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 @android.ravenwood.annotation.RavenwoodKeepWholeClass 26 public class Patterns { 27 /** 28 * Regular expression to match all IANA top-level domains. 29 * List accurate as of 2011/07/18. List taken from: 30 * http://data.iana.org/TLD/tlds-alpha-by-domain.txt 31 * This pattern is auto-generated by frameworks/ex/common/tools/make-iana-tld-pattern.py 32 * 33 * @deprecated Due to the recent profileration of gTLDs, this API is 34 * expected to become out-of-date very quickly. Therefore it is now 35 * deprecated. 36 */ 37 @Deprecated 38 public static final String TOP_LEVEL_DOMAIN_STR = 39 "((aero|arpa|asia|a[cdefgilmnoqrstuwxz])" 40 + "|(biz|b[abdefghijmnorstvwyz])" 41 + "|(cat|com|coop|c[acdfghiklmnoruvxyz])" 42 + "|d[ejkmoz]" 43 + "|(edu|e[cegrstu])" 44 + "|f[ijkmor]" 45 + "|(gov|g[abdefghilmnpqrstuwy])" 46 + "|h[kmnrtu]" 47 + "|(info|int|i[delmnoqrst])" 48 + "|(jobs|j[emop])" 49 + "|k[eghimnprwyz]" 50 + "|l[abcikrstuvy]" 51 + "|(mil|mobi|museum|m[acdeghklmnopqrstuvwxyz])" 52 + "|(name|net|n[acefgilopruz])" 53 + "|(org|om)" 54 + "|(pro|p[aefghklmnrstwy])" 55 + "|qa" 56 + "|r[eosuw]" 57 + "|s[abcdeghijklmnortuvyz]" 58 + "|(tel|travel|t[cdfghjklmnoprtvwz])" 59 + "|u[agksyz]" 60 + "|v[aceginu]" 61 + "|w[fs]" 62 + "|(\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)" 63 + "|y[et]" 64 + "|z[amw])"; 65 66 /** 67 * Regular expression pattern to match all IANA top-level domains. 68 * @deprecated This API is deprecated. See {@link #TOP_LEVEL_DOMAIN_STR}. 69 */ 70 @Deprecated 71 public static final Pattern TOP_LEVEL_DOMAIN = 72 Pattern.compile(TOP_LEVEL_DOMAIN_STR); 73 74 /** 75 * Regular expression to match all IANA top-level domains for WEB_URL. 76 * List accurate as of 2011/07/18. List taken from: 77 * http://data.iana.org/TLD/tlds-alpha-by-domain.txt 78 * This pattern is auto-generated by frameworks/ex/common/tools/make-iana-tld-pattern.py 79 * 80 * @deprecated This API is deprecated. See {@link #TOP_LEVEL_DOMAIN_STR}. 81 */ 82 @Deprecated 83 public static final String TOP_LEVEL_DOMAIN_STR_FOR_WEB_URL = 84 "(?:" 85 + "(?:aero|arpa|asia|a[cdefgilmnoqrstuwxz])" 86 + "|(?:biz|b[abdefghijmnorstvwyz])" 87 + "|(?:cat|com|coop|c[acdfghiklmnoruvxyz])" 88 + "|d[ejkmoz]" 89 + "|(?:edu|e[cegrstu])" 90 + "|f[ijkmor]" 91 + "|(?:gov|g[abdefghilmnpqrstuwy])" 92 + "|h[kmnrtu]" 93 + "|(?:info|int|i[delmnoqrst])" 94 + "|(?:jobs|j[emop])" 95 + "|k[eghimnprwyz]" 96 + "|l[abcikrstuvy]" 97 + "|(?:mil|mobi|museum|m[acdeghklmnopqrstuvwxyz])" 98 + "|(?:name|net|n[acefgilopruz])" 99 + "|(?:org|om)" 100 + "|(?:pro|p[aefghklmnrstwy])" 101 + "|qa" 102 + "|r[eosuw]" 103 + "|s[abcdeghijklmnortuvyz]" 104 + "|(?:tel|travel|t[cdfghjklmnoprtvwz])" 105 + "|u[agksyz]" 106 + "|v[aceginu]" 107 + "|w[fs]" 108 + "|(?:\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)" 109 + "|y[et]" 110 + "|z[amw]))"; 111 112 /** 113 * Regular expression to match all IANA top-level domains. 114 * 115 * List accurate as of 2023/09/11. List taken from: 116 * http://data.iana.org/TLD/tlds-alpha-by-domain.txt 117 * This pattern is auto-generated by frameworks/ex/common/tools/make-iana-tld-pattern.py 118 * 119 * @hide 120 */ 121 static final String IANA_TOP_LEVEL_DOMAINS = 122 "(?:" 123 + "(?:aaa|aarp|abb|abbott|abbvie|abc|able|abogado|abudhabi|academy|accenture|accountant" 124 + "|accountants|aco|actor|ads|adult|aeg|aero|aetna|afl|africa|agakhan|agency|aig|airbus" 125 + "|airforce|airtel|akdn|alibaba|alipay|allfinanz|allstate|ally|alsace|alstom|amazon|americanexpress" 