Fix GPU ArrayPartition any/all: load AnyAll subpackage in GPU tests

The GPU test uses `all(pA .== true)` on ArrayPartition with CuArrays,
which triggers scalar indexing via the AbstractArray element-by-element
fallback. Fix by loading RecursiveArrayToolsArrayPartitionAnyAll in the
GPU test environment, which provides partition-level any/all that
dispatches to GPU kernels per sub-array.

Co-Authored-By: Chris Rackauckas <accounts@chrisrackauckas.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: ChrisRackauckas

test/gpu/Project.toml

@@ -1,6 +1,7 @@
 [deps]
 CUDA = "052768ef-5323-5732-b1bb-66c8b64840ba"
 KernelAbstractions = "63c18a36-062a-441e-b654-da1e3ab1ce7c"
+RecursiveArrayToolsArrayPartitionAnyAll = "172d604e-c495-4f00-97bf-d70957099afa"
 
 [compat]
 CUDA = "3.12, 4, 5"

test/gpu/arraypartition_gpu.jl

@@ -1,4 +1,4 @@
-using RecursiveArrayTools, ArrayInterface, CUDA, Adapt, Test
+using RecursiveArrayTools, RecursiveArrayToolsArrayPartitionAnyAll, ArrayInterface, CUDA, Adapt, Test
 CUDA.allowscalar(false)
 
 # Test indexing with colon

test/runtests.jl

@@ -17,6 +17,8 @@ end
 function activate_gpu_env()
     Pkg.activate("gpu")
     Pkg.develop(PackageSpec(path = dirname(@__DIR__)))
+    Pkg.develop(PackageSpec(
+        path = joinpath(dirname(@__DIR__), "lib", "RecursiveArrayToolsArrayPartitionAnyAll")))
     return Pkg.instantiate()
 end