Re: [C++ PATCH] PR c++/79133

[Jason Merrill <jason at redhat dot com>]

On 08/07/2018 07:12 AM, Ville Voutilainen wrote:
> On 8 July 2018 at 02:08, Ville Voutilainen <ville.voutilainen@gmail.com> wrote:
> > On 8 July 2018 at 01:54, Paolo Carlini <paolo.carlini@oracle.com> wrote:
> > > > That would make this more consistent with such a shadow warning, but I
> > > > don't want
> > > > to use the shadowing wording (which would be easy to do; just set
> > > > 'shadowed' and do
> > > > a 'goto inform'), because this isn't shadowing in the precise sense;
> > > > the shadowing cases
> > > > are warnings, whereas this is more like the redeclaration errors in
> > > > the same function.
> > >
> > > ... indeed and that annoys me a bit. Not having studied at all c++/79133 so
> > > far (sorry) it seems a little weird to me that according to the standard we
> > > have to handle the two types of "shadowing" in different ways, one more
> > > strict, one less. Thus I would suggest double checking the details of that,
> > > eventually with Jason too in terms of the actual patch you would like to
> > > apply.
> >
> > Well. The PR is about DR 2211 which, in simple terms, says that lambda
> > parameters
> > and captures cannot have the same name. See
> > http://open-std.org/JTC1/SC22/WG21/docs/cwg_defects.html#2211
> >
> > That's stricter than -Wshadow, but otherwise equally strict as the
> > other error cases already handled
> > in check_local_shadow. So I'll make this error case more consistent
> > with the others. We already
> > handle redeclaration errors slightly differently from shadowing
> > warnings in that function.
>
> Here's an updated patch. Tested on Linux-PPC64, OK for trunk?
>
> Backports?
>
> 2018-08-06  Ville Voutilainen  <ville.voutilainen@gmail.com>
>
>      gcc/cp/
>
>      PR c++/79133
>      * name-lookup.c (check_local_shadow): Reject captures and parameters
>      with the same name.

>        if (DECL_CONTEXT (old) == current_function_decl
>  	  && TREE_CODE (decl) != PARM_DECL
> -	  && TREE_CODE (old) == PARM_DECL)
> +	  && TREE_CODE (old) == PARM_DECL
> +	  && !is_capture_proxy (decl))
>  	{
>  	  /* Go to where the parms should be and see if we find
>  	     them there.  */
> @@ -2665,6 +2667,21 @@ check_local_shadow (tree decl)
>  	      return;
>  	    }
>  	}
> +      /* DR 2211: check that captures and parameters
> +	 do not have the same name. */
> +      else if (is_capture_proxy (decl))

Maybe put this block first rather than add !is_capture_proxy to the if 
condition?

Jason