xref: /aosp_15_r20/art/test/094-pattern/src/Main.java (revision 795d594fd825385562da6b089ea9b2033f3abf5a)
1*795d594fSAndroid Build Coastguard Worker /*
2*795d594fSAndroid Build Coastguard Worker  * Copyright (C) 2012 The Android Open Source Project
3*795d594fSAndroid Build Coastguard Worker  *
4*795d594fSAndroid Build Coastguard Worker  * Licensed under the Apache License, Version 2.0 (the "License");
5*795d594fSAndroid Build Coastguard Worker  * you may not use this file except in compliance with the License.
6*795d594fSAndroid Build Coastguard Worker  * You may obtain a copy of the License at
7*795d594fSAndroid Build Coastguard Worker  *
8*795d594fSAndroid Build Coastguard Worker  *      http://www.apache.org/licenses/LICENSE-2.0
9*795d594fSAndroid Build Coastguard Worker  *
10*795d594fSAndroid Build Coastguard Worker  * Unless required by applicable law or agreed to in writing, software
11*795d594fSAndroid Build Coastguard Worker  * distributed under the License is distributed on an "AS IS" BASIS,
12*795d594fSAndroid Build Coastguard Worker  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13*795d594fSAndroid Build Coastguard Worker  * See the License for the specific language governing permissions and
14*795d594fSAndroid Build Coastguard Worker  * limitations under the License.
15*795d594fSAndroid Build Coastguard Worker  */
16*795d594fSAndroid Build Coastguard Worker 
17*795d594fSAndroid Build Coastguard Worker import java.util.regex.Matcher;
18*795d594fSAndroid Build Coastguard Worker import java.util.regex.Pattern;
19*795d594fSAndroid Build Coastguard Worker 
20*795d594fSAndroid Build Coastguard Worker public class Main {
21*795d594fSAndroid Build Coastguard Worker     // from android.util.Patterns
22*795d594fSAndroid Build Coastguard Worker     public static final String GOOD_IRI_CHAR =
23*795d594fSAndroid Build Coastguard Worker         "a-zA-Z0-9\u00A0-\uD7FF\uF900-\uFDCF\uFDF0-\uFFEF";
24*795d594fSAndroid Build Coastguard Worker 
25*795d594fSAndroid Build Coastguard Worker     public static final String TOP_LEVEL_DOMAIN_STR_FOR_WEB_URL =
26*795d594fSAndroid Build Coastguard Worker         "(?:"
27*795d594fSAndroid Build Coastguard Worker         + "(?:aero|arpa|asia|a[cdefgilmnoqrstuwxz])"
28*795d594fSAndroid Build Coastguard Worker         + "|(?:biz|b[abdefghijmnorstvwyz])"
29*795d594fSAndroid Build Coastguard Worker         + "|(?:cat|com|coop|c[acdfghiklmnoruvxyz])"
30*795d594fSAndroid Build Coastguard Worker         + "|d[ejkmoz]"
31*795d594fSAndroid Build Coastguard Worker         + "|(?:edu|e[cegrstu])"
32*795d594fSAndroid Build Coastguard Worker         + "|f[ijkmor]"
33*795d594fSAndroid Build Coastguard Worker         + "|(?:gov|g[abdefghilmnpqrstuwy])"
34*795d594fSAndroid Build Coastguard Worker         + "|h[kmnrtu]"
35*795d594fSAndroid Build Coastguard Worker         + "|(?:info|int|i[delmnoqrst])"
36*795d594fSAndroid Build Coastguard Worker         + "|(?:jobs|j[emop])"
37*795d594fSAndroid Build Coastguard Worker         + "|k[eghimnprwyz]"
38*795d594fSAndroid Build Coastguard Worker         + "|l[abcikrstuvy]"
39*795d594fSAndroid Build Coastguard Worker         + "|(?:mil|mobi|museum|m[acdeghklmnopqrstuvwxyz])"
40*795d594fSAndroid Build Coastguard Worker         + "|(?:name|net|n[acefgilopruz])"
41*795d594fSAndroid Build Coastguard Worker         + "|(?:org|om)"
42*795d594fSAndroid Build Coastguard Worker         + "|(?:pro|p[aefghklmnrstwy])"
43*795d594fSAndroid Build Coastguard Worker         + "|qa"
44*795d594fSAndroid Build Coastguard Worker         + "|r[eosuw]"
45*795d594fSAndroid Build Coastguard Worker         + "|s[abcdeghijklmnortuvyz]"
46*795d594fSAndroid Build Coastguard Worker         + "|(?:tel|travel|t[cdfghjklmnoprtvwz])"
47*795d594fSAndroid Build Coastguard Worker         + "|u[agksyz]"
48*795d594fSAndroid Build Coastguard Worker         + "|v[aceginu]"
49*795d594fSAndroid Build Coastguard Worker         + "|w[fs]"
50*795d594fSAndroid Build Coastguard Worker         + "|(?:\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)"
51*795d594fSAndroid Build Coastguard Worker         + "|y[et]"
52*795d594fSAndroid Build Coastguard Worker         + "|z[amw]))";
53*795d594fSAndroid Build Coastguard Worker 
54*795d594fSAndroid Build Coastguard Worker     public static final String WEB_URL_STR =
55*795d594fSAndroid Build Coastguard Worker         "((?:(http|https|Http|Https|rtsp|Rtsp):\\/\\/(?:(?:[a-zA-Z0-9\\$\\-\\_\\.\\+\\!\\*\\'\\(\\)"
56*795d594fSAndroid Build Coastguard Worker         + "\\,\\;\\?\\&\\=]|(?:\\%[a-fA-F0-9]{2})){1,64}(?:\\:(?:[a-zA-Z0-9\\$\\-\\_"
57*795d594fSAndroid Build Coastguard Worker         + "\\.\\+\\!\\*\\'\\(\\)\\,\\;\\?\\&\\=]|(?:\\%[a-fA-F0-9]{2})){1,25})?\\@)?)?"
