Re: Patch: FYI: PR20056 thinko fix

[Ranjit Mathew <rmathew at gmail dot com>]

Tom Tromey wrote:
> I'm checking this in.
> 
> My PR 20056 patch had a dumb thinko in it, which further testing
> (jonas) caught.  Whoops!  This patch fixes the bug.  It also adds a
> test case, the OP pointed out on irc that it was easy to write one in
> pure java (his came from a non-java language, I believe).
> 
> I ran this through all the previous tests plus jonas.

For some reason, I'm still getting a FAIL on
the testcase.

As are others:

http://gcc.gnu.org/ml/gcc-testresults/2005-02/msg00879.html
http://gcc.gnu.org/ml/gcc-testresults/2005-02/msg00867.html

Ranjit.


> Tom
> 
> Index: gcc/java/ChangeLog
> from  Tom Tromey  <tromey@redhat.com>
> 	PR java/20056:
> 	* verify-impl.c (types_equal): Fixed test.
> 
> Index: libjava/ChangeLog
> from  Tom Tromey  <tromey@redhat.com>
> 	PR java/20056:
> 	* verify.cc (type::equals): Fixed test.
> 	* testsuite/libjava.lang/PR20056.out: New file.
> 	* testsuite/libjava.lang/PR20056.java: New file.
> 
> Index: libjava/verify.cc
> ===================================================================
> RCS file: /cvs/gcc/gcc/libjava/verify.cc,v
> retrieving revision 1.66
> diff -u -r1.66 verify.cc
> --- libjava/verify.cc 19 Feb 2005 01:16:28 -0000 1.66
> +++ libjava/verify.cc 19 Feb 2005 02:44:45 -0000
> @@ -754,10 +754,10 @@
>      bool equals (const type &other, _Jv_BytecodeVerifier *vfy)
>      {
>        // Only works for reference types.
> -      if (key != reference_type
> -	  || key != uninitialized_reference_type
> -	  || other.key != reference_type
> -	  || other.key != uninitialized_reference_type)
> +      if ((key != reference_type
> +	   && key != uninitialized_reference_type)
> +	  || (other.key != reference_type
> +	      && other.key != uninitialized_reference_type))
>  	return false;
>        // Only for single-valued types.
>        if (klass->ref_next || other.klass->ref_next)
> Index: libjava/testsuite/libjava.lang/PR20056.java
> ===================================================================
> RCS file: libjava/testsuite/libjava.lang/PR20056.java
> diff -N libjava/testsuite/libjava.lang/PR20056.java
> --- /dev/null	1 Jan 1970 00:00:00 -0000
> +++ libjava/testsuite/libjava.lang/PR20056.java 19 Feb 2005 02:44:45 -0000
> @@ -0,0 +1,13 @@
> +public class PR20056 {
> +  int x;
> +  PR20056(int x) {}
> +  PR20056(PR20056 f) {
> +    // The verifier rejected the generated code in this case.
> +    this(f.x = 0);
> +  }
> +
> +  public static void main(String[] args)
> +  {
> +    System.out.println("maude");
> +  }
> +}
> Index: libjava/testsuite/libjava.lang/PR20056.out
> ===================================================================
> RCS file: libjava/testsuite/libjava.lang/PR20056.out
> diff -N libjava/testsuite/libjava.lang/PR20056.out
> --- /dev/null	1 Jan 1970 00:00:00 -0000
> +++ libjava/testsuite/libjava.lang/PR20056.out 19 Feb 2005 02:44:45 -0000
> @@ -0,0 +1 @@
> +maude
> Index: gcc/java/verify-impl.c
> ===================================================================
> RCS file: /cvs/gcc/gcc/gcc/java/verify-impl.c,v
> retrieving revision 1.5
> diff -u -r1.5 verify-impl.c
> --- gcc/java/verify-impl.c 19 Feb 2005 01:14:17 -0000 1.5
> +++ gcc/java/verify-impl.c 19 Feb 2005 02:46:10 -0000
> @@ -743,8 +743,9 @@
>  static bool
>  types_equal (type *t1, type *t2)
>  {
> -  if (t1->key != reference_type || t1->key != uninitialized_reference_type
> -      || t2->key != reference_type || t2->key != uninitialized_reference_type)
> +  if ((t1->key != reference_type && t1->key != uninitialized_reference_type)
> +      || (t2->key != reference_type
> +	  && t2->key != uninitialized_reference_type))
>      return false;
>    /* Only single-ref types are allowed.  */
>    if (t1->klass->ref_next || t2->klass->ref_next)
> 


-- 
Ranjit Mathew      Email: rmathew AT gmail DOT com

Bangalore, INDIA.    Web: http://ranjitmathew.hostingzero.com/