1 /*
2  * Copyright (c) 2015, 2018, Oracle and/or its affiliates. All rights reserved.
3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
4  *
5  * This code is free software; you can redistribute it and/or modify it
6  * under the terms of the GNU General Public License version 2 only, as
7  * published by the Free Software Foundation.
8  *
9  * This code is distributed in the hope that it will be useful, but WITHOUT
10  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
11  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
12  * version 2 for more details (a copy is included in the LICENSE file that
13  * accompanied this code).
14  *
15  * You should have received a copy of the GNU General Public License version
16  * 2 along with this work; if not, write to the Free Software Foundation,
17  * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
18  *
19  * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
20  * or visit www.oracle.com if you need additional information or have any
21  * questions.
22  */
23 
24 import java.io.UnsupportedEncodingException;
25 import java.nio.charset.Charset;
26 import java.util.HashMap;
27 import java.util.Map;
28 
29 import org.testng.annotations.BeforeClass;
30 
31 /*
32  * Base class of tests for Compact String.
33  *
34  */
35 public class CompactString {
36 
37     final Map<String, Map<String, String>> map = new HashMap<>();
38 
39     enum StringSources {
40         EMPTY(STRING_EMPTY, BYTE_ARRAY_EMTPY, CHAR_ARRAY_EMPTY,
41                 POINT_ARRAY_EMTPY), LDUPLICATE(STRING_LDUPLICATE,
42                 BYTE_ARRAY_LDUPLICATE, CHAR_ARRAY_LDUPLICATE,
43                 POINT_ARRAY_LDUPLICATE), LLONG(STRING_LLONG, BYTE_ARRAY_LLONG,
44                 CHAR_ARRAY_LLONG, POINT_ARRAY_LLONG), L1(STRING_L1,
45                 BYTE_ARRAY_L1, CHAR_ARRAY_L1, POINT_ARRAY_L1), L2(STRING_L2,
46                 BYTE_ARRAY_L2, CHAR_ARRAY_L2, POINT_ARRAY_L2), L4(STRING_L4,
47                 BYTE_ARRAY_L4, CHAR_ARRAY_L4, POINT_ARRAY_L4), UDUPLICATE(
48                 STRING_UDUPLICATE, BYTE_ARRAY_UDUPLICATE,
49                 CHAR_ARRAY_UDUPLICATE, POINT_ARRAY_UDUPLICATE), U1(STRING_U1,
50                 BYTE_ARRAY_U1, CHAR_ARRAY_U1, POINT_ARRAY_U1), U2(STRING_U2,
51                 BYTE_ARRAY_U2, CHAR_ARRAY_U2, POINT_ARRAY_U2), MDUPLICATE1(
52                 STRING_MDUPLICATE1, BYTE_ARRAY_MDUPLICATE1,
53                 CHAR_ARRAY_MDUPLICATE1, POINT_ARRAY_MDUPLICATE1), MDUPLICATE2(
54                 STRING_MDUPLICATE2, BYTE_ARRAY_MDUPLICATE2,
55                 CHAR_ARRAY_MDUPLICATE2, POINT_ARRAY_MDUPLICATE2), MLONG1(
56                 STRING_MLONG1, BYTE_ARRAY_MLONG1, CHAR_ARRAY_MLONG1,
57                 POINT_ARRAY_MLONG1), MLONG2(STRING_MLONG2, BYTE_ARRAY_MLONG2,
58                 CHAR_ARRAY_MLONG2, POINT_ARRAY_MLONG2), M11(STRING_M11,
59                 BYTE_ARRAY_M11, CHAR_ARRAY_M11, POINT_ARRAY_M11), M12(
60                 STRING_M12, BYTE_ARRAY_M12, CHAR_ARRAY_M12, POINT_ARRAY_M12), SUPPLEMENTARY(
61                 STRING_SUPPLEMENTARY, BYTE_ARRAY_SUPPLEMENTARY,
62                 CHAR_ARRAY_SUPPLEMENTARY, POINT_ARRAY_SUPPLEMENTARY), SUPPLEMENTARY_LOWERCASE(
63                 STRING_SUPPLEMENTARY_LOWERCASE,
64                 BYTE_ARRAY_SUPPLEMENTARY_LOWERCASE,
65                 CHAR_ARRAY_SUPPLEMENTARY_LOWERCASE,
66                 POINT_ARRAY_SUPPLEMENTARY_LOWERCASE);
67 
StringSources(String s, byte[] b, char[] c, int[] i)68         private StringSources(String s, byte[] b, char[] c, int[] i) {
69             str = s;
70             ba = b;
71             ca = c;
72             ia = i;
73         }
74 
getString()75         String getString() {
76             return str;
77         }
78 
getByteArray()79         byte[] getByteArray() {
80             return ba;
81         }
82 
getCharArray()83         char[] getCharArray() {
84             return ca;
85         }
86 
getIntArray()87         int[] getIntArray() {
88             return ia;
89         }
90 
91         private final String str;
