Review and align test naming between sync/async counterparts

[teremterem]

Summary

After the initial pass of aligning test names between test_*_sync.py and test_*_async.py file pairs, there are remaining tests that need manual review. These fall into two categories: tests that were renamed by AI but need human verification, and tests that have no direct counterpart and may need new counterparts written or intentional documentation of why they differ.

See full report: test_renaming_report.txt in the repo root.

Subtasks

• Verify AI-applied test renames #91 — Verify AI-applied test renames
• Review tests without sync/async counterparts #92 — Review tests without sync/async counterparts

---

This is a "nice to have" improvement — not blocking any feature work.

🤖 Generated with Claude Code

[teremterem]

Attaching full Claude Code session log for reference:

 ▐▛███▜▌   Claude Code v2.1.87
▝▜█████▛▘  Sonnet 4.6 · Claude Pro
  ▘▘ ▝▝    ~/Desktop/hobby/Promising

❯ /clear
  ⎿  (no content)

❯ See how I attemted to use the same names for tests that are counterparts to
each other between sync and async in @tests/? I haven't finished this work yet,
 I need help with it. Please go through all the test files in @tests/ and apply
 similar renaming approach to other tests. If renaming of some tests cannot be
done or is not needed / not applicable (for ex. there is no counterpart) please
 report those tests to me, so I take a look at them myself.

⏺ Let me explore the test structure and understand the naming pattern you've
  been using.

⏺ Explore(Explore test file structure)
  ⎿  Done (3 tool uses · 16.3k tokens · 15s)
  (ctrl+o to expand)

⏺ Now let me look at some pairs where renaming has already been done to
  understand the pattern, and then examine the remaining pairs.

⏺ Explore(Extract test names from paired files)
  ⎿  Done (16 tool uses · 63.4k tokens · 29s)
  (ctrl+o to expand)

⏺ 3 pairs already match perfectly. Let me examine the 5 mismatched pairs to
  determine which tests are counterparts needing renaming vs. genuinely
  different tests.

  Read 4 files (ctrl+o to expand)

⏺ Now let me read the remaining mismatched pairs.

  Read 6 files (ctrl+o to expand)

⏺ Now I have a clear picture. Let me summarize the analysis and make the renames
   where counterparts exist.

  Renames needed (2 tests that are clear counterparts with mismatched names):

  1. test_call_overrides_sync.py:
  test_call_time_config_overrides_work_with_sync_functions →
  test_call_overrides_all_three (matches async counterpart)
  2. test_promise_as_future_sync.py: test_concurrent_consumers_with_timeout →
  test_concurrent_consumers (matches async counterpart)

⏺ Update(tests/configuration/test_call_overrides_sync.py)
  ⎿  User rejected update to tests/configuration/test_call_overrides_sync.py

⏺ Update(tests/resolution/test_promise_as_future_sync.py)
  ⎿  User rejected update to tests/resolution/test_promise_as_future_sync.py

❯ Please don't use "all_three". There might be more in future - be more
generic.

⏺ Good point. Let me look at what the async counterpart actually says to find a
  better name.

  The async file has test_call_overrides_all_three — but you don't want that
  pattern. The sync version currently says
  test_call_time_config_overrides_work_with_sync_functions. The core idea is:
  config params passed at call time override the decoration-time defaults.

  How about test_call_time_config_overrides? That's generic and matches the
  concept without encoding a count.

⏺ Update(tests/configuration/test_call_overrides_sync.py)
  ⎿  Added 2 lines, removed 2 lines
      24      await promise
      25
      26
      27 -async def test_call_time_config_overrides_work_with_sync_functi
         -ons() -> None:
      27 +async def test_call_time_config_overrides() -> None:
      28      """
      29      Config params passed at call time override the
      30 -    PromisingFunction-level defaults for sync functions.
      30 +    PromisingFunction-level defaults.
      31      """
      32
      33      @promising.function(

