On 4/10/25 12:34 PM, Jason Merrill wrote:
On 4/8/25 10:29 AM, Patrick Palka wrote:
The template arguments aren't dependent however -- they're just
incomplete because when we deferred them we were in the middle
deduction, and we consider a NULL_TREE template argument as dependent.
I wonder what would break if that changed?
Interestingly, the answer seems to be nothing: returning false instead
of true for null in dependent_template_arg_p fixes the testcases in this
patch and causes no regressions.
So for 16, let's try doing that and bring back the assert?
Jason