92         private final byte[] ba;
93         private final char[] ca;
94         private final int[] ia;
95     }
96 
97     protected static final String DEFAULT_CHARSET_NAME = "UTF-8";
98     protected static final Charset DEFAULT_CHARSET = Charset
99             .forName(DEFAULT_CHARSET_NAME);
100 
101     protected static final String STRING_EMPTY = "";
102     protected static final byte[] BYTE_ARRAY_EMTPY = new byte[0];
103     protected static final char[] CHAR_ARRAY_EMPTY = new char[0];
104     protected static final int[] POINT_ARRAY_EMTPY = new int[0];
105 
106     protected static final String STRING_LDUPLICATE = "ABABABABAB";
107     protected static final byte[] BYTE_ARRAY_LDUPLICATE = new byte[] { 'A', 'B',
108             'A', 'B', 'A', 'B', 'A', 'B', 'A', 'B' };
109     protected static final char[] CHAR_ARRAY_LDUPLICATE = new char[] { 'A', 'B',
110             'A', 'B', 'A', 'B', 'A', 'B', 'A', 'B' };
111     protected static final int[] POINT_ARRAY_LDUPLICATE = new int[] { 'A', 'B',
112             'A', 'B', 'A', 'B', 'A', 'B', 'A', 'B' };
113 
114     protected static final String STRING_LLONG = "ABCDEFGH";
115     protected static final byte[] BYTE_ARRAY_LLONG = new byte[] { 'A', 'B', 'C',
116             'D', 'E', 'F', 'G', 'H' };
117     protected static final char[] CHAR_ARRAY_LLONG = new char[] { 'A', 'B', 'C',
118             'D', 'E', 'F', 'G', 'H' };
119     protected static final int[] POINT_ARRAY_LLONG = new int[] { 'A', 'B', 'C',
120             'D', 'E', 'F', 'G', 'H' };
121 
122     protected static final String STRING_L1 = "A";
123     protected static final byte[] BYTE_ARRAY_L1 = new byte[] { 'A' };
124     protected static final char[] CHAR_ARRAY_L1 = new char[] { 'A' };
125     protected static final int[] POINT_ARRAY_L1 = new int[] { 'A' };
126 
127     protected static final String STRING_L2 = "AB";
128     protected static final byte[] BYTE_ARRAY_L2 = new byte[] { 'A', 'B' };
129     protected static final char[] CHAR_ARRAY_L2 = new char[] { 'A', 'B' };
130     protected static final int[] POINT_ARRAY_L2 = new int[] { 'A', 'B' };
131 
132     protected static final String STRING_L4 = "ABCD";
133     protected static final byte[] BYTE_ARRAY_L4 = new byte[] { 'A', 'B', 'C', 'D' };
134     protected static final char[] CHAR_ARRAY_L4 = new char[] { 'A', 'B', 'C', 'D' };
135     protected static final int[] POINT_ARRAY_L4 = new int[] { 'A', 'B', 'C', 'D' };
136 
137     /*
138      * Because right now ASCII is the default encoding parameter for source code
139      * in JDK build environment, so we escape them. same as below.
140      */
141     protected static final String STRING_UDUPLICATE = "\uFF21\uFF22\uFF21\uFF22\uFF21\uFF22\uFF21\uFF22\uFF21\uFF22";
142     protected static final byte[] BYTE_ARRAY_UDUPLICATE = getBytes(STRING_UDUPLICATE);
143     protected static final char[] CHAR_ARRAY_UDUPLICATE = new char[] { '\uFF21',
144             '\uFF22', '\uFF21', '\uFF22', '\uFF21', '\uFF22', '\uFF21',
145             '\uFF22', '\uFF21', '\uFF22' };
146     protected static final int[] POINT_ARRAY_UDUPLICATE = new int[] { '\uFF21',
147             '\uFF22', '\uFF21', '\uFF22', '\uFF21', '\uFF22', '\uFF21',
148             '\uFF22', '\uFF21', '\uFF22' };
149 
150     protected static final String STRING_U1 = "\uFF21";
151     protected static final byte[] BYTE_ARRAY_U1 = getBytes(STRING_U1);
152     protected static final char[] CHAR_ARRAY_U1 = new char[] { '\uFF21' };
153     protected static final int[] POINT_ARRAY_U1 = new int[] { '\uFF21' };
154 
155     protected static final String STRING_U2 = "\uFF21\uFF22";
156     protected static final byte[] BYTE_ARRAY_U2 = getBytes(STRING_U2);
157     protected static final char[] CHAR_ARRAY_U2 = new char[] { '\uFF21', '\uFF22' };
158     protected static final int[] POINT_ARRAY_U2 = new int[] { '\uFF21', '\uFF22' };
159 
160     protected static final String STRING_MDUPLICATE1 = "\uFF21A\uFF21A\uFF21A\uFF21A\uFF21A";
161     protected static final byte[] BYTE_ARRAY_MDUPLICATE1 = getBytes(STRING_MDUPLICATE1);
