Sheehan Olver

Results 308 issues of Sheehan Olver

There should be nothing in the algorithms that depend on the parameters being Float64: BigFloat should in principle be allowed, though I suppose there may be assumptions on eps(T) for...

feature-request

```julia julia> CLArray{Float32,2}(10,10) ERROR: MethodError: CLArrays.CLArray{Float32,2}(::Int64, ::Int64) is ambiguous. Candidates: (::Type{CLArrays.CLArray{T,N}})(size, ptr) where {T, N} in CLArrays at /Users/solver/.julia/v0.6/CLArrays/src/array.jl:12 (T::Type{#s9} where #s9

I'm not sure if this should work, but nevertheless it throws an error: ```julia julia> A = CLArray{Float32}(randn(100,100)); julia> b = CLArray{Float32}(randn(10)); julia> view(A,1:10,1:10)*b ERROR: conversion to pointer not defined...

I'm in the process of replicating the `zeros` and `eye` convenience constructors in BandedMatrices.jl and other packages, see https://github.com/JuliaMatrices/BandedMatrices.jl/issues/42 Personally, the semantics of `zeros(CLArray{Float32},n,n)` seems wrong as it looks like...

I tried running Femtocleaner (both by creating an Issue and by bumping REQUIRE) on https://github.com/JuliaApproximation/SingularIntegralEquations.jl and I didn't receive a PR.

`SingularIntegral(S,0)` is _really_ different than `SingularIntegral(S,1)`: it's real valued, is not `(x-z)^0`, etc. I think `LogKernel` and `SingularIntegral` should be different types.

```julia n = 5; a ,b = 0.1, 0.2 _₂F₁(-n,n+a+b+1,a+1,z+eps()*im) ```

https://github.com/JuliaApproximation/SingularIntegralEquations.jl/blob/1e4cc4160447d0f638bda0e63d4a3e527b391821/src/singfuncauchy.jl#L170 `logkernel(1/sqrt(1-x^2),-5.0)`

```julia x = Fun(0 .. 20.0) α = 0.1 cauchy(x^α*exp(-x), 2.0+im) ```