almost a cosmetic bug

[Olivier]

On Sun, Feb 11, 2001 at 04:04:50PM -0800, Alexandre Petit-Bianco wrote:
> Thanks. Can you try this patch?
> 

Yep, it does fixe it. Thank you very much.


However I couldn't directly apply your patch, even after having
synchronized my tree. I am not very good at "patch" but if I
understand well 

          resolve_qualified_expression_name (wfl, 
						 EXPR_WFL_NODE (qual_wfl));

Should exist in the file and stay unchanged. However I have

	      field_decl = lookup_field_wrapper (type,
						 EXPR_WFL_NODE (qual_wfl));

in place in my orig file which is CVS 1.251 as well as your file???

Manually replacing:

	      if (!field_decl)

by 
	      if (!field_decl && TREE_CODE (decl) == TYPE_DECL)

did fixe the bug.




> Index: parse.y
> ===================================================================
> RCS file: /cvs/gcc/gcc/gcc/java/parse.y,v
> retrieving revision 1.251
> diff -u -p -r1.251 parse.y
> --- parse.y	2001/02/05 18:09:45	1.251
> +++ parse.y	2001/02/12 00:02:56
> @@ -9363,7 +9366,7 @@ resolve_qualified_expression_name (wfl, 
>  						 EXPR_WFL_NODE (qual_wfl));
>  
>  	      /* Maybe what we're trying to access an inner class. */
> -	      if (!field_decl)
> +	      if (!field_decl && TREE_CODE (decl) == TYPE_DECL)
>  		{
>  		  tree ptr, inner_decl;

Olivier

--
Olivier Louchart-Fletcher
Email: olivier@zipworld.com.au