cuda.parallel: Address Python<3.9 TODOs

[shwina]

Description

Following #3293, we can now use newer Python syntax and features (no newer than Python 3.9).

Checklist

• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[github-actions]

🟩 CI finished in 23m 17s: Pass: 100%/1 | Total: 23m 17s | Avg: 23m 17s | Max: 23m 17s

• 🟩 python: Pass: 100%/1 | Total: 23m 17s | Avg: 23m 17s | Max: 23m 17s

  🟩 cpu
    🟩 amd64              Pass: 100%/1   | Total: 23m 17s | Avg: 23m 17s | Max: 23m 17s
  🟩 ctk
    🟩 12.6               Pass: 100%/1   | Total: 23m 17s | Avg: 23m 17s | Max: 23m 17s
  🟩 cudacxx
    🟩 nvcc12.6           Pass: 100%/1   | Total: 23m 17s | Avg: 23m 17s | Max: 23m 17s
  🟩 cudacxx_family
    🟩 nvcc               Pass: 100%/1   | Total: 23m 17s | Avg: 23m 17s | Max: 23m 17s
  🟩 cxx
    🟩 GCC13              Pass: 100%/1   | Total: 23m 17s | Avg: 23m 17s | Max: 23m 17s
  🟩 cxx_family
    🟩 GCC                Pass: 100%/1   | Total: 23m 17s | Avg: 23m 17s | Max: 23m 17s
  🟩 gpu
    🟩 v100               Pass: 100%/1   | Total: 23m 17s | Avg: 23m 17s | Max: 23m 17s
  🟩 jobs
    🟩 Test               Pass: 100%/1   | Total: 23m 17s | Avg: 23m 17s | Max: 23m 17s
  

👃 Inspect Changes

Modifications in project?

	Project
	CCCL Infrastructure
	libcu++
	CUB
	Thrust
	CUDA Experimental
+/-	python
	CCCL C Parallel Library
	Catch2Helper

Modifications in project or dependencies?

	Project
	CCCL Infrastructure
	libcu++
	CUB
	Thrust
	CUDA Experimental
+/-	python
	CCCL C Parallel Library
	Catch2Helper

🏃‍ Runner counts (total jobs: 1)

#	Runner
1	linux-amd64-gpu-v100-latest-1

[NaderAlAwar]

Looks good!