Should method extension return the extended function?

[fredrikekre]

Consider the difference between these two function definitions

julia> Base.sum() = 1

julia> foo() = 1
foo (generic function with 1 method)

Should the method extension here return the sum function? I think it would be nice if the example above instead behaved like:

julia> Base.sum() = 1
sum (generic function with 16 methods)

Amusingly this is actually causing a doctest failure.

[Moelf]

I think the displaying is:

julia/base/show.jl

Lines 34 to 41 in 96147bb

	function show(io::IO, ::MIME"text/plain", f::Function)
	get(io, :compact, false)::Bool && return show(io, f)
	ft = typeof(f)
	name = ft.name.mt.name
	if isa(f, Core.IntrinsicFunction)
	print(io, f)
	id = Core.Intrinsics.bitcast(Int32, f)
	print(io, " (intrinsic function #$id)")

but I can't figure out what controls what to return when we have a statement like

julia> Base.sum() = 1

[vtjnash]

It is part of lowering (see return *):

julia> Meta.@lower Base.x() = 1
:($(Expr(:thunk, CodeInfo(
    @ none within `top-level scope`
1 ─ %1 = Base.getproperty(Base, :x)
│   %2 = Core.Typeof(%1)
│   %3 = Core.svec(%2)
│   %4 = Core.svec()
│   %5 = Core.svec(%3, %4, $(QuoteNode(:(#= REPL[1]:1 =#))))
│        $(Expr(:method, nothing, :(%5), CodeInfo(
    @ REPL[1]:1 within `none`
1 ─     return 1
)))
└──      return nothing
))))

julia> Meta.@lower x() = 1
:($(Expr(:thunk, CodeInfo(
    @ none within `top-level scope`
1 ─      $(Expr(:thunk, CodeInfo(
    @ none within `top-level scope`
1 ─     return $(Expr(:method, :x))
)))
│        $(Expr(:method, :x))
│   %3 = Core.Typeof(x)
│   %4 = Core.svec(%3)
│   %5 = Core.svec()
│   %6 = Core.svec(%4, %5, $(QuoteNode(:(#= REPL[2]:1 =#))))
│        $(Expr(:method, :x, :(%6), CodeInfo(
    @ REPL[2]:1 within `none`
1 ─     return 1
)))
└──      return x
))))