fix completion signature computation of starts_on, work around gcc9 ICE

[ericniebler]

Description

the cudax execution library is causing gcc9 to ICE in nightly CI. some implementation details of cudax::execution::get_completion_signatures were tripping up gcc9. this pr simplifies the code enough for gcc9 to accept it.

investigating this let me to an issue in how the starts_on sender is computing its completion signatures. this pr addresses that problem as well, which was contributing to the ICE.

Checklist

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

[copy-pr-bot]

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

[ericniebler]

/ok to test 19e8a10

[github-actions]

🟩 CI finished in 28m 34s: Pass: 100%/32 | Total: 4h 20m | Avg: 8m 08s | Max: 20m 27s | Hits: 83%/15498

• 🟩 cudax: Pass: 100%/28 | Total: 3h 55m | Avg: 8m 25s | Max: 20m 27s | Hits: 83%/15498

  🟩 cpu
    🟩 amd64              Pass: 100%/24  | Total:  3h 32m | Avg:  8m 50s | Max: 20m 27s | Hits:  83%/13110 
    🟩 arm64              Pass: 100%/4   | Total: 23m 28s | Avg:  5m 52s | Max:  6m 36s | Hits:  83%/2388  
  🟩 ctk
    🟩 12.0               Pass: 100%/3   | Total: 25m 07s | Avg:  8m 22s | Max: 14m 19s | Hits:  78%/1485  
    🟩 12.9               Pass: 100%/25  | Total:  3h 30m | Avg:  8m 25s | Max: 20m 27s | Hits:  83%/14013 
  🟩 cudacxx
    🟩 nvcc12.0           Pass: 100%/3   | Total: 25m 07s | Avg:  8m 22s | Max: 14m 19s | Hits:  78%/1485  
    🟩 nvcc12.9           Pass: 100%/25  | Total:  3h 30m | Avg:  8m 25s | Max: 20m 27s | Hits:  83%/14013 
  🟩 cudacxx_family
    🟩 nvcc               Pass: 100%/28  | Total:  3h 55m | Avg:  8m 25s | Max: 20m 27s | Hits:  83%/15498 
  🟩 cxx
    🟩 Clang14            Pass: 100%/2   | Total: 10m 42s | Avg:  5m 21s | Max:  5m 57s | Hits:  83%/1196  
    🟩 Clang15            Pass: 100%/1   | Total:  6m 48s | Avg:  6m 48s | Max:  6m 48s | Hits:  83%/597   
    🟩 Clang16            Pass: 100%/1   | Total:  5m 54s | Avg:  5m 54s | Max:  5m 54s | Hits:  83%/597   
    🟩 Clang17            Pass: 100%/1   | Total:  6m 15s | Avg:  6m 15s | Max:  6m 15s | Hits:  83%/597   
    🟩 Clang18            Pass: 100%/1   | Total:  6m 27s | Avg:  6m 27s | Max:  6m 27s | Hits:  83%/597   
    🟩 Clang19            Pass: 100%/4   | Total: 37m 26s | Avg:  9m 21s | Max: 20m 27s | Hits:  87%/2388  
    🟩 GCC10              Pass: 100%/2   | Total: 12m 08s | Avg:  6m 04s | Max:  6m 05s | Hits:  82%/1196  
    🟩 GCC11              Pass: 100%/1   | Total:  6m 38s | Avg:  6m 38s | Max:  6m 38s | Hits:  82%/597   
    🟩 GCC12              Pass: 100%/1   | Total:  6m 39s | Avg:  6m 39s | Max:  6m 39s | Hits:  82%/597   
    🟩 GCC13              Pass: 100%/8   | Total: 57m 36s | Avg:  7m 12s | Max: 12m 28s | Hits:  87%/4776  
    🟩 MSVC14.39          Pass: 100%/1   | Total: 14m 19s | Avg: 14m 19s | Max: 14m 19s | Hits:  62%/291   
    🟩 MSVC14.43          Pass: 100%/3   | Total: 44m 19s | Avg: 14m 46s | Max: 16m 16s | Hits:  62%/879   
    🟩 NVHPC25.7          Pass: 100%/2   | Total: 20m 39s | Avg: 10m 19s | Max: 10m 20s | Hits:  80%/1190  
  🟩 cxx_family
    🟩 Clang              Pass: 100%/10  | Total:  1h 13m | Avg:  7m 21s | Max: 20m 27s | Hits:  84%/5972  
    🟩 GCC                Pass: 100%/12  | Total:  1h 23m | Avg:  6m 55s | Max: 12m 28s | Hits:  85%/7166  
    🟩 MSVC               Pass: 100%/4   | Total: 58m 38s | Avg: 14m 39s | Max: 16m 16s | Hits:  62%/1170  
    🟩 NVHPC              Pass: 100%/2   | Total: 20m 39s | Avg: 10m 19s | Max: 10m 20s | Hits:  80%/1190  
  🟩 gpu
    🟩 h100               Pass: 100%/2   | Total: 13m 07s | Avg:  6m 33s | Max:  8m 09s | Hits:  91%/1194  
    🟩 rtx2080            Pass: 100%/26  | Total:  3h 42m | Avg:  8m 33s | Max: 20m 27s | Hits:  82%/14304 
  🟩 jobs
    🟩 Build              Pass: 100%/25  | Total:  3h 14m | Avg:  7m 47s | Max: 16m 16s | Hits:  81%/13707 
    🟩 Test               Pass: 100%/3   | Total: 41m 04s | Avg: 13m 41s | Max: 20m 27s | Hits:  99%/1791  
  🟩 sm
    🟩 90                 Pass: 100%/2   | Total: 13m 07s | Avg:  6m 33s | Max:  8m 09s | Hits:  91%/1194  
    🟩 90;90a             Pass: 100%/2   | Total: 20m 01s | Avg: 10m 00s | Max: 13m 59s | Hits:  76%/890   
    🟩 100;120            Pass: 100%/2   | Total: 20m 14s | Avg: 10m 07s | Max: 14m 04s | Hits:  76%/890   
  🟩 std
    🟩 17                 Pass: 100%/3   | Total: 21m 52s | Avg:  7m 17s | Max: 10m 19s | Hits:  82%/1789  
    🟩 20                 Pass: 100%/25  | Total:  3h 33m | Avg:  8m 33s | Max: 20m 27s | Hits:  83%/13709 
  

