Re: gij runs HelloWorld on PPC Linux

[Kevin B. Hendricks <khendricks at ivey dot uwo dot ca>]

Hi Mark and Jeff,

Thanks for that info.  I think I know why Primes is in an infinite loop and 
if I am right we are good to go on ppc linux.

Thanks again for your help!

Kevin

On Sunday 11 February 2001 16:05, jeff.sturm@appnet.com wrote:
> On Sun, 11 Feb 2001, Mark Wielaard wrote:
> > > gij Final        
> > > 
> > > (this returns the string "this should not happen" which is incorrect)
> > Same here:
> > $ gij Final 
> > This should not happen
> > 
> > > gij Primes
> > > 
> > > (this seems to run in an infinite loop never printing any Primes)
> > Although this class takes a while (more then 30 seconds) it does
> > print primes and finishes eventually. (output attached)
> 
> Those results match mine on Alpha.  The erroneous Final output is caused
> by a bug in the interpreter.  One possible fix is:
> 
> 
> Index: interpret.cc
> ===================================================================
> RCS file: /cvs/gcc/gcc/libjava/interpret.cc,v
> retrieving revision 1.22
> diff -u -r1.22 interpret.cc
> --- interpret.cc	2000/10/20 23:25:57	1.22
> +++ interpret.cc	2001/02/11 21:00:08
> @@ -684,7 +684,10 @@
>  	  {
>  	    // final methods do not appear in the vtable,
>  	    // if it does not appear in the superclass.
> -	    fun = (void (*)()) rmeth->method->ncode;
> +	    if (sp[0].o)
> +	      fun = (void (*)()) rmeth->method->ncode;
> +	    else
> +	      throw new ::java::lang::NullPointerException;
>  	  }
>  	else
>  	  {