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44 * expected to become out-of-date very quickly.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