Faster evaluation in normalized spaces

[jishnub]

This was hitting the fallback method previously, which was slow.
On master

julia> E = Evaluation(NormalizedUltraspherical(1), 0.5);

julia> @btime $E[1,1];
  2.391 μs (23 allocations: 704 bytes)

This PR

julia> @btime $E[1,1];
  183.369 ns (2 allocations: 128 bytes)

[codecov]

Codecov Report

Patch coverage: 92.85% and project coverage change: +0.88 🎉

Comparison is base (441f970) 70.36% compared to head (db4cad5) 71.24%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #269      +/-   ##
==========================================
+ Coverage   70.36%   71.24%   +0.88%     
==========================================
  Files          28       28              
  Lines        2686     2699      +13     
==========================================
+ Hits         1890     1923      +33     
+ Misses        796      776      -20     

Impacted Files	Coverage Δ
src/Spaces/Ultraspherical/DirichletSpace.jl	48.64% <0.00%> (ø)
src/Spaces/CurveSpace.jl	25.58% <100.00%> (+25.58%)	⬆️
src/Spaces/PolynomialSpace.jl	93.39% <100.00%> (+0.22%)	⬆️

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.