|post|praxi|press|pro|prod|productions|prof|properties|property" 172 + "|protection|pub|p[aefghklmnrstwy])" 173 + "|(?:qpon|quebec|qa)" 174 + "|(?:racing|realtor|realty|recipes|red|redstone|rehab|reise|reisen|reit|ren|rent|rentals" 175 + "|repair|report|republican|rest|restaurant|review|reviews|rich|ricoh|rio|rip|rocher|rocks" 176 + "|rodeo|rsvp|ruhr|run|rwe|ryukyu|r[eosuw])" 177 + "|(?:saarland|sakura|sale|samsung|sandvik|sandvikcoromant|sanofi|sap|sapo|sarl|saxo" 178 + "|sbs|sca|scb|schmidt|scholarships|school|schule|schwarz|science|scor|scot|seat|security" 179 + "|seek|sener|services|seven|sew|sex|sexy|shiksha|shoes|show|shriram|singles|site|ski" 180 + "|sky|skype|sncf|soccer|social|software|sohu|solar|solutions|sony|soy|space|spiegel|spreadbetting" 181 + "|srl|stada|starhub|statoil|stc|stcgroup|stockholm|studio|study|style|sucks|supplies" 182 + "|supply|support|surf|surgery|suzuki|swatch|swiss|sydney|systems|s[abcdeghijklmnortuvxyz])" 183 + "|(?:tab|taipei|tatamotors|tatar|tattoo|tax|taxi|team|tech|technology|tel|telefonica" 184 + "|temasek|tennis|thd|theater|theatre|tickets|tienda|tips|tires|tirol|today|tokyo|tools" 185 + "|top|toray|toshiba|tours|town|toyota|toys|trade|trading|training|travel|trust|tui|t[cdfghjklmnortvwz])" 186 + "|(?:ubs|university|uno|uol|u[agksyz])" 187 + "|(?:vacations|vana|vegas|ventures|versicherung|vet|viajes|video|villas|vin|virgin" 188 + "|vision|vista|vistaprint|viva|vlaanderen|vodka|vote|voting|voto|voyage|v[aceginu])" 189 + "|(?:wales|walter|wang|watch|webcam|website|wed|wedding|weir|whoswho|wien|wiki|williamhill" 190 + "|win|windows|wine|wme|work|works|world|wtc|wtf|w[fs])" 191 + "|(?:\u03b5\u03bb|\u0431\u0435\u043b|\u0434\u0435\u0442\u0438|\u043a\u043e\u043c|\u043c\u043a\u0434" 192 + "|\u043c\u043e\u043d|\u043c\u043e\u0441\u043a\u0432\u0430|\u043e\u043d\u043b\u0430\u0439\u043d" 193 + "|\u043e\u0440\u0433|\u0440\u0443\u0441|\u0440\u0444|\u0441\u0430\u0439\u0442|\u0441\u0440\u0431" 194 + "|\u0443\u043a\u0440|\u049b\u0430\u0437|\u0570\u0561\u0575|\u05e7\u05d5\u05dd|\u0627\u0631\u0627\u0645\u0643\u0648" 195 + "|\u0627\u0644\u0627\u0631\u062f\u0646|\u0627\u0644\u062c\u0632\u0627\u0626\u0631|\u0627\u0644\u0633\u0639\u0648\u062f\u064a\u0629" 196 + "|\u0627\u0644\u0645\u063a\u0631\u0628|\u0627\u0645\u0627\u0631\u0627\u062a|\u0627\u06cc\u0631\u0627\u0646" 197 + "|\u0628\u0627\u0632\u0627\u0631|\u0628\u06be\u0627\u0631\u062a|\u062a\u0648\u0646\u0633" 198 + "|\u0633\u0648\u062f\u0627\u0646|\u0633\u0648\u0631\u064a\u0629|\u0634\u0628\u0643\u0629" 199 + "|\u0639\u0631\u0627\u0642|\u0639\u0645\u0627\u0646|\u0641\u0644\u0633\u0637\u064a\u0646" 200 + "|\u0642\u0637\u0631|\u0643\u0648\u0645|\u0645\u0635\u0631|\u0645\u0644\u064a\u0633\u064a\u0627" 201 + "|\u0645\u0648\u0642\u0639|\u0915\u0949\u092e|\u0928\u0947\u091f|\u092d\u093e\u0930\u0924" 202 + "|\u0938\u0902\u0917\u0920\u0928|\u09ad\u09be\u09b0\u09a4|\u0a2d\u0a3e\u0a30\u0a24|\u0aad\u0abe\u0ab0\u0aa4" 203 + "|\u0b87\u0ba8\u0bcd\u0ba4\u0bbf\u0baf\u0bbe|\u0b87\u0bb2\u0b99\u0bcd\u0b95\u0bc8|\u0b9a\u0bbf\u0b99\u0bcd\u0b95\u0baa\u0bcd\u0baa\u0bc2\u0bb0\u0bcd" 204 + "|\u0c2d\u0c3e\u0c30\u0c24\u0c4d|\u0dbd\u0d82\u0d9a\u0dcf|\u0e04\u0e2d\u0e21|\u0e44\u0e17\u0e22" 205 + "|\u10d2\u10d4|\u307f\u3093\u306a|\u30b0\u30fc\u30b0\u30eb|\u30b3\u30e0|\u4e16\u754c" 206 + "|\u4e2d\u4fe1|\u4e2d\u56fd|\u4e2d\u570b|\u4e2d\u6587\u7f51|\u4f01\u4e1a|\u4f5b\u5c71" 207 + "|\u4fe1\u606f|\u5065\u5eb7|\u516b\u5366|\u516c\u53f8|\u516c\u76ca|\u53f0\u6e7e|\u53f0\u7063" 208 + "|\u5546\u57ce|\u5546\u5e97|\u5546\u6807|\u5728\u7ebf|\u5927\u62ff|\u5a31\u4e50|\u5de5\u884c" 209 + "|\u5e7f\u4e1c|\u6148\u5584|\u6211\u7231\u4f60|\u624b\u673a|\u653f\u52a1|\u653f\u5e9c" 210 + "|\u65b0\u52a0\u5761|\u65b0\u95fb|\u65f6\u5c1a|\u673a\u6784|\u6de1\u9a6c\u9521|\u6e38\u620f" 211 + "|\u70b9\u770b|\u79fb\u52a8|\u7ec4\u7ec7\u673a\u6784|\u7f51\u5740|\u7f51\u5e97|\u7f51\u7edc" 212 + "|\u8c37\u6b4c|\u96c6\u56e2|\u98de\u5229\u6d66|\u9910\u5385|\u9999\u6e2f|\ub2f7\ub137" 213 + "|\ub2f7\ucef4|\uc0bc\uc131|\ud55c\uad6d|xbox" 214 + "|xerox|xin|xn\\-\\-11b4c3d|xn\\-\\-1qqw23a|xn\\-\\-30rr7y|xn\\-\\-3bst00m|xn\\-\\-3ds443g" 215 + "|xn\\-\\-3e0b707e|xn\\-\\-3pxu8k|xn\\-\\-42c2d9a|xn\\-\\-45brj9c|xn\\-\\-45q11c|xn\\-\\-4gbrim" 216 + "|xn\\-\\-55qw42g|xn\\-\\-55qx5d|xn\\-\\-6frz82g|xn\\-\\-6qq986b3xl|xn\\-\\-80adxhks" 217 + "|xn\\-\\-80ao21a|xn\\-\\-80asehdb|xn\\-\\-80aswg|xn\\-\\-90a3ac|xn\\-\\-90ais|xn\\-\\-9dbq2a" 218 + "|xn\\-\\-9et52u|xn\\-\\-b4w605ferd|xn\\-\\-c1avg|xn\\-\\-c2br7g|xn\\-\\-cg4bki|xn\\-\\-clchc0ea0b2g2a9gcd" 219 + "|xn\\-\\-czr694b|xn\\-\\-czrs0t|xn\\-\\-czru2d|xn\\-\\-d1acj3b|xn\\-\\-d1alf|xn\\-\\-efvy88h" 220 + "|xn\\-\\-estv75g|xn\\-\\-fhbei|xn\\-\\-fiq228c5hs|xn\\-\\-fiq64b|xn\\-\\-fiqs8s|xn\\-\\-fiqz9s" 221 + "|xn\\-\\-fjq720a|xn\\-\\-flw351e|xn\\-\\-fpcrj9c3d|xn\\-\\-fzc2c9e2c|xn\\-\\-gecrj9c" 222 + "|xn\\-\\-h2brj9c|xn\\-\\-hxt814e|xn\\-\\-i1b6b1a6a2e|xn\\-\\-imr513n|xn\\-\\-io0a7i" 223 + "|xn\\-\\-j1aef|xn\\-\\-j1amh|xn\\-\\-j6w193g|xn\\-\\-kcrx77d1x4a|xn\\-\\-kprw13d|xn\\-\\-kpry57d" 224 + "|xn\\-\\-kput3i|xn\\-\\-l1acc|xn\\-\\-lgbbat1ad8j|xn\\-\\-mgb9awbf|xn\\-\\-mgba3a3ejt" 225 + "|xn\\-\\-mgba3a4f16a|xn\\-\\-mgbaam7a8h|xn\\-\\-mgbab2bd|xn\\-\\-mgbayh7gpa|xn\\-\\-mgbbh1a71e" 226 + "|xn\\-\\-mgbc0a9azcg|xn\\-\\-mgberp4a5d4ar|xn\\-\\-mgbpl2fh|xn\\-\\-mgbtx2b|xn\\-\\-mgbx4cd0ab" 