better show and a bit better compile times

[SimonDanisch]

No description provided.

[SimonDanisch]

• uses functors for map/on/onany, which makes it easier to spot them in listeners and also seems to help a little bit with compilation latency
• adds clear(obs), which disconnects the observable from any registered interactions as well as we can (TODO, can we do better?)
• 3x faster register_callback and therefore on & friends
• improves show:
  master:

julia> Observable(1)
Observable{Int64} with 0 listeners. Value:
1
julia> Observable{Any}(1)
Observable{Any} with 0 listeners. Value:
1
julia>  Observable{Any}()
Observable{Any} with 0 listeners. Value:
not assigned yet!
julia>  Observable{Any}(rand(1000))
Observable{Any} with 0 listeners. Value:
[0.8841847173085652, 0.17636229645481794, 0.1801141027029346, 0.8961865801112214, 0.9791553016060836, 0.09382024705822112, 0.5077761382502093, 0.7255823373854912, 0.43347224323187217, 0.13924684951959843  …  0.21019368689710627, 0.6806054284776022, 0.29723246950088134, 0.8230816367425313, 0.5785352956703508, 0.7225239242011102, 0.19871050884594854, 0.8674662591747743, 0.09680341443523366, 0.9864939526294473]
julia>  x = Observable(1); on(identity, x); x
Observable{Int64} with 1 listeners. Value:
1
julia> [Observable([1,2,3]) for i in 1:10]
10-element Vector{Observable{Vector{Int64}}}:
 Observable{Vector{Int64}} with 0 listeners. Value:
[1, 2, 3]
 Observable{Vector{Int64}} with 0 listeners. Value:
[1, 2, 3]
 Observable{Vector{Int64}} with 0 listeners. Value:
[1, 2, 3]
 Observable{Vector{Int64}} with 0 listeners. Value:
[1, 2, 3]
 Observable{Vector{Int64}} with 0 listeners. Value:
[1, 2, 3]
 Observable{Vector{Int64}} with 0 listeners. Value:
[1, 2, 3]
 Observable{Vector{Int64}} with 0 listeners. Value:
[1, 2, 3]
 Observable{Vector{Int64}} with 0 listeners. Value:
[1, 2, 3]
 Observable{Vector{Int64}} with 0 listeners. Value:
[1, 2, 3]
 Observable{Vector{Int64}} with 0 listeners. Value:
[1, 2, 3]

With this PR:

julia> Observable(1)
Observable(1)
julia> Observable{Any}(1)
Observable{Any}(1)
julia> Observable{Any}()
Observable{Any}(#undef)
julia> Observable{Any}(rand(1000))
Observable{Any}([0.8127470313464235, 0.735393878810991, 0.7095205370330129, 0.6043941038142925, 0.41500734502339887, 0.15371862494000232, 0.332946579126833, 0.23599271044932846, 0.18083317456374992, 0.46464445715753  …  0.13030068962485086, 0.3153281177326117, 0.08606749675082415, 0.45109560134589866, 0.10314979098549393, 0.4062614517373486, 0.8990326462007261, 0.09229643099020479, 0.43403034011145425, 0.1689857869888235])
julia> x = Observable(1); on(identity, x); x
Observable(1)
    0 => identity(x) in Base at operators.jl:526
julia> [Observable([1,2,3]) for i in 1:10]
10-element Vector{Observable{Vector{Int64}}}:
 Observable([1, 2, 3])
 Observable([1, 2, 3])
 Observable([1, 2, 3])
 Observable([1, 2, 3])
 Observable([1, 2, 3])
 Observable([1, 2, 3])
 Observable([1, 2, 3])
 Observable([1, 2, 3])
 Observable([1, 2, 3])
 Observable([1, 2, 3])