Problem converting C/C++ unsigned char to JAVA

[Ravi Kumar]

HI All,

I searched the forum for an answer, but could not get through.
Any help is greatly appreciated.

The problem with unsigned char.
I am reading a PPM image file which has data in ASCII.

For a character, eg. '†' ,
In JAVA, after reading it as char and typecasting into int its value is 8224.
In C/C++, after reading it as a unsigned char and typecasting into int its value is 160.

How would i read in JAVA so as to get value 160 ?

The followng C++ code

unsigned char ch1 ='†';
char ch2 = '†';

cout << (int) ch1 << "\n"; // prints 160
cout << (int) ch2 << "\n"; // prints -96

Thank you.

[Rob Spoor]

Ravi Kumar wrote:In JAVA, after reading it as char and typecasting into int its value is 8224.

There's your #1 mistake. A C char is only 1 byte in size, so you would need to use a Java byte for that. That would return (byte)160 which is actually the -96 you saw before. To turn that into a char you can add 256 to the byte:

char c = (char)(b + 256)

c will now have the char representation of ASCII 160.

[Ravi Kumar]

Thank you Rob for the prompt reply.

Normal typecasting works in JAVA for most of the characters.
The following code in JAVA

char ch1 = '^';
char ch2 = '†';
System.out.println (" value : " +  (int) ch1); // prints 94
System.out.println (" value :" +  (byte) ch1); // prints 94

System.out.println (" value : " +  (int) ch2); // prints 8224
System.out.println (" value :" +  (byte) ch2); // prints 32 

Following are some exceptions
8224 †
8226 •
8800 ≠
8482 ™
8710 ∆
8211 –
8221 ”
8216 ‘
9674 ◊
8260 ⁄
8249 ‹
8249 ‹
8734 ∞
8747 ∫
8364 €
8730 √
8804 ≤

Following are some good ones
94 ^
102 f
112 p
119 w
126 ~
196 Ä
122 z
197 Å
197 Å

Please suggest any help.
Thank you.

[Campbell Ritchie]

Ravi Kumar wrote: . . . data in ASCII.

For a character, eg. '†' . . .

That's not ASCII at all. ACSII only goes as far as 0x7f = 127.
The † character is 8224, well actually 0x2020, since Unicode characters are usually denominated in hexadecimal. You would find the casting much easier to understand in hex, so try the %x and %c tags after the printf method. Note you will have to cast to a char to use %c.

Anyway, this is much too difficult a question for "beginning Java", so I shall move it.

[Ravi Kumar]

Hi Campbell,
Thank you for the response.
I was wrong initially. The data is is ASCII/Extended ASCII, for the values are from 0-255.

In JAVA, how would i get value 160 out of the char † .

[Rob Spoor]

Cast it to int. In Java, char is nothing more than an unsigned 16-bit number with special support when printing to the screen.

[Ravi Kumar]

Thank you Rob.
That is where is hit the problem. When i simply cast the char † , I get a value of 8224.

I took help and got a solution:
I think this might help some others.

In JAVA the default encoding in UNICODE. for which the symbol † gets the value of 8224.
In MY CASE i need 160. So i need to find an appropriate charset.

I wrote a small method to get the charset. I found a hit for -96 (256-96=160).
Below is the function

public void findCharsets() {
		Map charSets = Charset.availableCharsets();
		Iterator it = charSets.keySet().iterator();
		String str = Character.toString('†');
		while (it.hasNext()) {
			try {
				String csName = (String) it.next();
				byte b[] = str.getBytes(Charset.forName(csName));
				if (b[0] == -96) {
					System.out.println("Found: " + csName);
				}
			} catch (Exception e) {
				// do nothing; go to next Charset
			}
		}
	}

The output i got is

Found: MacRoman
Found: x-MacCentralEurope
Found: x-MacCroatian
Found: x-MacCyrillic
Found: x-MacGreek
Found: x-MacRomania
Found: x-MacTurkish
Found: x-MacUkraine

Thanks everyone for your support.

[Campbell Ritchie]

Close examination of the hex value of † shows that it doesn't have 0xa0 (= 160 decimal) anywhere in. So I can't see how you expect to get 160 out of it. Most likely you are dealing with an encoding; maybe in UTF-8 there is 0xa0 in it somewhere. Look at this Joel Spolsky article for more about UTF-8.

[Campbell Ritchie]

" †"
: unicode-print
BEGIN
   DUP head
WHILE
  DUP head . CR 1+
REPEAT ; ok.
unicode-print 226 
128 
160 
ok.

It would appear there is in fact a 160 = 0xa0 in †. Try the program yourself; it is really easy to run . . . but it only works in FORTH

[Campbell Ritchie]

Oh, I forgot: you need the head word too

: head C@ ;

[Ravi Kumar]

Thank you Campbell, for your time and support.

[Campbell Ritchie]

You're welcome

And I presume you have worked out which encoding to use from the figures I showed earlier??