Tweaks

Author: gaurav-arya

docs/src/implementations.md

@@ -32,9 +32,12 @@ To define a new FFT implementation in your own module, you should
 
 * You can also define similar methods of `plan_rfft` and `plan_brfft` for real-input FFTs.
 
-* We offer an experimental `AdjointStyle` trait to enable automatic computation of adjoint plans via [`Base.adjoint`](@ref),
-(which `AbstractFFTs` uses to implement reverse-mode differentiation rules). To support adjoints in a new plan, define the trait `AbstractFFTs.AdjointStyle(::MyPlan)`. This should return a subtype of `AS <: AbstractFFTs.AdjointStyle` supporting `AbstractFFTs.adjoint_mul(::AdjointPlan, ::AbstractArray, ::AS)`. `AbstractFFTs` pre-implements [`AbstractFFTs.FFTAdjointStyle`](@ref), [`AbstractFFTs.RFFTAdjointStyle`](@ref),
-[`AbstractFFTs.BRFFTAdjointStyle`](@ref), and [`AbstractFFTs.UnitaryAdjointStyle`](@ref).
+* To enable automatic computation of adjoint plans via [`Base.adjoint`](@ref), `AbstractFFTs` offers an experimental `AdjointStyle` trait.
+  To support adjoints in a new plan, define the trait `AbstractFFTs.AdjointStyle(::MyPlan)`, returning an instance of type
+  `AS <: AbstractFFTs.AdjointStyle` supporting `AbstractFFTs.adjoint_mul(::AdjointPlan, ::AbstractArray, ::AS)`.
+
+  `AbstractFFTs` pre-implements the following adjoint styles: [`AbstractFFTs.FFTAdjointStyle`](@ref),
+  [`AbstractFFTs.RFFTAdjointStyle`](@ref), [`AbstractFFTs.BRFFTAdjointStyle`](@ref), [`AbstractFFTs.UnitaryAdjointStyle`](@ref).
 
 The normalization convention for your FFT should be that it computes ``y_k = \sum_j x_j \exp(-2\pi i j k/n)`` for a transform of
 length ``n``, and the "backwards" (unnormalized inverse) transform computes the same thing but with ``\exp(+2\pi i jk/n)``.

src/definitions.jl

@@ -608,7 +608,7 @@ to exploit its conjugate symmetry (see [`rfft`](@ref)).
 struct RFFTAdjointStyle <: AdjointStyle end 
 
 """
-    BRFFTAdjointStyle(d::Dim)
+    BRFFTAdjointStyle(d::Int)
 
 Projection style for complex to real discrete Fourier transforms, for plans that 
 expect an input with a halved dimension analogously to [`irfft`](@ref), where `d`