Is there defined ordering in multiple declarations of partial record?

[ufcpp]

As a comparison, having fields in multiple declarations of struct results in a warning CS0282.

partial struct Struct // Warning: CS0282, There is no defined ordering between fields in multiple declarations of partial struct
{
    public int X;
}

partial struct Struct
{
    public int Y;
}

Don't records need the same type of warning for their properties? Ordering of properties in records has a visible effect on the synthesized methods such as ToString.

// in the same file
partial record class Record
{
    public int X { get; init; }
}

partial record class Record
{
    public int Y => X * X;
}
// WriteLine(new Record { X = 2 });
// Record { X = 2, Y = 4 }

// in the same file
partial record class Record
{
    public int Y => X * X;
}

partial record class Record
{
    public int X { get; init; }
}
// WriteLine(new Record { X = 2 });
// Record { Y = 4, X = 2 }

// A.cs
partial record class Record
{
    public int X { get; init; }
}

// B.cs
partial record class Record
{
    public int Y => X * X;
}
// WriteLine(new Record { X = 2 });
// Record { X = 2, Y = 4 }

// A.cs
partial record class Record
{
    public int Y => X * X;
}

// B.cs
partial record class Record
{
    public int X { get; init; }
}
// WriteLine(new Record { X = 2 });
// Record { Y = 4, X = 2 }

Should this behavior be defined in the spec, especially for cross-file partial declarations?

[RikkiGibson]

It seems like this warning is specific to structs since it more frequently interacts with issues of field layout in memory. We do give this warning with record structs. SharpLab.

I think it's probably fine to explicitly say in the spec that an implementation-defined ordering is used when determining the order in which to call printable members in ToString for record classes and record structs, for example. But it doesn't seem worthy of a warning to me.

cc @jcouv for visibility.

[jcouv]

It's not a new problem or specific to records, but it seems surprising that the spec doesn't call out the order in which parts of a partial type are stitched together.
The compiler certainly has an order. If I recall correctly, it's based on the syntactic order and order of inputs given to the compilation.

@RikkiGibson Do you want to start a thread with LDM?