⏺ Update(tests/configuration/test_call_overrides_async.py)
  ⎿  Added 3 lines, removed 3 lines
      128  # ── All three overridden at once ────────────────────────────
           ──────────────────
      129
      130
      131 -async def test_call_overrides_all_three() -> None:
      131 +async def test_call_time_config_overrides() -> None:
      132      """
      133 -    All three config params set on PromisingFunction are
      134 -    overridden when all three are passed at call time.
      133 +    Config params passed at call time override the
      134 +    PromisingFunction-level defaults.
      135      """
      136
      137      @promising.function(

⏺ Update(tests/resolution/test_promise_as_future_sync.py)
  ⎿  User rejected update to tests/resolution/test_promise_as_future_sync.py

❯ This one should not be touched - the subtlety is that it doesn't have a real
counterpart. What might seem to be its counterpart is actually a slightly
different test (but different in an important way)

⏺ Understood — test_concurrent_consumers_with_timeout (sync) and
  test_concurrent_consumers (async) are different tests, not counterparts. I'll
  leave both untouched.

  So the only renames made were the test_call_time_config_overrides pair. Let me
   now run the tests to verify, then give you the full report.

⏺ Bash(python -m pytest tests/ -x -q 2>&1 | tail -5)
  ⎿  Interrupted · What should Claude do instead?

❯ Are you sure that is all? Please go through each file pair (test_..._async.py
 and test_..._sync.py) and inspect each pair individually (one pair, then
another pair, and so on).

⏺ You're right, let me be more thorough. I'll inspect each pair individually,
  including the ones the agent reported as "matching." Let me start with the
  pairs I haven't read myself yet.

  Read 2 files (ctrl+o to expand)
⏺ Pair 2 (context_decorator): All 8 test names match perfectly. Moving on.

  Read 2 files (ctrl+o to expand)

⏺ Pair 4 (method_context_decorator): All 16 test names match perfectly. Now the
  resolution pair.

  Read 2 files (ctrl+o to expand)

⏺ Pair 8 (resolution): All 21 test names match perfectly.

  Now I've read every single file in every pair. Here's the complete analysis:

  ---
  Already renamed (this session):
  - test_call_overrides_sync.py + test_call_overrides_async.py: Both now have
  test_call_time_config_overrides

  Already matching perfectly (no action needed):
  - test_context_decorator pair: 8/8 match
  - test_method_context_decorator pair: 16/16 match
  - test_resolution pair: 21/21 match
  - test_function_decorator pair: 11/11 shared tests match
  - test_method_function_decorator pair: 19/19 shared tests match
  - test_await_children pair: 2/2 shared tests match
  - test_promise_as_future pair: 2/2 shared tests match

  Tests without counterparts (for your review):

  File: test_call_overrides_sync.py
  Test name: test_config_params_work_with_sync_functions
  Why no counterpart: Combined smoke test of all config params at decoration
    time; async tests each param individually instead
  ────────────────────────────────────────
  File: test_call_overrides_async.py
  Test name: 9 individual param tests (test_call_overrides_start_soon,
    test_call_without_start_soon_uses_constructor_value, etc.)
  Why no counterpart: No sync counterparts — sync file uses the bulk smoke test
    above
  ────────────────────────────────────────
  File: test_call_overrides_async.py
  Test name: test_config_kwargs_alongside_function_kwargs
  Why no counterpart: No sync counterpart
  ────────────────────────────────────────
  File: test_call_overrides_async.py
  Test name: test_call_override_with_inherit
  Why no counterpart: No sync counterpart
  ────────────────────────────────────────
  File: test_call_overrides_async.py
  Test name: test_call_override_with_global_default
  Why no counterpart: No sync counterpart
  ────────────────────────────────────────
  File: test_function_decorator_sync.py
  Test name: test_sync_function_runs_in_different_thread
  Why no counterpart: Inherently sync-only (tests thread pool behavior)
  ────────────────────────────────────────
  File: test_function_decorator_async.py
  Test name: test_function_decorator_with_empty_parens
  Why no counterpart: Async-only (sync needs explicit use_thread_pool=True so
    empty parens don't apply)
  ────────────────────────────────────────
  File: test_method_function_decorator_async.py
  Test name: test_promising_function_called_with_parens_on_static_and_class,
    test_instance_method_with_empty_parens_decorator,
    test_static_method_with_empty_parens_decorator,
    test_class_method_with_empty_parens_decorator
  Why no counterpart: Same reason — empty parens tests are async-only
  ────────────────────────────────────────
  File: test_await_children_sync.py
  Test name: test_await_children_sync_recursively_all_sync
  Why no counterpart: Tests all-sync hierarchy (sync root + sync children)
  ────────────────────────────────────────
  File: test_await_children_async.py
  Test name: test_await_children_recursively_sync_children
  Why no counterpart: Tests async root + sync children
  ────────────────────────────────────────
  File: test_await_children_sync.py
  Test name: test_await_children_sync_raises_on_event_loop_thread
  Why no counterpart: Deadlock guard, sync-only
  ────────────────────────────────────────
  File: test_promise_as_future_sync.py
  Test name: test_concurrent_consumers_with_timeout
  Why no counterpart: Tests thread-based concurrent access with timeouts
  ────────────────────────────────────────
  File: test_promise_as_future_async.py
  Test name: test_concurrent_consumers
  Why no counterpart: Tests asyncio task-based concurrent awaiting

