On 05/23/2011 10:05 AM, Richard Guenther wrote: > On Mon, May 23, 2011 at 3:53 PM, Nathan Froyd <froy...@codesourcery.com> > wrote: >> +/* Return the Nth argument type from FNTYPE. */ >> + >> +tree >> +nth_arg_type (const_tree fntype, int n) >> +{ >> + function_args_iterator iter; >> + tree t; >> + int i; >> + >> + gcc_assert (fntype != NULL_TREE); >> + gcc_assert (n >= 0); > > Please merge the asserts and do s/gcc_assert/gcc_checking_assert/
Ack. > And if n should be >= 0 why not pass it in as unsigned? Consistency with all the other interfaces where n is logically unsigned but passed in as int? > The patch is ok with both changes. Thanks for the review. -Nathan