Re: New egcs/gcc/testsuite/g77.f-torture/noncompile/980615-0.f

[Craig Burley <burley at gnu dot org>]

>>       CaLL foo(W)
>>       END
>>       SUBROUTINE foo(W)
>>       yy(I)=A(I)Q(X)
>
>It might be necessary to add the required "END" at the end :-)
>
>Otherwise this will keep failing to be compiled, even if the crash  
>will be repaired (I tried this by replacing the conjunction A(I)Q(X)  
>by A(I)*Q(X) [bonus points for a non-fortran type why this is OK]  
>and it produced a non-zero status without the conclusive END)

The subject line says `noncompile' -- this test is *supposed* to fail
to be compiled, but is not supposed to crash.  I already submitted
a patch to fix the crash, but no patch should be submitted to cause
g77 to compile the code without error.  ;-)

In particular, it's precisely the invalid statement-function-definition
expression that stimulates the crash in question.

        tq vm, (burley)