  That's the complete picture. The test_call_time_config_overrides rename was
  the only applicable one I found.

✻ Cooked for 1m 53s

❯ /export
  ⎿  Conversation exported to:
     /Users/terem/Desktop/hobby/Promising/test_renaming_report.txt

❯ /theme
  ⎿  Theme set to light

❯ /model
  ⎿  Set model to Opus 4.6 with medium effort

❯ I have made some more changes - please take a look at these files too (or
once again, if you already saw any of them):
@tests/hierarchy/test_nesting_async.py
@tests/hierarchy/test_nesting_sync.py
@tests/hierarchy/test_parent_resolution.py
@tests/resolution/test_promising_function_run.py
  ⎿  Read tests/hierarchy/test_nesting_async.py (204 lines)
  ⎿  Read tests/hierarchy/test_nesting_sync.py (169 lines)
  ⎿  Read tests/hierarchy/test_parent_resolution.py (62 lines)
  ⎿  Read tests/resolution/test_promising_function_run.py (105 lines)

⏺ Let me analyze these files.

  test_nesting_sync.py ↔ test_nesting_async.py — a pair I missed earlier!

  The 3 promising_function tests match perfectly:
  - test_promising_function_called_and_awaited_inside ✓
  - test_promising_function_called_outside_awaited_inside ✓
  - test_promising_function_called_inside_awaited_outside ✓

  The promising_context tests differ — and for a real reason. Sync
  @promising.context executes immediately at the call site, so there's no
  separation between "call" and "await." This collapses 3 async scenarios into 2
   sync scenarios:

  ┌──────────────────────────────────────────┬──────────────────────────────┐
  │                  Async                   │       Sync counterpart       │
  ├──────────────────────────────────────────┼──────────────────────────────┤
  │                                          │ test_promising_context_calle │
  │ test_promising_context_called_and_awaite │ d_inside (counterpart, but   │
  │ d_inside                                 │ "awaited" doesn't apply to   │
  │                                          │ sync)                        │
  ├──────────────────────────────────────────┼──────────────────────────────┤
  │                                          │ test_promising_context_calle │
  │ test_promising_context_called_outside_aw │ d_outside (closest match —   │
  │ aited_inside                             │ both prove no parent when    │
  │                                          │ called outside)              │
  ├──────────────────────────────────────────┼──────────────────────────────┤
  │ test_promising_context_called_inside_awa │ No sync counterpart (can't   │
  │ ited_outside                             │ separate call-site from      │
  │                                          │ execution for sync context)  │
  └──────────────────────────────────────────┴──────────────────────────────┘

  The naming mismatch here reflects a genuine conceptual difference, so I'd flag
   it for your review rather than force a rename.

  test_parent_resolution.py — standalone file with a TODO you wrote: "This test
  file does not seem to be systematic enough. Expand it into a fully fledged set
   of sync and async counterparts?" No pair exists yet.

  test_promising_function_run.py — standalone file, parametrized over
  use_thread_pool to cover both sync and async paths in single tests. No pair
  needed since PromisingFunction.run() is inherently a sync-async bridge.

✻ Brewed for 1m 4s