162     protected static final char[] CHAR_ARRAY_MDUPLICATE1 = new char[] { '\uFF21',
163             'A', '\uFF21', 'A', '\uFF21', 'A', '\uFF21', 'A', '\uFF21', 'A' };
164     protected static final int[] POINT_ARRAY_MDUPLICATE1 = new int[] { '\uFF21',
165             'A', '\uFF21', 'A', '\uFF21', 'A', '\uFF21', 'A', '\uFF21', 'A' };
166 
167     protected static final String STRING_MDUPLICATE2 = "A\uFF21A\uFF21A\uFF21A\uFF21A\uFF21";
168     protected static final byte[] BYTE_ARRAY_MDUPLICATE2 = getBytes(STRING_MDUPLICATE2);
169     protected static final char[] CHAR_ARRAY_MDUPLICATE2 = new char[] { 'A',
170             '\uFF21', 'A', '\uFF21', 'A', '\uFF21', 'A', '\uFF21', 'A',
171             '\uFF21' };
172     protected static final int[] POINT_ARRAY_MDUPLICATE2 = new int[] { 'A',
173             '\uFF21', 'A', '\uFF21', 'A', '\uFF21', 'A', '\uFF21', 'A',
174             '\uFF21' };
175 
176     protected static final String STRING_MLONG1 = "A\uFF21B\uFF22C\uFF23D\uFF24E\uFF25F\uFF26G\uFF27H\uFF28";
177     protected static final byte[] BYTE_ARRAY_MLONG1 = getBytes(STRING_MLONG1);
178     protected static final char[] CHAR_ARRAY_MLONG1 = new char[] { 'A', '\uFF21',
179             'B', '\uFF22', 'C', '\uFF23', 'D', '\uFF24', 'E', '\uFF25', 'F',
180             '\uFF26', 'G', '\uFF27', 'H', '\uFF28' };
181     protected static final int[] POINT_ARRAY_MLONG1 = new int[] { 'A', '\uFF21',
182             'B', '\uFF22', 'C', '\uFF23', 'D', '\uFF24', 'E', '\uFF25', 'F',
183             '\uFF26', 'G', '\uFF27', 'H', '\uFF28' };
184 
185     protected static final String STRING_MLONG2 = "\uFF21A\uFF22B\uFF23C\uFF24D\uFF25E\uFF26F\uFF27G\uFF28H";
186     protected static final byte[] BYTE_ARRAY_MLONG2 = getBytes(STRING_MLONG2);
187     protected static final char[] CHAR_ARRAY_MLONG2 = new char[] { '\uFF21', 'A',
188             '\uFF22', 'B', '\uFF23', 'C', '\uFF24', 'D', '\uFF25', 'E',
189             '\uFF26', 'F', '\uFF27', 'G', '\uFF28', 'H' };
190     protected static final int[] POINT_ARRAY_MLONG2 = new int[] { '\uFF21', 'A',
191             '\uFF22', 'B', '\uFF23', 'C', '\uFF24', 'D', '\uFF25', 'E',
192             '\uFF26', 'F', '\uFF27', 'G', '\uFF28', 'H' };
193 
194     protected static final String STRING_M11 = "A\uFF21";
195     protected static final byte[] BYTE_ARRAY_M11 = getBytes(STRING_M11);
196     protected static final char[] CHAR_ARRAY_M11 = new char[] { 'A', '\uFF21' };
197     protected static final int[] POINT_ARRAY_M11 = new int[] { 'A', '\uFF21' };
198 
199     protected static final String STRING_M12 = "\uFF21A";
200     protected static final byte[] BYTE_ARRAY_M12 = getBytes(STRING_M12);
201     protected static final char[] CHAR_ARRAY_M12 = new char[] { '\uFF21', 'A' };
202     protected static final int[] POINT_ARRAY_M12 = new int[] { '\uFF21', 'A' };
203 
204     protected static final String STRING_SUPPLEMENTARY = "\uD801\uDC00\uD801\uDC01\uFF21A";
205     protected static final byte[] BYTE_ARRAY_SUPPLEMENTARY = getBytes(STRING_SUPPLEMENTARY);
206     protected static final char[] CHAR_ARRAY_SUPPLEMENTARY = new char[] {
207             '\uD801', '\uDC00', '\uD801', '\uDC01', '\uFF21', 'A' };
208     protected static final int[] POINT_ARRAY_SUPPLEMENTARY = new int[] {
209             '\uD801', '\uDC00', '\uD801', '\uDC01', '\uFF21', 'A' };
210 
211     protected static final String STRING_SUPPLEMENTARY_LOWERCASE = "\uD801\uDC28\uD801\uDC29\uFF41a";
212     protected static final byte[] BYTE_ARRAY_SUPPLEMENTARY_LOWERCASE = getBytes(STRING_SUPPLEMENTARY_LOWERCASE);
213     protected static final char[] CHAR_ARRAY_SUPPLEMENTARY_LOWERCASE = new char[] {
214             '\uD801', '\uDC28', '\uD801', '\uDC29', '\uFF41', 'a' };
215     protected static final int[] POINT_ARRAY_SUPPLEMENTARY_LOWERCASE = new int[] {
216             '\uD801', '\uDC28', '\uD801', '\uDC29', '\uFF41', 'a' };
217 
218     protected static final String SRC_BYTE_ARRAY_WITH_CHARSETNAME = "source from byte array with charset name";
