Shorter stacktrace

[YingboMa]

Now the stacktrace looks like this instead of the whole type printing with
bytes.

julia> f1(du,u,p,t)
ERROR:
Stacktrace:
 [1] error() at ./error.jl:42
 [2] macro expansion at ./REPL[36]:2 [inlined]
 [3] macro expansion at /Users/scheme/.julia/packages/RuntimeGeneratedFunctions/fIcZp/src/RuntimeGeneratedFunctions.jl:80 [inlined]
 [4] macro expansion at ./none:0 [inlined]
 [5] generated_callfunc at ./none:0 [inlined]
 [6] (::RuntimeGeneratedFunction{(:_du, :_u, :_p, :_t)})(::Array{Float64,1}, ::Array{Float64,1}, ::Nothing, ::Nothing) at /Users/scheme/.julia/dev/RuntimeGeneratedFunctions/src/RuntimeGeneratedFunctions.jl:77
 [7] top-level scope at REPL[42]:1

julia> typeof(f1)
RuntimeGeneratedFunction{var"#_RuntimeGeneratedFunctions_ModTag", (0x47, 0x4a, 0x02, 0xf7, 0xc0, 0xd0, 0xdf, 0x8d, 0x8c, 0xac, 0x18, 0x52, 0x0f, 0x01, 0xbf, 0x94, 0xc0, 0xf6, 0xf1, 0x73, 0x89, 0x1b, 0xe5, 0x1d, 0x94, 0x75, 0xa9, 0xbf, 0x06, 0xa0, 0x96, 0xd5, 0xbb, 0xd3, 0xfa, 0x0d, 0xfd, 0xb4, 0x1f, 0xec, 0xfd, 0xb3, 0x6a, 0x05, 0x53, 0x92, 0x3a, 0x59, 0x0e, 0x35, 0x51, 0x05, 0xe8, 0xd4, 0x28, 0xec, 0x95, 0x97, 0x9b, 0x78, 0xe0, 0x1b, 0xc9, 0x9c), (:_du, :_u, :_p, :_t)}

[codecov]

Codecov Report

Merging #15 into master will decrease coverage by 8.18%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master      #15      +/-   ##
==========================================
- Coverage   90.00%   81.81%   -8.19%     
==========================================
  Files           1        1              
  Lines          40       44       +4     
==========================================
  Hits           36       36              
- Misses          4        8       +4     

Impacted Files	Coverage Δ
src/RuntimeGeneratedFunctions.jl	81.81% <ø> (-8.19%)	⬇️

---

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 2953c18...9b9426e. Read the comment docs.

[ChrisRackauckas]

Thank god.

[c42f]

Note that overriding show(::IO, ::Type{SomeType}) has a bit of a bad reputation of causing instability. See for example: PainterQubits/Unitful.jl#321

Unfortunately I don't know exactly what's wrong with doing this, I just remember it being discussed on a few occasions.

[devmotion]

Interesting 👍 Maybe it would be helpful to change the order of the type parameters such that the argnames are displayed more prominently in stacktraces, instead of implementing show(::IO, ::Type)?

[c42f]

Could be done. Or I thought sha512 seems like overkill anyway and we could just use sha1. And also shorten the tag name to get something like

julia> typeof(@RuntimeGeneratedFunction(:((x,)->x+1)))
RuntimeGeneratedFunctions.RuntimeGeneratedFunction{var"#_RGF_ModTag",(0x19cc2a7a, 0xeba7d29f, 0x2c764791, 0x3c9e5a40, 0xfadc08ad),(:x,)}

[c42f]

change the order of the type parameters such that the argnames are displayed more prominently in stacktraces

I tried this in #16, along with the other ideas to make the types a bit less horrible.