Question on concatenation

[Veena Pointi]

According to of rules of concatination of operands that not primitive numeric type "One operand must be a string object,otherwise the expression is illegal"
But the following code compiles & runs fine

 
class ex{
public static void main(String args[]){
char a='b'+'c';
System.out.println(a);
}
}
 

Char is neither a primitive numeric type nor it is String object.Then how the above expression is valid?
Thanks
Veena

[Rattan Mann]

probably the reason is that a char is an integer
under the hood.
Rattan

[Ravi Anamalay]

Veena,
I think what has happened here is an integer addition, not a concatenation.
char a ='b' + 'c' is treated as an arithmetic expression and is evaluated as
a = 98 + 99 = 197
The character value for 197 is stored in 'a'.

[Ravi Anamalay]

I forgot to mention that char is an integral type (unsigned 16 bit) and hence you can perform arithmetic expressions on them.

[Veena Pointi]

Thanks.I read in R&H even char is numeric type.
Veena

[Paul Anilprem]

Originally posted by Veena Point:
Char is neither a primitive numeric type nor it is String object.Then how the above expression is valid?

char IS a primitive type. In addition, it is also a numeric type (byte, short, int, long). So what you are doing is not a concatination but simple addition.
Further, since 'a' and 'b' are constants and the summation fits into a char, the compiler is allowing you to assign the sum to a char without a cast. But try this:
char a = 'a';
char b = 'b';
char c = a + b;
Then try this:
char ch1 = '\uFFF0'+'\u000f';
char ch2 = '\uFFF0'+'\u000f';
ch2++;
char ch3 = '\uFFF0'+'\u001f';<== ??
System.out.println((int)ch1);
System.out.println((int)ch2);
What will be the output?
Have fun

[Veena Pointi]

Originally posted by Paul Anil:

char a = 'a';
char b = 'b';
char c = a + b;

I tried above code.Why it doesn't give error when a&b are constants,but gives error when they are variables?

[peter greaves]

char a = 'a';
char b = 'b';
char c = a + b;
"I tried above code.Why it doesn't give error when a&b are constants,but gives error when they are variables? "
because assignment with constants includes range-aware conversion to type. but when they are variables, then there's a widening conversion to an int (5.6.2 Binary Numeric Promotion) and therefore the expression will be an int - so that will require an explicit cast.