227 + "|xn\\-\\-mk1bu44c|xn\\-\\-mxtq1m|xn\\-\\-ngbc5azd|xn\\-\\-node|xn\\-\\-nqv7f|xn\\-\\-nqv7fs00ema" 228 + "|xn\\-\\-nyqy26a|xn\\-\\-o3cw4h|xn\\-\\-ogbpf8fl|xn\\-\\-p1acf|xn\\-\\-p1ai|xn\\-\\-pgbs0dh" 229 + "|xn\\-\\-pssy2u|xn\\-\\-q9jyb4c|xn\\-\\-qcka1pmc|xn\\-\\-qxam|xn\\-\\-rhqv96g|xn\\-\\-s9brj9c" 230 + "|xn\\-\\-ses554g|xn\\-\\-t60b56a|xn\\-\\-tckwe|xn\\-\\-unup4y|xn\\-\\-vermgensberater\\-ctb" 231 + "|xn\\-\\-vermgensberatung\\-pwb|xn\\-\\-vhquv|xn\\-\\-vuq861b|xn\\-\\-wgbh1c|xn\\-\\-wgbl6a" 232 + "|xn\\-\\-xhq521b|xn\\-\\-xkc2al3hye2a|xn\\-\\-xkc2dl3a5ee0h|xn\\-\\-y9a3aq|xn\\-\\-yfro4i67o" 233 + "|xn\\-\\-ygbi2ammx|xn\\-\\-zfr164b|xperia|xxx|xyz)" 234 + "|(?:yachts|yamaxun|yandex|yodobashi|yoga|yokohama|youtube|y[et])" 235 + "|(?:zara|zip|zone|zuerich|z[amw]))"; 236 237 /** 238 * Kept for backward compatibility reasons. 239 * 240 * @deprecated Deprecated since it does not include all IRI characters defined in RFC 3987 241 */ 242 @Deprecated 243 public static final String GOOD_IRI_CHAR = 244 "a-zA-Z0-9\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF"; 245 246 private static final String IP_ADDRESS_STRING = 247 "((25[0-5]|2[0-4][0-9]|[0-1][0-9]{2}|[1-9][0-9]|[1-9])\\.(25[0-5]|2[0-4]" 248 + "[0-9]|[0-1][0-9]{2}|[1-9][0-9]|[1-9]|0)\\.(25[0-5]|2[0-4][0-9]|[0-1]" 249 + "[0-9]{2}|[1-9][0-9]|[1-9]|0)\\.(25[0-5]|2[0-4][0-9]|[0-1][0-9]{2}" 250 + "|[1-9][0-9]|[0-9]))"; 251 public static final Pattern IP_ADDRESS = Pattern.compile(IP_ADDRESS_STRING); 252 253 /** 254 * Valid UCS characters defined in RFC 3987. Excludes space characters. 255 */ 256 private static final String UCS_CHAR = "[" + 257 "\u00A0-\uD7FF" + 258 "\uF900-\uFDCF" + 259 "\uFDF0-\uFFEF" + 260 "\uD800\uDC00-\uD83F\uDFFD" + 261 "\uD840\uDC00-\uD87F\uDFFD" + 262 "\uD880\uDC00-\uD8BF\uDFFD" + 263 "\uD8C0\uDC00-\uD8FF\uDFFD" + 264 "\uD900\uDC00-\uD93F\uDFFD" + 265 "\uD940\uDC00-\uD97F\uDFFD" + 266 "\uD980\uDC00-\uD9BF\uDFFD" + 267 "\uD9C0\uDC00-\uD9FF\uDFFD" + 268 "\uDA00\uDC00-\uDA3F\uDFFD" + 269 "\uDA40\uDC00-\uDA7F\uDFFD" + 270 "\uDA80\uDC00-\uDABF\uDFFD" + 271 "\uDAC0\uDC00-\uDAFF\uDFFD" + 272 "\uDB00\uDC00-\uDB3F\uDFFD" + 273 "\uDB44\uDC00-\uDB7F\uDFFD" + 274 "&&[^\u00A0[\u2000-\u200A]\u2028\u2029\u202F\u3000]]"; 275 276 /** 277 * Valid characters for IRI label defined in RFC 3987. 278 */ 279 private static final String LABEL_CHAR = "a-zA-Z0-9" + UCS_CHAR; 280 281 /** 282 * Valid characters for IRI TLD defined in RFC 3987. 283 */ 284 private static final String TLD_CHAR = "a-zA-Z" + UCS_CHAR; 285 286 /** 287 * RFC 1035 Section 2.3.4 limits the labels to a maximum 63 octets. 