1 // URLEncoder.java - Provides a method for encoding strings according to
2 // application/x-www-form-urlencoded MIME type.
4 /* Copyright (C) 1999 Cygnus Solutions
6 This file is part of libgcj.
8 This software is copyrighted work licensed under the terms of the
9 Libgcj License. Please consult the file "LIBGCJ_LICENSE" for
13 import java
.io
.UnsupportedEncodingException
;
16 * @author Warren Levy <warrenl@cygnus.com>
17 * @date April 22, 1999.
21 * Written using on-line Java Platform 1.2 API Specification, as well
22 * as "The Java Class Libraries", 2nd edition (Addison-Wesley, 1998).
23 * Status: Believed complete and correct.
26 public class URLEncoder
28 // This method, per the JCL, is conservative in that it encodes
29 // some "allowable" characters as % triplets.
30 public static String
encode(String s
)
32 // Get the bytes in ISO-Latin-1 (i.e. 8859_1) per the JCL.
33 // Even though it is the default in most cases, it's specified here
34 // just in case System.getProperty("file.encoding") is not "8859_1".
38 byte[] buf
= s
.getBytes("8859_1");
40 for (int i
= 0; i
< buf
.length
; i
++)
41 // For efficiency, check the byte in order of most likely
42 // possibility so as to minimize the number of comparisons.
43 // Hence, exclude all the alphanumeric & allowed special chars first.
44 if ((buf
[i
] >= 'a' && buf
[i
] <= 'z') ||
45 (buf
[i
] >= 'A' && buf
[i
] <= 'Z') ||
46 (buf
[i
] >= '0' && buf
[i
] <= '9') ||
47 buf
[i
] == '-' || buf
[i
] == '_' || buf
[i
] == '.' || buf
[i
] == '*')
48 ; // This is the most likely case so exclude first for efficiency.
49 else if (buf
[i
] == ' ')
50 buf
[i
] = (byte) '+'; // Replace space char with plus symbol.
53 result
= result
+ new String(buf
, start
, i
- start
, "8859_1") +
54 "%" + Integer
.toHexString(((int) buf
[i
]) & 0xFF);
58 // Append remainder of allowable chars from the string, if any.
59 if (start
< buf
.length
)
61 new String(buf
, start
, buf
.length
- start
, "8859_1");
63 catch (UnsupportedEncodingException ex
)
65 // This should never happen as "8859_1" is the default encoding.