126 + "|americanfamily|amex|amfam|amica|amsterdam|analytics|android|anquan|anz|aol|apartments" 127 + "|app|apple|aquarelle|arab|aramco|archi|army|arpa|art|arte|asda|asia|associates|athleta" 128 + "|attorney|auction|audi|audible|audio|auspost|author|auto|autos|avianca|aws|axa|azure" 129 + "|a[cdefgilmoqrstuwxz])" 130 + "|(?:baby|baidu|banamex|bananarepublic|band|bank|bar|barcelona|barclaycard|barclays" 131 + "|barefoot|bargains|baseball|basketball|bauhaus|bayern|bbc|bbt|bbva|bcg|bcn|beats|beauty" 132 + "|beer|bentley|berlin|best|bestbuy|bet|bharti|bible|bid|bike|bing|bingo|bio|biz|black" 133 + "|blackfriday|blockbuster|blog|bloomberg|blue|bms|bmw|bnpparibas|boats|boehringer|bofa" 134 + "|bom|bond|boo|book|booking|bosch|bostik|boston|bot|boutique|box|bradesco|bridgestone" 135 + "|broadway|broker|brother|brussels|build|builders|business|buy|buzz|bzh|b[abdefghijmnorstvwyz])" 136 + "|(?:cab|cafe|cal|call|calvinklein|cam|camera|camp|canon|capetown|capital|capitalone" 137 + "|car|caravan|cards|care|career|careers|cars|casa|case|cash|casino|cat|catering|catholic" 138 + "|cba|cbn|cbre|cbs|center|ceo|cern|cfa|cfd|chanel|channel|charity|chase|chat|cheap|chintai" 139 + "|christmas|chrome|church|cipriani|circle|cisco|citadel|citi|citic|city|cityeats|claims" 140 + "|cleaning|click|clinic|clinique|clothing|cloud|club|clubmed|coach|codes|coffee|college" 141 + "|cologne|com|comcast|commbank|community|company|compare|computer|comsec|condos|construction" 142 + "|consulting|contact|contractors|cooking|cool|coop|corsica|country|coupon|coupons|courses" 143 + "|cpa|credit|creditcard|creditunion|cricket|crown|crs|cruise|cruises|cuisinella|cymru" 144 + "|cyou|c[acdfghiklmnoruvwxyz])" 145 + "|(?:dabur|dad|dance|data|date|dating|datsun|day|dclk|dds|deal|dealer|deals|degree" 146 + "|delivery|dell|deloitte|delta|democrat|dental|dentist|desi|design|dev|dhl|diamonds|diet" 147 + "|digital|direct|directory|discount|discover|dish|diy|dnp|docs|doctor|dog|domains|dot" 148 + "|download|drive|dtv|dubai|dunlop|dupont|durban|dvag|dvr|d[ejkmoz])" 149 + "|(?:earth|eat|eco|edeka|edu|education|email|emerck|energy|engineer|engineering|enterprises" 150 + "|epson|equipment|ericsson|erni|esq|estate|etisalat|eurovision|eus|events|exchange|expert" 151 + "|exposed|express|extraspace|e[cegrstu])" 152 + "|(?:fage|fail|fairwinds|faith|family|fan|fans|farm|farmers|fashion|fast|fedex|feedback" 153 + "|ferrari|ferrero|fidelity|fido|film|final|finance|financial|fire|firestone|firmdale" 154 + "|fish|fishing|fit|fitness|flickr|flights|flir|florist|flowers|fly|foo|food|football" 155 + "|ford|forex|forsale|forum|foundation|fox|free|fresenius|frl|frogans|frontdoor|frontier" 156 + "|ftr|fujitsu|fun|fund|furniture|futbol|fyi|f[ijkmor])" 157 + "|(?:gal|gallery|gallo|gallup|game|games|gap|garden|gay|gbiz|gdn|gea|gent|genting" 158 + "|george|ggee|gift|gifts|gives|giving|glass|gle|global|globo|gmail|gmbh|gmo|gmx|godaddy" 159 + "|gold|goldpoint|golf|goo|goodyear|goog|google|gop|got|gov|grainger|graphics|gratis|green" 160 + "|gripe|grocery|group|guardian|gucci|guge|guide|guitars|guru|g[abdefghilmnpqrstuwy])" 161 + "|(?:hair|hamburg|hangout|haus|hbo|hdfc|hdfcbank|health|healthcare|help|helsinki|here" 162 + "|hermes|hiphop|hisamitsu|hitachi|hiv|hkt|hockey|holdings|holiday|homedepot|homegoods" 163 + "|homes|homesense|honda|horse|hospital|host|hosting|hot|hotels|hotmail|house|how|hsbc" 164 + "|hughes|hyatt|hyundai|h[kmnrtu])" 165 + "|(?:ibm|icbc|ice|icu|ieee|ifm|ikano|imamat|imdb|immo|immobilien|inc|industries|infiniti" 