diff --git a/src/Evaluate.jl b/src/Evaluate.jl
index 5b5df1a8..f139fc03 100644
--- a/src/Evaluate.jl
+++ b/src/Evaluate.jl
@@ -81,7 +81,7 @@ function eval_tree_array(
             error("Please load the LoopVectorization.jl package to use this feature.")
     end
     if (v_turbo isa Val{true} || v_bumper isa Val{true}) && !(T <: Number)
-        error("Bumper feature only works with numbers")
+        error("Bumper and LoopVectorization features are only compatible with numeric element types")
     end
     if v_bumper isa Val{true}
         return bumper_eval_tree_array(tree, cX, operators, v_turbo)
@@ -97,7 +97,7 @@ function eval_tree_array(
     operators::OperatorEnum;
     kws...
 ) where {T}
-    return eval_tree_array(tree, reshape(cX, (size(cX)[1], 1))::AbstractMatrix{T}, operators; kws...)
+    return eval_tree_array(tree, reshape(cX, (size(cX, 1), 1)), operators; kws...)
 end
 
 function eval_tree_array(
diff --git a/test/test_non_number_eval_tree_array.jl b/test/test_non_number_eval_tree_array.jl
index f7ec071d..088dba02 100644
--- a/test/test_non_number_eval_tree_array.jl
+++ b/test/test_non_number_eval_tree_array.jl
@@ -8,7 +8,7 @@ struct SVM{T}
     scalar :: T
     vector :: Vector{T}
     matrix :: Matrix{T}
-    SVM{T}() where {T} = new(Int8(0), zero(T), T[], T[;;])
+    SVM{T}() where {T} = new(Int8(0), zero(T), T[], Array{T}(undef, 0, 0))
     SVM{T}(scalar :: W) where {T, W <: Number} = new(Int8(0), Base.convert(T, scalar), T[], T[;;])
     SVM{T}(vector :: Vector{W}) where {T, W <: Number} = new(Int8(1), zero(T), Vector{T}(vector) , T[;;])
     SVM{T}(matrix :: Matrix{W}) where {T, W <: Number} = new(Int8(2), zero(T), T[], Matrix{T}(matrix))
@@ -29,8 +29,7 @@ end
 function Base.:(==)(x::SVM{T}, y::SVM{T}) where T
     if x.dims !== y.dims
         return false
-    end
-    if x.dims == 0
+    elseif x.dims == 0
         return x.scalar == y.scalar
     elseif val.dims == 1
         return x.vector == y.vector
@@ -61,7 +60,7 @@ Base.invokelatest(() -> begin
     @test !hasmethod(a, Tuple{Node{SVM{Float32}}, Node{SVM{Float32}}})
 
     tree = a(Node{SVM{Float64}}(; feature=1), SVM{Float64}(3.0))
-    results = tree([SVM{Float64}(1.0);; SVM{Float64}(2.0);; SVM{Float64}(3.0)])
+    results = tree([SVM{Float64}(1.0) SVM{Float64}(2.0) SVM{Float64}(3.0)])
     @test results == [SVM{Float64}(4), SVM{Float64}(5), SVM{Float64}(6)]