[dcl.contract.func]/4 Missing IFNDR

[BlowingWind314]

[dcl.contract.func]/4 says

A declaration D of a function or function template f that is not a first declaration shall have either no function-contract-specifier-seq or the same function-contract-specifier-seq (see below) as any first declaration F reachable from D.
If D and F are in different translation units, a diagnostic is required only if D is attached to a named module.
If a declaration F₁ is a first declaration of f in one translation unit and a declaration F₂ is a first declaration of f in another translation unit, F₁ and F₂ shall specify the same function-contract-specifier-seq, no diagnostic required ([ifndr:dcl.contract.func.mismatched.contract.specifiers]).

There are two IFNDR cases, one about not-the-first declarations and another is about first declarations.

However, G.6.4 is only about the latter case.

If two different first declarations of a function (which must therefore not be reachable from one another) do not have equivalent function contract specifiers the program is ill-formed, no diagnostic required.

BTW, [dcl.contract.func]/5 has a totally missing IFNDR

... If this condition is not met solely due to the comparison of two lambda-expressions that are contained within P₁ and P₂, no diagnostic is required.

[jensmaurer]

@notadragon , please add to your omnibus paper.

[notadragon]

Noted. The rationale for the first IFFNDR (a redecaration of a function not attached to a named module) somewhat escapes me right now. I will add both entries though and we can remove one if that one really doesn't prove needed. (It won't make the mailing, but i'll put up a draft update of D4284R1 in a day or two.)

[BlowingWind314]

@notadragon, [module.unit]/3 is in a similar situation

A module partition is a module unit whose module-declaration contains a module-partition. A named module shall not contain multiple module partitions with the same module-partition. All module partitions of a module that are module interface units shall be directly or indirectly exported by the primary module interface unit ([module.import]). No diagnostic is required for a violation of these rules ([ifndr:module.unit.unexported.module.partition]).

where [ifndr:module.unit.unexported.module.partition] also forgets the former sentence (i.e., a named module containing multiple module partitions with the same module-partition).

[BlowingWind314]

@notadragon [except.spec]/4 also contains a missing IFNDR.

... If a noexcept-specifier is specified in an explicit instantiation, the exception specification shall be the same as the exception specification of all other declarations of that function. A diagnostic is required only if the exception specifications are not the same within a single translation unit.

[notadragon]

Added to the omnibus draft (https://isocpp.org/files/papers/D4284R1.pdf) Didn't fill in examples for everything yet, but will before the next mailing (though reviews of what is there and suggestions for any of the TODOs are always welcome). (Let me know if/how you'd like to be acknowledged in the paper).