Explicit cast of primitive - some strange things

[Ranjan Sarangi]

All, This one may look easy for some. But at first sight it is confuses. Try the following.
int i=10;
// Following need explicit cast in order to get compiled.
char c=i; byte b=i; short s=i;
//But Following works fine - compiles well.
char c=10; byte b=10; short s=10;
My school of thought would say : As long as the value of RHS is in range of the LHS type, it should not need explicit cast no matter RHS is a variable or value.
Would someone please justify this behaviour of JVM!!!
Another thing - in a mockexam (http://valiveru.tripod.com/java/jvaltest.html) the following declaration is said valid. (which I tested does NOT compile)
int i=10; char c=i;
Are these site authenticated? This site has very tough Qs and added to that couple of wrong answers like above. Testing yourself at these site might be frustrating. Just a breather to concerned ones.

[William Brogden]

//But Following works fine - compiles well.
char c=10; byte b=10; short s=10;
The compiler is allowed to make decisions like this
to simplify the life of the programmer.
Don't confuse what happens at compile-time with what happens at runtime. If the runtime JVM had to do a range check every time an int value was assigned to a byte, it would slow things down considerably.
Bill

[Sanjeet Karamchandani]

Hi,
int i=10;
// Following need explicit cast in order to get compiled.
char c=i; byte b=i; short s=i;
//But Following works fine - compiles well.
char c=10; byte b=10; short s=10;
Well, try this
final int i=10;
char c=i;byte=i;short=i;
All these assignments work well coz some optimization done by the compilier( Im not very sure what they mean by optimization, but u should be aware abt this behaviour). So it seems that if u assign a final value or a final variable it will work fine.

Another thing - in a mockexam (http://valiveru.tripod.com/java/jvaltest.html) the following declaration is said valid. (which I tested does NOT compile)
int i=10; char c=i;
It will work but with a cast.
Anything wrong?
Thanks
Sanjeet

[This message has been edited by Sanjeet Karamchandani (edited August 19, 2000).]

[Ranjan Sarangi]

Thanks Bill and Sanjeet for drawing line between the compile time and run time checks.
Bill, Currently I'm reading your cram book. It's a great resource. I'll appear the exam shortly. It's a pleasure that you responded to my questoin.
Thanks.