Correct some typos

Signed-off-by: Alexander Seiler <seileralex@gmail.com>

Author: goggle

src/rulesets/Base/array.jl

@@ -72,7 +72,7 @@ materialises each zero `ẋ` to be `zero(x)`.
 _instantiate_zeros(ẋs, xs) = map(_i_zero, ẋs, xs)
 _i_zero(ẋ, x) = ẋ
 _i_zero(ẋ::AbstractZero, x) = zero(x)
-# Possibly this won't work for partly non-diff arrays, sometihng like `gradient(x -> ["abc", x][end], 1)`
+# Possibly this won't work for partly non-diff arrays, something like `gradient(x -> ["abc", x][end], 1)`
 # may give a MethodError for `zero` but won't be wrong.
 
 # Fast paths. Should it also collapse all-Zero cases?

src/rulesets/Base/base.jl

@@ -175,7 +175,7 @@ end
 
 ChainRulesCore.derivatives_given_output(Ω, ::typeof(identity), x) = tuple(tuple(true))
 
-# rouding related,
+# rounding related,
 # we use `zero` rather than `ZeroTangent()` for scalar, and avoids issues with map etc
 @scalar_rule round(x) zero(x)
 @scalar_rule floor(x) zero(x)

src/rulesets/Base/fastmath_able.jl

@@ -1,7 +1,7 @@
 let
     # Include inside this quote any rules that should have FastMath versions
     # IMPORTANT:
-    # Do not add any rules here for functions that do not have varients in Base.FastMath
+    # Do not add any rules here for functions that do not have variants in Base.FastMath
     # e.g. do not add `foo` unless `Base.FastMath.foo_fast` exists.
     fastable_ast = quote
         #  Trig-Basics

src/rulesets/Base/indexing.jl

@@ -11,7 +11,7 @@ function frule((_, ẋ), ::typeof(getindex), x::Tuple, i)
     return y, Tangent{typeof(y)}(ẋ[i]...)
 end
 
-"for a given typle type, returns a Val{N} where N is the length of the tuple"
+"for a given tuple type, returns a Val{N} where N is the length of the tuple"
 _tuple_N(::Type{<:Tuple{Vararg{Any, N}}}) where {N} = Val(N)
 
 function rrule(::typeof(getindex), x::T, i::Integer) where {T<:Tuple}

src/rulesets/Base/mapreduce.jl

@@ -194,7 +194,7 @@ end
 
 # Fix dispatch for this pidgeon-hole optimization,
 # Rules with RuleConfig dispatch with priority over without (regardless of other args).
-# and if we don't specify what do do for one that HasReverseMode then it is ambigious
+# and if we don't specify what do do for one that HasReverseMode then it is ambiguous
 for Config in (RuleConfig, RuleConfig{>:HasReverseMode})
     @eval function rrule(
         ::$Config, ::typeof(sum), ::typeof(abs2), x::AbstractArray{T}; dims=:,

src/rulesets/LinearAlgebra/dense.jl

@@ -120,7 +120,7 @@ end
 
 function frule((_, Δx), ::typeof(det), x::AbstractMatrix)
     Ω = det(x)
-    # TODO Performance optimization: probably there is an efficent
+    # TODO Performance optimization: probably there is an efficient
     # way to compute this trace without during the full compution within
     return Ω, Ω * tr(x \ Δx)
 end

test/rulesets/Base/fastmath_able.jl

@@ -1,4 +1,4 @@
-# Add tests to the quote for functions with  FastMath varients.
+# Add tests to the quote for functions with  FastMath variants.
 function jacobian_via_frule(f,z)
     du_dx, dv_dx = reim(frule((ZeroTangent(), 1),f,z)[2])
     du_dy, dv_dy = reim(frule((ZeroTangent(),im),f,z)[2])
@@ -38,7 +38,7 @@ function complex_jacobian_test(f, z)
 end
 
 # IMPORTANT:
-# Do not add any tests here for functions that do not have varients in Base.FastMath
+# Do not add any tests here for functions that do not have variants in Base.FastMath
 # e.g. do not add `foo` unless `Base.FastMath.foo_fast` exists.
 const FASTABLE_AST = quote
     @testset "Trig" begin

test/rulesets/Base/indexing.jl

@@ -40,7 +40,7 @@
             test_rrule(getindex, x, CartesianIndex(2, 3))
         end
 
-        @testset "slice/index postions" begin
+        @testset "slice/index positions" begin
             test_rrule(getindex, x, 2:3)
             test_rrule(getindex, x, 3:-1:2)
             test_rrule(getindex, x, [3,2])

test/rulesets/LinearAlgebra/factorization.jl

@@ -510,7 +510,7 @@ end
             @testset "$op" for op in (det, logdet)
                 @testset "$T" for T in (Float64, ComplexF64)
                     n = 5
-                    # rand (not randn) so det will be postive, so logdet will be defined
+                    # rand (not randn) so det will be positive, so logdet will be defined
                     A = 3 * rand(T, (n, n))
                     X = Cholesky(A * A' + I, p, 0)
                     X̄_acc = Tangent{typeof(X)}(; factors=Diagonal(randn(T, n))) # sensitivity is always a diagonal
@@ -547,7 +547,7 @@ end
                 C = cholesky(A)
                 B = randn(T, sz)
                 # because the rule calls the rrule for getproperty, its rrule is not
-                # completely type-inferrable
+                # completely type-inferable
                 test_rrule(\, C, B; check_inferred=false)
             end
         end
@@ -559,7 +559,7 @@ end
                 C = cholesky(A)
                 B = randn(T, 5, n)
                 # because the rule calls the rrule for getproperty, its rrule is not
-                # completely type-inferrable
+                # completely type-inferable
                 test_rrule(/, B, C; check_inferred=false)
             end
         end

test/rulesets/LinearAlgebra/matfun.jl

@@ -47,7 +47,7 @@
 
             A = randn(T, n, n)
             A *= nrm / opnorm(A, 1)
-            # rrule is not inferrable, but pullback should be
+            # rrule is not inferable, but pullback should be
             tols = nrm == 0.1 ? (atol=1e-8, rtol=1e-8) : NamedTuple()
             test_rrule(exp, A; check_inferred=false, tols...)
             Y, back = rrule(exp, A)