add documentation for overwriting behaviour of in-place application of real-input fft plans

[trahflow]

As noted in JuliaMath/FFTW.jl#158, application of inverse plans of real-input ffts via mul! or ldiv! potentially overwrites the input array.
This might be quite unexpected to many people (including me), especially since this is not the case for the complex-input versions.

This PR is an attempt at documenting this behaviour.

[codecov]

Codecov Report

Patch and project coverage have no change.

Comparison is base (a52b2ab) 87.08% compared to head (45f3531) 87.08%.

Additional details and impacted files

@@           Coverage Diff           @@
##           master      #88   +/-   ##
=======================================
  Coverage   87.08%   87.08%           
=======================================
  Files           3        3           
  Lines         209      209           
=======================================
  Hits          182      182           
  Misses         27       27           

Impacted Files	Coverage Δ
src/definitions.jl	74.03% <ø> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

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

[navidcy]

I thought that irfft destroying its input was an FFTW library thing and not an issue of FFTW.jl. Right?

[trahflow]

Yes I think it's an FFTW library thing.
Still worth documenting here, I believe.

[trahflow]

bump on review

[navidcy]

lgtm

is it only for multidimensional transforms that the overwrite occurs?

[trahflow]

is it only for multidimensional transforms that the overwrite occurs?

at least according to my observations, yes.