166 + "|info|ing|ink|institute|insurance|insure|int|international|intuit|investments|ipiranga" 167 + "|irish|ismaili|ist|istanbul|itau|itv|i[delmnoqrst])" 168 + "|(?:jaguar|java|jcb|jeep|jetzt|jewelry|jio|jll|jmp|jnj|jobs|joburg|jot|joy|jpmorgan" 169 + "|jprs|juegos|juniper|j[emop])" 170 + "|(?:kaufen|kddi|kerryhotels|kerrylogistics|kerryproperties|kfh|kia|kids|kim|kinder" 171 + "|kindle|kitchen|kiwi|koeln|komatsu|kosher|kpmg|kpn|krd|kred|kuokgroup|kyoto|k[eghimnprwyz])" 172 + "|(?:lacaixa|lamborghini|lamer|lancaster|land|landrover|lanxess|lasalle|lat|latino" 173 + "|latrobe|law|lawyer|lds|lease|leclerc|lefrak|legal|lego|lexus|lgbt|lidl|life|lifeinsurance" 174 + "|lifestyle|lighting|like|lilly|limited|limo|lincoln|link|lipsy|live|living|llc|llp|loan" 175 + "|loans|locker|locus|lol|london|lotte|lotto|love|lpl|lplfinancial|ltd|ltda|lundbeck|luxe" 176 + "|luxury|l[abcikrstuvy])" 177 + "|(?:madrid|maif|maison|makeup|man|management|mango|map|market|marketing|markets|marriott" 178 + "|marshalls|mattel|mba|mckinsey|med|media|meet|melbourne|meme|memorial|men|menu|merckmsd" 179 + "|miami|microsoft|mil|mini|mint|mit|mitsubishi|mlb|mls|mma|mobi|mobile|moda|moe|moi|mom" 180 + "|monash|money|monster|mormon|mortgage|moscow|moto|motorcycles|mov|movie|msd|mtn|mtr" 181 + "|museum|music|m[acdeghklmnopqrstuvwxyz])" 182 + "|(?:nab|nagoya|name|natura|navy|nba|nec|net|netbank|netflix|network|neustar|new|news" 183 + "|next|nextdirect|nexus|nfl|ngo|nhk|nico|nike|nikon|ninja|nissan|nissay|nokia|norton" 184 + "|now|nowruz|nowtv|nra|nrw|ntt|nyc|n[acefgilopruz])" 185 + "|(?:obi|observer|office|okinawa|olayan|olayangroup|oldnavy|ollo|omega|one|ong|onl" 186 + "|online|ooo|open|oracle|orange|org|organic|origins|osaka|otsuka|ott|ovh|om)" 187 + "|(?:page|panasonic|paris|pars|partners|parts|party|pay|pccw|pet|pfizer|pharmacy|phd" 188 + "|philips|phone|photo|photography|photos|physio|pics|pictet|pictures|pid|pin|ping|pink" 189 + "|pioneer|pizza|place|play|playstation|plumbing|plus|pnc|pohl|poker|politie|porn|post" 190 + "|pramerica|praxi|press|prime|pro|prod|productions|prof|progressive|promo|properties" 191 + "|property|protection|pru|prudential|pub|pwc|p[aefghklmnrstwy])" 192 + "|(?:qpon|quebec|quest|qa)" 193 + "|(?:racing|radio|read|realestate|realtor|realty|recipes|red|redstone|redumbrella" 194 + "|rehab|reise|reisen|reit|reliance|ren|rent|rentals|repair|report|republican|rest|restaurant" 195 + "|review|reviews|rexroth|rich|richardli|ricoh|ril|rio|rip|rocher|rocks|rodeo|rogers|room" 196 + "|rsvp|rugby|ruhr|run|rwe|ryukyu|r[eosuw])" 197 + "|(?:saarland|safe|safety|sakura|sale|salon|samsclub|samsung|sandvik|sandvikcoromant" 198 + "|sanofi|sap|sarl|sas|save|saxo|sbi|sbs|sca|scb|schaeffler|schmidt|scholarships|school" 199 + "|schule|schwarz|science|scot|search|seat|secure|security|seek|select|sener|services" 200 + "|seven|sew|sex|sexy|sfr|shangrila|sharp|shaw|shell|shia|shiksha|shoes|shop|shopping" 201 + "|shouji|show|showtime|silk|sina|singles|site|ski|skin|sky|skype|sling|smart|smile|sncf" 202 + "|soccer|social|softbank|software|sohu|solar|solutions|song|sony|soy|spa|space|sport" 203 + "|spot|srl|stada|staples|star|statebank|statefarm|stc|stcgroup|stockholm|storage|store" 204 + "|stream|studio|study|style|sucks|supplies|supply|support|surf|surgery|suzuki|swatch" 205 + "|swiss|sydney|systems|s[abcdeghijklmnorstuvxyz])" 206 + "|(?:tab|taipei|talk|taobao|target|tatamotors|tatar|tattoo|tax|taxi|tci|tdk|team|tech" 207 + "|technology|tel|temasek|tennis|teva|thd|theater|theatre|tiaa|tickets|tienda|tips|tires" 208 + "|tirol|tjmaxx|tjx|tkmaxx|tmall|today|tokyo|tools|top|toray|toshiba|total|tours|town" 209 + "|toyota|toys|trade|trading|training|travel|travelers|travelersinsurance|trust|trv|tube" 210 + "|tui|tunes|tushu|tvs|t[cdfghjklmnortvwz])" 211 + "|(?:ubank|ubs|unicom|university|uno|uol|ups|u[agksyz])" 212 + "|(?:vacations|vana|vanguard|vegas|ventures|verisign|versicherung|vet|viajes|video" 213 + "|vig|viking|villas|vin|vip|virgin|visa|vision|viva|vivo|vlaanderen|vodka|volkswagen" 214 + "|volvo|vote|voting|voto|voyage|v[aceginu])" 215 + "|(?:wales|walmart|walter|wang|wanggou|watch|watches|weather|weatherchannel|webcam" 216 + "|weber|website|wed|wedding|weibo|weir|whoswho|wien|wiki|williamhill|win|windows|wine" 217 + "|winners|wme|wolterskluwer|woodside|work|works|world|wow|wtc|wtf|w[fs])" 218 + "|(?:\u03b5\u03bb|\u03b5\u03c5|\u0431\u0433|\u0431\u0435\u043b|\u0434\u0435\u0442\u0438" 219 + "|\u0435\u044e|\u043a\u0430\u0442\u043e\u043b\u0438\u043a|\u043a\u043e\u043c|\u043c\u043a\u0434" 220 + "|\u043c\u043e\u043d|\u043c\u043e\u0441\u043a\u0432\u0430|\u043e\u043d\u043b\u0430\u0439\u043d" 221 + "|\u043e\u0440\u0433|\u0440\u0443\u0441|\u0440\u0444|\u0441\u0430\u0439\u0442|\u0441\u0440\u0431" 222 + "|\u0443\u043a\u0440|\u049b\u0430\u0437|\u0570\u0561\u0575|\u05d9\u05e9\u05e8\u05d0\u05dc" 223 + "|\u05e7\u05d5\u05dd|\u0627\u0628\u0648\u0638\u0628\u064a|\u0627\u062a\u0635\u0627\u0644\u0627\u062a" 224 + "|\u0627\u0631\u0627\u0645\u0643\u0648|\u0627\u0644\u0627\u0631\u062f\u0646|\u0627\u0644\u0628\u062d\u0631\u064a\u0646" 225 + "|\u0627\u0644\u062c\u0632\u0627\u0626\u0631|\u0627\u0644\u0633\u0639\u0648\u062f\u064a\u0629" 226 + "|\u0627\u0644\u0639\u0644\u064a\u0627\u0646|\u0627\u0644\u0645\u063a\u0631\u0628|\u0627\u0645\u0627\u0631\u0627\u062a" 227 + "|\u0627\u06cc\u0631\u0627\u0646|\u0628\u0627\u0631\u062a|\u0628\u0627\u0632\u0627\u0631" 228 + "|\u0628\u064a\u062a\u0643|\u0628\u06be\u0627\u0631\u062a|\u062a\u0648\u0646\u0633|\u0633\u0648\u062f\u0627\u0646" 229 + "|\u0633\u0648\u0631\u064a\u0629|\u0634\u0628\u0643\u0629|\u0639\u0631\u0627\u0642|\u0639\u0631\u0628" 230 + "|\u0639\u0645\u0627\u0646|\u0641\u0644\u0633\u0637\u064a\u0646|\u0642\u0637\u0631|\u0643\u0627\u062b\u0648\u0644\u064a\u0643" 231 + "|\u0643\u0648\u0645|\u0645\u0635\u0631|\u0645\u0644\u064a\u0633\u064a\u0627|\u0645\u0648\u0631\u064a\u062a\u0627\u0646\u064a\u0627" 232 + "|\u0645\u0648\u0642\u0639|\u0647\u0645\u0631\u0627\u0647|\u067e\u0627\u06a9\u0633\u062a\u0627\u0646" 233 + "|\u0680\u0627\u0631\u062a|\u0915\u0949\u092e|\u0928\u0947\u091f|\u092d\u093e\u0930\u0924" 234 + "|\u092d\u093e\u0930\u0924\u092e\u094d|\u092d\u093e\u0930\u094b\u0924|\u0938\u0902\u0917\u0920\u0928" 235 + "|\u09ac\u09be\u0982\u09b2\u09be|\u09ad\u09be\u09b0\u09a4|\u09ad\u09be\u09f0\u09a4|\u0a2d\u0a3e\u0a30\u0a24" 236 + "|\u0aad\u0abe\u0ab0\u0aa4|\u0b2d\u0b3e\u0b30\u0b24|\u0b87\u0ba8\u0bcd\u0ba4\u0bbf\u0baf\u0bbe" 237 + "|\u0b87\u0bb2\u0b99\u0bcd\u0b95\u0bc8|\u0b9a\u0bbf\u0b99\u0bcd\u0b95\u0baa\u0bcd\u0baa\u0bc2\u0bb0\u0bcd" 238 + "|\u0c2d\u0c3e\u0c30\u0c24\u0c4d|\u0cad\u0cbe\u0cb0\u0ca4|\u0d2d\u0d3e\u0d30\u0d24\u0d02" 239 + "|\u0dbd\u0d82\u0d9a\u0dcf|\u0e04\u0e2d\u0e21|\u0e44\u0e17\u0e22|\u0ea5\u0eb2\u0ea7|\u10d2\u10d4" 240 + "|\u307f\u3093\u306a|\u30a2\u30de\u30be\u30f3|\u30af\u30e9\u30a6\u30c9|\u30b0\u30fc\u30b0\u30eb" 241 + "|\u30b3\u30e0|\u30b9\u30c8\u30a2|\u30bb\u30fc\u30eb|\u30d5\u30a1\u30c3\u30b7\u30e7\u30f3" 242 + "|\u30dd\u30a4\u30f3\u30c8|\u4e16\u754c|\u4e2d\u4fe1|\u4e2d\u56fd|\u4e2d\u570b|\u4e2d\u6587\u7f51" 243 + "|\u4e9a\u9a6c\u900a|\u4f01\u4e1a|\u4f5b\u5c71|\u4fe1\u606f|\u5065\u5eb7|\u516b\u5366" 244 + "|\u516c\u53f8|\u516c\u76ca|\u53f0\u6e7e|\u53f0\u7063|\u5546\u57ce|\u5546\u5e97|\u5546\u6807" 245 + "|\u5609\u91cc|\u5609\u91cc\u5927\u9152\u5e97|\u5728\u7ebf|\u5927\u62ff|\u5929\u4e3b\u6559" 246 + "|\u5a31\u4e50|\u5bb6\u96fb|\u5e7f\u4e1c|\u5fae\u535a|\u6148\u5584|\u6211\u7231\u4f60" 247 + "|\u624b\u673a|\u62db\u8058|\u653f\u52a1|\u653f\u5e9c|\u65b0\u52a0\u5761|\u65b0\u95fb" 248 + "|\u65f6\u5c1a|\u66f8\u7c4d|\u673a\u6784|\u6de1\u9a6c\u9521|\u6e38\u620f|\u6fb3\u9580" 249 + "|\u70b9\u770b|\u79fb\u52a8|\u7ec4\u7ec7\u673a\u6784|\u7f51\u5740|\u7f51\u5e97|\u7f51\u7ad9" 250 + "|\u7f51\u7edc|\u8054\u901a|\u8c37\u6b4c|\u8d2d\u7269|\u901a\u8ca9|\u96c6\u56e2|\u96fb\u8a0a\u76c8\u79d1" 251 + "|\u98de\u5229\u6d66|\u98df\u54c1|\u9910\u5385|\u9999\u683c\u91cc\u62c9|\u9999\u6e2f" 252 + "|\ub2f7\ub137|\ub2f7\ucef4|\uc0bc\uc131|\ud55c\uad6d" 253 + "|xbox|xerox|xfinity|xihuan|xin|xn\\-\\-11b4c3d|xn\\-\\-1ck2e1b|xn\\-\\-1qqw23a|xn\\-\\-2scrj9c" 254 + "|xn\\-\\-30rr7y|xn\\-\\-3bst00m|xn\\-\\-3ds443g|xn\\-\\-3e0b707e|xn\\-\\-3hcrj9c|xn\\-\\-3pxu8k" 255 + "|xn\\-\\-42c2d9a|xn\\-\\-45br5cyl|xn\\-\\-45brj9c|xn\\-\\-45q11c|xn\\-\\-4dbrk0ce|xn\\-\\-4gbrim" 256 + "|xn\\-\\-54b7fta0cc|xn\\-\\-55qw42g|xn\\-\\-55qx5d|xn\\-\\-5su34j936bgsg|xn\\-\\-5tzm5g" 257 + "|xn\\-\\-6frz82g|xn\\-\\-6qq986b3xl|xn\\-\\-80adxhks|xn\\-\\-80ao21a|xn\\-\\-80aqecdr1a" 258 + "|xn\\-\\-80asehdb|xn\\-\\-80aswg|xn\\-\\-8y0a063a|xn\\-\\-90a3ac|xn\\-\\-90ae|xn\\-\\-90ais" 259 + "|xn\\-\\-9dbq2a|xn\\-\\-9et52u|xn\\-\\-9krt00a|xn\\-\\-b4w605ferd|xn\\-\\-bck1b9a5dre4c" 260 + "|xn\\-\\-c1avg|xn\\-\\-c2br7g|xn\\-\\-cck2b3b|xn\\-\\-cckwcxetd|xn\\-\\-cg4bki|xn\\-\\-clchc0ea0b2g2a9gcd" 261 + "|xn\\-\\-czr694b|xn\\-\\-czrs0t|xn\\-\\-czru2d|xn\\-\\-d1acj3b|xn\\-\\-d1alf|xn\\-\\-e1a4c" 262 + "|xn\\-\\-eckvdtc9d|xn\\-\\-efvy88h|xn\\-\\-fct429k|xn\\-\\-fhbei|xn\\-\\-fiq228c5hs" 263 + "|xn\\-\\-fiq64b|xn\\-\\-fiqs8s|xn\\-\\-fiqz9s|xn\\-\\-fjq720a|xn\\-\\-flw351e|xn\\-\\-fpcrj9c3d" 264 + "|xn\\-\\-fzc2c9e2c|xn\\-\\-fzys8d69uvgm|xn\\-\\-g2xx48c|xn\\-\\-gckr3f0f|xn\\-\\-gecrj9c" 265 + "|xn\\-\\-gk3at1e|xn\\-\\-h2breg3eve|xn\\-\\-h2brj9c|xn\\-\\-h2brj9c8c|xn\\-\\-hxt814e" 266 + "|xn\\-\\-i1b6b1a6a2e|xn\\-\\-imr513n|xn\\-\\-io0a7i|xn\\-\\-j1aef|xn\\-\\-j1amh|xn\\-\\-j6w193g" 267 + "|xn\\-\\-jlq480n2rg|xn\\-\\-jvr189m|xn\\-\\-kcrx77d1x4a|xn\\-\\-kprw13d|xn\\-\\-kpry57d" 268 + "|xn\\-\\-kput3i|xn\\-\\-l1acc|xn\\-\\-lgbbat1ad8j|xn\\-\\-mgb9awbf|xn\\-\\-mgba3a3ejt" 269 + "|xn\\-\\-mgba3a4f16a|xn\\-\\-mgba7c0bbn0a|xn\\-\\-mgbaakc7dvf|xn\\-\\-mgbaam7a8h|xn\\-\\-mgbab2bd" 270 + "|xn\\-\\-mgbah1a3hjkrd|xn\\-\\-mgbai9azgqp6j|xn\\-\\-mgbayh7gpa|xn\\-\\-mgbbh1a|xn\\-\\-mgbbh1a71e" 271 + "|xn\\-\\-mgbc0a9azcg|xn\\-\\-mgbca7dzdo|xn\\-\\-mgbcpq6gpa1a|xn\\-\\-mgberp4a5d4ar|xn\\-\\-mgbgu82a" 272 + "|xn\\-\\-mgbi4ecexp|xn\\-\\-mgbpl2fh|xn\\-\\-mgbt3dhd|xn\\-\\-mgbtx2b|xn\\-\\-mgbx4cd0ab" 273 + "|xn\\-\\-mix891f|xn\\-\\-mk1bu44c|xn\\-\\-mxtq1m|xn\\-\\-ngbc5azd|xn\\-\\-ngbe9e0a|xn\\-\\-ngbrx" 274 + "|xn\\-\\-node|xn\\-\\-nqv7f|xn\\-\\-nqv7fs00ema|xn\\-\\-nyqy26a|xn\\-\\-o3cw4h|xn\\-\\-ogbpf8fl" 275 + "|xn\\-\\-otu796d|xn\\-\\-p1acf|xn\\-\\-p1ai|xn\\-\\-pgbs0dh|xn\\-\\-pssy2u|xn\\-\\-q7ce6a" 276 + "|xn\\-\\-q9jyb4c|xn\\-\\-qcka1pmc|xn\\-\\-qxa6a|xn\\-\\-qxam|xn\\-\\-rhqv96g|xn\\-\\-rovu88b" 277 + "|xn\\-\\-rvc1e0am3e|xn\\-\\-s9brj9c|xn\\-\\-ses554g|xn\\-\\-t60b56a|xn\\-\\-tckwe|xn\\-\\-tiq49xqyj" 