58*795d594fSAndroid Build Coastguard Worker         + "((?:(?:[" + GOOD_IRI_CHAR + "][" + GOOD_IRI_CHAR + "\\-]{0,64}\\.)+"   // named host
59*795d594fSAndroid Build Coastguard Worker         + TOP_LEVEL_DOMAIN_STR_FOR_WEB_URL
60*795d594fSAndroid Build Coastguard Worker         + "|(?:(?:25[0-5]|2[0-4]" // or ip address
61*795d594fSAndroid Build Coastguard Worker         + "[0-9]|[0-1][0-9]{2}|[1-9][0-9]|[1-9])\\.(?:25[0-5]|2[0-4][0-9]"
62*795d594fSAndroid Build Coastguard Worker         + "|[0-1][0-9]{2}|[1-9][0-9]|[1-9]|0)\\.(?:25[0-5]|2[0-4][0-9]|[0-1]"
63*795d594fSAndroid Build Coastguard Worker         + "[0-9]{2}|[1-9][0-9]|[1-9]|0)\\.(?:25[0-5]|2[0-4][0-9]|[0-1][0-9]{2}"
64*795d594fSAndroid Build Coastguard Worker         + "|[1-9][0-9]|[0-9])))"
65*795d594fSAndroid Build Coastguard Worker         + "(?:\\:\\d{1,5})?)" // plus option port number
66*795d594fSAndroid Build Coastguard Worker         + "(\\/(?:(?:[" + GOOD_IRI_CHAR + "\\;\\/\\?\\:\\@\\&\\=\\#\\~"  // plus option query params
67*795d594fSAndroid Build Coastguard Worker         + "\\-\\.\\+\\!\\*\\'\\(\\)\\,\\_])|(?:\\%[a-fA-F0-9]{2}))*)?"
68*795d594fSAndroid Build Coastguard Worker         + "(?:\\b|$)"; // and finally, a word boundary or end of
69*795d594fSAndroid Build Coastguard Worker                         // input.  This is to stop foo.sure from
70*795d594fSAndroid Build Coastguard Worker                         // matching as foo.su
71*795d594fSAndroid Build Coastguard Worker 
72*795d594fSAndroid Build Coastguard Worker     public static final Pattern WEB_URL = Pattern.compile(WEB_URL_STR);
73*795d594fSAndroid Build Coastguard Worker 
74*795d594fSAndroid Build Coastguard Worker     public static final String testStr1 =
75*795d594fSAndroid Build Coastguard Worker         "http://www.google.com/blah?client=tablet-android&source=android-home";
76*795d594fSAndroid Build Coastguard Worker     public static final String testStr2 = "http:///www.google.com/";
77*795d594fSAndroid Build Coastguard Worker     public static final String testStr3 =
78*795d594fSAndroid Build Coastguard Worker         "http://www.google.com/search?hl=en&redir_esc=&client=tablet-android-verizon&source=android-browser-type&v=141000000&qsubts=1327020479959&action=devloc&q=cnn";
79*795d594fSAndroid Build Coastguard Worker 
main(String[] args)80*795d594fSAndroid Build Coastguard Worker     public static void main(String[] args) {
81*795d594fSAndroid Build Coastguard Worker         System.out.println("str1 matches: " + WEB_URL.matcher(testStr1).matches());
82*795d594fSAndroid Build Coastguard Worker         System.out.println("str2 matches: " + WEB_URL.matcher(testStr2).matches());
83*795d594fSAndroid Build Coastguard Worker         System.out.println("str3 matches: " + WEB_URL.matcher(testStr3).matches());
84*795d594fSAndroid Build Coastguard Worker     }
85*795d594fSAndroid Build Coastguard Worker 
getStringAsHex(String text)86*795d594fSAndroid Build Coastguard Worker     static String getStringAsHex(String text) {
87*795d594fSAndroid Build Coastguard Worker         StringBuilder sb = new StringBuilder(text.length() * 4);
88*795d594fSAndroid Build Coastguard Worker 
89*795d594fSAndroid Build Coastguard Worker         for (int i = 0; i < text.length(); i++) {
90*795d594fSAndroid Build Coastguard Worker             sb.append(Integer.toHexString((int) text.charAt(i)));
91*795d594fSAndroid Build Coastguard Worker         }
92*795d594fSAndroid Build Coastguard Worker 
93*795d594fSAndroid Build Coastguard Worker         return sb.toString();
94*795d594fSAndroid Build Coastguard Worker     }
95*795d594fSAndroid Build Coastguard Worker }
96