288 */ 289 private static final String IRI_LABEL = 290 "[" + LABEL_CHAR + "](?:[" + LABEL_CHAR + "_\\-]{0,61}[" + LABEL_CHAR + "]){0,1}"; 291 292 /** 293 * RFC 3492 references RFC 1034 and limits Punycode algorithm output to 63 characters. 294 */ 295 private static final String PUNYCODE_TLD = "xn\\-\\-[\\w\\-]{0,58}\\w"; 296 297 private static final String TLD = "(" + PUNYCODE_TLD + "|" + "[" + TLD_CHAR + "]{2,63}" +")"; 298 299 private static final String HOST_NAME = "(" + IRI_LABEL + "\\.)+" + TLD; 300 301 private static final String DOMAIN_NAME_STR = "(" + HOST_NAME + "|" + IP_ADDRESS_STRING + ")"; 302 public static final Pattern DOMAIN_NAME = Pattern.compile(DOMAIN_NAME_STR); 303 304 private static final String PROTOCOL = "(?i:http|https|rtsp)://"; 305 306 /* A word boundary or end of input. This is to stop foo.sure from matching as foo.su */ 307 private static final String WORD_BOUNDARY = "(?:\\b|$|^)"; 308 309 private static final String USER_INFO = "(?:[a-zA-Z0-9\\$\\-\\_\\.\\+\\!\\*\\'\\(\\)" 310 + "\\,\\;\\?\\&\\=]|(?:\\%[a-fA-F0-9]{2})){1,64}(?:\\:(?:[a-zA-Z0-9\\$\\-\\_" 311 + "\\.\\+\\!\\*\\'\\(\\)\\,\\;\\?\\&\\=]|(?:\\%[a-fA-F0-9]{2})){1,25})?\\@"; 312 313 private static final String PORT_NUMBER = "\\:\\d{1,5}"; 314 315 private static final String PATH_AND_QUERY = "[/\\?](?:(?:[" + LABEL_CHAR 316 + ";/\\?:@&=#~" // plus optional query params 317 + "\\-\\.\\+!\\*'\\(\\),_\\$])|(?:%[a-fA-F0-9]{2}))*"; 318 319 /** 320 * Regular expression pattern to match most part of RFC 3987 321 * Internationalized URLs, aka IRIs. 322 */ 323 public static final Pattern WEB_URL = Pattern.compile("(" 324 + "(" 325 + "(?:" + PROTOCOL + "(?:" + USER_INFO + ")?" + ")?" 326 + "(?:" + DOMAIN_NAME_STR + ")" 327 + "(?:" + PORT_NUMBER + ")?" 328 + ")" 329 + "(" + PATH_AND_QUERY + ")?" 330 + WORD_BOUNDARY 331 + ")"); 332 333 /** 334 * Regular expression that matches known TLDs and punycode TLDs 335 */ 336 private static final String STRICT_TLD = "(?:" + 337 IANA_TOP_LEVEL_DOMAINS + "|" + PUNYCODE_TLD + ")"; 338 339 /** 340 * Regular expression that matches host names using {@link #STRICT_TLD} 341 */ 342 private static final String STRICT_HOST_NAME = "(?:(?:" + IRI_LABEL + "\\.)+" 343 + STRICT_TLD + ")"; 344 345 /** 346 * Regular expression that matches domain names using either {@link #STRICT_HOST_NAME} or 347 * {@link #IP_ADDRESS} 348 */ 349 private static final String STRICT_DOMAIN_NAME = "(?:" + STRICT_HOST_NAME + "|" 350 + IP_ADDRESS_STRING + ")"; 351 352 /** 353 * Regular expression that matches domain names without a TLD 354 */ 355 private static final String RELAXED_DOMAIN_NAME = 356 "(?:" + "(?:" + IRI_LABEL + "(?:\\.