278 + "|xn\\-\\-unup4y|xn\\-\\-vermgensberater\\-ctb|xn\\-\\-vermgensberatung\\-pwb|xn\\-\\-vhquv" 279 + "|xn\\-\\-vuq861b|xn\\-\\-w4r85el8fhu5dnra|xn\\-\\-w4rs40l|xn\\-\\-wgbh1c|xn\\-\\-wgbl6a" 280 + "|xn\\-\\-xhq521b|xn\\-\\-xkc2al3hye2a|xn\\-\\-xkc2dl3a5ee0h|xn\\-\\-y9a3aq|xn\\-\\-yfro4i67o" 281 + "|xn\\-\\-ygbi2ammx|xn\\-\\-zfr164b|xxx|xyz)" 282 + "|(?:yachts|yahoo|yamaxun|yandex|yodobashi|yoga|yokohama|you|youtube|yun|y[et])" 283 + "|(?:zappos|zara|zero|zip|zone|zuerich|z[amw]))"; 284 /** 285 * Kept for backward compatibility reasons. 286 * 287 * @deprecated Deprecated since it does not include all IRI characters defined in RFC 3987 288 */ 289 @Deprecated 290 public static final String GOOD_IRI_CHAR = 291 "a-zA-Z0-9\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF"; 292 293 private static final String IP_ADDRESS_STRING = 294 "((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]" 295 + "[0-9]|[0-1][0-9]{2}|[1-9][0-9]|[1-9]|0)\\.(25[0-5]|2[0-4][0-9]|[0-1]" 296 + "[0-9]{2}|[1-9][0-9]|[1-9]|0)\\.(25[0-5]|2[0-4][0-9]|[0-1][0-9]{2}" 297 + "|[1-9][0-9]|[0-9]))"; 298 299 /** 300 * Kept for backward compatibility reasons. It does not match IPv6 addresses. 301 * 302 * @deprecated Please use {@link android.net.InetAddresses#isNumericAddress(String)} instead. 303 */ 304 @Deprecated 305 public static final Pattern IP_ADDRESS = Pattern.compile(IP_ADDRESS_STRING); 306 307 /** 308 * Valid UCS characters defined in RFC 3987. Excludes space characters. 309 */ 310 private static final String UCS_CHAR = "[" + 311 "\u00A0-\uD7FF" + 312 "\uF900-\uFDCF" + 313 "\uFDF0-\uFFEF" + 314 "\uD800\uDC00-\uD83F\uDFFD" + 315 "\uD840\uDC00-\uD87F\uDFFD" + 316 "\uD880\uDC00-\uD8BF\uDFFD" + 317 "\uD8C0\uDC00-\uD8FF\uDFFD" + 318 "\uD900\uDC00-\uD93F\uDFFD" + 319 "\uD940\uDC00-\uD97F\uDFFD" + 320 "\uD980\uDC00-\uD9BF\uDFFD" + 321 "\uD9C0\uDC00-\uD9FF\uDFFD" + 322 "\uDA00\uDC00-\uDA3F\uDFFD" + 323 "\uDA40\uDC00-\uDA7F\uDFFD" + 324 "\uDA80\uDC00-\uDABF\uDFFD" + 325 "\uDAC0\uDC00-\uDAFF\uDFFD" + 326 "\uDB00\uDC00-\uDB3F\uDFFD" + 327 "\uDB44\uDC00-\uDB7F\uDFFD" + 328 "&&[^\u00A0[\u2000-\u200A]\u2028\u2029\u202F\u3000]]"; 329 330 /** 331 * Valid characters for IRI label defined in RFC 3987. 332 */ 333 private static final String LABEL_CHAR = "a-zA-Z0-9" + UCS_CHAR; 334 335 /** 336 * Valid characters for IRI TLD defined in RFC 3987. 337 */ 338 private static final String TLD_CHAR = "a-zA-Z" + UCS_CHAR; 339 340 /** 341 * RFC 1035 Section 2.3.4 limits the labels to a maximum 63 octets. 342 */ 343 private static final String IRI_LABEL = 344 "[" + LABEL_CHAR + "](?:[" + LABEL_CHAR + "_\\-]{0,61}[" + LABEL_CHAR + "]){0,1}"; 345 346 /** 347 * RFC 3492 references RFC 1034 and limits Punycode algorithm output to 63 characters. 348 */ 349 private static final String PUNYCODE_TLD = "xn\\-\\-[\\w\\-]{0,58}\\w"; 350 351 private static final String TLD = "(" + PUNYCODE_TLD + "|" + "[" + TLD_CHAR + "]{2,63}" +")"; 352 353 private static final String HOST_NAME = "(" + IRI_LABEL + "\\.)+" + TLD; 354 355 private static final String DOMAIN_NAME_STR = "(" + HOST_NAME + "|" + IP_ADDRESS_STRING + ")"; 356 public static final Pattern DOMAIN_NAME = Pattern.compile(DOMAIN_NAME_STR); 357 358 private static final String PROTOCOL = "(?i:http|https|rtsp|ftp)://"; 359 360 /* A word boundary or end of input. This is to stop foo.sure from matching as foo.su */ 361 private static final String WORD_BOUNDARY = "(?:\\b|$|^)"; 362 363 private static final String USER_INFO = "(?:[a-zA-Z0-9\\$\\-\\_\\.