Remove mention of (now unsupported) kwargs (#162)

Author: gdalle

docs/src/faq.md

@@ -58,7 +58,7 @@ See the examples for a demonstration.
 
 ### Multiple inputs | Derivatives not needed
 
-If your forward mapping (or conditions) takes multiple inputs but you don't care about derivatives, then you can add further positional and keyword arguments beyond `x`.
+If your forward mapping (or conditions) takes multiple inputs but you don't care about derivatives, then you can add further positional arguments beyond `x`.
 It is important to make sure that the forward mapping and conditions accept the same set of arguments, even if each of these functions only uses a subset of them.
 
 ```julia

ext/ImplicitDifferentiationChainRulesCoreExt.jl

@@ -14,13 +14,9 @@ using ImplicitDifferentiation:
 ImplicitDifferentiation.chainrules_suggested_backend(rc::RuleConfig) = AutoChainRules(rc)
 
 function ChainRulesCore.rrule(
-    rc::RuleConfig,
-    implicit::ImplicitFunction,
-    x::AbstractVector,
-    args::Vararg{Any,N};
-    kwargs...,
+    rc::RuleConfig, implicit::ImplicitFunction, x::AbstractVector, args::Vararg{Any,N};
 ) where {N}
-    y, z = implicit(x, args...; kwargs...)
+    y, z = implicit(x, args...)
 
     suggested_backend = chainrules_suggested_backend(rc)
     Aᵀ = build_Aᵀ(implicit, x, y, z, args...; suggested_backend)

ext/ImplicitDifferentiationForwardDiffExt.jl

@@ -5,10 +5,10 @@ using ForwardDiff: Dual, Partials, partials, value
 using ImplicitDifferentiation: ImplicitFunction, build_A, build_B
 
 function (implicit::ImplicitFunction)(
-    x_and_dx::AbstractVector{Dual{T,R,N}}, args...; kwargs...
+    x_and_dx::AbstractVector{Dual{T,R,N}}, args...
 ) where {T,R,N}
     x = value.(x_and_dx)
-    y, z = implicit(x, args...; kwargs...)
+    y, z = implicit(x, args...)
 
     suggested_backend = AutoForwardDiff()
     A = build_A(implicit, x, y, z, args...; suggested_backend)