219     protected static final String SRC_BYTE_ARRAY_WITH_CHARSET = "source from byte array with charset";
220     protected static final String SRC_CHAR_ARRAY = "source from char array";
221     protected static final String SRC_POINT_ARRAY = "source from code point array";
222     protected static final String SRC_STRING = "source from String";
223     protected static final String SRC_STRINGBUFFER = "source from StringBuffer";
224     protected static final String SRC_STRINGBUILDER = "source from StringBuilder";
225     protected static final String SRC_COPYVALUEOF = "source from copyValueOf from char array";
226     protected static final String SRC_VALUEOF = "source from valueOf from char array";
227 
228     static {
229         System.out
println(String .format(R, Charset.defaultCharset().name(), DEFAULT_CHARSET_NAME))230                 .println(String
231                         .format("====== The platform's default charset is \"%s\", we're using \"%s\" for testing.",
232                                 Charset.defaultCharset().name(),
233                                 DEFAULT_CHARSET_NAME));
234     }
235 
getBytes(String str)236     private static byte[] getBytes(String str) {
237         byte[] res = null;
238         try {
239             res = str.getBytes(DEFAULT_CHARSET_NAME);
240         } catch (UnsupportedEncodingException e) {
241             e.printStackTrace();
242             throw new RuntimeException("caught UnsupportedEncodingException!!!", e);
243         }
244         return res;
245     }
246 
setUpOneString(String content, byte[] ba, char[] ca, int[] cpa)247     private void setUpOneString(String content, byte[] ba, char[] ca, int[] cpa)
248             throws UnsupportedEncodingException {
249         final Map<String, String> m = new HashMap<>();
250         m.put(SRC_BYTE_ARRAY_WITH_CHARSETNAME, new String(ba,
251                 DEFAULT_CHARSET_NAME));
252         m.put(SRC_BYTE_ARRAY_WITH_CHARSET, new String(ba, DEFAULT_CHARSET));
253         m.put(SRC_CHAR_ARRAY, new String(ca));
254         m.put(SRC_POINT_ARRAY, new String(cpa, 0, cpa.length));
255         m.put(SRC_STRING, new String(content));
256         m.put(SRC_STRINGBUFFER, new String(new StringBuffer(content)));
257         m.put(SRC_STRINGBUILDER, new String(new StringBuilder(content)));
258         m.put(SRC_COPYVALUEOF, String.copyValueOf(ca));
259         m.put(SRC_VALUEOF, String.valueOf(ca));
260         map.put(content, m);
261     }
262 
263     /*
264      * Set up the test data, use 9 ways to construct one String.
265      *
266      * @throws UnsupportedEncodingException
267      *         If the named charset is not supported in setUpOneString(xxx).
268      */
269     @BeforeClass
setUp()270     public void setUp() throws UnsupportedEncodingException {
271         for (StringSources src : StringSources.values()) {
272             setUpOneString(src.getString(), src.getByteArray(),
273                     src.getCharArray(), src.getIntArray());
274         }
275     }
276 
277     /*
278      * Escape non-ASCII characters since not all systems support them.
279      */
escapeNonASCIIs(String str)280     protected String escapeNonASCIIs(String str) {
281         StringBuilder sb = new StringBuilder();
282         for (int i = 0; i < str.length(); i++) {
283             char c = str.charAt(i);
284             if (c > 0x7F) {
285                 sb.append("\\u").append(Integer.toHexString((int) c));
286             } else {
287                 sb.append(c);
288             }
289         }
290         return sb.toString();
291     }
292 
293     /*
294      * Escape non-ASCII characters since not all systems support them.
295      */
escapeNonASCII(char c)296     protected String escapeNonASCII(char c) {
297         StringBuilder sb = new StringBuilder();
298         if (c > 0x7F) {
299             sb.append("\\u").append(Integer.toHexString((int) c));
300         } else {
301             sb.append(c);
302         }
303         return sb.toString();
304     }
305 }
306