(?=\\S))" +"?)+" + "|" + IP_ADDRESS_STRING + ")"; 357 358 /** 359 * Regular expression to match strings that do not start with a supported protocol. The TLDs 360 * are expected to be one of the known TLDs. 361 */ 362 private static final String WEB_URL_WITHOUT_PROTOCOL = "(" 363 + WORD_BOUNDARY 364 + "(?<!:\\/\\/)" 365 + "(" 366 + "(?:" + STRICT_DOMAIN_NAME + ")" 367 + "(?:" + PORT_NUMBER + ")?" 368 + ")" 369 + "(?:" + PATH_AND_QUERY + ")?" 370 + WORD_BOUNDARY 371 + ")"; 372 373 /** 374 * Regular expression to match strings that start with a supported protocol. Rules for domain 375 * names and TLDs are more relaxed. TLDs are optional. 376 */ 377 private static final String WEB_URL_WITH_PROTOCOL = "(" 378 + WORD_BOUNDARY 379 + "(?:" 380 + "(?:" + PROTOCOL + "(?:" + USER_INFO + ")?" + ")" 381 + "(?:" + RELAXED_DOMAIN_NAME + ")?" 382 + "(?:" + PORT_NUMBER + ")?" 383 + ")" 384 + "(?:" + PATH_AND_QUERY + ")?" 385 + WORD_BOUNDARY 386 + ")"; 387 388 /** 389 * Regular expression pattern to match IRIs. If a string starts with http(s):// the expression 390 * tries to match the URL structure with a relaxed rule for TLDs. If the string does not start 391 * with http(s):// the TLDs are expected to be one of the known TLDs. 392 * 393 * @hide 394 */ 395 public static final Pattern AUTOLINK_WEB_URL = Pattern.compile( 396 "(" + WEB_URL_WITH_PROTOCOL + "|" + WEB_URL_WITHOUT_PROTOCOL + ")"); 397 398 /** 399 * Regular expression for valid email characters. Does not include some of the valid characters 400 * defined in RFC5321: #&~!^`{}/=$*?| 401 */ 402 private static final String EMAIL_CHAR = LABEL_CHAR + "\\+\\-_%'"; 403 404 /** 405 * Regular expression for local part of an email address. RFC5321 section 4.5.3.1.1 limits 406 * the local part to be at most 64 octets. 407 */ 408 private static final String EMAIL_ADDRESS_LOCAL_PART = 409 "[" + EMAIL_CHAR + "]" + "(?:[" + EMAIL_CHAR + "\\.]{0,62}[" + EMAIL_CHAR + "])?"; 410 411 /** 412 * Regular expression for the domain part of an email address. RFC5321 section 4.5.3.1.2 limits 413 * the domain to be at most 255 octets. 414 */ 415 private static final String EMAIL_ADDRESS_DOMAIN = 416 "(?=.{1,255}(?:\\s|$|^))" + HOST_NAME; 417 418 /** 419 * Regular expression pattern to match email addresses. It excludes double quoted local parts 420 * and the special characters #&~!^`{}/=$*?| that are included in RFC5321. 421 * @hide 422 */ 423 public static final Pattern AUTOLINK_EMAIL_ADDRESS = Pattern.compile("(" + WORD_BOUNDARY + 424 "(?:" + EMAIL_ADDRESS_LOCAL_PART + "@" + EMAIL_ADDRESS_DOMAIN + ")" + 425 WORD_BOUNDARY + ")" 426 ); 427 428 public static final Pattern EMAIL_ADDRESS 429 = Pattern.compile( 430 "[a-zA-Z0-9\\+\\.