Overhaul name_of_type

[oxinabox]

• Correctly handle types that are in another name-space
• Correctly handle Tuple{}
• Correctly handle Symbols in typeparams (e.g. Val{:foo})
• I think this also has some other changes, to support other things, but not sure what they are. Hopefully code-cov will tell me if I have added any features i forget about.
• Some restructuring.

Here are tests that failed before

julia> @test_signature empty_tuple_constraint(x::Tuple{}) = 2
Test Failed at REPL[48]:5
  Expression: target[:args] == get(candidate, :args, nothing)
   Evaluated: Any[:(x::Tuple{})] == Expr[:(x::Tuple)]
ERROR: There was an error during testing

julia> @test_signature qualified_constraint(x::Base.Threads.SpinLock) = 2
Test Failed at REPL[48]:5
  Expression: target[:args] == get(candidate, :args, nothing)
   Evaluated: Any[:(x::Base.Threads.SpinLock)] == Expr[:(x::var"Base.Threads.SpinLock")]
ERROR: There was an error during testing

julia> @test_signature f_symbol_param(x::Val{:foobar}) where T = 2x
Test Failed at REPL[48]:5
  Expression: target[:args] == get(candidate, :args, nothing)
   Evaluated: Any[:(x::Val{:foobar})] == Expr[:(x::Val{foobar})]
ERROR: There was an error during testing

[oxinabox]

Added these literals rather than name_of_type(x)=x so i can be sure nothing slipped through

[codecov]

Codecov Report

Merging #12 into master will increase coverage by 0.11%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #12      +/-   ##
==========================================
+ Coverage   99.31%   99.42%   +0.11%     
==========================================
  Files           4        4              
  Lines         145      173      +28     
==========================================
+ Hits          144      172      +28     
  Misses          1        1              

Impacted Files	Coverage Δ
src/method.jl	100.00% <100.00%> (ø)
src/function.jl	98.79% <0.00%> (+0.06%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4b175b0...cba2e15. Read the comment docs.

[oxinabox]

This rename is because it doesn't return list of where parameters
just formats 1 of them

[mattBrzezinski]

I feel like we need something a bit more descriptive here.

[oxinabox]

I can't think of a better one.
Its not unsual to have:

foos(xs) = map(foo1, xs)

This is the processing for single thing that kind of maps to the multiple things that whereparams is trying to process.

[nickrobinson251]

Its not unsual to have:
foos(xs) = map(foo1, xs)

I've never seen this before (in Julia, Python or Scala that i've seen)

[nickrobinson251]

but i've no better name suggestion - sorry

[mattBrzezinski]

I've never seen this before (in Julia, Python or Scala that i've seen)

Just a different way of doing things.

julia> @time [x^2 for x in [1,2,3]]
  0.022936 seconds (41.61 k allocations: 2.128 MiB)
3-element Array{Int64,1}:
 1
 4
 9

julia> @time map(x->x^2, [1,2,3])
  0.023335 seconds (48.81 k allocations: 2.499 MiB)
3-element Array{Int64,1}:
 1
 4
 9

[nickrobinson251]

I meant the 1 suffix naming convention

[mattBrzezinski]

That makes more sense 😆

[mattBrzezinski]

I feel like we need something a bit more descriptive here.

[oxinabox]

The better name was in reprospect obvious
as long as we are ok with the idea of no constraints being considered a form of constraint

[mattBrzezinski]

I'm logging off for the night, but I will take a look at this first thing tomorrow morning!