Re: Require canonical type comparison for typedefs again.

[Jason Merrill <jason@redhat.com>]

On 10/20/2010 05:37 PM, Dodji Seketeli wrote:
> +	  /* Only substitute into template non-type parms.  */
> +	  if (parameter != NULL_TREE
> +	      && TREE_CODE (parameter) == TREE_LIST
> +	      && (TREE_CODE (TREE_VALUE (parameter)) == TYPE_DECL
> +		  || TREE_CODE (TREE_VALUE (parameter)) == TEMPLATE_DECL))
> +	    continue;
> +
> +	  if (!dependent_type_p (TREE_TYPE (TREE_VALUE (parameter))))
> +	    continue;

These optimizations seem unsafe:

template <class T, template <class U = T> class V> ...
template <class T, template <int I = T()> class U> ...

Let's handle all parms, and check in fixup_template_parm* to see if a 
parm has the right NUM_SIBLINGS before we try to adjust it; I would 
expect that the parms of a template template parm will already have 
NUM_SIBLINGS set properly, we just need to substitute in the outer parms.

> +	 Consider the level of the parms of TT; T and U both have
> +	 level 2; TT has no template parm of level 1. So in this case
> +	 the first element of full_template_args is NULL_TREE. If we
> +	 leave it like this TMPL_ARG_DEPTH on args returns 1 instead
> +	 of 2. This will make tsubst wrongly consider that T and U
> +	 have level 1. Instead, let's create a dummy vector as the
> +	 first element of full_template_args so that TMPL_ARG_DEPTH
> +	 returns the correct depth for args.
> +      */

The */ should go on the last line of text, not on a line by itself.

Jason