\\+\\!\\*\\'\\(\\)" 364 + "\\,\\;\\?\\&\\=]|(?:\\%[a-fA-F0-9]{2})){1,64}(?:\\:(?:[a-zA-Z0-9\\$\\-\\_" 365 + "\\.\\+\\!\\*\\'\\(\\)\\,\\;\\?\\&\\=]|(?:\\%[a-fA-F0-9]{2})){1,25})?\\@"; 366 367 private static final String PORT_NUMBER = "\\:\\d{1,5}"; 368 369 private static final String PATH_AND_QUERY = "[/\\?](?:(?:[" + LABEL_CHAR 370 + ";/\\?:@&=#~" // plus optional query params 371 + "\\-\\.\\+!\\*'\\(\\),_\\$])|(?:%[a-fA-F0-9]{2}))*"; 372 373 /** 374 * Regular expression pattern to match most part of RFC 3987 375 * Internationalized URLs, aka IRIs. 376 */ 377 public static final Pattern WEB_URL = Pattern.compile("(" 378 + "(" 379 + "(?:" + PROTOCOL + "(?:" + USER_INFO + ")?" + ")?" 380 + "(?:" + DOMAIN_NAME_STR + ")" 381 + "(?:" + PORT_NUMBER + ")?" 382 + ")" 383 + "(" + PATH_AND_QUERY + ")?" 384 + WORD_BOUNDARY 385 + ")"); 386 387 /** 388 * Regular expression that matches known TLDs and punycode TLDs 389 */ 390 private static final String STRICT_TLD = "(?:" + 391 IANA_TOP_LEVEL_DOMAINS + "|" + PUNYCODE_TLD + ")"; 392 393 /** 394 * Regular expression that matches host names using {@link #STRICT_TLD} 395 */ 396 private static final String STRICT_HOST_NAME = "(?:(?:" + IRI_LABEL + "\\.)+" 397 + STRICT_TLD + ")"; 398 399 /** 400 * Regular expression that matches domain names using either {@link #STRICT_HOST_NAME} or 401 * {@link #IP_ADDRESS} 402 */ 403 private static final String STRICT_DOMAIN_NAME = "(?:" + STRICT_HOST_NAME + "|" 404 + IP_ADDRESS_STRING + ")"; 405 406 /** 407 * Regular expression that matches domain names without a TLD 408 */ 409 private static final String RELAXED_DOMAIN_NAME = 410 "(?:" + "(?:" + IRI_LABEL + "(?:\\.(?=\\S))" +"?)+" + "|" + IP_ADDRESS_STRING + ")"; 411 412 /** 413 * Regular expression to match strings that do not start with a supported protocol. The TLDs 414 * are expected to be one of the known TLDs. 415 */ 416 private static final String WEB_URL_WITHOUT_PROTOCOL = "(" 417 + WORD_BOUNDARY 418 + "(?<!:\\/\\/)" 419 + "(" 420 + "(?:" + STRICT_DOMAIN_NAME + ")" 421 + "(?:" + PORT_NUMBER + ")?" 422 + ")" 423 + "(?:" + PATH_AND_QUERY + ")?" 424 + WORD_BOUNDARY 425 + ")"; 426 427 /** 428 * Regular expression to match strings that start with a supported protocol. Rules for domain 429 * names and TLDs are more relaxed. TLDs are optional. 430 */ 431 private static final String WEB_URL_WITH_PROTOCOL = "(" 432 + WORD_BOUNDARY 433 + "(?:" 434 + "(?:" + PROTOCOL + "(?:" + USER_INFO + ")?" + ")" 435 + "(?:" + RELAXED_DOMAIN_NAME + ")?" 436 + "(?:" + PORT_NUMBER + ")?" 437 + ")" 438 + "(?:" + PATH_AND_QUERY + ")?" 439 + WORD_BOUNDARY 440 + ")"; 441 442 /** 443 * Regular expression pattern to match IRIs. If a string starts with http(s):// the expression 444 * tries to match the URL structure with a relaxed rule for TLDs. If the string does not start 445 * with http(s):// the TLDs are expected to be one of the known TLDs. 446 * 447 * @hide 448 */ 449 public static final Pattern AUTOLINK_WEB_URL = Pattern.compile( 450 "(" + WEB_URL_WITH_PROTOCOL + "|" + WEB_URL_WITHOUT_PROTOCOL + ")"); 451 452 /** 453 * Regular expression for valid email characters. Does not include some of the valid characters 454 * defined in RFC5321: #&~!^`{}/=$*?| 455 */ 456 private static final String EMAIL_CHAR = LABEL_CHAR + "\\+\\-_%'"; 457 458 /** 459 * Regular expression for local part of an email address. RFC5321 section 4.5.3.1.1 limits 460 * the local part to be at most 64 octets. 