A possible bug in JEP 482: Flexible Constructor Bodies

[Remi Forax]

> From: "Ella Ananeva" <ella.ananeva at oracle.com>
> To: "amber-dev" <amber-dev at openjdk.org>
> Sent: Monday, May 20, 2024 11:12:55 PM
> Subject: A possible bug in JEP 482: Flexible Constructor Bodies

> Hi,

> I found a discrepancy in the behavior of the JEP 482 compiler regarding definite
> assignment of a final non-static blank field in a prologue of a constructor:

> Case 1

> Case 2

> Case 3

> class Q {
> final int x ;
> int y ;
> { y = x ; }
> public Q ( int a ) {
> x = a ;
> super ();

> }
> }

> class Q {
> final int x ;
> int y ;
> { y = x ; }
> public Q ( int a ) {
> x = a ;
> this ();
> }

> public Q () {}
> }

> class Q {
> final int x ;
> int y ;
> { y = x ; }
> public Q ( int a ) {
> x = a ;
> }
> }

> Compilation succeeds

> Compilation fails

> Compilation fails

> Shouldn’t the behavior in this 3 cases be similar?

Hello, 
case 3 is equivalent to 

class Q { 
final int x ; 
int y ; 
public Q ( int a ) { 
y = x ; 
x = a ; 
} 
} 

so 'y' can not be initialized by 'x' given that 'x' is not initialized. 
case 1 is equivalent to 
class Q { 
final int x ; 
int y ; 
public Q ( int a ) { 
x = a ; 
super (); 
y = x ; 
} 
} 
so it compiles correctly, when y = x is executed, x is already initialized. 

For me, case 2 is a compiler bug, the call to this() is indirectly a super constructor call. 

> Thank you,

> Ella

regards, 
Rémi 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/amber-dev/attachments/20240520/b075411b/attachment-0001.htm>