Add FFT convolution kernels

[wheeheee]

This is a draft for adding FFT convolution kernels (that are also plans), which are useful for repeated convolutions.
A lot has been adapted from the code here, courtesy of @JanJereczek
Here's what I have done so far:

• created an abstract type for FFT convolution kernels tentatively named FFTConvKernel, subtypes R/CFFTConvKernel
• working code for real and complex FFT convolution kernels and Array, FFTW flag defaults to FFTW.MEASURE
• in the process, found that we don't check if the output indices are valid (i.e. out[output_indices] doesn't throw) before doing work except for :direct, so I added that. Don't remember if we have discussed this before, so left it in just in case
• IIUC the kernel/plans can be reused for smaller x, maybe larger depending on nffts
• not very tested, will add tests later if the feature is wanted

A typical workflow goes like this:

1. Calculate outsize, the size of the intended output size(out) .+ size(ker) .- 1
2. Create kernel with RFFTConvKernel or CFFTConvKernel(kernel, outsize)
3. Run something like conv!(out[i], arrs[i], kernel) in a loop

Feedback would be much appreciated

[codecov]

Codecov Report

Attention: Patch coverage is 15.21739% with 39 lines in your changes missing coverage. Please review.

Project coverage is 96.98%. Comparing base (0660d55) to head (f26843a).

Files with missing lines	Patch %	Lines
src/dspbase.jl	15.21%	39 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #638      +/-   ##
==========================================
- Coverage   98.13%   96.98%   -1.16%     
==========================================
  Files          19       19              
  Lines        3277     3320      +43     
==========================================
+ Hits         3216     3220       +4     
- Misses         61      100      +39     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.