Fix #38

[dlfivefifty]

No description provided.

[dlfivefifty]

@kburns

This now works:

julia> W = Weighted(ChebyshevT())
Weighted(ChebyshevT())

julia> x = axes(W, 1)
Inclusion(-1.0 .. 1.0 (Chebyshev))

julia> f = expand(W, x -> exp(x) / sqrt(1-x^2))
Weighted(ChebyshevT()) * [1.2660658777520084, 1.1303182079849703, 0.27149533953407656, 0.044336849848663776, 0.0054742404420936785, 0.0005429263119139354, 4.497732295427654e-5, 3.198436462529006e-6, 1.992124804817033e-7, 1.1036771880698548e-8  …  ]

julia> log.(abs.(0.3 .- x')) * f
-3.4376226814884023

[dlfivefifty]

Here's a 10k by 10k sample from one interval to another:

lia> n = 10_000; @time logkernel(Weighted(T), range(2,3,n))[:,1:n]
  0.473830 seconds (26 allocations: 1.491 GiB, 3.24% gc time)
10000×10000 Matrix{Float64}:
 1.95976  -0.841787  -0.112778   -0.0201459   -0.00404855   …  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0
 1.95994  -0.841739  -0.112765   -0.0201424   -0.00404762      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0
 1.96012  -0.84169   -0.112752   -0.0201389   -0.00404668      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0
 1.9603   -0.841641  -0.112739   -0.0201354   -0.00404575      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0
 1.96048  -0.841593  -0.112726   -0.0201319   -0.00404481      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0
 1.96067  -0.841544  -0.112713   -0.0201284   -0.00404388   …  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0
 ⋮                                                          ⋱                 ⋮                        ⋮                   
 3.3598   -0.539088  -0.046253   -0.00529126  -0.000680973  …  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0
 3.35991  -0.539069  -0.0462497  -0.00529069  -0.000680877     0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0
 3.36003  -0.53905   -0.0462465  -0.00529013  -0.000680781     0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0
 3.36014  -0.539031  -0.0462432  -0.00528957  -0.000680684     0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0
 3.36025  -0.539012  -0.0462399  -0.00528901  -0.000680588     0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0