CWG2942 [dcl.fct] packs in parameter-type-list

[cor3ntin]

in decl.fct p4, we say

After producing the list of parameter types, any top-level cv-qualifiers modifying a parameter type are deleted when forming the function type. The resulting list of transformed parameter types and the presence or absence of the ellipsis or a function parameter pack is the function's parameter-type-list.

Emphasis mine.

It's not clear what we mean by "or a function parameter pack".
A function can have a number of parameter packs in arbitrary position, and the fact that only the trailing pack is a deduced context is irrelevant when we are not deducing anything.

Are these functions equivalent?

void f(auto..., auto) {}
void f(auto..., auto) {}

Proposed resolution (which needs further massaging)

After producing the list of parameter types, any top-level cv-qualifiers modifying a parameter type are deleted when forming the function type. The resulting list of transformed parameter types and whether each corresponding parameter-declaration is a function parameter pack and the presence or absence of the ellipsis or a function parameter pack is the function's parameter-type-list.

[Endilll]

Regarding the proposed resolution. Given

The type of each parameter (including function parameter packs) is determined from its own parameter-declaration ([dcl.decl]).

I think whether a given parameter declaration is a pack or not is covered by the preceding wording. We can just strike the part about parameter packs where we talk about ellipsis.

[jensmaurer]

CWG2942

[cor3ntin]

I think whether a given parameter declaration is a pack or not is covered by the preceding wording. We can just strike the part about parameter packs where we talk about ellipsis.

I am not sure this is correct. the "packness" is part of the declaration, not the type. So presumably, there needs to be some wording to propagate that.

[jensmaurer]

I disagree. (Note that we're at the template level here, so "type" might be a bit more than what the rest of the type system, after instantiation, knows about.)

"The type of each parameter (including function parameter packs) is determined from its own parameter-declaration (9.3)'"

The packness is part of the parameter-declaration; if we feel something is amiss, [dcl.meaning.general] would need amendment somewhere. Suggestions welcome.