Faster gaussjacobi and gausslobatto, and refactoring

[hyrodium]

Benchmark

Current master branch↓

This branch↓

About twice faster and less memory usage:tada:

Type unstability

There were type unstabilities in some functions.
For example:

But, in this branch, this problem is solved!

The trap of Adjoint

The type unstability was because of Adjoint.
Here's a small example:

Related issue in Julia↓
JuliaLang/julia#32289

Refactored things

• Update error-handling
• More comments

[codecov]

Codecov Report

Merging #85 (1a975de) into master (ba2da14) will increase coverage by 1.36%.
The diff coverage is 99.47%.

@@            Coverage Diff             @@
##           master      #85      +/-   ##
==========================================
+ Coverage   94.71%   96.08%   +1.36%     
==========================================
  Files           9        9              
  Lines        1116     1123       +7     
==========================================
+ Hits         1057     1079      +22     
+ Misses         59       44      -15     

Impacted Files	Coverage Δ
src/FastGaussQuadrature.jl	83.33% <ø> (ø)
src/gaussjacobi.jl	97.10% <99.02%> (+0.39%)	⬆️
src/besselroots.jl	92.68% <100.00%> (+2.68%)	⬆️
src/gausschebyshev.jl	100.00% <100.00%> (+10.00%)	⬆️
src/gausshermite.jl	100.00% <100.00%> (+3.19%)	⬆️
src/gausslaguerre.jl	92.04% <100.00%> (+0.85%)	⬆️
src/gausslegendre.jl	98.07% <100.00%> (+0.98%)	⬆️
src/gausslobatto.jl	100.00% <100.00%> (+6.66%)	⬆️
src/gaussradau.jl	100.00% <100.00%> (ø)

---

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 ba2da14...1a975de. Read the comment docs.