towards release 2.5 to-do list

[ahbarnett]

• cuda type-1 output-driven (meth=3) cuFINUFFT ouput driven #661
• adjoints Add the capability to do adjoint transforms #633 some issues of API to resolve (see discussion on PR).
• constness For illustration only: add more const-ness and access control #647
• consider whether the big array fwbatch should be malloced at plan or at execute. #480 allocating fwbatch workspace on the fly each exec call. Always do it (saves RAM in multiple plans).
• kernel math on CPU: switch onedim_ft to use kernel
• on-the-fly kernel coeff generation - use Robert's inverse Vandermonde evaluator to avoid LSQ solve lin alg dependency?
• finally replace GPL-licensed Gauss quadrature code with Apache-2-compatible code.  #689