Coverage Report

Coverage Report - de.glossmaker.common.io.UnicodeInputStream

Classes in this File

Line Coverage

Branch Coverage

Complexity

UnicodeInputStream

52%

23/44

22%

8/36

4,5

 package de.glossmaker.common.io;
 
 import java.io.*;
 
 /**
  * This inputstream will recognize unicode BOM marks and will skip bytes if
  * getEncoding() method is called before any of the read(...) methods.
  * 
  * @author Aki Nieminen
  */
 public class UnicodeInputStream extends InputStream {
         PushbackInputStream internalIn;
         boolean isInited = false;
         String defaultEnc;
         String encoding;
 
         private static final int BOM_SIZE = 4;
 
         public UnicodeInputStream(InputStream in, String defaultEnc) {
                 internalIn = new PushbackInputStream(in, BOM_SIZE);
                 this.defaultEnc = defaultEnc;
         }
 
         public String getDefaultEncoding() {
                 return defaultEnc;
         }
 
         public String getEncoding() {
                 if (!isInited) {
                         try {
                                 init();
                         } catch (IOException ex) {
                                 IllegalStateException ise = new IllegalStateException(
                                                 "Init method failed.");
                                 ise.initCause(ise);
                                 throw ise;
                         }
                 }
                 return encoding;
         }
 
         /**
          * Read-ahead four bytes and check for BOM marks. Extra bytes are unread
          * back to the stream, only BOM bytes are skipped.
          */
         protected void init() throws IOException {
                 if (isInited)
                         return;
 
                 byte bom[] = new byte[BOM_SIZE];
                 int n, unread;
                 n = internalIn.read(bom, 0, bom.length);
 
                 if ((bom[0] == (byte) 0x00) && (bom[1] == (byte) 0x00)
                                 && (bom[2] == (byte) 0xFE) && (bom[3] == (byte) 0xFF)) {
                         encoding = "UTF-32BE";
                         unread = n - 4;
                 } else if ((bom[0] == (byte) 0xFF) && (bom[1] == (byte) 0xFE)
                                 && (bom[2] == (byte) 0x00) && (bom[3] == (byte) 0x00)) {
                         encoding = "UTF-32LE";
                         unread = n - 4;
                 } else if ((bom[0] == (byte) 0xEF) && (bom[1] == (byte) 0xBB)
                                 && (bom[2] == (byte) 0xBF)) {
                         encoding = "UTF-8";
                         unread = n - 3;
                 } else if ((bom[0] == (byte) 0xFE) && (bom[1] == (byte) 0xFF)) {
                         encoding = "UTF-16BE";
                         unread = n - 2;
                 } else if ((bom[0] == (byte) 0xFF) && (bom[1] == (byte) 0xFE)) {
                         encoding = "UTF-16LE";
                         unread = n - 2;
                 } else {
                         // Unicode BOM mark not found, unread all bytes
                         encoding = defaultEnc;
                         unread = n;
                 }
 
                 if (unread > 0)
                         internalIn.unread(bom, (n - unread), unread);
 
                 isInited = true;
         }
 
         public void close() throws IOException {
                 isInited = true;
                 internalIn.close();
         }
 
         public int read() throws IOException {
                 isInited = true;
                 return internalIn.read();
         }
 }

1		package de.glossmaker.common.io;
2
3		import java.io.*;
4
5		/**
6		* This inputstream will recognize unicode BOM marks and will skip bytes if
7		* getEncoding() method is called before any of the read(...) methods.
8		*
9		* @author Aki Nieminen
10		*/
11		public class UnicodeInputStream extends InputStream {
12		PushbackInputStream internalIn;
13	74	boolean isInited = false;
14		String defaultEnc;
15		String encoding;
16
17		private static final int BOM_SIZE = 4;
18
19	74	public UnicodeInputStream(InputStream in, String defaultEnc) {
20	74	internalIn = new PushbackInputStream(in, BOM_SIZE);
21	74	this.defaultEnc = defaultEnc;
22	74	}
23
24		public String getDefaultEncoding() {
25	0	return defaultEnc;
26		}
27
28		public String getEncoding() {
29	74	if (!isInited) {
30		try {
31	74	init();
32	0	} catch (IOException ex) {
33	0	IllegalStateException ise = new IllegalStateException(
34		"Init method failed.");
35	0	ise.initCause(ise);
36	0	throw ise;
37	74	}
38		}
39	74	return encoding;
40		}
41
42		/**
43		* Read-ahead four bytes and check for BOM marks. Extra bytes are unread
44		* back to the stream, only BOM bytes are skipped.
45		*/
46		protected void init() throws IOException {
47	74	if (isInited)
48	0	return;
49
50	74	byte bom[] = new byte[BOM_SIZE];
51		int n, unread;
52	74	n = internalIn.read(bom, 0, bom.length);
53
54	74	if ((bom[0] == (byte) 0x00) && (bom[1] == (byte) 0x00)
55		&& (bom[2] == (byte) 0xFE) && (bom[3] == (byte) 0xFF)) {
56	0	encoding = "UTF-32BE";
57	0	unread = n - 4;
58	74	} else if ((bom[0] == (byte) 0xFF) && (bom[1] == (byte) 0xFE)
59		&& (bom[2] == (byte) 0x00) && (bom[3] == (byte) 0x00)) {
60	0	encoding = "UTF-32LE";
61	0	unread = n - 4;
62	74	} else if ((bom[0] == (byte) 0xEF) && (bom[1] == (byte) 0xBB)
63		&& (bom[2] == (byte) 0xBF)) {
64	0	encoding = "UTF-8";
65	0	unread = n - 3;
66	74	} else if ((bom[0] == (byte) 0xFE) && (bom[1] == (byte) 0xFF)) {
67	0	encoding = "UTF-16BE";
68	0	unread = n - 2;
69	74	} else if ((bom[0] == (byte) 0xFF) && (bom[1] == (byte) 0xFE)) {
70	0	encoding = "UTF-16LE";
71	0	unread = n - 2;
72		} else {
73		// Unicode BOM mark not found, unread all bytes
74	74	encoding = defaultEnc;
75	74	unread = n;
76		}
77
78	74	if (unread > 0)
79	74	internalIn.unread(bom, (n - unread), unread);
80
81	74	isInited = true;
82	74	}
83
84		public void close() throws IOException {
85	0	isInited = true;
86	0	internalIn.close();
87	0	}
88
89		public int read() throws IOException {
90	0	isInited = true;
91	0	return internalIn.read();
92		}
93		}