On 2/19/26 11:08 PM, Marek Polacek wrote:
On Thu, Feb 19, 2026 at 02:53:29PM +0900, Jason Merrill wrote:
On 2/18/26 8:33 AM, Marek Polacek wrote:
A function template is supposed to be wrapped in an OVERLOAD.
I don't object to this change, but I've noticed that expectation in the
reflection code (e.g. using really_overloaded_ to test for a template), and
wondered why it's necessary? It would seem simpler to do without it.
I think since the rest of the front end works with that assumption,
the Reflection code should work that way too. Otherwise we crash like
here when splicing.
Yeah, I just wonder how much trouble it would be to change that
assumption elsewhere.
Jason