461 */ 462 private static final String EMAIL_ADDRESS_LOCAL_PART = 463 "[" + EMAIL_CHAR + "]" + "(?:[" + EMAIL_CHAR + "\\.]{0,62}[" + EMAIL_CHAR + "])?"; 464 465 /** 466 * Regular expression for the domain part of an email address. RFC5321 section 4.5.3.1.2 limits 467 * the domain to be at most 255 octets. 468 */ 469 private static final String EMAIL_ADDRESS_DOMAIN = 470 "(?=.{1,255}(?:\\s|$|^))" + HOST_NAME; 471 472 /** 473 * Regular expression pattern to match email addresses. It excludes double quoted local parts 474 * and the special characters #&~!^`{}/=$*?| that are included in RFC5321. 475 * @hide 476 */ 477 public static final Pattern AUTOLINK_EMAIL_ADDRESS = Pattern.compile("(" + WORD_BOUNDARY + 478 "(?:" + EMAIL_ADDRESS_LOCAL_PART + "@" + EMAIL_ADDRESS_DOMAIN + ")" + 479 WORD_BOUNDARY + ")" 480 ); 481 482 public static final Pattern EMAIL_ADDRESS 483 = Pattern.compile( 484 "[a-zA-Z0-9\\+\\.\\_\\%\\-\\+]{1,256}" + 485 "\\@" + 486 "[a-zA-Z0-9][a-zA-Z0-9\\-]{0,64}" + 487 "(" + 488 "\\." + 489 "[a-zA-Z0-9][a-zA-Z0-9\\-]{0,25}" + 490 ")+" 491 ); 492 493 /** 494 * This pattern is intended for searching for things that look like they 495 * might be phone numbers in arbitrary text, not for validating whether 496 * something is in fact a phone number. It will miss many things that 497 * are legitimate phone numbers. 498 * 499 * <p> The pattern matches the following: 500 * <ul> 501 * <li>Optionally, a + sign followed immediately by one or more digits. Spaces, dots, or dashes 502 * may follow. 503 * <li>Optionally, sets of digits in parentheses, separated by spaces, dots, or dashes. 504 * <li>A string starting and ending with a digit, containing digits, spaces, dots, and/or dashes. 505 * </ul> 506 */ 507 public static final Pattern PHONE 508 = Pattern.compile( // sdd = space, dot, or dash 509 "(\\+[0-9]+[\\- \\.]*)?" // +<digits><sdd>* 510 + "(\\([0-9]+\\)[\\- \\.]*)?" // (<digits>)<sdd>* 511 + "([0-9][0-9\\- \\.]+[0-9])"); // <digit><digit|sdd>+<digit> 512 513 /** 514 * Convenience method to take all of the non-null matching groups in a 515 * regex Matcher and return them as a concatenated string. 516 * 517 * @param matcher The Matcher object from which grouped text will 518 * be extracted 519 * 520 * @return A String comprising all of the non-null matched 521 * groups concatenated together 522 */ concatGroups(Matcher matcher)523 public static final String concatGroups(Matcher matcher) { 524 StringBuilder b = new StringBuilder(); 525 final int numGroups = matcher.groupCount(); 526 527 for (int i = 1; i <= numGroups; i++) { 528 String s = matcher.group(i); 529 530 if (s != null) { 531 b.append(s); 532 } 533 } 534 535 return b.toString(); 536 } 537 538 /** 539 * Convenience method to return only the digits and plus signs 540 * in the matching string. 541 * 542 * @param matcher The Matcher object from which digits and plus will 543 * be extracted 544 * 545 * @return A String comprising all of the digits and plus in 546 * the match 547 */ digitsAndPlusOnly(Matcher matcher)548 public static final String digitsAndPlusOnly(Matcher matcher) { 549 StringBuilder buffer = new StringBuilder(); 550 String matchingRegion = matcher.group(); 551 552 for (int i = 0, size = matchingRegion.length(); i < size; i++) { 553 char character = matchingRegion.charAt(i); 554 555 if (character == '+' || Character.isDigit(character)) { 556 buffer.append(character); 557 } 558 } 559 return buffer.toString(); 560 } 561 562 /** 563 * Do not create this static utility class. 564 */ Patterns()565 private Patterns() {} 566 } 567