• 🟩 packaging: Pass: 100%/4 | Total: 24m 49s | Avg: 6m 12s | Max: 8m 17s

  🟩 cpu
    🟩 amd64              Pass: 100%/4   | Total: 24m 49s | Avg:  6m 12s | Max:  8m 17s
  🟩 ctk
    🟩 12.0               Pass: 100%/2   | Total: 14m 21s | Avg:  7m 10s | Max:  8m 17s
    🟩 12.9               Pass: 100%/2   | Total: 10m 28s | Avg:  5m 14s | Max:  7m 35s
  🟩 cudacxx
    🟩 nvcc12.0           Pass: 100%/2   | Total: 14m 21s | Avg:  7m 10s | Max:  8m 17s
    🟩 nvcc12.9           Pass: 100%/2   | Total: 10m 28s | Avg:  5m 14s | Max:  7m 35s
  🟩 cudacxx_family
    🟩 nvcc               Pass: 100%/4   | Total: 24m 49s | Avg:  6m 12s | Max:  8m 17s
  🟩 cxx
    🟩 Clang14            Pass: 100%/1   | Total:  8m 17s | Avg:  8m 17s | Max:  8m 17s
    🟩 Clang19            Pass: 100%/1   | Total:  2m 53s | Avg:  2m 53s | Max:  2m 53s
    🟩 GCC12              Pass: 100%/1   | Total:  6m 04s | Avg:  6m 04s | Max:  6m 04s
    🟩 GCC13              Pass: 100%/1   | Total:  7m 35s | Avg:  7m 35s | Max:  7m 35s
  🟩 cxx_family
    🟩 Clang              Pass: 100%/2   | Total: 11m 10s | Avg:  5m 35s | Max:  8m 17s
    🟩 GCC                Pass: 100%/2   | Total: 13m 39s | Avg:  6m 49s | Max:  7m 35s
  🟩 gpu
    🟩 rtx2080            Pass: 100%/4   | Total: 24m 49s | Avg:  6m 12s | Max:  8m 17s
  🟩 jobs
    🟩 Test               Pass: 100%/4   | Total: 24m 49s | Avg:  6m 12s | Max:  8m 17s
  

👃 Inspect Changes

Modifications in project?

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

Modifications in project or dependencies?

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

🏃‍ Runner counts (total jobs: 32)

#	Runner
17	linux-amd64-cpu16
6	linux-amd64-gpu-rtx2080-latest-1
4	linux-arm64-cpu16
4	windows-amd64-cpu16
1	linux-amd64-gpu-h100-latest-1