This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch] Fix PR c++/27668
Lee Millward wrote:
> Index: gcc/cp/pt.c
> ===================================================================
> --- gcc/cp/pt.c (revision 115597)
> +++ gcc/cp/pt.c (working copy)
> @@ -4069,7 +4069,8 @@ coerce_template_parms (tree parms,
> in_decl);
>
> if (arg == error_mark_node)
> - lost++;
> + return error_mark_node;
> +
> TREE_VEC_ELT (new_inner_args, i) = arg;
> }
>
I'm not sure why the code is the way it presently is -- but how does
your patch help? Before, once we exit the loop, we do:
if (lost)
return error_mark_node;
So, I'm confused as to how your patch makes us return error_mark_node
where we before we did not. (Except in the pathological case of INT_MAX
+ 1 template parameters, which is not the case that you posted.)
I'm sure I'm being dreadfully dense here.
Thanks,
--
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-3385 x713