\\_\\%\\-\\+]{1,256}" + 431 "\\@" + 432 "[a-zA-Z0-9][a-zA-Z0-9\\-]{0,64}" + 433 "(" + 434 "\\." + 435 "[a-zA-Z0-9][a-zA-Z0-9\\-]{0,25}" + 436 ")+" 437 ); 438 439 /** 440 * This pattern is intended for searching for things that look like they 441 * might be phone numbers in arbitrary text, not for validating whether 442 * something is in fact a phone number. It will miss many things that 443 * are legitimate phone numbers. 444 * 445 * <p> The pattern matches the following: 446 * <ul> 447 * <li>Optionally, a + sign followed immediately by one or more digits. Spaces, dots, or dashes 448 * may follow. 449 * <li>Optionally, sets of digits in parentheses, separated by spaces, dots, or dashes. 450 * <li>A string starting and ending with a digit, containing digits, spaces, dots, and/or dashes. 451 * </ul> 452 */ 453 public static final Pattern PHONE 454 = Pattern.compile( // sdd = space, dot, or dash 455 "(\\+[0-9]+[\\- \\.]*)?" // +<digits><sdd>* 456 + "(\\([0-9]+\\)[\\- \\.]*)?" // (<digits>)<sdd>* 457 + "([0-9][0-9\\- \\.]+[0-9])"); // <digit><digit|sdd>+<digit> 458 459 /** 460 * Convenience method to take all of the non-null matching groups in a 461 * regex Matcher and return them as a concatenated string. 462 * 463 * @param matcher The Matcher object from which grouped text will 464 * be extracted 465 * 466 * @return A String comprising all of the non-null matched 467 * groups concatenated together 468 */ 469 public static final String concatGroups(Matcher matcher) { 470 StringBuilder b = new StringBuilder(); 471 final int numGroups = matcher.groupCount(); 472 473 for (int i = 1; i <= numGroups; i++) { 474 String s = matcher.group(i); 475 476 if (s != null) { 477 b.append(s); 478 } 479 } 480 481 return b.toString(); 482 } 483 484 /** 485 * Convenience method to return only the digits and plus signs 486 * in the matching string. 487 * 488 * @param matcher The Matcher object from which digits and plus will 489 * be extracted 490 * 491 * @return A String comprising all of the digits and plus in 492 * the match 493 */ 494 public static final String digitsAndPlusOnly(Matcher matcher) { 495 StringBuilder buffer = new StringBuilder(); 496 String matchingRegion = matcher.group(); 497 498 for (int i = 0, size = matchingRegion.length(); i < size; i++) { 499 char character = matchingRegion.charAt(i); 500 501 if (character == '+' || Character.isDigit(character)) { 502 buffer.append(character); 503 } 504 } 505 return buffer.toString(); 506 } 507 508 /** 509 * Do not create this static utility class. 510 */ 511 private Patterns() {} 512 } 513
